2 cfL ANNALS OF BOTANY VOL. XII Annals of Botany EDITED BY ISAAC BAYLEY BALFOUR, M.A., M.D., F.R.S. QUEEN’S BOTANIST IN SCOTLAND, PROFESSOR OF BOTANY IN THE UNIVERSITY AND KEEPER OF THE ROYAL BOTANIC GARDEN, EDINBURGH SYDNEY HOWARD VINES, M.A., D.Sc., F.R.S. FELLOW OF MAGDALEN COLLEGE, AND SHERARDIAN PROFESSOR OF BOTANY IN THE UNIVERSITY OF OXFORD D. H. SCOTT, M.A., Ph.D., F.R.S. HONORARY KEEPER OF THE JODRELL LABORATORY, ROYAL 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 XII fiott&on HENRY FROWDE, M.A., AMEN CORNER, E.C. OXFORD: CLARENDON PRESS DEPOSITORY, 116 HIGH STREET 1898 Oxfor b PRINTED AT THE CLARENDON PRESS BY HORACE HART, M.A. PRINTER TO THE UNIVERSITY SD.5M'2- AW\ CONTENTS. No. XLV, March, 1898. PAGE Campbell, D. H. — The Development of the Flower and Embryo in Lilaea subulata, H. B. K. (With Plates I— III) ... 1 West, W., and West, G. S. — Observations on the Conjugatae. (With Plates IV and V) 29 Ward, H. M.- — A Violet Bacillus from the Thames. (With Plate VI) 59 Church, A. H. — The Polymorphy of Cutleria multifida, Grev. (With Plates VII-IX) 75 Dawson, M. — On the Structure of an Ancient Paper . . . .nr NOTES. Townsend, C. O. — Correlation of Growth under the Influence of Injuries 1 1 7 Dixon, H. H. — Gelatine as a Fixative .117 Groom, P. — Lathraea Squamaria . . . . . . .118 No. XLVI, June, 1898. Johnson, D. S. — On the Development of the Leaf and Sporocarp in Marsilia quadrifolia, L. (With Plates X-XII) . . . 119 Parkin, J. — On some points in the Histology of Monocotyledons. (With Plate XIII) 147 Magnus, P. — On Aecidium graveolens (Shuttlew.). (With Plate XIV) 155 Biffen, R. H. — The Coagulation of Latex 165 Phillips, R. W. — The Development of the Cystocarp in Rhody- meniales : IT. Delesseriaceae. (With Plates XV and XVI) . 173 Worsdell, W. C. — The Vascular Structure of the Sporophylls of the Cycadaceae. (With Plates XVII and XVIII) . . . 203 Reid, C. — Further Contributions to the Geological History of the British Flora 243 NOTES. Lang, W. H. — On Apogamy and the Development of Sporangia upon Fern-Prothalli .251 Maslen, A. J. — The Ligule in Lepidostrobus. (With Woodcut 1) . 256 VI Contents. No. XL VII, September, 1898. Shaw, W. R.— *The Fertilization of Onoclea. (With Plate XIX) Ward, H. M. — Some Thames Bacteria. (With Plates XX and XXI) Hill, T. G.— On the Roots of Bignonia. (With Plate XXII) . Barber, C. A. — Cupressinoxylon vectense. (With Plates XXIII and XXIV) Ewart, A. J. — The Action of Cold and of Sunlight upon Aquatic Plants . . Scott, R., and Sargant, E. — On the Development of Arum maculatum from the Seed. (With Plate XXV) . . . . . NOTES. PAGE 26l 287 323 329 363 399 Ewart, A. J. — The Action of Chloroform on C02-assimilation . . 415 Lewis, F. J. — The Action of Light on Mesocarpus . . . .418 No. XL VIII, December, 1898. Ganong, W. F. — Contributions to a Knowledge of the Cactaceae : II. The Comparative Morphology of the Embryos and Seedlings. (With Plate XXVI) 423 Pearson, H. H. W. — Anatomy of the Seedling of Bowenia spectabilis. (With Plates XXVII and XXVIII) . . . 475 Green, J. R. — The Alcohol-producing Enzyme of Yeast . . .491 Wager,. H. — The Nucleus of the Yeast-Plant (With Plates XXIX and XXX) 499 Vines, S. H. — The Proteolytic Enzyme of Nepenthes, II . . . 545 NOTES. Burkill, I. H. — Changes in the Sex of Willows .... 557 Jones, C. E. — Anatomy of the Stem of Species of Lycopodium . . 558 Williams, J. Lloyd. — Reproduction in Dictyota dichotoma . . 559 Huie, L. II. — Changes in the Gland-Cells of Drosera produced by various Food-materials ........ 560 Ward, H. M. — A Potato-Disease . 561 Penicillium as a Wood-destroying Fungus 565 Ellis, W. G. P. — A Method of obtaining Material for illustrating Smut in Barley 566 Errera, L. — Structure of the Yeast-Cell 567 Osmotic Optimum and Measurements 568 Phillips, R. W. — The Form of the Protoplasmic Body in certain Florideae . 569 Klebs, G. — Alternation of Generations in the Thallophytes . . 570 Lang, W. H. — Alternation of Generations in the Archegoniatae . . 583 Hartog, M. — Alternation of Generations 593 CONTENTS AND INDEX i-viii Hooker, Sir J. D. — Biographical Memoir of George Bentham. (With Portrait) ix-xxx INDEX. A. ORIGINAL PAPERS AND NOTES. Barber, C. A. — Cupressinoxylon vectense. (With Plates XXIII and XXIV) . . . . . Biffen, R. H. — The Coagulation of Latex Burkill, I. H. — Changes in the Sex of Willows .... Campbell, D. H. — The Development of the Flower and Embryo in Lilaea subulata, H. B. K. (With Plates I-III) . Church, A. H. — The Polymorphy of Cutleria multifida, Grev. (With Plates VII-IX) Dawson, M. — On the Structure of an Ancient Paper .... Dixon, H. H. — Gelatine as a Fixative Ellis, W. G. P.---A Method of obtaining Material for illustrating Smut in Barley Errera, L. — Structure of the Yeast-Cell ...... Osmotic Optimum and Measurements ...... Ewart, A. J. — The Action of Cold and of Sunlight upon Aquatic Plants The Action of Chloroform on C02-assimilation .... Ganong, W. F. — Contributions to a Knowledge of the Cactaceae : II. The Comparative Morphology of the Embryos and Seedlings. (With Plate XXVI) Green, J. R. — The Alcohol-producing Enzyme of Yeast . Groom, P. — Lathraea Squamaria Hartog, M. — Alternation of Generations ...... Hill, T. G. — On the Roots of Bignonia. (With Plate XXII) . Hooker, J. D. — Biographical Memoir of George Bentham. (With Portrait) Huie, L. H. — Changes in the Gland-Cells of Drosera produced by various Food-materials Johnson, D. S. — On the Development of the Leaf and Sporocarp in Marsilia quadrifolia, L. (With Plates X-XII) Jones, C. E. — Anatomy of the Stem of Species of Lycopodium . Klebs, G. — Alternation of Generations in the Thallophytes Lang, W. H. — On Apogamy and the Development of Sporangia upon Fern-Prothalli ........ Alternation of Generations in the Archegoniatae . Lewis, F. J. — The Action of Light on Mesocarpus . Magnus, P. — On Aecidium graveolens (Shuttlew.). (With Plate XIV) Maslen, A. J. — The Ligule in Lepidostrobus. (With Woodcut i) . PAGE 329 165 557 75 111 11 7 566 567 568 363 4i5 423 49 1 118 593 323 ix 560 119 558 570 251 583 418 155 256 Vlll Index. PAGE Parkin, J. — On some points in the Histology of Monocotyledons. (With Plate XIII) 147 PEARSON, H. H. W. — Anatomy of the Seedling of Bowenia spectabilis. (With Plates XXVII and XXVIII) 475 Phillips, R. W. — The Development of the Cystocarp in Rhody- meniales: II. Delesseriaceae. (With Plates XV and XVI) . 173 The Form of the Protoplasmic Body in certain Florideae . . 569 Reid, C. — Further Contributions to the Geological History of the British Flora 243 Scott, R., and Sargant, E. — On the Development of Arum maculatum from the Seed. (With Plate XXV) ..... 399 Shaw, W. R. — The Fertilization of Onoclea. (With Plate XIX) . 261 Townsend, C. O. — Correlation of Growth under the Influence of Injuries 1 1 7 Vines, S. H. — The Pro teolytic Enzyme of Nepenthes, II . . . 545 Wager, H. — The Nucleus of the Yeast-Plant. (With Plates XXIX and XXX) 499 Ward, H. M. — A Violet Bacillus from the Thames. (With Plate VI). 59 Some Thames Bacteria. (With Plates XX and XXI) . . . 287 A Potato-Disease 561 Penicillium as a Wood-destroying Fungus 565 West, W., and West, G. S. — Observations on the Conjugatae. (With Plates IV and V) 29 Williams, J. Lloyd. — Reproduction in Dictyota dichotoma . . 559 Worsdell, W. C. — The Vascular Structure of the Sporophylls of the Cycadaceae. (With Plates XVII and XVIII) . . . 203 B. LIST OF ILLUSTRATIONS. a. Plates. Portrait of George Bentham (Frontispiece). I, II, III. Development of Flower and Embryo in Lilaea subulata (Campbell). IV, V. Observations on Conjugatae (West, W., and West, G. S.). VI. Violet Bacillus from the Thames (Ward). VII, VIII, IX. Polymorphy of Cutleria multifida (Church). X, XI, XII. Development of Leaf and Sporocarp in Marsilia quadrifolia (Johnson). XIII. Some points in the Histology of Monocotyledons (Parkin). XIV. Aecidium graveolens (Magnus). XV, XVI. Development of Cystocarp in Delesseriaceae (Phillips). XVII, XVIII. Vascular Structure of Sporophylls of Cycadaceae (Worsdell). XIX. Fertilization of Onoclea (Shaw). XX, XXI. Some Thames Bacteria (Ward). XXII. Roots of Bignonia (Hill). XXIII, XXIV. Cupressinoxylon vectense (Barber). XXV. Development of Arum maculatum (Scott and Sargant). XXVI. Comparative Morphology of Embryos and Seedlings of Cactaceae (Ganong). XXVII, XXVIII. Anatomy of Seedling of Bowenia spectabilis (Pearson). XXIX, XXX. Nucleus of the Yeast-Plant (Wager). b. Woodcut. 1. Ligule in Lepidostrobus (Maslen). GEORGE BENTHAM, F.R.S. ( With Portrait .) The following account of the life and labours of George Bentham is based on an obituary notice which I communi- cated to Nature (Vol. XXX, October 1884, p. 359). In reproducing it in the present form, I have enlarged it con- siderably, and further availed myself of four subsequent accounts, namely, of Mr. Thiselton-Dyer’s Eulogium, read before the Linnean Society (Proceedings, Sessions 1887-1889); of Prof. Gray’s Memorial, presented to the American Academy of Arts and Sciences (Journal, Vol. XXIX, February, 1885) ; of Prof. Oliver’s Obituary notice (Proc. Royal Society, 1885), and of Mr. Daydon Jackson’s notice (Proc. Linn. Soc., Session 1884-5). The reminiscences of his very early life are taken from an autobiography which he commenced very shortly before his death, but which he was unable to continue. The life of George Bentham presents such variety, such startling changes of conditions, and a combination of so many natural and acquired mental powers of a high order, that it cannot be perused without the question arising, how far heredity and environments had influenced his career. Such being the case, I think no apology is needed for com- mencing this sketch with some account of his parentage. He was born on September 22, 1800, in the village of Stoke, near Portsmouth. His father, afterwards Sir Samuel Bentham, who was the son of a wealthy scrivener in the Minories, and the only brother of Jeremy Bentham the publicist, devoted himself as a youth to the study of Naval Architecture, and at the age of 22, at the suggestion of Lord Howe, went to Russia with the view of further instructing X George Bent ham, F.R.S. himself in that science. Then he travelled to the Crimea, visited the naval establishments in the Baltic and the Black Sea, and thence went on to Siberia (penetrating to the frontiers of China) for the purpose of making himself ac- quainted with the mines, foundries, and other great industries of that country. Meanwhile he had gained the friendship of Prince Potemkin, who, impressed by his genius and ability, induced him to enter the service of the Empress Catherine II, who gave him a Lieutenant-Coloners commission, without requiring him to pass through the subaltern grades of the army. In this capacity he was sent to the Crimea, where, amongst many other engineering feats, he built a flotilla of gun-boats, in command of which (under Prince Potemkin) he gained a signal victory over the whole Turkish fleet in the Black Sea. For this he received from the Empress the cross of St. George, conferring Knighthood, a sword of honour, and promotion to the rank of Colonel in command of a cavalry regiment in Siberia, which country he re-traversed from the Obi to the Amur, engaged chiefly in the construction of boats for the navigation of the Siberian rivers. After the death of the Empress he returned to England, left the Russian service, and entered that of the Admiralty, by whom he was commissioned to return to Russia, and there superintend the building of some ships for the British Navy. Thither he went with his wife and family, including George, and remained, till the declaration of war with that country required his recall. Finally he rose to be Inspector of all our dockyards, in which capacity he introduced a multitude of improvements, including steam saw- and other mills, the replacement of water-casks in ships by iron tanks ; and with Sir Isambard Mark Brunei, whom he brought over from the Continent, he constructed the eccentric machinery for turning elliptic blocks1. G. Bentham’s mother was the daughter of Dr. G. Fordyce, F.R.S. , an eminent London physician, and lecturer on chemistry, author of various works on medicine and agri- 1 Sir Samuel Bentham’s portrait hangs in Greenwich Hospital. XI George Benthani , F.R.S . culture. She was a woman of remarkable power of mind, who aided her father in his scientific labours, and her husband in preparing his voluminous official reports to the Admiralty. At the age of 80 she wrote a beautiful hand, and during the Crimean war, when considerably over 90, she commenced a series of letters to the Times , urging the adoption of guns of a large calibre, and other improvements in war-material, the inventions of her late husband, whose Life she published in 1 862. Not less influential on George Bentham’s career was the teaching of his uncle Jeremy, who imbued him with that love for methodical and logical analysis which is so con- spicuous in all his nephew’s writings. As has been well remarked in this relation, ‘ The same inherited aptitude and contemporary influences which produced a great publicist in Jeremy, yielded, by an almost accidental deflection, a great systematic botanist in his nephew ’ (Eulogium, p. 8). Environments were as favourable to Bentham in his scientific career, as were the qualities of his progenitors. He was one of five children (three of them girls), all of them precocious. They were taught to read by words, not by syllables or letters, and the two brothers commenced learning Latin before they were five years old. In 1805 the whole family accompanied the father to Russia, where their education was entrusted to a talented Russian lady who could speak no English, whilst in Latin the boys were instructed by a Russian priest, of whom George in after life always spoke with great regard. Music, of which the latter became passionately fond, was not neglected, and it resulted in his becoming an accomplished pianist. Thus, having a remarkable facility for acquiring languages, Bentham could, at seven years old, converse fluently in English, French, German, and Russian, to which, by hard work, he added Swedish, during a detention of some weeks at Carls- crona on the voyage back to England. The said voyage proved a tempestuous and perilous one. Embarking at Revel in a Russian frigate, with a crew, few of whom had ever before seen the sea, they were tossed about in the Baltic xii George Bentham , F.R.S. for five weeks before arriving at Carlscrona ; and there were as many equally stormy weeks in their passage thence in an ill-found merchantman to Harwich, where the family arrived in a half-starved condition, having been reduced to picking up stray scraps of biscuit in out-of-the-way corners of the cabins. In England Sir Samuel Bentham took up his residence at Hampstead in the summer months, daily visiting his office at the Admiralty. In winter he resorted to a small house and property called Berry Lodge1, between Alverstoke and Gosport, which was in convenient proximity to Portsmouth Dockyard. The boys meanwhile pursued their studies under private tutors, a plan continued throughout the whole course of George’s education. It was a life-long source of regret with George that he had never been sent to school or college, which may account for a shyness and reserve, attributed by those that did not know him to a want of sympathy. In 1812-13, the invasion of Russia by Napoleon, and the burning of Moscow, naturally caused great excitement in the Bentham family. It led to the first appearance of George, then only 13, before the public; he, with his brother and sisters, contributing to the London Magazine a series of papers, gleaned from Russian sources, detailing the operations of the armies, and glorying in the reverses and final abdication of Napoleon. After the proclamation of peace, Sir Samuel Bentham took his family to France, and resided successively at Tours, Saumur, and Paris. During this period, which extended from Napoleon’s return from Russia to his final overthrow, young Bentham kept a full journal of all that passed, interspersed with anecdotes relating to the forced exile of Louis, the restoration of the Bourbons, the execution of Ney and 1 It was from here that, while George was still in his teens, his father took him on a visit to Lady Spencer at Ryde, at whose house he met John Stuart Mill, a lad of six, dressed in a scarlet jacket buttoned over nankeen trousers, and con- sidered to be a prodigy. Bentham has described him to me as having been wonderfully precocious, a Greek and Latin scholar, historian and logician, whom he heard discussing with Lady Spencer the relative merits of her ancestor, the Duke of Marlborough, and of Wellington, young Mill taking the part of the latter. George Bentham , F.R.S . xiii Labedoyere, the condition of the city of Paris, and to Walter Savage Landor, who was intimate with the family. Even at this age he could take his part in the society of the leaders of the Paris Salons in literature and science, making the acquaintance of the Due de Richelieu, Talleyrand, Dumas, Jean Baptiste Say, the aged Madame Andelau (daughter of Helvetius), and Alexander Humboldt. Of these the latter took an especial interest in him, encouraging him in the prosecution of a work he had begun on the data of physical geography, by advice and by procuring him introductions to libraries and to individuals who could aid him. Unfor- tunately this projected work was not continued. In 1816 Sir Samuel Bentham organized at Paris a caravan- tour in France for himself and family. The caravan consisted of a two-horse coach fitted up as a sleeping- room, a one-horse spring van furnished with a library and piano, for himself and Mrs. Bentham, and another for his daughters and their governess. Thus equipped they travelled by day, visiting friends and places of interest, bivouacking by night in gipsy fashion in the gardens of friends, or in the precincts of the prefectures, to which he brought credentials from Paris. In this way he visited Orleans, Tours, Angouleme, Bordeaux, Toulouse, Montpellier, and finally Montauban, where the caravan having broken down, the tour was continued by ordinary conveyances to Carcassonne, Narbonne, Nimes, Tarascon, Marseilles, Toulon, and Hyeres. The most interesting incident of this tour occurred at Angouleme, for there G. Bentham’s attention was first directed to botany. His mother, who was fond of plants, and a friend of Aiton of Kew, had purchased a copy of De Candolle’s just then issued ‘Flore Fran£aise.’ Young Bentham acci- dentally taking it up was interested in the analytical tables for determining the affinities and names of the plants described, which fitted in with the ideas he had derived from his uncle Jeremy’s works, when constructing his own geographical tables. He at once went into the yard of the house, gathered the first plant he found, and after spending XIV George Bent ham, B.R.S. the morning in studying its structure, with the aid of the introductory chapter of the c Flore/ succeeded in referring it to its order, genus, and species. The plant, Salvia pratensis , is not an easy one for a beginner. Encouraged by his success with it, he pursued the study of the native flora as a diversion, naming every plant he subsequently met with. At Montauban, where his father had purchased a country house, which the family occupied for about two years, young George Bentham passed what he always regarded as the most enjoyable period of his life. He was entered as a student in a Protestant theological college, and followed with ardour the courses of mathematics, Hebrew, and comparative philo- logy, the latter a favourite study in after life. At home, during the holidays, he occupied his time with drawing plants, learning Spanish, and with music, society, and dancing, of which latter he was passionately fond. It was a favourite boast of his, that at Montauban he had danced at thirty-four balls between Twelfth-night and Mardi-gras, of which thirteen were consecutive, and lasted from 9 p.m. to 9 a.m. Here, too, his mind was first opened to scientific and exotic botany, to which he was led by the works of De Candolle, by the appearance of the ‘ Dictionnaire d’Histoire Naturelle/ and by a course of lectures under Benedict Prevost. This was fol- lowed by a devotion to ornithology, including shooting and stuffing birds, and that again by entomology, tabulating the phenomena of insect-life. Here, too, probably inspired by John Stuart Mill, who resided for some time with the family, his mind was turned to philosophy and the study of Lamarck’s works, beginning with the ‘ Systeme analytique des Connais- sances positives de Thomme/ only to give it up with disgust on reading that ‘ Dieu crea d’abord la matiere,5 followed by the statement that Nature was the second thing created, and that this produced everything else. More to the purpose was his translation into French of his uncle’s Chrestomathia, which was a prelude to his becoming secretary to the great publicist at a later period. From Montauban Sir Samuel Bentham moved to a large XV George Bentham , F.R.S. estate of 2,000 acres, which he had purchased, near Montpellier, and the management of which he made over to his now only- son, the eldest having died some years previously. The estate consisted of farms and vineyards, to the improvement of which George devoted himself with alacrity and success. They became very profitable, and throughout the remainder of his life in England an excellent St. George Burgundy (the produce of the Restinalieres estate), and a rare and luscious Lunel from a neighbouring vineyard, were familiar to the guests at his table. All his spare time was devoted to botanical excursions in the Cevennes and Pyrenees, and to making a French translation of his uncle’s Essay on Nomen- clature and Classification. Here, too, he wrote his first important work, c Essai sur la Nomenclature et Classification des Arts et Sciences,’ which was published in Paris, and which established his position as an acute analyzer, clear expositor, and cautious reasoner. Half a century after its appearance it was praised by Professor Stanley Jevons in his History of the Sciences. In 1823 G. Bentham was sent to England for the purpose of obtaining agricultural implements and information as to improved methods of farming. On his arrival in London he was well received by his uncle, and introduced into the best literary and scientific society of the capital. He was invited to the breakfasts and receptions of Sir Joseph Banks, and studied in his library and herbarium, where he commenced a life-long friendship with their curator, Robert Brown, ‘ Botani- corum facile princeps.’ There, and at the Horticultural and Linnean Societies, he met the elite of the naturalists of the day. From London he made a tour into Scotland, where he was hospitably entertained by the Professors of Botany in Edinburgh (Graham) and Glasgow (Hooker)1, and 1 It was from this visit that Bentham was wont to date his permanent adherence to Botany. I, then six years old, remember him and the enthusiasm with which he received from my father a collection of Alpine Scotch plants, the first examples of the Northern European Flora he had ever seen. The intimate friendship between my father and Bentham, which lasted forty-two years, dates from this period. b XVI George Bent ham, F.R.S. by Arnott, of Arlary, in Kinross-shire, who was subsequently Professor of Botany in Glasgow. With the latter he arranged to make an extended botanical excursion in the Pyrenees, which was carried out in 1824, and which resulted in his first botanical work, ‘ Catalogue des Plantes Indigenes des Pyrenees et du Bas Languedoc, avec des notes et observations ’ (Paris, 1826). Another result of his Pyrenean exploration was . the publication in the London Magazine for 1827 of two articles entitled ‘ Sketches of Manners in the South of France,’ wherein, amongst much curious philological and other matter relating to the Roussillonais, Catalonians, and Languedociens, an account is given of a visit to the Lilliputian Republic of Andorra, its physical features, people, government, agriculture, and productions. These sketches are masterpieces of their kind. In 1826 the Restinalieres estate had to be abandoned, owing to provincial jealousy, which threw every obstacle in the way of improvements, and the Bentham family returned to England for good. Here a new, and as it proved, a very uncongenial career was opened to George. His uncle Jeremy, gratified by the translation of Chrestomathia, invited his nephew to be his aid in arranging his MSS. for the press, accompanying it with the assurance that he would provide for him at his death. This invitation was accepted, but not the offered provision, for he desired to follow an independent profes- sion, and the result of many interviews was that he determined to enter Lincoln’s Inn as a student of Law, whilst giving some morning hours to his uncle’s work, dining with him twice a week, and writing for him after dinner, from 8 to 1 1 p.m. In one capacity or another he acted as his uncle’s secretary until 1832, when the death of the latter, in many of whose ideas he did not participate, released him from his irksome labour, without however fulfilling his just expectations of reward ; for, owing to the many fruitless speculations of the great jurist, the sums squandered by his executors on the posthumous publications of his works, and some irregularities in his will, Bentham benefited chiefly by coming into pos- XVII George Bent ham, F.R.S. session of the house in Queen Square Place1. Here, after his marriage in 1833 to a daughter of the Rt. Hon. Sir Harford Brydges, Bart., formerly H. M. Envoy at the Court of the Shah of Persia, he resided until 1843. He had, however, been rendered independent through his fathers death two years previously. The amount and variety of mental work achieved by Bentham, during the years of bondage to his uncle, is very remarkable in many ways. Over and above his duties as his uncle’s secretary, he had to arrange, often rewrite, and edit, his father’s voluminous papers on the administration of the dockyards and other naval matters, and to study law. His legal studies were finally abandoned for logic and juris- prudence, but not till after he had published three notable papers ; one on codification, on which subject he entirely dis- agreed with his uncle, but the paper attracted the attention of Brougham, Hume, and O’Connel. Another paper was on the laws affecting larceny, apropos of Sir Robert Peel’s bill for the consolidation of the Criminal Law. Of this Peel thought so highly that he complimented its author, and informed him that it should be submitted to Sir John Richardson, to whom the bill was referred ; a copy of it being shown by his uncle to Lord Brougham, the latter wrote a letter of eighteen pages of remarks upon it. The third was a pamphlet on the Law of Real Property. But his most considerable work of this period received scant attention from those most interested in its subject, and passed from its birth directly into an oblivion from which it was rescued only in later years, yet without word or sign from its author. This was his 4 Outlines of a new system of Logic, with a critical ex- amination of Dr. Whately’s Elements of Logic,’ published in 1827. In it the Quantification of the Predicate2 was 1 It overlooked St. James’s Park and the parade ground of Wellington Barracks, and its site is now approximately occupied by the ‘ Bentham wing ’ of the Queen Anne’s Mansions. It had been in possession of the family for upwards of a century, having been purchased by Bentham’s paternal grandfather. 2 The following history of this episode in Bentham’s career is, I think, too interesting and too important to be omitted in his obituary. It is taken from the b 2 XVlll George Bent ham, F.R.S. first systematically applied in such wise that Professor Stanley Jevons declared it to be £ undoubtedly the most fruitful discovery made in abstract logical science since the time of Aristotle.’ Meanwhile Bentham had been called to the Bar by Lincoln’s Inn, but his career as a barrister was the briefest. As counsel in a case he broke down in court through nervousness, and thenceforth wisely abandoned the practice of the profession. In botany Bentham was more at home than in the Law- courts. In 1828 his herbarium arrived from France, and in the same year he was elected a Fellow of the Linnean Memorial by Professor Gray : ‘ Before sixty copies of the work had been sold, the publishers became bankrupt, and the whole impression of this work of a young and unknown author was sold for waste paper. One of the extant copies, however, came into the hands of the distinguished philosopher, Sir W. Hamilton, to whom the discovery of the quantification of the predicate was credited, and who, in claim- ing it, brought an acrimonious charge of plagiarism against Professor De Morgan upon this subject : yet this very book of Mr. Bentham’s is one of the ten placed by the title at the head of Sir William Hamilton’s article on Logic in the Edinburgh Review for April, 1833, and is once or twice referred to in that article ; and a dozen years later, in the course of the controversy with De Morgan, Sir William alluded to the article, as containing the germs of his discovery. We may imagine the avidity with which De Morgan, injuriously attacked, would have seized upon Mr. Bentham’s book if he had known of it. It is not so easy to understand how Mr. Bentham, although then absorbed in botanical researches, could have over- looked the controversy in the Athenaeum, or how, if he knew of it, he could have kept silence. It was only at the close of 1850 that Mr. Warlow sent from the coast of Wales a letter to the Athenaeum, in which he refers to Bentham’s book as one that had long before anticipated this interesting discovery. Although Hamilton himself never offered an explanation of his now unpleasant position (for the note obliquely referring to the matter in the second edition of his Discussions is not an explanation), Mr. Bain did (in the Athenaeum for Feb. 1, 1851); he immediately endeavoured to discredit the importance of Bentham’s work, and again in 1873 (Contemporary Review, Vol. xxi) in reply to Herbert Spencer’s reclamation of Bentham’s discovery. To this Stanley Jevons made reply in the same volume (pp. 821-824) ; and later in his Principles of Science (ii. p. 387) this competent and impartial judge, in speaking of the connexion of Bentham’s work “ with the great discovery of the Quantification of the Predicate ” adds, “ I must continue to hold that the principle of quantification is explicitly stated by Mr. Bentham ; and it must be regarded as a remarkable fact in the history of logic, that Hamilton, while vindicating in 1847 his own claims to originality and priority as against the scheme of De Morgan, should have overlooked the much earlier and more closely related discoveries of Bentham. It must be that Hamilton reviewed Bentham’s book without reading it through, or that its ideas did not at the time leave any conscious impression upon the reviewer’s mind, yet may have fructified afterwards.” ’ XIX George Bent ham , F.R.S. Society, the meetings of which, the anniversary dinners, and those of its club, he punctually attended. Soon after this Robert Brown proposed his name for election by the Royal Society, but withdrew it before the day fixed for election, to mark the dissatisfaction on the part of the scientific Fellows with the management of the Society, when a Royal Duke was made President. It was not until 1862 that he was again proposed and elected. In 1829, at the joint solicitation of his friend Mr. Joseph Sabine, the Hon. Secretary, and Dr. Lindley, Assistant Secretary, who were at issue as to the management of the Horticultural Society, he accepted the honorary secretaryship himself, and held it until 1840. On his entering office the Society was in a perilous position from debt and dissensions, from which, with Bindley ’s active co-operation, he rescued it. It was during his term of office that the celebrated Chiswick Horticultural fetes were inaugurated, which gave a new life to the science. At the first of these, held on April 3, 1832, seventeen hundred persons were present. It was during the same period that so many of our most popular garden-plants were introduced, especially from California, through collectors sent out by the Society (Douglas, Hartweg, and others). These plants were named, and many novelties amongst them described, by Bentham in the Society’s publications ; to which he also contributed a translation of Targioni-Tozetti’s ‘ His- torical Notes on Cultivated Plants,’ in which he added much valuable matter to the author’s work. At about this time Dr. Wallich returned from India with his enormous collection of Himalayan, Burmese, and Indian plants, destined to be named and distributed to the principal herbaria in Europe by the Honourable East India Company. In fur- therance of this great work Bentham offered his aid to Dr. Wallich, with whom he co-operated zealously for several successive years. Over and above his gratuitous labours as an assistant, he undertook the naming and distributing of the orders Euphorbiaceae and Gramineae, lithographing the tickets of the latter with his own hand. This marks an epoch in XX George Bentham, F.R.S. his career, for it was his introduction to a tropical flora (which he was never privileged to see), and afforded the chief materials for his first work on exotic plants, the ‘ Scrophu- larineae Indicae,’ published in 1835. This was followed, in 18 36, by the completion of his great work, ‘ Labiatarum genera et species, or a description of the genera and species of the order Labiatae, with their general history, characters, affinities, and geographical distribution,’ which gave him a position in the very foremost ranks of taxonomic botanists1. It was followed, in 1837, by his enumeration of plants collected in the Swan River district by Baron Hugel (Vienna, 1837), which is remarkable as showing the ease and rapidity with which he mastered a Flora totally different from those he had previously studied. From the summer of 1836 till the early part of 1837 he resided with his wife in Germany, visiting the principal Botanic Gardens and Herbaria, especially engaged in studying the order Leguminosae ; his account of these and of the botanists whom he met was communicated in letters to Sir W. Hooker, and published (anonymously) in the Companion to the Botanical Magazine (Vol. ii) and Journal of Botany (Vol. i). They are very interesting as contributions to the History of Botany in the first third of the century. During the winter at Vienna he published his masterly ‘ Commen- tationes de Leguminosarum generibus ’ in the ‘ Annalen des Wiener Museums ’ (Vol. ii). In 1846-7 he undertook, accompanied by his wife, an extended tour in Europe. Commencing with Hamburg, they visited Copenhagen, Stockholm, St. Petersburg, Moscow, Odessa, Constantinople, Trieste, Bologna, Florence, Leghorn, Naples, Rome, Palermo, and Geneva. What he saw in these towns and their environs of botanical, horticultural, and other 1 To give some idea of the thoroughness of Bentham’s methods, it is well to state, that in prosecution of this work he visited the following herbaria : — in 1830 Hamburg and Berlin; in 1831 Paris, Geneva, Avignon, and Montpellier ; in 1832 Hamburg again, Copenhagen, Leipzig, Dresden, Prague, Vienna, and Munich ; in 1833 Paris and Montpellier again; and in 1834 Bonn, Frankfort, Geneva again, Pavia and Turin (Labiat. Gen. and S., Pref. p. v). George Bentham , F.R.S. xxi interest was communicated in thirty-one anonymous letters to the Gardener’s Chronicle (Vols. 1846-7). Like those men- tioned above from Germany, they are of great value as con- tributions to the history of botany and horticulture during the period in which they were written. With the view of providing better accommodation for his library and herbarium, and devoting himself exclusively to science, Bentham removed in 1842 to Pontrilas, an Eliza- bethan manor house belonging to his brother-in-law, Colonel Scudamore. Here his chief occupation was providing material for the continuation of Auguste De Candolle’s ‘ Prodromus systematis naturalis Regni Vegetabilis,’ which had been undertaken by his (Bentham’s) intimate friend, Alphonse De Candolle. In this work he contributed the Ericeae, Polemoniaceae, Scrophularineae, Eriogoneae, and a greatly enlarged revision of the Labiatae, amounting in all to over 4,730 species. During the same interval he published the Botany of the Voyage of the Sulphur in the Malayan seas and Pacific ocean, a quarto work with 60 plates. Whilst resident at Pontrilas he also did his duty as Justice of the Peace for the county of Hereford with punctuality and efficiency. In 1854, finding that the cost of keeping up his library and herbarium threatened to exceed his income, he determined to offer these to the Government, with the stipulation that they should form part of the establishment of Kew1, he himself abandoning botany, and removing from Pontrilas to London. This munificent offer was of course gladly accepted by the Government, and the materials were placed in what is now called the Herbarium building of Kew, to be subsequently amalgamated with the richer collections of books and plants 1 Where there were at that time no other library or herbarium than the private ones of Sir W. Hooker, which he had been permitted to deposit in a house previously in the occupation of one of the Royal Family. The said house had, at the advice of Sir Joseph Banks, been purchased by George III for the very purpose which it now serves, and one room was actually shelved for the books. On the death of the king and his scientific counsellor, in the same year, the house was otherwise appropriated. xxii George Bent ham, F.RS \ of Sir W. Hooker. On the other hand, the idea of Bentham’s giving up botany was a shock to his scientific friends at home and abroad, and especially to the oldest and most intimate of them, Sir W. Hooker, who begged him to reconsider his intention, combated his own modest estimate of himself as a mere amateur systematise and pointed out to him that a residence in London offered the means of study at Kew, where a room, containing his own herbarium, should be devoted to his use, and where he would be in proximity to the garden, museum, and collections already at Kew. Fortunately Sir W. Hooker’s counsels, backed by those of other friends, especially Dr. Lindley, prevailed. In 1855 he took up his residence in London, for the first few years in Victoria Street, Westminster, and for the remainder of his life in 25 Wilton Place, between Hyde Park and Belgrave Square. From London he went to Kew daily (a few weeks of autumn holidays excepted) for five days a week, with perfect regularity, arriving at 10 a.m. and leaving at 4 p.m., devoting the evenings to writing out the notes of his day’s work, and never breaking the long fast between breakfast at 8 or 8.30 and dinner at 7.30 or 8. c With such methodical habits, with freedom from professional or administrative functions, which consume the time of most botanists, with steady devotion to his chosen work, and with nearly all authentic material and needful appliances at hand or within reach, it is not surprising that he should have undertaken and have so well accomplished such a vast amount of work, and he has the crowning merit and happy fortune of having completed all that he undertook ’ (A. Gray, Memorial). No sooner was Bentham settled within reach of Kew than he was induced by Sir W. Hooker to inaugurate a series of Colonial Floras, which had been planned by the former, and of which the first is that of Hong-Kong1. It was followed by the ‘Flora Australiensis,’ in seven volumes, which is the first flora of any large continental area that has ever been 1 The ‘Flora Hong-Kongensis,’ published in 1861, one vol. 8vo, pp. 455, contains 1,003 species. George Bentham , B.R.S . xxiii finished. It was commenced in 1861, and was concluded in 1870; it comprises about 7,000 species. He was aided in it by valuable notes and preliminary studies supplied by Baron Mueller, but every description, generic and specific, was strictly his own. As has been well said, ‘ it is a work which would alone found a reputation.’ But Bentham’s magnum opus is unquestionably the ‘ Genera Plantarum1,’ issued under the joint authorship of himself and the contributor of this memoir to the Annals ; but which, whether for the overwhelming share of the work which Bentham undertook, or for the aid he gave his partner in certain Orders elaborated by the latter, may justly be regarded as on the whole the product of one botanist. In the planning and execution of the work only two points were contested between us, whether his or my name should take precedence on the title-page, and whether in the headings of the pages the author’s name should be given with that of the Order described. On the first point my opinion prevailed, his on the second. The only other separate work published by Bentham during this period was an Illustrated Handbook of the British Flora, for the use of beginners and amateurs, including a series of wood-engravings, with dissections, by W. H. Fitch, F.L.S. The first edition of the Handbook appeared in 1858, and has been succeeded by five others. This work, on its appearance, was criticized on the false assumption that its author had no knowledge of plants in the field, supported by the fact that he took a much wider view of the variations under which species present themselves in nature, than do authors who have that knowledge. It was unknown to, or forgotten by, 1 Genera Plantarum ad Exemplaria, imprimis in Herbariis Kewensibus servata, definita, auctoribus G. Bentham et J. D. Hooker, 1862-1883. The last article which Bentham wrote was a communication to the Linnean Society (Proc. xx, 1883, p. 304): On the joint and separate work of the authors of Bentham and Hooker’s ‘ Genera Plantarum ’ ; where a full account is given of the part each author took. It may be mentioned that with the view of reducing the price of the work' to the public as far as possible, the expense of production was defrayed by the authors, Mr. Bentham guaranteeing that his fellow author should not lose. XXIV George Bent ham, F.R.S. his critics, that he had for half-a-century observed, collected, and preserved most of the species of the British Flora over a great part of Continental Europe, as well as in the British Isles, that his views were founded on wide experience, and that the results of them in terms of genera and species were drawn up from examination, in almost every case, of living examples. The Handbook was a great favourite of Mr. Dar- win, whose admiration of the masterly way in which the author dealt with the main features of the British' Flora, drew from him the exclamation, * Good Heavens ! to think of British botanists turning up their noses and saying that he knows nothing of British plants V Bentham’s labours at Kew on the above three works were interrupted by several serious demands on his time and energies, his response to which places in a striking light his disinterested devotion to the progress of science. Of these the most important was his acceptance of the Presidency of the Linnean Society. This he accepted in 1863, and threw himself into the duties of the office, which he dis- charged for eleven years, with energy, wisdom, and single- ness of purpose ; and, it should be recorded, with no small expenditure of his means. It made no difference in respect of the time devoted to his work at Kew ; for the one day of the week which he had reserved for his own affairs was thereafter devoted wholly, or in part, together with much of his evening hours, to the Society’s affairs. During the years in which he held office he took the chair at the evening and council meetings, with all but unbroken punctuality. On the transference of the Society’s library, collections, and portraits, to the apartments in Burlington House provided by the Government, he personally superintended the arrange- ments, classifying the books, and literally with his own hands placing them on the book-shelves ; and he himself indexed 1 Life and Letters, Vol. ii, p. 363. It must not be supposed that Bentham disparaged the labours of those who aimed at what he considered the multiplication of species. No naturalist was more appreciative of accurate work in this depart- ment of botany, and of its value. XXV George Bentham , F.R.S. the first ten volumes of the Transactions. He constituted himself the editor of most of the botanical papers published in the Transactions and Journal, in some cases earning the gratitude of the authors by rearranging their matter (with their approval), and himself rewriting their papers. His annual presidential addresses were remarkable for their wide range of knowledge, and those who knew him only as a systematist and descriptive botanist recognized with surprise the power of analysis and sound judgment which he displayed in these addresses, wherein he discussed evolu- tion in all its bearings, the writings of Haeckel, geographical distribution, the prospects of fossil botany, deep sea life, abiogenesis, methods of biological study, the histories and labours of the principal Natural History Societies, and periodicals of every civilized country on the globe. In respect of evolution, perhaps the most important of his addresses is that of 1863, dealing with discussions on the Origin of Species. Alluding to his own subsequent tardy adoption of the theory of the Survival of the Fittest, he says : ‘ I scarcely think that due allowance is made for those who, like myself, through a long course of study of the phenomena of organic life, had been led more or less to believe in the immutability of species within certain limits, and have now felt their theories rudely shaken by the new light opened on the field by Mr. Darwin, but who cannot surrender at discretion so long as many important outworks remained contestable.’ In correspondence with Mr. Darwin on some of these outworks, the latter in a letter dated June 19, 1863, alluding to the effect of the address as a whole, wrote 1 : ‘I verily believe that your address, written as it is, will do more to shake the unshaken and bring in those leaning to our side than anything written in favour of trans- mutation.’ It is interesting to find in later addresses, a frank acceptance of evolution, in such passages as those in which he recognizes ‘ the coexistence of indefinite permanency, and of gradual or rapid change in different races in the same 1 Life and Letters, Vol. iii, p. 26. XXVI George Bent ham, F.R.S. area, and under the same physical conditions ’ ; and, ‘ we must now test our species as well as genera or other groups, by such evidence as we can collect of affinity derived from consanguinity.’ In short, as with Lyell in the later editions of his famous Principles of Geology, when dealing with the history of life on the globe, Bentham had to underpin his edifices, and replace their old foundations with new. Happily in the cases of both philosophers, this was effected without injury to the superstructures. This brief notice of Bentham’s final adhesion to Darwin’s views may be supplemented by the following interesting extract from a letter he wrote to Francis Darwin1, May 2, 1882 (two years before his death). It says : — ‘ I have always been throughout one of his (Darwin’s) most sincere admirers, and fully adopted his theories and conclusions, notwith- standing the severe pain and disappointment they at first occasioned me. On the day that his celebrated paper was read at the Linnean Society, July 1, 1858, a long paper of mine had been chosen for reading, in which, in commenting on the British Flora, I had collected a number of observations and facts illustrating what I then believed to be a fixity of species, however difficult it might be to assign their limits, and showing a tendency of abnormal forms produced by cultivation or otherwise to withdraw within their original limits when left to themselves. Most fortunately my paper had to give way to Mr. Darwin’s, and when once that was read, I felt bound to defer mine for reconsideration ; I began to entertain doubts on the subject ; and on the appearance of the Origin of Species I was forced, however reluctantly, to give up my long cherished convictions, the results of much labour and study, and I cancelled all that part of my paper which urged original fixity.’ This paper of Bentham’s was never published. Of the many laborious tasks undertaken and gratuitously performed by Bentham, chiefly at Kew, the following deserve especial notice. First and greatest was the equipment of Life and Letters, Vol. ii, p. 293. XXV11 George Bent ham, F.R.S. the University of Cambridge with an authentically named consulting herbarium. This consisted for the most part of that of his friend, Dr. C. Leman, F. L.S., a zealous collector, especially by purchase, which he was disposed to leave by will to Bentham. The latter, on the other hand, urged its being left to Cambridge, of which Dr. Leman was a graduate in medicine. It was finally arranged between them, that on his friend's death the collections should be sent to Bentham, who should select from them any specimens which he might want for his own herbarium, whilst the remainder (the much larger portion), augmented by duplicates from Bentham’s herbarium, should go to Cambridge. Aided by a small grant from the University for the purchase of paper, for the expenses of mounting and poisoning the specimens, and for other contingencies, Bentham classified, named, had fastened down and enclosed in genus-covers, a consulting herbarium of 30,000 species. This great labour occupied more or less of ten years of his life. Other gratuitous tasks were the ticketing, and dividing into sets for sale, of the collections of Robert and Richard Schomburgk in Guiana and Brazil, and of Hartweg1 in British Columbia, California, and Mexico. A still greater service to science was his undertaking the distribution and sale of the magnificent collections of the distinguished traveller Richard Spruce in the Amazon region and Peru. These, amounting in all to 6,500 numbers, were sent to him as collected, to be arranged, named, and divided into twenty to thirty sets, for which he obtained subscribers in the principal public and private museums in Europe and America. He further collected the money due by the subscribers, transmitting it to Spruce, who depended on it for the prosecution of his thirteen years of exploration, thus saving to all parties the expenses of agency and commission. Bentham s last work was the ‘ Genera Plantarum,’ of 1 Of Hartweg’s plants he published a catalogue, with descriptions of new genera and species under the title of ‘ Plantae Hartwegianae.’ It enumerates about 2,000 species. xxviii George Bent ham, F.R.S. which the first part appeared in July 1862, the concluding in April 1883. The closing years of his life are feelingly described by Mr. Thiselton-Dyer in his Eulogium, in the following words : — ‘ In the latter years of his life Bentham was not less imbued with affection for his task, though the sense of the precariousness of life chiefly affected him with anxiety as to its completion. The flame of his intellectual powers never burnt more brightly, too brightly perhaps for a frame which slowly but perceptibly enfeebled. During the last years of what was a supreme effort, it was impossible not to feel a degree of awe for the intense devotion with which he pursued, without intermission, his self-imposed labour. Towards the last it appeared to one that by mere effort of will he actually sustained his bodily vitality. When the last revise of the last sheet 1 was returned to the printer, the stimulus was withdrawn. Nature, so long indulgent, would no longer be withstood. He came once or twice again to Kew, but found no task that he could settle to. At home he commenced a brief autobiography. The pen2 with which he had written his two greatest works broke in his hand in the middle of a page. He accepted the omen, laid aside the unfinished manuscript, and patiently awaited the not distant end.’ I cannot better conclude this attempt to convey an adequate idea of the value and amount of Bentham’s labours, than by citing a passage from his intimate friend’s, Dr. Asa Gray’s, Memorial, premising that the latter botanist most nearly approached him of all his scientific contem- poraries in the qualities he alludes to and the range of his work. He writes : — 6 It will have been seen that Mr. Bentham confined himself to the Phanerogams, to morphological, 1 This was the general index, which, as those of each successive part, he made himself, so scrupulous was he in his e fforts to avoid error, even in so mechanical an operation. 2 This was one of Mordan’s gold pens, which have I think iridium nibs. It ‘ wrote ’ not only the seven volumes of the ‘ Flora Australiensis,’ and the three of the ‘ Genera Plantarum,’ but a vast number of botanical papers and letters. The pen -holder is preserved at the Herbarium, Kew. XXIX George Bent ham, F.RS. taxonomical, and descriptive work, not paying attention to the Cryptogams below the Ferns, nor to vegetable anatomy, physiology or palaeontology. He was what may be called a botanist of the old school. Up to middle age, and beyond, he used rather to regard himself as an amateur, pursuing botany as an intellectual exercise. “ There are diversities of gifts,” perhaps no professional naturalist made more of his, certainly no one ever laboured more diligently, nor indeed more successfully over so wide a field, within these chosen lines. For extent and variety of good work accomplished, for an intuitive sense of method, for lucidity and accuracy, and for insight, George Bentham may fairly be compared with Linnaeus, De Candolle, and Robert Brown.’ This is a just tribute to his memory, to which I would add my own, that method, grasp of subject, and thoroughness, were his watchwords. It remains to allude to his personal characteristics. He was tall, of spare habit, with a slight stoop in his gait ; his features were strongly marked, his complexion rather dark, his hair black and eyebrows bushy. The likeness 1 accompanying this memoir is an excellent one. In dieting himself, he was extraordinarily abstemious, taking but two meals a day, and those most sparing. Though shy and reserved in manner, he was a most amiable, warm- hearted man, the kindest of help-mates, and the most disinterested of friends. As a companion or guest he was charming, high bred, and courteous, communicative of stores of anecdotes and reminiscences of the events he had wit- nessed, the interesting people he had known, and the places he had seen all over Europe. To which must be added his musical gift, which was at the service of whoever asked for it. To recognitions and honours he was indifferent. He gratefully received the Royal Medal from the Royal Society, awarded in 1859, and the Clarke Medal of the Royal Society 1 It is a reproduction of the portrait, painted in 1870 by Lowes Dickinson, in the possession of the Linnean Society of London, by whose kind permission it has been here reproduced. XXX George Bent ham, F.R.S. of New South Wales in 1878 ; but it was with great difficulty that he was prevailed upon to receive the Companionship of St. Michael and St. George, conferred on him by Her Majesty. He was a correspondent of the Institute of France, and member of the Academies of Science of Berlin, St. Petersburg, and America, and of many other scientific Societies that cultivate Natural History. Bentham died of old age at his house in Wilton Place, December 10, 1884, shortly after his eighty-fourth birthday, retaining his faculties to the last. His wife predeceased him by four years ; he never had any family. The bulk of his modest fortune went to his only relative, a great-niece residing in France, after liberal bequests to the Linnean Society and the scientific relief fund of the Royal Society, and of a considerable sum under trust to be expended in the interest of the herbarium at Kew, especially in continu- ing the publication of Hooker’s leones Plantarum, a work in which he took a great interest, having indeed provided the plates and letter-press of several volumes at his own expense. J. D. HOOKER. The Development of the Flower and Embryo in Lilaea subulata, H.B.K. BY D. H. CAMPBELL, Ph.D., Professor of Botany in the Stanford University, California , U.S.A. With Plates I-III. HE genus Lilaea is a peculiar monotypic one, the single X species, L . subulata, being widely distributed through the western part of the American continent. According to Hieronymus1, who has made the most careful study of the plant, its range is from Oregon, throughout the coast region of California and Mexico, into South America. In the latter it has been collected in Colombia, Chile, Argen- tina, and Uruguay, and has been found at various elevations, from sea-level to a height of 3,000 metres. There is much diversity of opinion as to the systematic position of the plant. In Engler and Prantl’s Natiirliche Pflanzenfamilien 2, it is classed with the Juncaginaceae ; but Schumann 3 is inclined to consider it as representing a special family, Lilaeaceae, proposed originally by Hieronymus, and 1 Engler and Prantl, Die Naturlichen Pflanzenfamilien, II, 1, p. 225. 2 Loc. cit. 3 Morphologische Studien, Leipzig, 1892, Heft I, p. 187. [Annals of Botany, Vol. XII. No. XLV. March, 1898.] 2 Campbell. — The Development of the Flower and this probably is more in accordance with the peculiar char- acters of the plant. Our knowledge of the morphology of the plant is mainly derived from the elaborate monograph of Hieronymus x, which proposed to give a very full account of the morphology, but was unfortunately left incomplete. Schumann1 2 has given some details as to the relation of the flower to the axis of the plant, but these simply confirm the earlier observations of Hieronymus. Beyond the work of these observers, so far as the writer knows, the plant has been described only in a superficial way. The writer has been engaged for some time upon a study of the flower and embryo in a number of the simpler Monocotyledons, and among the forms which have engaged his attention is Lilaea , which is common in the region about San Francisco Bay. The results of these studies are given in the following pages. The material upon which these were made was for the most part collected in the neighbourhood of Stanford Uni- versity, where the plant is a common one. In this neighbour- hood the plant grows either in shallow water, or completely exposed upon the mud. More rarely the plant is completely submerged except the flowers. It is an annual, germinating with the advent of the winter-rains, and flowering within a few weeks of germination. Flowers continue to form as long as the plant grows, but the plant is finally killed by the drying up of the mud in which it is rooted. The ripened fruits remain in the dried mud during the summer and autumn, and germinate as soon as the rains have soaked the ground. Most of the plants collected by the writer grew in the tenacious black clay (‘ adobe ’) characteristic of much of the land in the immediate vicinity of the University. The favourite localities for the plant were depressions in the fields 1 Monografia de Lilaea: Actas de la Academia Nacional de Ciencias en Cordoba. Buenos Aires, 1892. 2 Loc. cit. 3 Embryo in Lilaea subulata , H.B.K. where the water collected and formed shallow pools. These seemed to offer the most favourable conditions for the germination of the seeds. Of course the time of germination varies with the time of the rains, but in January, 1897, the plants were found in various stages of germination ; and in 1895, an unusually rainy winter, the plants were well in flower in February. The general aspect of the plant is curiously like Isoetes, the subulate leaves forming dense tufts from the short, bulb-like stem. The leaves are strictly two-ranked, and numerous white unbranched roots fasten the plant into the soft mud. The arrangement of the leaves and roots, as well as that of the flowers, has been given in detail by Hieronymus1, and will be discussed only briefly here. The flowers are ex- ceedingly simple in structure, but nevertheless show much variation, at least in the pistillate flowers which exhibit heterostylism in a very remarkable degree. After the young plant has produced several — usually from five to seven — leaves, the axis becomes transformed into a shaft bearing a spike of flowers; while in the axil of the last-formed leaf is formed a shoot which, after producing a single leaf, is transformed also into an inflorescence like that derived from the original stem-apex, and in the axil of the leaf is formed another shoot which behaves in the same way. Thus the branching of the plant is sympodial. In addition to the flowers borne upon the spike, there are usually found two close to the base of the shaft but not outgrowths of it. These basal flowers are always pistillate, and are distinguished by their extremely long styles, which may reach a length of several centimetres (14 c. in some cases, according to Hieronymus). The lowest flowers of the spike, where these are pistillate, are destitute of any subtending bract or leaf, but the other flowers are borne in the axil of a small bract. Of the pistillate flowers borne upon the spike, the lower ones are Monografia de Lilaea. 4 Campbell. — The Development of the Flower and intermediate in character between the very long-styled basal flowers and the very short-styled flowers near the apex, in which the stigma is nearly or quite sessile. In the axils of all the foliage-leaves are borne a number of the delicate scales ( squamulae intrav agin ales) so characteristic of most of the simple aquatic Monocotyledons. The histology of the vegetative organs of the plant has been worked out very carefully by Hieronymus : but un- fortunately only the roots and leaves are fully described, his memoir ending abruptly before the description of the stem was completed. Numerous beautiful and accurate figures of the histology are however given, which make it possible to follow out most of the structural details. In common with other aquatics, the leaves contain large lacunae, which here are irregularly arranged, and separated from one another by single layers of cells. There are numerous vascular bundles of the ordinary collateral type. The number of these bundles varies much with the size of the leaves. Hieronymus 1 states that in the most vigorous leaves he studied there were sometimes twenty-one. Of these one occupies a nearly median position, near the outer side of the leaf, and he considers this to represent a midrib. The epidermis of the leaves shows the usual elongated cells of similar leaves, and stomata are found, characterized by the presence of accessory cells not unlike those of many Grasses. The authors own observations were based mainly upon series of microtome-sections. The material was fixed with chromic acid, and, after thorough washing, stained in toto with Czokor’s alum-cochineal, and afterwards stained on the slide with alcoholic Bismark-brown. For the study of the embryo-sac, safranin and haematoxylin were also used to some extent as nuclear stains. 1 Monografi'a de Lilaea, p. 39. Embryo in Lilaea subnlata , H. B, K. 5 General Morphology of the Young Plant. In Fig. 2 is shown a median longitudinal section of a young plant with the apex prolonged into the first floral spike ( x ), and the two basal pistillate flowers (?) already formed. The stem itself is very short, and made up of nearly uniform parenchymatous tissue, which is traversed by the vascular bundles running into the leaves and the young inflorescence ; short bundles are also given off to the basal flowers. The arrangement of the parts in a more advanced stage can be seen in Fig. i, where several secondary in- florescences have formed. The number of roots is very large. According to Hiero- nymus, there is usually a secondary root on each side of the cotyledon, and the number formed from the later nodes varies with the size of the plant. No special study was made of the origin of these secondary roots, beyond noting that they form deep down within the tissue of the stem near the base of the leaves and close to a vascular bundle. They are consequently well developed before they finally break through the overlying tissues. The structure of the roots and the arrangement of the primary tissues are only briefly treated by Hieronymus 1, but he gives very accurate figures which make it perfectly clear that the arrangement of the tissues is that of the typical Monocotyledons, and these points have been confirmed by the writer. The very distinct plerome-cylinder, showing about five rows of cells in longitudinal section, is separated from the root-cap by a group of cells which are the common initials for epidermis and cortex. There is a separate group of initials for the root-cap. Hieronymus’ figures would indicate a single initial cell for the plerome, but he has not stated this in the text ; my own preparations make this seem not improbable, but it was not possible to decide the matter positively. The number of squamulae is variable, and they do not 1 Monografia de Lilaea, PL IV. 6 Campbell. — The Development of the Flower and differ in any marked degree from similar structures in other Monocotyledons. The cells have delicate walls, and densely granular protoplasm with a distinct nucleus. The whole aspect of the cells is that of secretory ones, but this point was not further investigated. Development of the Inflorescence. The primary inflorescence, as we have seen, is the direct prolongation of the original stem-apex of the young plant, and the later ones have a similar relation to the secondary shoots, which first produce a single leaf and then elongate at once to form the inflorescence. The apex gives rise to a stamen, while the other flowers — except the basal pistillate ones — are formed as lateral appendages. The lowermost flowers are usually female, with a moderately long style, and without any subtending bract ; but the other flowers have below them a small bract, which probably is the equivalent of the leaf at the base of the main shoots. The arrangement of the flowers upon the inflorescence shows a good deal of variation. The commonest arrangement upon well-developed spikes is that in which the lowest flowers are pistillate, the central ones hermaphrodite — or probably a secondary inflorescence made up of two flowers, male and female, — and those at the apex are staminate only. The writer has, however, seen cases where all the flowers were staminate, each stamen being subtended by a bract ; and Hieronymus 1 figures a specimen where only female flowers were developed. In the latter case, to judge from his figures, all the flowers were destitute of bracts. Where a single flower only is produced in the axil of the bract, the primordium or young shoot is transformed directly into the carpel or stamen, as the case may be ; but when the carpel and stamen are formed together, there is a division of the primordium into two equal parts, and this appears 1 Monografia de Lilaea, PI. I, Fig. 6. Embryo in Lilaea subulata , H.B.K. 7 to be a true dichotomy like that which occurs in Naias previous to the formation of the flower. In Fig. 3 is shown a median section through the floral complex from the middle region of a very young inflorescence. The arrangement of the organs suggests very strongly that found in Naias at a similar stage of development h This consists of two nearly equal superimposed protuberances, respectively J and ? , which apparently are formed by a true dichotomy of a common primordium. The latter has below it a very slightly projecting ridge, /, the rudiment of the subtending bract found in the older flowers. Of the two parts into which the primordium is divided, the upper is bluntly conical in form, the lower more pointed when seen in section, but really considerably flattened in the plane of division of the primor- dium. The upper prominence is the young carpel, the lower the stamen. Each of these structures, from analogy with the very similar ones in Naias and Z annichellia, may be con- sidered as representing shoots of equal morphological value, resulting from the dichotomy of a common primordium. If this view is accepted, both the stamen and carpel must be considered as truly axial structures. The young primordium is composed of nearly similar cells, which are arranged in the manner characteristic of the stem- apex of the Angiosperms. There is a definite dermatogen below which is a pretty clearly defined layer of periblem which separates the dermatogen from the central plerome- mass. This arrangement is especially clear in the staminal rudiment. The Male Flower. The development of the single stamen which constitutes the male flower, whether formed alone in the axil of the bract, or secondarily from the dichotomy of the axillary primordium, follows the same line of development. 1 Magnus, Beitrage zur Kenntniss der Gattung Naias; Berlin, 1870. Campbell, A Morphological Study of Naias and Zannichellia, Proc. California Academy of Sciences, 1897. 8 Campbell . — 77^ Development of the Flower and The very young stamen, cut parallel to its broad face (Fig. 5), is broadly conical, slightly constricted at the base, thus indicating the differentiation of anther and filament. In this stage it is composed of nearly isodiametric paren- chyma. but shows the definite divisions of the tissues into dermatogen, periblem, and plerome, referred to in connexion with the original primordium. The plerome in this view has a broadly conical form, with two lateral outgrowths marking the position of the future loculi. A cross-section of a similar stage shows that there are really four of these. A cross-section of a somewhat older stage is shown in Fig. 7. While there is in general the same arrangement of the tissues seen in the younger stamen, the loculi are more clearly defined. The two upon the outer side are somewhat broader than the others, and between them there is a small sterile lobe. Each loculus shows much the same arrangement of the tissues as the whole of the young staminal rudiment, and the origin of the sporogenous tissue is plainly traceable in all cases to the plerome. The latter is usually well defined, and forms a conical mass, usually three or four cells wide at the base, and narrowing above. The periblem, which at first is but one cell thick, later, by the formation of periclinal walls, becomes thicker, and finally the limits between it and the plerome are no longer distinguishable. In cross-section at this stage, there may be seen two vascular bundles near the base of the inner loculi. The exact origin of the sporogenous cells is difficult to trace, but there seems no question that they originate from the outer cells of the plerome, and that the archesporium is not of hypodermal origin as is usually the case. In this respect Lilaea agrees with Naias jlexilis , and probably also with Zannichellia , although in the latter the question is not quite clear 1. At first the young sporogenous cells are not readily distin- guishable from those surrounding them, and it is quite impossible to trace them back certainly to the division of 1 Campbell, 1. c., pp. 13 and 41. 9 Embryo in Lilaea subulata , H. B. K . a single primary archesporial cell ; indeed it is quite impro- bable that they all originate from a common mother-cell. As they become older, however, they become noticeably larger than their neighbours, and show the usual dense contents and large nuclei (Figs, 9, 10). However, even in these stages the transition from genuine sporogenous tissue to the tapetal cells is somewhat gradual. The latter are derived mainly from the periblem, but functionally, at least, some of the plerome-cells must be regarded as tapetal. As the stamen grows, the four masses of sporogenous tissue become very clearly defined. Each group of sporogenous cells is surrounded by about three, or in places four or five, layers of cells which separate it from the epidermis. The cells, which at first are much alike, later show considerable differentiation. Fig. io shows a longitudinal section of a loculus shortly before the separation of the sporogenous cells. The epidermal cells are now much elongated, but are very little deeper than they were in the very young stamen. Under the epidermis is a layer of somewhat swollen cells which, with the epidermis, persists as the wall of the mature loculus. Within this second layer is a third one, composed of very much compressed cells (c), which with a fourth layer ( d ) make up the tapetum. The layer d has larger cells which resemble the sporogenous cells in the character of their contents. The sporogenous cells ( sp .) have the usual characters of such cells. They are thin-walled, isodiametric, with abundant granular cytoplasm and large nuclei. The nucleoli are con- spicuous and the chromatin abundant. The nucleus, in material fixed with chromic acid, usually shows a conspicuous areola about the nucleolus. On the inner side of the loculus especially, the tapetal cells encroach upon the sporogenous area, and there are cells which are intermediate in character between the perfect sporogenous cells and those of the tapetum. These are probably to be considered as potentially sporogenous cells which do not, however, develop into spores, but become broken down and serve, like the true tapetal cells, ro Campbell. — The Development of the Flower and to nourish the developing spores. A similar disintegration of part of the sporogenous tissue has been observed by the writer in Z annichellia , and is very much like what takes place in Equisetum. The sporogenous cells after separation are imbedded in a nucleated mass of protoplasm derived from the tapetal cells and the imperfect sporogenous ones (Fig. n). As in most Monocotyledons the spores are of the bilateral type, i. e. there are two complete successive cell-divisions of the pollen- mother- cell. The spores do not long remain united in tetrads, but separate completely and assume an almost perfectly globular form. The young spore contains but a single nucleus, but there is later a division into two cells of very unequal size. The ripe spore (Figs. 12-14) shows a double wall, the outer one marked with fine reticulations, the inner one being the delicate endospore. The spore before maturity contains very little granular contents, but these increase rapidly as the spore ripens. The exact nature of the reserve substances in the ripe spore was not investigated. The structure of the anther-wall (Fig. 13) is of the usual type. The nuclei of the two cells in the germinating spore are very different in appearance. That of the large vegetative cell (Fig. 12, v) is large, with but little chromatin and a large nucleolus. The nucleus of the small generative or antheridial cell (g), on the other hand, is small, staining strongly and having an inconspicuous nucleolus. No indication of a further division of the antheridial cell, such as occurs in many Mono- cotyledons, was seen, although it is almost certain that this does occur in the pollen-tube after germination. The Female Flower. The homologies of the two pistillate flowers which usually occur near the base of the shaft of the inflorescence are not entirely clear, but they probably represent shoots of the same nature as the innovations which occur in the larger plants, in addition to the shoots formed in the axils of the leaves. Embryo in Lilaea subulata , H.B.K. n These basal flowers arise on either side of the floral axis (Fig. 2, ? ), and are very early recognizable. In these the young ovule is already evident as a slender prominence whose axial nature is unmistakable, and it represents, with little question, the end of the metamorphosed shoot which constitutes the flower. Sometimes, and perhaps always, the flower has at its base a small bract, in which respect it differs from the pistillate flowers of the spike, which are destitute of a similar subtending bract. The central ovular rudiment (Fig. 1 5, o) is enclosed by a cup-shaped envelope arising from the growth of the surrounding tissue, and this body, the carpel, which is formed precisely as in Naias , is probably to be considered as a foliar member. It is of equal height on all sides, and shows no evident dorsi-ventral structure. The pistillate flowers upon the spike are of two kinds, those which stand alone at the base of the spike, and those which are associated with the male flowers. The former are longer-styled, and are in this respect intermediate between the extremely long-styled ones and the upper short-styled ones. In the lower flowers of the spike, the whole primordium is transformed into the flower ; in the upper ones, as we have seen, there is first a dichotomy of the primordium, only one member of which forms the pistillate flower. In both cases, however, the flower is to be looked upon as a transformed shoot, whose apex develops into the ovule, while the carpel represents a foliar appendage of the floral axis. In a section of the young pistillate flower from the spike (Fig. 17), the ovule is not nearly so conspicuous as in the long-styled flowers. Here the young carpel is developed much more strongly upon the inner side, while upon the outer side it is not clearly distinguishable from the ovular rudiment, which is much less noticeable at this stage than it is in the long-styled flowers. The ovule is not so slender as that of the basal flowers, and more rounded at the end ; but by comparing it with a somewhat older one (Fig. 19), it is evident that here too the ovule is the transformed apex of the shoot. The young short-styled flower, at this stage, is 1 2 Campbell. — The Development of the Flower and very much like the corresponding stage in Zannichellia ; indeed the whole development of the flower is very much like that of the latter. In the long-styled flower, the slender ovular rudiment soon shows a broader and somewhat pointed form (Fig. 16). The pointed appearance is due to a stronger growth at this place, by which the original apex is forced over to one side, this being the first indication of the anatropous form of the older ovule. In the short-styled flower, the young ovule (Figs. 17, 19, o) is much blunter from the first, but here also the original apex is soon bent over by the excessive growth upon the outer side of the young ovule. In both forms of flowers, the growth of the carpel is rapid ; it soon grows up beyond the top of the ovule, and the margins, which at first are quite free, so that the carpel forms an open cup, approach and finally meet, the ovary-cavity thus becoming completely closed. The carpel is closely appressed to the ovule, and its upper part is extended into the tubular style, the final development of the latter varying much as we have already indicated. The young style is traversed by a narrow open canal, but this later becomes entirely closed by the cohesion of the cells lining it. In the long-styled flower, the style and stigma are perfectly symmetrical ; but in the short-styled one, the stronger development of the carpel upon the inner side of the young flower persists, so that the stigma is much more strongly developed upon this side (Fig. 25). As the ovule develops, the much stronger growth on one side forces the original apex over until it assumes nearly a horizontal position (Figs. 18, 20), and finally it becomes perfectly anatropous (Fig. 25). The first integument becomes evident at a very early period, and forms a shallow cup- shaped structure, more strongly developed on the upper side of the ovule (Figs. 18, 20, in.). Not long after this, the second (outer) integument is differentiated, but this is fully developed only upon the outer free side of the ovule (Fig. 21, in2); while upon the inner side, which is in contact with the funiculus, it is imperfect. The inner integument is Embryo in Lilaea subulata , H. B. K. 1 3 composed, for the most part, of two layers of cells and soon reaches to the top of the nucellus, and later closes over it to form the micropyle. The outer integument remains less developed than the inner one, and barely reaches to its level. The Style and Stigma. While there are some minor differences in the development of the style in the long- and short-styled flowers, the structure of the fully developed parts is essentially the same. At an early period in the growth of the flower, the free superficial cells of the upper margins of the carpel become enlarged into papillae (Fig. 25, which later reach a great size, and are distinguished also by very dense granular cytoplasm and large nuclei. When the pistil is mature, they form a dense tuft of conspicuous stigmatic hairs. The narrow canal, which is present in the very young style, soon becomes completely obliterated by the coalescence of the cells forming its walls. These cells, which are the con- tinuation of the epidermal cells which form the stigmatic papillae, much resemble the latter in the character of their contents, although the nuclei are smaller. There are, how- ever, transitional forms in the upper part of the style. Occasional indications of more than one nucleus in these cells were seen, but the matter was not further investigated. These cells together form a very distinct strand of con- ducting-tissue, which in cross-section is oval, and very clearly defined. The arrangement of the tissues in the style may be readily seen by comparing cross and longitudinal sections. The epidermis is composed of cells with a thickened outer wall, and within this are several layers of loose parenchyma. In the sections examined there were three vascular bundles, but whether this number is constant cannot now be stated. The bundles were of the usual collateral type with a few annular and spiral tracheids in the xylem. These bundles were separated from the central cylinder of conducting-tissue by 14 Campbell — The Development of the Flower and several layers of more compact parenchyma than that lying between the bundles and the epidermis. The ovule fills the ovarian cavity almost completely, and there is no development of conducting papillae such as occur in Naias , either at the top of the ovary where the conducting tissue of the style terminates, or at the base of the funiculus near the micropyle. This may be accounted for by the fact that a pollen-tube which has reached the ovary can hardly fail to reach the micropyle, as its course along the wall would almost certainly bring it to the opening of the ovule. The Embryo-Sac. Unfortunately, not a sufficient number of specimens of the earlier stages of the embryo-sac were found to make it certain just how uniform the course of development is. With few exceptions there was nothing to indicate any marked de- parture from the ordinary type. In the earliest stage in which the archesporial cells could be recognized with cer- tainty, there were two cells, evidently the product of the division of a primary hypodermal cell (Fig. 19). The outer of them was the larger, and is probably to be considered the real archesporium, and from it is apparently next cut off the primary tapetal cell (Fig. 18, t ). There is thus formed a row of three cells in the axis of the nucellus. The form of the primary tapetal cell was quite different in different specimens examined (Figs. 18, 20), but in all cases it undergoes repeated divisions so that the sporogenous cell becomes more deeply sunk in the nucellus (Fig. 22). The further history of the sporogenous cell must also be left somewhat incomplete, owing to the small number of satisfactory preparations of the next stage. It is extremely unlikely that the primary sporogenous cell ever develops at once into the embryo-sac, although such a form as that shown in Fig. 21 might possibly be so interpreted. In some- what later stages (Fig. 22) there were found two or three cells derived from transverse divisions of the primary sporo- x5 Embryo in Lilaea subulata , II. B. K. genous cell, one of which by its subsequent growth destroys the others, and becomes the embryo-sac. In the case figured the lower cell already shows signs of disintegration, while in the upper one it is difficult to say whether we have the young embryo-sac showing the first nuclear division, or what seems more likely, from a comparison with other more advanced stages, the second division of the sporogenous.cell, with a suppression of the cell-wall. In other cases where the young embryo-sac was found with two nuclei, it was much larger, and there were the remains of one, and in some cases of two sporogenous cells above it. Whether in any instances there are four complete sporogenous cells formed can only be determined by further investigations. After it is once formed, the growth of the embryo-sac proceeds rapidly and the other sporogenous cells are destroyed. The youngest stages at which the embryo-sac could certainly be identified, already showed two nuclei. The cytoplasm did not fill the cell, but there was a large ventral vacuole. Near each end was a conspicuous but not very large nucleus, surrounded by a mass of granular protoplasm. The actual divisions of these nuclei were not seen, but there is no reason to suppose that they differ from other similar ones. As usual the four nuclei derived from each occupy either end of the embryo-sac. Those of the upper end are perhaps a little larger than those of the antipodal region, but the difference is very slight, and perhaps not constant. The nuclei are usually distinct with a single large nucleolus. The granular cytoplasm is now confined to a very thin layer at the sides of the embryo-sac, but is more abundant at the ends where the nuclei are situated. The remains of the sister-cell (or cells ?) of the embryo-sac are still evident as a structureless mass lying above it (Figs. 24, 26). There now begins the differentiation of the antipodal cells and egg-apparatus. The former, which are later very con- spicuous, become invested with evident membranes, probably of cellulose, while the two synergidae and the egg soon become easily recognizable at the upper end. These are 1 6 Campbell. — The Development of the Flower and distinct and bounded by a well-marked protoplasmic mem- brane. The two polar nuclei move toward the centre of the sac, where they finally fuse. There is usually no appreciable difference in the size or structure of the two, in which respect Lilaea agrees with Naias and Z annichellia. In his recent paper on the development of Sagittaria variabilis , Schaffner states1 that the upper polar nucleus is by far the largest nucleus of the embryo-sac, and in the case figured in Fig. 24 the upper polar nucleus is slightly larger, but this is probably not constant. The embryo-sac increases rapidly in size after the formation of the egg-apparatus and antipodal cells, and with this increase in size there is a change in the various nuclei, as well as marked growth in both the antipodal cells and the egg- apparatus. The former increase very much in size, and are usually arranged in a manner which often strongly suggests the arrangement of the cells of the egg-apparatus. The uppermost cell projects strongly into the cavity of the embryo-sac, and its nucleus becomes decidedly larger than those of the two lower cells (Fig. 28). In all of the cells the protoplasm is very granular, and there are often aggrega- tions which stain strongly, and look almost like nuclei. The nuclei stain readily, and possess a single large nucleolus. In the upper part of the embryo-sac the three nuclei of the egg-apparatus, which at first were nearly alike, soon show a decided difference in appearance. The two nearest the apex of the sac, the nuclei of the synergidae, increase very little in size, but remain clearly defined. The lower, or egg- nucleus, however, becomes many times larger than at first. The polar nuclei also increase very much in size, and the endosperm-nucleus resulting from the fusion (Fig. 27, en.) is the largest of all the nuclei in the embryo-sac. The syner- gidae, have a large vacuole, but the upper part, including the nucleus, is filled with granular protoplasm. The egg-cell extends some distance below the synergidae, and its granular 1 Schaffner, The Life History of Sagittaria variabilis, Bot. Gazette, April, 1897. Embryo in Lilaea subulata , H. B. K. 17 protoplasm is more abundant, but contains numerous vacuoles. The nucleus is, as we have said, much larger than those of the synergidae, and contains more chromatin. The nucleolus is large, and in stained sections shows a vacuolated appearance. The large endosperm-nucleus is lenticular in form and has very little chromatin, but the very large nucleolus stains strongly and is much like that of the egg-cell. In the specimen figured there were two small bodies {cen.) lying near the nucleus which may possibly have been centrospheres ; but they were not very conspicuous, and it is doubtful, at least, whether they can really be considered as such. They were not seen associated with the other nuclei in the embryo- sac, so that it must be considered questionable whether they were really centrospheres or only granules belonging to the cytoplasm. While nearly all the embryo-sacs examined showed the normal structure just described, evidences of a deviation from this were seen in a few cases. The most marked was the one shown in Fig. 30. Unfortunately the structure of the lower part of the sac could not be clearly made out, as the series of sections was not complete. In the upper part of the sac, which was blunter than in the normal form, there was an irregular cellular mass, showing imperfect cell-walls. Eight nuclei could be certainly made out, but no trace of the definite egg-apparatus or other special structures usually found in the embryo-sac. Whether in the missing sections there were more nuclei than those seen, cannot be stated, but it is not impossible. Whether there were more than the normal number of nuclei, or not, the filling of the upper part of the sac with a cellular structure is a marked departure from the normal structure. Similar abnormal cases have been observed by the writer in Naias Jiexilis , Zannichellia palustris , and Sparganium eurycarpum ; but otherwise, exceptions of this kind seem to have escaped observation. C 1 8 Campbell,— The Development of the Flower and Pollination. The large stigmatic papillae and the conducting-tissue of the style, which is a continuation of the same epidermal tissue whose cells form the stigmatic papillae, have much the same appearance. The dense granular cytoplasm and large nuclei indicate that it is a secretory tissue, and with little question these cells are mainly concerned in forming the substances which serve to nourish the pollen-tube on its way to the ovary. The ripe pollen-spore is nearly globular, and its finely reticulate exospore is ruptured by the pollen-tube (Fig. 34). The latter grows along the side of the papilla to which it is closely appressed. The growth of the pollen-tube through the conducting-tissue is not easily followed, and no especial study of this point was made, nor was the actual penetration of the pollen-tube into the ovule studied. There was nothing, however, to indicate anything peculiar in this respect. One of the synergidae is probably destroyed by the growth of the pollen-tube, but one of these can often be detected even after the first division of the embryo (Fig. 29 b, s). The Embryo. The development of the embryo was studied by Hieronymus, apparently with a good deal of care, and he gives numerous accurate figures of the different stages of development in his monograph. Unfortunately there is no account given in the text, nor is any explanation appended to the plates. On the whole his figures correspond closely to my own preparations. After the egg becomes invested with its cellulose membrane as a result of fertilization, it elongates and divides, as most other Monocotyledons, by a transverse wall into two cells, a basal suspensor-cell, in contact with the upper end of the embryo-sac, and a terminal embryo-cell, which projects 19 Embryo in Lilaea subulata, H. B. K. into the cavity. The embryo-cell, as in Naias and Zanni- chellia , alone divides, the suspensor-cell remaining permanently undivided. Schaffner’s recent observations on Sagittaria and Alisma 1 show this to be the case in these forms also, although Hanstein 2 supposed that the primary suspensor-cell underwent subsequent divisions. The two cells arising from the first division of the egg in Lilaea are almost equal in size (Fig. 29), and much alike in the character of their cell-contents. With the elongation of the embryo, the free end becomes somewhat enlarged, and a transverse wall is formed in the embryo-cell (Fig. 31, b). The next division, at least in the few cases where this was seen, is in the terminal cell and is nearly vertical (Fig. 32). Following this is a transverse division in the middle cell, and next a further division, by a vertical wall, of each of the two terminal cells. The young embryo at this stage (Fig. 33) consists of six cells exclusive of the suspensor — the four terminal quadrants, only two of which show in the longi- tudinal section figured, and two cells between these and the suspensor. The latter has undergone no division, but there is a noticeable increase in the size of the nucleus which later becomes still more marked. In these early divisions of the embryo Lilaea agrees closely with Zannichellia and Naias , from which it differs mainly in the embryo being relatively shorter and the suspensor decidedly smaller. As in other similar embryos, the cells contain large vacuoles, the granular cytoplasm being principally confined to the neighbourhood of the nucleus and the periphery of the cell. There is no absolute uniformity in the next divisions. It not infrequently happens that the first wall in the terminal cell is oblique, and it is possible that sometimes the second vertical walls may be suppressed. This seems to have been the case in the embryo shown in Fig. 36. Probably the next wall to form, following the quadrant-division in the terminal cell, is in ordinary cases a median vertical wall in the cell 1 Schaffner, The Embryo-Sac of Alisma Plantago, Bot. Gazette, March, 1896. 2 Sachs, Text-book of Botany, 1882, p. 589. C 2 20 Campbell. — The Development of the Flower and immediately below (Fig. 33), and this is later followed by a quadrant-wall in the same cell. The cell next above the suspensor divides by a varying number of transverse walls into a short row of cells, the uppermost of which undergo quadrant-divisions by vertical walls, much like those described in the cell next the terminal quadrants. The lowermost of the series of cells above the suspensor remains for some time, at least, undivided, but may finally undergo division by vertical walls. The suspensor-cell may finally become much enlarged, but this takes place at a later period than is usual in Naias or Zannichellia ; as in them, the nucleus finally reaches a very large size (Fig. 39). The later divisions in the terminal cells vary a good deal, and the first division in each quadrant-cell may be either approximately vertical or longitudinal (Figs. 35, 37, 38). The second divisions are also more or less variable, but the result of the early divisions is usually the formation of a central group of four cells surrounded by a single layer of peripheral ones. The four inner cells form the primary group of plerome-cells for the cotyledon ; while the outer ones, by subsequent periclinal divisions, develop the dermatogen and periblem. While these divisions have been taking place in the terminal group of cells, a similar separation of a central group of plerome-cells also occurs in the segments lying just below the terminal group (Figs. 35, 37) ; and later, similar but less regular divisions occur in some or all of the basal segments. The number of these basal segments is usually three, and, as a rule, all of them sooner or later show vertical divisions ; but transverse divisions are only found at a later period of develop- ment, and then only sparingly, so that the limits of the original segments may be made out for a long time (Fig. 39). Finally the limits become indistinguishable, and it is difficult to tell exactly how far the primary segments contribute to the different members of the older embryo. In this respect Lilaea differs from Naias and Zannichellia , where the relation Embryo in Lilaea subulata , //. B. K. 21 of the members of the embryo to the young segments is quite evident. Owing to the formation of vertical walls in all of the lower segments, there is usually no secondary suspensor, such as usually occurs in other forms, but the enlarged vesicular primary suspensor-cell is in direct contact with the basal cells of the embryo (Fig. 43). The whole of the cotyledon, and possibly also the stem- apex, is derived from the terminal segment of the young embryo ; but owing to the late period at which the stem-apex is first recognizable, it is impossible to decide positively whether, as in Naias , it originates from the second segment, i. e. the cell immediately below the terminal quadrant-cells, or whether, as in Z annichellia^ it is the product of the terminal quadrants (Figs. 40, 41). The cell-divisions in the terminal region proceed with great rapidity, while in the basal segments growth proceeds much more slowly. The embryo in con- sequence becomes pear-shaped, the lower narrow portion being the product of the basal segments, while the enlarged upper part is derived entirely from the terminal segment. Fig. 39 shows a nearly median longitudinal section of an embryo where the primary tissue-systems are clearly defined, but the external differentiation is not yet indicated. The central plerome-strand is well marked, and shows in longi- tudinal section two rows of cells, separated from the dermatogen by the periblem, which is for the most part composed of two layers. In an older stage (Fig. 40) the plerome-cells have under- gone longitudinal divisions, and the periblem, especially in the cotyledon, has become very much more massive, owing to cell-divisions in all directions. Cross-sections of the young embryo are oval in outline (Fig. 44), and in the middle region {c) show a somewhat evident differentiation of the primary tissues, which is not clear in the basal region (a, b). The first evidence of external differentiation is a slight depression on one side of the embryo, near the base. This marks the position of the stem-apex : but it is difficult to 22 Campbell. — The Development of the Flower and tell whether it is derived from cells originating from the terminal segment of the embryo, or from the segment im- mediately below, as at this time the limits of the original segments can no longer be recognized with certainty. From its strictly lateral origin, however, and a comparison with other forms where the origin of the stem-apex is undoubtedly from the second segment, it is probable that in Lilaea the terminal segment gives rise to the cotyledon only, and that the stem is the product of the next segment. The rapid increase in size in the embryo which now takes place is mainly due to the growth of the cotyledon, while the lower part of the embryo remains short. In the cotyledon the plerome-strand is easily seen (Fig. 41), but it is much less evident in the basal part of the embryo. The epidermis is well defined in all parts of the embryo. In cross-sections of the older embryo made through the region of the stem-apex, the latter is seen lying in a shallow indentation formed by the base of the cotyledon, whose margins are beginning to form the sheath which later com- pletely encloses it. The point in which the embryo of Lilaea differs most markedly from that of other Monocotyledons which have been examined, is the origin of the primary root. This, instead of lying with the apex in direct contact with the suspensor, is decidedly lateral in position. In the earlier stages of the embryo it is impossible to make out clearly the relation of the tissues of the root to the other parts of the embryo. As soon as the root can be recognized as such, its axis is almost coincident with that of the stem (Figs. 41-43), and forms a marked angle with that of the cotyledon. It was not possible to decide positively, from the sections which were examined, what was the exact origin of the different root-tissues. The plerome is evidently continuous with the original axial plerome-cylinder of the young embryo, and is probably derived from the central cells of the second, and perhaps the third segment ; i. e. it is derived, in part at least, from the same segment as the stem. In the older embryo Embryo in Lilaea subulata , H.B.K . 23 (Fig. 43) the broad plerome-cylinder of the root is conspicuous, and the other tissues at the apex of the root begin to show the arrangement found in the fully developed apex. At this stage in the specimens examined, no single initial cell could be made out for the plerome, whose apex was covered by a single layer of periblem-cells which back of the apex become divided by periclinal walls. Outside were two layers of cells, apparently formed by a periclinal division of the original dermatogen, and these give rise to the root-cap. In the root of the full-grown embryo, the arrangement of the tissues is exactly like that of the roots of the adult plant. The plerome-cylinder shows about five rows of cells in longitudinal section, the central row being the largest, and probably later forming a central vessel. A single cell, some- what larger than its neighbours, was seen at the apex of the plerome, and may possibly be a single initial, but this point needs further examination. A single layer of cells lies between the plerome and the root-cap, and this group of initials, by the periclinal division of its segments, gives rise to the epidermis and cortex of the root. The inner of the two layers of cells derived from the primary dermatogen (see Fig. 43) becomes the calyptrogen, and from it arise all the later layers of the root-cap. As the embryo approaches maturity, a second leaf, much like the cotyledon, is developed opposite it, and later a third one at the base of the cotyledon1. In this condition the stem-axis has assumed a nearly vertical position, and with this displacement of the stem-apex there is a corresponding change in the position of the primary root, which comes to lie in nearly the same plane as the cotyledon. Hieronymus does not figure any intermediate conditions between quite young stages and the mature embryo, and to judge from his figures (nothing is given on the subject in the text of his memoir) he apparently supposed that the origin of the root was terminal, as in other Monocotyledons. 1 See Hieronymus, PI. IV, Fig. 42. 24 Campbell. — The Development of the Flower and The Endosperm. The formation of the endosperm begins shortly after fertili- zation. The primary endosperm-nucleus divides in the upper part of the embryo-sac, and the derivative nuclei distribute themselves in the layer of protoplasm lining its wall. The number of nuclei is large, and the protoplasmic layer becomes a good deal thickened, but no cell-divisions were seen in the endosperm. The nuclei are distinct, each with a single con- spicuous nucleolus (Fig. 44, a), and vary a good deal in size. The embryo finally fills the embryo-sac completely, and in the mature seed there is no trace of the endosperm. Summary. 1. The flowers of Lilaea are of strictly terminal origin, both anther and ovule being formed directly from the trans- formed apex of the shoot. 2. The sporogenous tissue of the stamen is not hypodermal in its origin, but arises from the plerome, as in Naias and Zannichellia. 3. The ripe pollen-spore has two cells. The generative nucleus remains undivided in the ripe spore. 4. The archesporium of the ovule is hypodermal, and a tapetal cell is cut off from it. 5. The primary sporogenous cell of the ovule divides usually into three, of which the middle one becomes the embryo-sac. 6. The embryo-sac usually develops in the manner typical of the Angiosperms, but there may be a suppression of a definite egg-apparatus, and a formation of cellular tissue in the upper part of the embryo-sac before fertilization. This is probably accompanied by an increase in the number of nuclei, such as has been observed in other low Mono- cotyledons. 7. The first division of the embryo is the typical one into Embryo in Lilaea subulcita , II. B. K. 25 two cells, a basal suspensor-cell which remains permanently undivided, and a terminal embryo-cell. 8. The cotyledon is derived entirely from the terminal one of the primary segments into which the embryo-cell first divides. 9. The stem probably always originates from the second embryonal segment, but this point is still somewhat doubtful. Its position is strongly lateral. 10. The root is of lateral origin, in this respect differing from other Monocotyledons which have been studied. 11. The root of the mature embryo is entirely like that of the older plant. Conclusions. From the study made of the development of the flower and embryo of Lilaea , it is clear that it shows resemblances to the other low Monocotyledons which have been studied. The apical origin of the sporangia is very much like that of Naias and Zannichellia , and it is quite probable that this will be found to be the case in other low types, such as Sparganium and the Potamogetonaceae ; but further investi- gations are necessary to determine this. The origin of the stamen and carpel in cases where they occur together, from the dichotomy of a common primordium, makes it probable that the resulting complex should be con- sidered as a secondary inflorescence composed of two flowers, rather than as an hermaphrodite flower. While the development of the embryo-sac is normally that of the typical Angiosperms, the occurrence of exceptional cases with a probable multiplication of the nuclei and the development of cellular tissue before fertilization is signifi- cant, especially in view of the similar phenomena in Naias , Z annichellia, and Sparganium , and suggests a possible case of reversion to a more primitive condition. The development of the embryo itself is most remarkable for the peculiar lateral origin of the root, which is quite 26 Campbell. — The Development of the Flower and different from that in the typical Monocotyledons. Just what the significance of this is, is hard to determine. This lateral position is rather suggestive of the root in Isoetes , and possibly the basal segments of the embryo with the suspensor might be interpreted as equivalent to the foot in the embryo of the Pteridophytes. A study of the embryo in other simple Mono- cotyledons may yield some further information upon this point. While it must be admitted that these investigations do not throw much light upon the question of the origin of the simpler Monocotyledons from pteridophytic ancestors, such as Isoetes , it may also be said that there is no further evidence for the view commonly held that they are degenerate forms, descended from more specialized ancestors. There is certainly no evidence that the flowers are derived from any type found among the higher Monocotyledons, and the writer is strongly inclined to believe that the simplicity of the flowers is really primitive. It is hoped that further investigations, which it is proposed to make, may possibly help to elucidate this very interesting subject. Embryo in Lilaea subulata, H. B. K. 27 EXPLANATION OF FIGURES IN PLATES I, II, AND III. Illustrating Professor Campbell’s Paper on Lilaea subulata. PLATE I. Fig. 1. Median longitudinal section of a young plant; x 40; 1,1, leaves; 9, female; 5, male flowers; sq., squamulae intravaginales ; i, lacunae in the leaves. Fig. 2. A similar section of a younger plant with the primary axis transformed into the first inflorescence, x ; lettering as in Fig. 1. Fig. 3. Median section through a very young floral complex from the middle region of the inflorescence ; x 600 (about) ; /, the subtending bract ; £, male ; 9, female flower. Fig. 4. A similar section of an older floral complex ; x 100. Fig. 5. Longitudinal section of the young stamen, showing the arrangement of the primary tissues ; x 350. Fig. 6. A similar section of an older stamen; x 100; l, the subtending bract. Fig. 7. Cross-section of a young stamen, showing the four loculi ; s, sterile lobe ; x 350- Fig. 8. A single loculus from another of about the same age as that shown in Fig. 6 ; x 600. The limits of the plerome are indicated by a heavy line. Fig. 9. Cross-section of a loculus showing the young sporogenous tissue; x6oo. The sporogenous cells have the nuclei shown. Fig. 10. Longitudinal section of the anther shortly before the isolation of the sporogenous cells ; x 600 ; c, d, the outer tapetal cells ; t , the cells of the inner tapetum. Fig. 11. Young spore-tetrad imbedded in the nucleated protoplasm derived from the disintegrated tapetal and sterile sporogenous cells. Fig. 12. Section of a nearly ripe pollen-spore, with the generative cell, g, and the nucleus, v, of the large vegetative cell ; x 600. Fig. 13. Cross-section of the wall of the ripe anther ; x 600. Fig. 14. Germinating pollen-spore upon one of the stigmatic papillae, p ; x 600. Fig. 15. Longitudinal section of a very young basal (long-styled) female flower ; x 350 ; 0, the ovule. Fig. 16. The ovule from a somewhat older flower; v, the apex; x 600. Fig. 17. Longitudinal section through a very young short-styled flower; car. the carpel ; 0 , the ovule ; x 600. Fig. 18. Longitudinal section of an ovule after the differentiation of the first integument, in. ; t., the primary tapetal cell ; x 600. PLATE II. Fig. 19. Longitudinal section of young short-styled flower; x 600; 0, ovule; car., carpel. Fig. 20. Longitudinal section of the young ovule, showing the beginning of the first integument, in. ; x 600. 28 Campbell . — - Development in Lilaea subulata. Fig. 21. An older ovule with the tapetum (/.) already divided; in1, first, in 2, second integument ; x 600. Fig. 22. A still older ovule. The lower cell of the axial row is becoming disorganized, and the one above it has two nuclei, but no division-wall has formed between the latter. Fig. 23. The nucellus of an older ovule, showing the young embryo-sac with two nuclei ; the section is cut somewhat obliquely ; m, the remains of the upper archesporial cell. Fig. 24. An older embryo-sac with eight nuclei; p, pl, the polar nuclei. The egg-nucleus lies immediately below the upper polar nucleus. Fig. 25. Longitudinal section of a flower with style of medium length, showing the stigmatic papillae, si. ; x 100. Fig. 26. Embryo-sac with egg-apparatus and antipodal cells fully formed, but the polar nuclei not yet united ; only two antipodal cells show in the section ; X about 600. Fig. 27. Upper part of a fully developed embryo-sac ; en., the endosperm-nucleus ; x 650. Fig. 28. The antipodal end of the fully developed embryo-sac ; x 650. Fig. 29. Two sections of an embryo-sac with a two-celled embryo ; s, the remains of one of the synergidae ; x 300. Fig. 30. Upper part of an abnormal embryo-sac. There were eight nuclei, two of which do not show in this section. Figs. 31-35. Successive stages in the development of the embryo, in longitudinal section ; x 600. The order of the transverse divisions is indicated by the lettering. The embryo shown in Fig. 33 had the terminal cell divided into four. Fig. 36. A young embryo in which the second wall (2) in the terminal cell was oblique, instead of being formed at right angles to the first one. Figs. 37, 38. Two sections of an older embryo. PLATE III. Fig. 39. A somewhat advanced embryo, seen in median section, showing the enlarged suspensor-cell with its nucleus. The limits of the first transverse walls are still visible. Fig. 40. A similar section of an older embryo, showing the first trace of the Stem-apex, si., and the root, r. Fig. 41. A still older embryo, showing the strongly lateral position of the primary root, r. ; si., stem-apex ; cot., cotyledon. Fig. 42. Section of an embryo from a nearly ripe seed ; x 100. Fig. 43. The basal part of a similar embryo ; x 350. Fig. 44. Three transverse sections of a young embryo ; x 350 \ a is a section just above the suspensor, and shows several of the endosperm-nuclei. I D.H.Camp’bell, del. CAMPBELL. L I L A E A SUBULATA. VOL.X/I; PL./ Vol.Xf Pt.I. Jnnxxls of Botany D.H.Camp'bell, del. CAMPBELL. — L1LAEA SUBULATA University Press, Oxford. ftrmals of Botany irv 22. 23. D-H. Campbell, del. CAMPBELL. — Ll LAE A SU BU LATA. Voi.xn,PUi. University Press, Oxford. 34. Yol.XII,PUI Asuvals of Botany University Press, Oxford. D.H.Cam-p'bell.del. SIBULATA CAMPBELL Voi.xn,pi.ui. Annals of Botany i V, i iM 9 D.H. Campbell, del. University Press, Oxford. CAMPBELL.— L1LAEA S U B U LATA . Observations on the Conjugatae. BY W. WEST, F.L.S., & G. S. WEST, A.R.C.S. With Plates IV and V. DURING a prolonged study of Freshwater Algae from all parts of the world, many more or less interesting observations concerning the Conjugatae have accumulated : in this paper we propose to set forth some of them, together with certain conclusions derived therefrom. This group of Algae has been extensively studied by many previous botanists, amongst whom particular mention may be made of De Bary, Wittrock, Nordstedt, Lagerheim, Klebs, Bennett, and others; and we have here attempted to correlate with each other and with our own observations, a few of the facts described by these several observers, and from this to ascertain, as nearly as possible, the relationship existing between the various members of the group. We classify the Conjugatae into the three following families : Zygnemaceae, Temnogametaceae, and Desmidiaceae, and in this paper we think it advisable to deal with them separately. Many authors regard the Mesocarpeae as a separate family, owing to the peculiar formation of the spores ; but we think [Annals of Botany, Vol. XII. No. XLV. March, 1898.] 30 West & West . — Observations on the Conjugatae. it is better regarded as a sub-family of the Zygnemaceae, as Pyxispora has the same method of formation of its spores, although the chromatophores are similar to those of Zygnema . These plants may occur as solitary cells, or they may be filaments which at some or all periods of their existence more or less easily dissociate into the separate cells of which they are composed. The genus Gonatozygon may be taken as an illustration ; it is sometimes found in long filaments of about thirty or more cells ; but on being subjected to the least dis- turbing influences these filaments break up, and in some species of the genus the filamentous condition is rarely attained. There is a tendency in many of the small species of Cos- marium , very noticeable in C. moniliforme and C. Regnellii , to assume a filamentous condition, and this may have induced Rabenhorst 1 to place C. pygmaeum under Sphaerozosma. We have also noticed this tendency in Euastrum binale (cf. Fig. 38). In Micrasterias foliacea , a representative of a genus the species of which normally occur as solitary plants, this filamentous condition has been attained by a remarkable degree of specialization of the polar lobes of the semi-cells, which possess an arrangement of apical teeth which interlock so firmly with those of the adjoining cell, that the connexion is too rigid to allow of hardly any flexibility in the filaments. A filamentous condition of the genus Mesotaenium (which is generally unicellular) is found in the Arctic plant named by Berggren Ancylonema N ordenskioldii. The filaments of Hyalo - theca dissociate into separate cells just prior to conjugation, and the dissociated cells remain imbedded in a mucus derived from that which surrounded the original filaments. Conjuga- tion is very soon general throughout the mass, as can readily be seen in a conjugating example of H. dissiliens. The frag- menting of the old filaments into individual cells is well known as a method of reproduction in some genera of Zygnemaceae. It can therefore be considered that a strictly filamentous condition is of no essential importance to the life of the Conjugatae. 1 Flor. Europ. Algar. Ill, p. 150. West & West . — Observations on the Conjugatae. 31 All Conjugates are surrounded by a definite mucilaginous envelope. In the great majority this covering is very thin, but in others it is profusely developed, e. g. Zygnema ano- mcthim , Hyalotheca mucosa , Staurastrum tumidum , 5. longi - spinum , &c. ; and even in those species in which it is normally almost absent it is occasionally developed to a large extent. We have seen very extensive mucilaginous envelopes round Closterium Lumda , Penium Libellula , and Cosmarium ovale , species which are normally destitute of such extensive invest- ments. No doubt this mucus serves in many cases as a means of attachment, as we have seen as many as a dozen specimens of Staurastrum tumidum attached to one leaf of Utricularia minor , and in the case of many of the species which occur on dripping rocks, this mucus is absolutely necessary for the purpose of attachment ; but an equally important use is probably that of protection from epiphytes and parasites (Chytridiaceae, &c.). With regard to the nature of this gelatinous investment, it must be considered either as a secre- tion or a mucous condition of the outer layers of the cell- membranes. Klebs regards it as quite independent of the substance of the cell-wall ; and as the cell-membrane of most Desmids is perforated by a large number of minute pores — excessively minute in some and not visible although perhaps present in others — one would be inclined to regard this mucous envelope as a secretion. But in some species of Zygnema it seems to us to be partly if not entirely due to the diffluent outer layers of the cell-wall, and this may be also true for some Desmids. In an almost pure gelatinous gathering of Cosmarium cymatopleurum , var. tyrolicum , many of the specimens were seen casting the outer coats of the cell- membrane, and in some cases many such successive coats could be seen round each individual gradually fading into the mass of jelly in which the plants were imbedded. This gathering was from the vertical face of a dripping rock. We have also noticed Cosmarium pyramidatum , when imbedded in a gelatinous mass of Desmids, casting its outer cell- membrane in a similar manner. 32 West & West . — Observations on the Conjugatae . That this mucus is often of a tough nature is proved by a consideration of the genus Spondylosium , in which the individuals are united into long filaments by a layer of mucus between the apposed ends of the cells ; and that the connexion is by no means a weak one is shown when the filament is fractured, the cells more often breaking across the isthmus than coming apart at the apical attachment. Organs for attacimjent are occasionally developed in young plants of Spirogyrax (Fig. 20) and Mougeotia (Fig. 16), but have not been noticed in any other genus of Conjugatae; they are special outgrowths (simple or branched) at the base of the shoot, and are homologous to those organs of attach- ments found amongst other Algae and usually termed rhizoids , but more recently known as haptera. We have also noticed them to be developed in Spirogyra , as a result of the modi- fication of a conjugating-tube protruded by a cell some distance removed from those cells of the filament engaged in conjuga- tion (Fig. 21). Branching amongst members of the Conjugatae is ab- normal and of somewhat rare occurrence. When present in the Zygnemaceae it is generally limited to lateral out- growths consisting of a few cells ; we have only noticed it in the genera Zygnema1 2 and Mougeotia (Figs. 17-19). In the Desmidieae one semi-cell occasionally undergoes a partial lateral 3 or dichotomous 4 branching (Fig. 40). The apical cells of filamentous Zygnemaceae are generally rounded at the free end, but they often become elongate and irregular. The filamentous forms consist of but one series of cells, longitudinal septa rarely making their appearance ; they have 1 Borge, Ueber die Rhizoidenbildung einig. fadenform. Chloroph., Upsala, Nya Tidnings Aktieb., Tr. 1894. Wolle, Freshw. Alg. U. S., PI. CXLII, f. 7, 8. 2 Cf. Zygnema fiachydermum, West, Alg. from W. Indies, Jonrn. Linn. Soc. Bot., Vol. xxx, PI. XIII, Figs. 12-15. 3 Reinsch, Contrib. Alg. et Fung., T. xviii, f. 12 and 15. 4 Jacobsen in Journ. de Botanique, Copenhague, 1874, t. viii, f. 31. West & West. — Observations on the Conjugatae. 33 been noticed in an incomplete form in Zygnema pachydermum , var. conferv aides1. With regard to the effect of temperature on the Conjugatae, a paper by Alfred J. Ewart, entitled ‘ On Assimilatory Inhibi- tion in Plants,’ has recently appeared 2, in which the author states (p. 395) that ‘ freshwater Algae . . . are not very resistant to cold, all those examined being killed by being frozen. This statement we cannot agree with, as we have found them to be very resistant to cold, and as a large number of the plants belonging to the Alpine Algal flora are Conjugates we illustrate the matter by a few examples. We have melted out of the ice from Mitcham Common, Surrey, excellent examples of Spirogyra catcieniformis in a state of conjugation, the vitality of which was in no way impaired. From Frizinghall, W. Yorks., we have also melted out of the ice hundreds of specimens of Closterium Leibleinii , which subsequently remained in a perfectly healthy and normally active condition (moving to that side of a vessel exposed to most light, just as we find all other species of Desmids to act), and in each of these cases the specimens examined had been frozen for over fourteen days. These facts alone disprove the generality of Mr. Ewart’s statement ; but let us now consider some still more convincing ones. Many of the upland tarns of Yorkshire, the Lake District, the Scotch Highlands, and other places, are situated at altitudes of over 2,000 feet, some of them being much higher, and the water in them is of a relatively low temperature even in summer. For many months in the winter these tarns are frozen, and the small ones often buried in deep snow drifts, although by the middle of summer they are fairly crowded with filamentous Algae, of which the most abundant forms are small species of Mougeotia. We have never yet seen these species of Mougeotia in conjugation from these altitudes3, 1 Cf. West, l.c., PI. XIV, Fig. 5. 2 Joum. Linn. Soc. Bot., Vol. xxxi, 1896, No. 217. 3 We have examined sterile species of this genus, obtained at 3,000 feet in the Scotch Highlands and 6,500 feet in Switzerland. D 34 West & West. — Observations on the Conjugatae. and in ordinary seasons in all probability they never do conjugate. How, then, are they preserved throughout the winter? It must be by means of the survival of some of the plants (without the formation of spores) through the prolonged freezing they have to undergo, which is followed by their active division in the spring. We can also mention instances of this in the Desmidieae. There is a small peaty ditch in Eldwick, on the edge of Rombald’s Moor, W. Yorks., in which count- less numbers of Micrasterias denticulata have occurred in a perfectly pure state for very many years. We have examined this ditch carefully at all times of the year, and always find some specimens of this Desmid, even when it is frozen ; but never once have we come across a single zygo- spore from this locality, although constant search for them has been made. Here, then, the perpetuation of the species must be dependent upon the survival of some of the ordinary vegetative plants through the winter ; and we may mention that the locality is a bleak one, its altitude being near 1,000 feet, and the water is generally frozen for some weeks during the winter. It is also worthy of note that in all the contributions to the Algal Flora of the Arctic regions yet published, the occur- rence of the zygospores of Desmids has seldom been mentioned, though many species are recorded from places noted for their intense cold in winter, for instance, Greenland, Spitzbergen, Nova Zembla, and northern Siberia. We have a species of Oscillatoria from a valley in the Davos Platz district in Switzerland, collected by Mr. A. Howard in August, 1897, from a stream at 8,000 ft. elevation with the temperature of the water at 5°C. This is the summer condition, and the winter one may be easily imagined ; we must therefore reject Mr. Ewart’s statement that ‘ owing to their slight powers of resistance to cold, the temperatures to which they can be exposed without being permanently injured are necessarily relatively high.’ Stationary masses of water, such as pools and small lakes, even at this altitude, attain during summer a comparatively warmer temperature than the streams ; a lake West & West. — Observations on the Conjugatae. 35 close to the above-mentioned locality, and at the same alti- tude, had a water-temperature of 20°C. In this lake fine examples of Staurastrum Meriani occurred, and it is evident that they must be capable of withstanding frost for a few months during the winter. We have found specimens of Closterium striolatum and Cylindrocystis Brebissonii 1 in mate- rial collected on the top of Green Hill, Clova Mts., at 2,700 ft., from water which was derived from melting snow close by, and which could not be more than i° or 2°C. These speci- mens were collected by Mr. J. H. Burkill in May, 1897. We also point out the four following papers dealing entirely with snow-floras : — S. Berggren : Alger fran Gronlands inlandis (Ofvers. K. Vet.-Akad. Forh. 1871, No. 2). V. B. Wittrock : Om snons och isens flora, sarskeldt i de arktiska trakterna (A. E. Nordenskiold, Studier och forskningar foranleda af mina resor i hoga norden), Stockholm, 1883. G. Lagerheim : Bidrag till kannedomen om snofloran i Lulea Lappmark(BotaniskerNotiser, 1883, Heft 6), Lund, ] 883. G. Lagerheim : Die Schneeflora des Pichineha, ein Beitrag zur Kenntniss der Nivalen Algen und Pilze (Bericht. d. Deutsch. Botan. Gesellsch., Jahrg. 1892, Bd. x, Heft 8), Berlin, 1892. In the same paper by Mr. Ewart we also find (p. 439) the following statement : ‘ It is well known that prolonged ex- posure to direct sunlight is fatal to . . . many Algae.’ From our own experience we should at once say that nothing could be more beneficial to Freshwater Algae than prolonged exposure to direct sunlight, provided they remain under natural conditions. Round the margins of the two ponds on Frensham Common, Surrey, there is a belt of very shallow water, which is the home of large numbers of Algae, and these plants on bright 1 This species occurs in pure gelatinous masses (during early spring before Easter) on the peat at the extreme summit (2,346 feet) of Great Shunnor Fell in N. Yorks. D 2 36 West & West,— Observations on the Conjugatae. days are not unfrequently exposed to direct sunlight from almost the rising to the setting of the sun. We have noticed this shallow water become quite warm. What is the effect of this prolonged exposure to sunlight and the increase in the temperature of the water ? It is certainly not a detrimental one, because there is an acceleration in the growth of the lower green and blue-green Algae1, and the Conjugatae form zygospores much more abundantly than they otherwise would do : we could multiply instances indefinitely, but the following one will suffice. From Vehar Lake, Parel, Bombay, we have examined the finest specimens of Clathrocystis aeruginosa we have yet seen, and these are exposed to direct sunlight every day for weeks ; moreover the atmospheric (shade) temperature was 96° Fahr., and that of the water 87° Fahr. The material was collected for us in 1895 by Mr. S. Tomlinson, C.E., the Government Engineer to the Waterworks. Yet Mr. Ewart would inform us that prolonged exposure to direct sunlight is fatal ! It is so (as we well know) in the small vessels of the laboratory, but not in nature 2. There are four methods of reproduction in the Conjugatae : by fragmentation of the filaments (asexual) ; rarely in some genera by resting-cells or cysts (asexual) ; by conjugation with formation of zygospores or carpospores ; and by aplano- spores (asexual). Temperature and climatic conditions affect reproduction only so far as to promote or prevent it ; they have little effect on the method, although an increase of temperature considerably helps conjugation, and so far as we have observed, a higher altitude (which is usually accompanied by a lower temperature) favours the formation of £ cysts.’ During conjugation the activity of the filament is increased ; even those cells which take no part in it show greater vigour. 1 Specially noticeable were Clathrocystis aeruginosa and Crucigenia rectangu- laris , the latter with single families of 1 28 cells, the normal number being 16 or 32. 2 The reader should consult the excellent work by Klebs entitled ‘ Bedingungen der Fortpflanzung bei einigen Algen und Pilzen,’ chapter on Conjugatae, in which he shows (among other things) that they bear intense light very well, and that bright light is necessary for conjugation. West & West, — Observations on the Conjugatae, 37 We have often noticed these cells begin to divide actively and ultimately produce new filaments (Figs. 62 and 65). That their activity is increased is also proved by the extraordinary development from these cells of swellings and processes, which so often occurs as an accompaniment to conjugation (cf. Fig. 58). Moreover, as previously mentioned, some cells by reason of this activity are induced to put out conjugating- tubes, which, not meeting with others, and not being able to fulfil their proper function, ultimately become rhizoids or organs of attachment. FAM. I . — ZYGNEM ACEAE. Sub-fam. 1. — Mesocarpeae. This sub-family includes two genera, Gonatonema , com- prising but four species, and Mougeotia , comprising upwards of thirty species. 1. Mougeotia. This genus now includes, and we think quite correctly, the genera Mesocarpus, Craterospermum, Plagio- spermum and Staurospermum , all those characters regarded in the past as generic distinctions having been found by Wittrock 1 to be present in one species (M. calcarea). Many other observations also tend to prove the identity of these so-called genera. The conformation of the young zygospores of M. uberosperma (not taking into consideration the four outer processes) is decidedly that of a Staurospermum , whereas the adult zygospores are almost globose (cf. Figs. 42 and 43). In this genus an axile plate-like chromatophore is present in each cell ; and so far as our observations go, there is but one exception to this, M. capucina having an axile sub- irregular rod of chlorophyll connected to the lining primordial utricle by fine colourless threads of protoplasm. The rest of the cell-cavity between these meshes of protoplasm is filled with purple-coloured cell-sap 2 ; the nucleus also stands out 1 V. B. Wittrock, Om Gott. och 01. Sotv. Alg., Bih. till K. Sv. Vet.-Akad. Handl., Bd. i, No. i, Stockholm, 1872. 2 Cf. remarks on this species by Lagerheim, Ueber das Phy coporphyrin, Vidensk.-Selsk. Skrift., I. Mathem.-natur. Kl., Kristiania, 1895, No. 5, p. 6 (Sep.). 38 West & West. — Observations on the Conjugatae . very plainly, being opposed to the axile rod of chlorophyll towards the centre of the cell. The method of conjugation and the formation of the rudi- mentary sporocarp are very well known, but we wish to point out a few irregularities which are occasionally met with. It is no uncommon thing for conjugation to take place through the end of one of the cells, the latter cell forming no con- jugating-tube ; we have observed this in M. parvula (Fig. 44) and M. nummuloides. We have also seen a hybrid example (Fig. 55), corresponding to Spirogyra maxima , var. inaequalis and others (Figs. 70 and 71), in which conjugation has taken place between two species of different thickness. Fig. 45 is an example of M. recurva in which three cells were con- jugating to form one spore (analogous to other cases in Spirogyra and Zygnema ; cf. Fig. 66). There is a most noticeable disparity in size between the carpospores of different species in relation to the size of the sterile cells of the sporocarp (cf. Figs. 47 and 46 of M. nummu- loides and M. angolensis , also similar remarks relating to the aplanospores of Gonatonema). An example of M. capucina from the New Forest is figured, in which there are two carpospores present in the same sporocarp (Fig. 48). This is analogous to the double zygospores of Closterium lineatum and certain abnormal cases of Spirogyra (Figs. 75 and 76). The carpospores of M. irregularis are worthy of note for the extreme irregularity of their spore-membrane (Figs. 56 and 57). Spores resembling aplanospores are occasionally found in Mougeotia , but we have not been so fortunate as to meet with any. They are spores produced by the division of the original cell1, and not by a rounding off of the contents as in Gonatonema ; they may be regarded as carpospores formed from sporocarps (consisting of two or three cells) produced without conjugation, but possibly in consequence of the 1 Wittrock, 1. c., t. ii, f. 7 s, s (pseudospora tripartitione (more Staurospermi) sine copulatione formata), et 8 m, m (pseudospora bipartitione (more Mesocarpi) sine copulatione formata). West & West. — Observations on the Conjugatae. 39 stimulus which has already caused conjugation to take place in a distant part of the filament. Indications of sexuality are to be found in the Mesocarpeae, but they are much less marked than in the Zygnemeae. The spores are often seen to be nearer one filament, and the conjugating-tubes of that filament to be thicker and shorter than those of the other (cf. Fig. 47) ; hence the former may be looked upon as a female and the latter as a male filament As these scarcely appreciable indications of sexuality are often absent, we may regard the Mesocarpeae as having lost almost all traces of differentiation into male and female gametes. 1. Gonatonema. The sterile specimens of this genus are undistinguishable from those of Mougeotia , although the chromatophore is more an axile rod (as in Mougeotia capucina ) than an axile plate ; the species of this genus are also of very much rarer occurrence than those of Mougeotia. The spores are asexual and parthenogenetic, and the whole con- tents of the cell are utilized in their formation. During the formation of the spore and just before the appearance of the thin membrane round the cell-contents, we have noticed, both in G. Boodlei and G. tropicum , that in a few of the cells a more or less indistinct division of the cell-contents into two portions takes place. As to the precise import of this we cannot at present offer an opinion. Is it merely a chance arrangement of the cell-contents, or may it not be some slight retention of the last traces of ancestral sexual characters ? Much is yet to be observed from the study of living Gonatonema during the active formation of spores. It is also noticeable that the great difference in size between the spores of G. Boodlei and G. tropicum is more than can be accounted for by the difference in cubical capacity of the vegetative cells and contained cell-contents, the latter being almost the same in each case. Figs, 1-15 illustrate the spore-formation in two species of Gonatonema which as yet have not been figured. 40 West & West. — Observations on the Conjugatae. Sub-fam. 2. — Pyxisporeae . This family is represented solely by the genus Pyxispora obtained from West Central Africa1. The vegetative cells, which are about 12-13*5 m in thickness, contain two chroma- tophores very similar to those present in Zygnema , and in the sterile condition the plant could not be distinguished from the vegetative filaments of a species of the latter genus ; each of these chromatophores has a small central pyrenoid. The conjugation is scalariform and similar to that present in the Mesocarpeae, resulting in an immediate tripartition into a sporocarp consisting of two sterile cells and an intervening carpospore. The characters of this carpospore are unique, and sharply demarcate this genus from any other in the Zygnemaceae. It is broadly elliptical with rounded poles : it is disposed transversely to the longitudinal axes of the conjugating filaments, and around its edge, in the plane of its shorter diameter, is a small annular ridge marked by a circumscissile crack. Some further figures of this interesting genus are given (Figs. 53 and 54). Sub-fam. 3. — Zygnemeae. This is the largest family of filamentous Conjugatae, and includes the five genera, Zygnema , Pleurodiscus , Spirogyra, Sirogonium , and Debarya. The chromatophores of the genus Spirogyra , according to some botanical text-books, ‘take the form of green spiral bands with toothed edges ’ ; this is often true, but throughout the genus they exhibit much variation, there being every gradation between the slender, perfectly smooth spirals of S. neglecta with their axile uniform series of pyrenoids, and the broad serrated spirals of S', nitida and S. porticalis , containing scattered pyrenoids of various sizes. In fact, the 1 West and G. S. West, Welw. Afric. Algae, Journ. Bot. 1897, p. 39. West & West . — Observations on the Conjugatae . 41 characters of the chromatophores are not only remarkably constant but also widely different in many of the common species of the genus ; those species with toothed edges to the chromatophores are however the most frequent. The presence of straight chromatophores in the genus Sirogonium is in itself of no generic value, as those of Spiro- gyra majuscula are quite as straight, if not straighter, but the method of conjugation seems to us quite distinctive. Owing to the somewhat irregular thickening of the walls of some species of Zygnema , such as Z. ericetorum and Z. pachydermumy and the more or less non-stellate condition of their chromatophores, they can be readily mistaken in the sterile condition for species of Rhizoclonium (a genus of Confervaceae Isogamae), and the short, few-celled branches of Z. pachydermum1 render it still more liable to an error of this nature. There are two modes of conjugation, scalariform and lateral , the details of which have been minutely followed out. In the former the cells of two or more filaments take part in the formation of the zygospores, but in the latter, conjugation takes place between the adjoining cells 2 of one filament only. If conjugation is affecting only a portion of a filament, the increased activity along its whole length (as previously mentioned) often causes the cells of its free portions to de- velope conjugating-tubes, which, after making futile attempts to meet with a fellow, become more or less irregularly branched 3 ; such is also the case in many examples in which conjugation has been interrupted. On examining a large number of conjugated examples of Spirogyra or Zygnema , there is one prominent feature which at once strikes the observer, and on this point we cannot 1 West, Algae from the West Indies, Journ. Linn. Soc. Bot. Vol. xxx, PI. XIII, Figs. 12-15. 2 In Cooke’s Brit. Freshw. Alg., PI. XXXI, f. 3 c, an example of lateral conju- gation is shown between two non-adjacent cells. 3 Cf. West, Sulla Conj. delle Zygn., Notarisia, 1891, Vol. vi, t. 12, Figs. 3, 5-7, and 9. 42 West & West . — Observations on the Conjugatae . do better than quote Bennett and Murray 1. £ As De Bary has pointed out — and his statement is confirmed by nearly all more recent observers — the direction of conjugation is clearly governed by some physiological law, the movement of the protoplasm between the two filaments almost invariably taking place in one direction only, so that one of the two conjugating filaments is entirely emptied, while the other is filled with zygosperms.’ In this paper we shall refer to the filament filled with zygosperms as the female, and the emptied one as the male filament. As a rule a zygospore is formed by the fusion of the contents of two conjugating cells, but very rarely it is seen that three cells (two male and one female) have participated in its formation 2 (vide Fig. 66) ; in this way even three filaments may be concerned in the production of one zygo- spore. That this manner of conjugation is abnormal is proved by the larger number of failures than of completed attempts (vide Figs. 67 and 69). In those species belonging to the sub-genus Zygogonium , in which the zygospore is formed in the conjugating-tube, conjugation between three cells entails the production of two somewhat smaller zygospores, as in the example figured (Fig. 63). Two filaments are generally concerned in an example of scalariform conjugation, but three, four, five, and even six are not uncommonly seen 3. In such cases we have to deal with either polygamy or polyandry, and after the examination of hundreds of examples, we can confirm Bennett’s statement that the former is rather more frequent, the ratio of the frequency of polygamy to polyandry being about i-6 : 1. During conjugation the filaments frequently assume a darker colour, this being most marked in Spirogyra angolensis , in which species they become blackish- or brownish-purple. 1 Bennett and Murray, A Handbook of Cryptogamic Botany, p. 2 66. 2 Cf. Z. cruciatum in West, Sulla Conj. delle Zygn., 1. c., t. 13, f. 13; also Spirogyra , sp. in Borge, Siber. Chlorophy., Bih. t. Sv. Vet.-Akad. Handl., Bd. 17, Afd. 3, No. 2 (1891), t. i, f. 2. 8 West, 1. c., t. 12, f. 1. West & West . — Observations on the Conjugatae . 43 Normal conjugation depends to a certain extent on the general surroundings of the filaments, many hindered and consequently irregular examples being met with in every gathering in an active state of conjugation. On rare occasions hybrids are produced, one species of Spirogyra conjugating with another of different thickness ; examples of this are N. maxima , var. inaequalis 1, and some smaller species gathered in 1893 on Mitcham Common, Surrey (Figs. 70 and 71). Several abortive attempts at hybridism were seen in this gathering, the two examples figured being the only two observed with thick- walled zygospores. These spores were of variable form and dimensions, and were present in both filaments. In contradiction to Cooke, we find scalariform conjugation to be much commoner than lateral conjugation, as may be gathered from the following table. No. of gatherings No. of gatherings of scalariform containing lateral Species. conjugation examined conjugation during the past examined in few years. same period. S. ajftnis ... 1 3 S. angolensis ... 1 S. arcta ... 2 S. bellis ... 6 1 S. calospora 1 S. cataeniformis ... 4 S. communis 5 S. condensata 9 S. crassa ... 9 1 S. cylindrospora ... S. decimina 3 S. dubia ... 1 S.fusco-atra 1 S. gracilis , and v.Jlavescens ••• 13 i S. Grevilleana 4 S. injlata ... 1 ... 2 S. insignis 4 S. Jurgensii ... 2 ... 1 S.longata... ... 13 S. Lutetiana 1 S. majuscula {S. ortho spira) ... 2 S. maxima (S. orbicularis') 1 1 Wolle, Freshw. Alg. U. S., PI. CXXXVIII, Figs. 5 and 6, and PI. CXLII, Figs. 5 and 6. 44 West & West . — Observations on the Conjugatae . No. of gatherings No. of gatherings of scalariform containing lateral Species. conjugation examined conjugation during the past examined in few years. same period. S. negleda v. ternata 1 1 S. nitida ... ... 17 S. porticalis 3 S. setiformis S. Spreeiana 2 S. tenuissima ... 16 S. varians ... • 6 4 S. velata ... 1 S. Weberi 7 1 S. Welwitschii ... 1 Total ... 123 32 Lateral conjugation is much rarer in Zygnema than in Spirogyra , and although it is figured in various works; we have never yet seen an example of it. It is figured by Schmidle 1 in this genus in a species which he names Zygnema (Zygogoniuni) Heydrichii , but of which he gives no proper diagnosis 2. Scalariform conjugation being far more predominant, we may say that lateral conjugation is the exception rather than the rule, and in view of this we may regard it with some amount of truth, as brought about by conditions unfavourable to conjugation in the natural or scalariform way. It may be thus considered to a certain extent as abnormal, and to what extent this abnormality is carried may be gathered from a consideration of Fig. 68 (drawn from an example from Mitcham Common, Surrey), in which specimen conjugation has taken place through the ends of the cells, and the con- jugating cells have become genuflexed near their junction. 1 W. Schmidle, Zur Entwickelung einer Zygnema und Calothrix, Flora, 1897, Bd. 84, Heft 2. 2 This species seems to us to be only a Zygnema spontaneum with lateral conju- gation. Nordstedt only found aplanospores when he described the species, but we have since found the zygospores (cf. Figs. 60 and 61) produced by scalariform conjugation (Journ. Bot., Feb. 1897, p. 40). The zygospores seen by Schmidle and produced by lateral conjugation agree in every way with those we saw in Zygnema spontaneum \ moreover, the plants are of the same dimensions. West & West, — Observations on the Conjugatae. 45 That it is not a perfectly normal condition is also proved by the numbers of failures where lateral conjugation was attempted h Against the sexuality of the Zygnemeae only two plausible objections can be raised ; these are the phenomena of lateral and cross-conjugation . It is clear that in the case of lateral conjugation sexual differentiation of the individual cells and not of the whole filaments must have taken place, and con- cerning this Bennett and Murray2 state, ‘that there is some differentiation of this kind would appear from the fact that when lateral conjugation takes place in a group of four cells the zygospores are formed in the two centre cells, which may be regarded as female.’ This we find to be the case, as may be seen from the figures of 5. Jurgensii and .S', inflata (Figs. 72 and 73). Under certain conditions why should not this individual differentiation take place ? Why should not the cells in a filament of Spirogyra or Zygnema be considered in a sense as only partially developed, further physiological changes taking place just antecedent to conjugation, which give the cells the characters either of a germ-cell or a sperm- cell ? We see no reason for regarding the filaments as sexual until conjugation is about to commence, and then instead of the reproductive cells being specially cut off (as in Temnoga - metum ), the contents of the individual cells undergo a profound physiological change, being imperceptibly converted into iso- gamous gametes. Also if the conditions of environment be such (as in an isolated filament) as to render it impossible for the whole of the cells of one filament to become of one sex, why should not individual sexuality of each cell be assumed ? In some cases scalariform and lateral conjugation occur in the same filament 3, and within a few cells of each other. We have mentioned that a filament engaged in conjugation has the vigour of all its cells largely augmented, and may not the activity of the changes converting the ordinary vegetative 1 West, Sulla Conj. delle Zygn., 1. c., t. 12, f. 8. 2 Bennett and Murray, Cryptogamic Botany, p. 267. 3 Petit, Spirogyra des envir. de Paris (Paris, 1880), Plate I, f. 13. 4.6 West & West. — Observations on the Conjugatae. cells into reproductive cells be so far modified at different parts of the same filaments that differentiation of sex is brought about? Regarding the cells in this light, each one may be considered as an individual plant ; and why not ? Each individual cell is capable of living apart from its neigh- bours, obtaining its own nourishment from the surrounding medium, and its life is in no way dependent upon the other cells of the filament. Moreover, if we allow that the Zygne- meae are comparable to the filamentous Desmidieae, we find that the latter readily dissociate into separate cells, which are not at all affected by their isolation (cf. p. 30 supra). The only important function of the assumption of a filamentous condition seems to us to be in the greater facility for conjuga- tion afforded by the entanglement of the gregarious filaments. Before considering cross-conjugation it will be as well to consider some examples of interrupted conjugation. A keen observer is continually coming across instances in which conjugation has by some means been brought to an abrupt termination before the proper formation of the zygospores has taken place, and in these cases the spores formed are very variable. It may be that something has caused a cessation of the activity along the whole filament, or that conjugation has been stopped between two cells only. Fig. 69 is an illus- tration of the latter, the forcible pressure of a second male conjugating-tube having narrowed the channel of communica- tion to such an extent that union of the contents of the gametes was rendered impossible. The former is, however, much the most frequent. Sometimes the spore in the germ-cell is not of its true form l, and occasionally two spores, one large and one small, are present in place of the normal one2 (Figs. 75 and 76). When the conjugation has by some influence been hastened, a zygospore is often produced from only a portion 1 Cf. Spirogyra Groenlcmdica , Rosenvinge in Ofvers. K. Vet.-Akad. Forh., 1883, No. 8, t. viii, f. 1-11. 2 West, Sulla Conj. delle Zygn., 1. c., T. XIII, Figs. 27, 28; binate spores in Spirogyra communis, A. Hansgirg, in Hedwigia, 1888, Hefte 9 u. 10, T. X, f. 6; binate spore in Spirogyra Weberi. West & West. — Observations on the Conjugatae . 47 of the contents of the cells, and in these cases the spore is generally in the female filament (Fig. 74). Some examples have a spore in each conjugating cell (Figs. 77-80), and as a rule that in the female cell is of larger size. Occasionally the spores are of equal size (Fig. 78), and in rare cases the largest spore is in the male (?) cell. All this leads up to cross-conjugation, which is the only other objection to sexuality. By cross-conjugation we mean scalariform conjugation with the formation of perfectly normal zygospores in each of the conjugating filaments, and we have seen but a solitary example of this amongst the thousands of conjugating specimens examined. This was a specimen of Spirogyra gracilis (Fig. 81) found in a gathering of Desmids obtained from a mass of Utricularia minor in a bog near Bowness, Westmoreland. As will be seen from the figure, there are two female cells and one male cell in one filament, and two male cells and one female in the other ; moreover, the zygospores in each filament are perfectly normal, and the con- jugation is complete and also normal. Now this is explicable, as in the case of lateral conjugation, by supposing that each individual cell has assumed sexuality. That the sexual condition of the filaments is the same in both lateral and cross-conjugation is proved by the occurrence of the former in both male and female filaments, which are also conjugating in a scalariform manner1. As a rule examples with zygospores in both filaments only exhibit a false cross-conjugation (Fig. 64), the zygospores in one filament being smaller than those in the other. This fact tends to prove that numerous attempts at cross-conjugation result in failures, normal zygospores not being produced, and together with its extreme rarity serves to show to what degree it is abnormal. From the foregoing statements we have shown that lateral and cross-conjugation are explicable from a sexual point of view, and that there is no reason to regard the Zygnemeae as otherwise than sexual. 1 Petit, in Bull. Soc. Botan. France, fevr. 1874, t. xxi, PI. I, f. 2. 48 West & West. — Observations on the Conjugatae. Other minor observations have also been brought forward as a proof of this sexuality, such as the comparative lengths of the cells and the relative thickness of the male and female conjugating-tubes. These are, however, of little value, although it is certainly a fact that in the majority of instances the female conjugating-tube is shorter and thicker than the male (Figs. 66, 67, 79-81); also the female cells are often so swollen that all trace of a conjugating-tube is lost1 (Figs. 67 and 81). In a gathering of Spirogyra velata from a stream at Baildon, W. Yorks., several zygospores were noticed which did not assume a thick wall, but germinated immediately after their formation (Figs. 84 and 85). The formation of spores without conjugation takes place not uncommonly in the Zygnemeae ; these asexually-produced spores are called aplanospores, and are produced from the contents of a single cell. We have noticed them in Zygnema leiospermum (Fig. 83), Z . pachydermum 2 and Spirogyra varians (Fig. 82). They have been observed by Nordstedt 3 in Zygnema spoilt aneum and by Wille 4 in if. cruciatum . In all cases the aplanospores are somewhat smaller than the zygospores, have a thinner membrane, and as a rule they are spherical, no matter what the form of the zygospore. We may also quote a remark made by Petit 5 concerning Spirogyra mirabilis , in which species the spores are produced without conjugation. He writes: ‘Cette tres curieuse espece ne conjugue pas et ne laisse voir aucun tube copulateur ; a une certaine epoque de la vie de la plante, les cellules se renflent vers le milieu, 1’endo- chrome se partage en deux parties qui se concentrent sous forme de globule aux deux extremites de la cellule ; il se forme ainsi une differentiation entre les parties de l’endo- 1 Petit, Spirogyra des envir. de Paris, PI. IX, f. 10 ; West, Fresh w. Alg. W. Ireland, Journ. Linn. Soc. Bot., Vol. xxix, T. XVIII, f. 5. 2 West, Alg. W. Indies, Jonrn. Linn. Soc. Bot., Vol. xxx, p. 266, PL XIII, f. 9, 10. 3 O. Nordstedt, De Alg. aq. dulc. et Char. Sandvic, p. 17, T. I, f. 23, 24. 4 N. Wille, Ferskv.-alg. Nov. Semlj. Ofvers. K. Vet.-Akad. Forh., 1897, No. 5, p. 63, T. XIV, f. 87. 5 Petit, 1. c. p. 14. West & West. — Observations on the Conjngatae. 49 chrome. Bientot les deux globules se rapprochent vers la partie renflee de la cellule et finissent par se reunir en constituant ainsi la zygospore.’ The vegetative cells of the genus Debarya are like those of Mougeotia. The conj ugating-tubes of D . glyptosperma are long, some of them very long, and when they do not happen to meet with a fellow they often become club-shaped. As the cell-contents pass into the conjugating-tube the chro- matophore takes the form of a loop, and a peculiar change comes over the empty cells as the zygospore is being formed ; they become very clear and refractive, and a series of striations parallel to the transverse septa become visible. After this has taken place the cells have the appearance of solidity, this ap- pearance being possibly due to annular thickenings deposited inside the cell-wall on the receding of the protoplasm during conjugation ; in any case this character stands out distinctly in both old preserved specimens and in living ones. The large size of the zygospore is also a noticeable feature. In D. laevis the conj ugating-tubes are shorter and thicker and the spores are proportionately smaller. We give a figure of a specimen of this plant which has conjugated in a re- markable manner (Fig. 58). It has two to four pyrenoids in each chromatophore. Mougeotiopsis , a genus recently described by Palla1, seems to us to differ in no way from Debarya , except in the absence of pyrenoids. This is certainly in itself an insufficient generic character, and might probably be caused by the conditions under which the plants were growing. Lagerheim2 states that Mougeotia laevis belongs to Mougeotiopsis ; it is, how- ever, a true species of Debarya , and as stated above certainly contains pyrenoids. The zygospores of both Debarya laevis and Mougeotiopsis calospora are scrobiculate, which seems to further indicate the identity of these two genera. 1 E. Palla, Ueber eine neue, pyrenoidlose Art und Gattung der Conjugaten, Ber. der Deutsch. Botan. Gesellsch., Jahrg. xii (1894), Heft 8, pp. 228-236, T. XVIII. 2 G. Lagerheim, Ueber das Phycoporphyrin, Vidensk.-Selsk. Skrift., I. mathem.- natur. Kl., Kristiania, 1895, No. 5, p. 16 (Sep.). E 50 West & West . — Observations on the Conjugatae . Fam. II. — * Temnogametaceae. This order, defined as follows, ‘ Ordo novus Conjugatarum , conjugatio solum inter cellulas speciatim abstrictas,’ was insti- tuted a short time ago1 to include a West African plant differing in a marked way from all the genera of Conjugatae. The sole representative plant is Temnogametum he ter osp or urn, which has a great superficial resemblance to some species of Mougeotia. The vegetative cells are precisely like those of the latter genus, each cell being provided with a more or less plate-like chromatophore, in which a single series of from one to six small globose pyrenoids is embedded. The conjugation is remarkable, owing to the fact that the reproductive cells are specially cut off from the rest of the plant ; they are short, isogamous gametes, being about a quarter or a sixth part the length of the ordinary vegetative cells, and are cut off at intervals along the filaments. Some are cut off singly and others in pairs ; in the former case the conjugation is scalariform, in the latter it is lateral. In scalariform conjugation the contiguous faces of the gametes become swollen, these swellings being merely short, rounded conjugating-tubes which finally unite (cf. Fig. 49), their union being followed by the bending towards each other and ultimate coalescence of the gametes to form a somewhat cruciate zygo- spore (Fig. 50). As previously mentioned 2, this zygospore at first sight very much resembles the central cell (or carpospore) of the five cells constituting the sporocarp of those species of Mougeotia belonging to the section Staurospermeae, but on closer examination the four contiguous cells are seen to possess their complete cell-contents, and to have taken no part in the formation of the zygospore. In the case of lateral conjugation, the pairs of cells become a little oblique or somewhat swollen on one side and then unite, this coalescence giving rise to an obliquely subcylindrical zygospore 1 West and G. S. West, Welw. Afric. Algae, Journ. Bot., Feb. 1897, p. 37. 2 West and G. S. West, 1. c. West & West. — Observations on the Conjugatae. 51 (Fig. 52), which has a considerable resemblance to the aplano- spore of some species of Gonatonema (e. g. G. notabile). Soon after the coalescence of the gametes the wall of the zygospore increases much in thickness. One case was noticed in which a solitary gamete in one filament was conjugating with one of a pair in another filament (Fig. 51). There is no perceptible sexual differentiation between these gametes, but owing to the fact that they are specially cut off, this family must be regarded as considerably removed from the other families of the Conjugatae, though it is not so highly specialized as the Mesocarpeae. Fam. III. — Desmidiaceae. The chromatophores in the Desmidiaceae are disposed more or less symmetrically in the two halves of the cell, either as central masses of chlorophyll arranged in relation to certain pyrenoids, or as parietal and somewhat pulvinate masses containing scattered pyrenoids. In some genera these pyrenoids are definite ; the majority of Cosmaria have either one or two in each semi-cell, and the great majority of Staurastra contain one large one in the centre of each semi-cell. In a paper published by Lutke- muller1 entitled ‘ Beobachtungen liber die Chlorophyllkorper einiger Desmidiaceen ’ the author demonstrates the irregu- larity of the pyrenoids in certain species. No doubt many irregularities are to be found in most Desmids, but as a general rule we find the central pyrenoids very constant in character. One of the species mentioned by Lutkemuller as very variable in this respect is Cosmarium pyramidatum , normal specimens of which should contain two pyrenoids in each semi-cell. Our experience of this species confirms his observations, but we may add that in this respect it is the most variable species that we have yet examined. Oesterreich. Botan. Zeitschr., 43. Jahrg. 1893, No. 1. 52 West & West . — Observations on the Conjugatae. Specimens of Cosmarium ornatum occasionally have irregular pyrenoids, but we have not yet seen more than one example in a hundred. Amongst an immense number of examples of Cosmarium sphagnicolum , collected in early spring in N. Yorks, from moorland pools nearly filled with Sphagnum cuspidatum and Ptilidium ciliare , many variations were observed in the chromatophores, and the pyrenoids which were normally one in each semi-cell, varied from one to three (cf. Figs. 34-36). This variability of the chromatophores has been described in Penium minutum 1. We figure four cells of Hyalotheca neglecta (Figs. 30-33), which show considerable variation in the pyrenoids, but these were only found after the examination of a very large number of filaments. This irregularity is usually found after rapid division, in the same way that abnormality of form is probably caused (cf. Fig. 39 of Euastrum didelta). This subject is one concerning which much work is yet desirable, as several doubtful genera have been founded on the structure and arrangement of the chromatophores ; e. g. Pleurotaeniopsis , Pleuren teriu m , &c. Sexual (?) reproduction is by conjugation and formation of zygospores, the conjugating cells generally not being differentiated. Double spores are formed in Closterium lineatum , Cylindrocystis diplospora and Penium didym oca rpu m , and are analogous to those in the Mesocarpeae and homolo- gous with those in the Zygnemeae. In some rare cases three (or even four) cells have participated in the formation of a zygospore2. The zygospore is formed between the conju- gating cells in all Desmids except Desmidium cylindricum. In this species it is formed within the female cell 3 as in Spirogyra 1 Cf. remarks by Lutkemuller under Docidium baculum. Also Journ. Bot., March, 1895, p. 65. 2 Cf. Staurastrum teliferum in West, Freshw. Alg. W. Ireland, Journ. Linn. Soc. Bot., Vol. xxix, PI. XXIV, f. 5 ; Cosmarium rectisporum} W. B. Turner, Freshw. Alg. E. India, K. Sv. Vet.-Akad. Handl., Bd. xxv, No. 5, T. X, f. 16 e; also Closterimn Pritchardianum , West and G. S. West, Freshw. Alg. S. of England, Journ. Royal Micr. Soc., PI. VI, Fig. 5, December, 1897. 3 Ralfs, Brit. Desm., T. II, f. 1, e, f, g, h, i, k ; Wolle, Desm. U. S., PI. Ill, f. 4 ; West & West. — Observations on the Conjugatae . 53 and Zygnema , although, as in most filamentous Desmidieae, the filaments break up before conjugation. So far as we know, this is the only case of differentiation of the conjugating cells met with in the whole of the Desmidiaceae. Boldt 1 figures a ‘ forma monstrosa 5 of Hycilotheca dissiliens with the zygospore in one of the cells, and Joshua2 mentions a case where Hyalo- theca dissiliens was conjugated like Desmidmm cylindricum . Fig. 37 is also an approximation to this stage in an example of Hyalotheca dissiliens. What is this c monstrous form 5 of conjugation in this species? Abnormal it certainly is as compared with ordinary conjugated examples, but is it not a case of reversion to some ancestral type of conjugation, represented at present by the Zygnemeae, and which the Desmidieae have almost lost, the lingering remains of which are still found in Desmidium cylindricum ? Thus degeneration and loss of sexual differentiation of the conjugating cells have gone on hand in hand with the loss of the filamentous condition, the majority of filamentous forms dissociating before conjugation. An extreme morphological specialization has accompanied this loss of the filamentous condition, causing the large majority of this family of unicellular plants to be remarkable for their beauty and variety of form. In the genus Desmidium conjugating-tubes are formed, and we have noticed rudimentary conjugating-tubes in some species of Closteriumz and in Arthrodesmus octocornis. Conjugation seems to take place in many Desmids immediately after division and before the young semi-cells have had time to attain maturity4: for sexuality to exist West and G. S. West, N. Amer. Desm., Trans. Linn. Soc. Bot., ser. 2, Vol. v, Pt. v, PI. XII, f. 29. 1 R. Boldt, Desm. fran Gronl., Bih. till Sv. Vet.-Akad. Handl., Bd. xiii, Afd. 3, No. 5, T. II, f. 33. 2 W. Joshua, Notes on Brit. Desm., Journ. Bot., Vol. xx (1882). 3 In Closterium Ehrenbergii they are perforated protuberances at the base of the younger semi-cells; cf. West and G. S. West, Journ. Roy. Micr. Soc., 1896, p. 151. 4 West and G. S. West, 1. c., pp. 151 and 153, PI. Ill, f. 29 ; also W. Archer, in Quart. Journ. Micr. Soc., Vol. ii, p. 251. 54 West & West. — Observations on the Conjugatae. under these conditions, the physiological change (previously referred to) from the vegetative to the reproductive cell must be immediately antecedent to conjugation. Lateral conjugation is not unknown amongst filamentous Desmids. Ralfs describes 1 the conjugation of two adjacent cells in a filament of Sphaerozosma excavatum as taking place between their flat ends, and we have seen an example of this in Spondylosium pulchrum , var. planum , from Orono, Maine, U.S.A. In these instances the filament does not fragment before conjugation, the zygospore filling up the space originally occupied by the two adjacent semi-cells of the conjugating cells. Aplanospores are occasionally found in the Desmidieae ; Bennett 2 mentions the occurrence of some spore-like bodies produced without conjugation in Closterium , and they are figured by Wallich 3 and Turner4 in Spondylosium nitens. In a gathering of Desmids from the New Forest in which Hyalotheca neglecta was abundant, many of the cells con- tained aplanospores (Cf. Figs. 23-27) ; these were produced by the rounding off of the cell-contents and final assumption of a thick cell-wall. They differ in form from the globose zygospores, being, elliptical, with rounded poles, and when mature their walls turn yellowish-brown. Phylogeny. In all probability the Zygnemaceae have arisen along two distinct lines from some ancestral filamentous sexual Conjugates. The Mesocarpeae may have been developed through Debarya along one of these lines, and from them Temnogametum probably struck off at some early stage. Along the other line the remainder of the existing Conjugates 1 Ralfs, Brit. Desm., p. 67. 2 A. W. Bennett, in Annals of Botany, Vol. vi, No. 21, April, 1892. 3 G. C. Wallich, Desm. Low. Bengal, Ann. Mag. Nat. Hist., Ser. iii, Vol. v, i860, T. VII, f. 10, 11. 4 W. B. Turner, Freshw. Alg. E. India, K. Sv. Vet.-Akad. Hand!., Bd. xxv, No. 5, T. XVIII, f. 7, 8. West & West. — Observations on the Conjugatae. 55 were probably developed, and at a short period before the Zygnemeae became differentiated into the two distinct groups represented by Spirogyra and Zygnema , the Desmidieae were probably evolved by retrogression. The view that the latter may have been evolved in this way is confirmed by the occasional reversion of Hyalotheca to its ancestral mode of conjugation, the remains of which are still found in Desmidium Table of Phylogeny. cylindricum . Pleurodiscus , which conjugates like a Zygnema , was probably evolved from the Zygnema group just after its differentiation from the Spirogyra group, Pyxispora being evolved from Zygnema by the assumption of a rudi- mentary sporophyte-generation, thus placing it on an equal level of specialization with the Mesocarpeae. Zygnema can be connected with Debarya by the subgenus Zygogonium , and 56 West & West. — Observations on the Conjugatae. a parallelism of modification has gone on in the reproduction of Spirogyra and certain filamentous Desmids, as shown by the obsolete Rhynchonema and the laterally conjugated Spoil- dylosium. We cannot but regard the Mesocarpeae and the Pyxisporeae as the most highly specialized families of the Conjugatae, the formation of the sporocarp being a faint indication of an ‘ alternation of generations V There is a little retrogression in the Mesocarpeae, certain plants of this family (placed under a distinct genus— Goiiato- nenia) producing spores only asexually. The accompanying phylogenetic table has been drawn up to graphically illustrate the conclusions we have arrived at. DESCRIPTION OF PLATES IV AND V. Illustrating Messrs. West’s paper on Conjugatae. PLATE IV. Figs. 1-9. Gonatonema Boodlei, West and G. S. West. From Mitcham Common, Surrey, x 520. Figs. 10-1 5. Gonatonema tropicum , West and G. S. West. From Huilla, Angola, W. Africa, x 520. Fig. 16. Mougeotia sp. Showing organ of attachment, x 120. Figs. 17-19. Mougeotia sp. Showing short lateral branches, x 120. From Frizinghall, W. Yorks. Figs. 20, 21. Spirogyra sp. With rhizoids ; from Hanka Deela, Somaliland. X 120. Figs. 22-33. Hyalotheca neglect a, Racib. x 520. From the New Forest, Hants. 22, vegetative filament with wide gelatinous sheath; 23-27, showing formation of aplanospores ; 28-29, zygospores; 30-33, single cells with irregu- larity of pyrenoids. Figs. 34-36. Cosmarium sphagnicolum, West and G. S. West, x 520. From Mossdale Moor, Widdale Fell, N. Yorks. 34, with normal pyrenoids ; 35 and 36, with irregular pyrenoids. 1 The Mesocarpeae afford a better example amongst the lower Algae of a sporophyte-generation and a rudimentary alternation of generations than that shown by the Oedogoniaceae. West & West. — Observations on the Conjugatae. 57 Fig. 37. Hyalotheca dissiliens (Sm.), Breb. A peculiarly conjugated example from Thursley Common, Surrey, x 250. Fig* 38* Euastrum binale (Turp.), Ehrenb. Showing tendency to assume a filamentous condition; from Thursley Common, Surrey, x 250. Fig. 39. Euastrum didelta (Turp.), Ralfs. Specimen from Wrynose, Lake District, abnormally divided. X 220. Fig. 40. Tetmemorus granulatus (Breb.), Ralfs. Specimen from the New Forest, Hants, one semi-cell branched, x 250. Fig. 41. Mougeotia sp. Showing branching; from near Lindley Reservoir, W. Yorks, x 250. Fig. 42. Mougeotia uberosperma , West and G. S. West. With immature spores, x 520. Fig. 43. Mougeotia uberosperma, West and G. S. West. With mature spores, x 520. Fig. 44. Mougeotia parvula , Hass. From Black Hill, near Settle, W. Yorks, x 520. Fig. 45. Mougeotia recurva (Hass.), DeToni. Three cells conjugating together ; from Borrowdale, Lake District. x 520. Fig. 46. Mougeotia angolensis , West and G. S. West. From Pungo Andongo, Angola, W. Africa, x 250. Fig. 47. Mougeotia nummuloides, Hass. From Scarf Gap Pass, Lake District. X 250. Fig. 48. Mougeotia capucina (Bory.), Ag. Example from New Forest, Hants, with two carpospores. x 250. PLATE V. Figs. 49-52. Temnogametum heterosporum, West and G. S. West. From Huilla, Angola, W. Africa, x 250. Figs. 53, 54. Pyxispora mirabilis zst and G. S. West. From Huilla, Angola, W. Africa, x 520. Fig. 55. A hybrid specimen from Strensall Common, N. Yorks. ; conjugation taking place between two species of Mougeotia. x 400. Figs. 56, 57. Mougeotia irregularis , West and G. S. West. From Pungo Andongo, Angola, W. Africa, x 350. Fig. 58. Debarya laevis (Kutz.), West and G. S. West. Peculiarly conjugated example from Mitcham Common, Surrey, x 220. Fig. 59. Debarya laevis. Mature zygospore showing the scrobiculations. x 520. Figs. 60, 61. Zygnema spontarieum , Nordst. Mature zygospores produced by scalariform conjugation ; specimens from Huilla, Angola, W. Africa, x 520. Figs. 62,63. Zygnema pectinatum (Vauch.), Kutz. 62, x 120; 63, x 220. Fig. 64. Spirogyra condensata (Vauch.), Kutz. x 120. Fig. 65. Zygnema pectinatum (Vauch.), Kutz. x 120. Fig. 66. Spirogyra maxima (Hass.), Wittr. x 100. Conjugation between three cells. Fig. 67. Spirogyra sp. From Huilla, Angola, W. Africa, x 120. Fig. 68. Spirogyra inflata (Vauch.), Rabenh. x 220. From Mitcham Common, Surrey. 58 West & West . — Observations on the Conjugatae . Fig. 69. Spirogyra condensate/, (Vauch.), Kutz. x 250. Figs. 70, 71. Hybrids from Mitcham Common, Surrey, x 250. Conjugation between two species of Spirogyra of different thickness. Fig. 72. Spirogyra inflata (Vauch.), Rabenh. x 100. Fig. 73. Spirogyra Jurgentii, Rabenh. x 100. Figs. 74, 75. Spirogyra bellis (Hass.), Crouan. x 140. Fig. 76. Spirogyra velata , Nordst. x 220. Figs. 77, 78. Spirogyra bellis (Hass.), Crouan. x 140. Figs. 79, 80. Spirogyra velata, Nordst. x 220. Fig. 81. Spirogyra gracilis (Hass.), Kutz. forma, x 350. Fig. 82. Spirogyra varians (Hass.), Kutz. x 520. Aplanospore. Fig. 83. Zygnema leiospermum , De Bary. x 520. Showing aplanospore. Figs. 84, 85. Spirogyra velata , Nordst. Two zygospores germinating imme- diately after formation. x 220. G.S.West ad nat.&el. WEST. CONJUGATAE. Vol. XII ; Pl.1V University Press, Oxford. BtuzoZs of Bolasiy G. S.West ad nat. del. WEST. — CONJUGATAE. voi.xn,pi. v. University Press, Oxford.. rummi A Violet Bacillus from the Thames. BY H. MARSHALL WARD, Sc.D., F.R.S. Professor of Botany in the University of Cambridge. With Plate VI. NE of the most interesting forms of Schizomycete I have isolated from the Thames 1 is a bacillar or filamentous one observed more especially in the winter, though probably present also in the summer with altered characters. It is remarkable for producing in vigorous cultures a mag- nificent violet pigment of a peculiarly pure hue ; in fact almost exactly the colour of dark blue violets, or of a strong solution of gentian-violet. This species grows well and is easily cultivated at all temperatures from about io°C. to 250 C., on all the ordinary solid media, and in milk and broth, though difficulties are occasionally met with in potato-cultures. Specimens in gelatine and broth-drops grow out as long filaments, often 50 to 60 jut and longer, which move with an oscillating and sinuous motion, or progress slowly across the field with similar movements. These filaments are obviously segmented, 1 See Proc. Roy. Soc., Vol. lxi, 1897, p. 415. The form here described is that constituting the type of Group II on p. 417. [Annals of Botany, Vol. XII. No. XLV. March, 1898.] 6o Ward. — -A Violet Bacillus from the Thames. and one can see them break at the septa into bacillar rodlets from 3 to 5 or even 7 fx long by about 075 to o-8 n broad or a little more (Fig. 3). They are perfectly colourless, and move more rapidly than the filaments, often progressing straight forward with fairly rapid motions right across the field — say a distance of 100 to 200 times their own length, — and then suddenly stop and dart backwards a short distance, and then on again. Sometimes these progressive movements are associated with a spinning on the axis, or about a central point, but more often they are straight forwards. Specimens stained with methyl-violet show similar sizes, but if taken from older gelatine-cultures the rodlets may be very short — about 1 [x long by 075 to o-8 ju, broad — and almost like cocci (Fig. 1). After a year in culture, the rods are often thinner, o*6 fx or so (Fig. 2), but soon thicken up in the broth-drop at ordinary temperatures. No trace of spores could be detected ; but from some experiments where old cultures were maintained at 5°° C. to 6o° C. for several hours, it appears not improbable that the coccus-like joints act in the capacity of spores, and can withstand fairly high temperatures for a short time. If this turns out to be correct, this Schizomycete may have to be classed as a Bacterium and not as a Bacillus. The rods often show three or four spore-like, bright, denser spots (Figs. 1, 2). Numerous attempts to obtain any further development in culture-cells led to negative results only. In gelatine-drops at 18-22° C. the growth to filaments occurs as in broth, but more slowly, and in twenty-four hours the rods and filaments are motile, slowly writhing or jerking in the liquefying drop ; the movements are obviously restricted. Plate-cultures in gelatine at 20° C. show colonies in forty- eight hours — visible to the unaided eye on the third day — which are small, slightly yellowish discs, with a granular texture under a one-third objective and sometimes traces of zonal structure where submerged. Those nearer the surface break through in a conical shape, and spread their margins irregularly over the gelatine : the much more extended Ward. — A Violet Bacillus from the Thames. 61 surfaces of these colonies appear granulated and rugose, as if marked with a delicate and complicated series of contours and mounds, giving a characteristic aspect, which, however, is by no means peculiar to this specific form — e. g. such colonies occur in Typhoid (Fig. 5). In four or five days the colonies appear as milk-white, opaque, thin, discoid, smooth and glistening expansions to the unaided eye, though the microscope shows the same characters as described above. They grow very slowly, and even after ten days at 20° C. may be still white, though somewhat thicker, like an opaque, flattened milky drop, about 4 to 5 mm. in diameter, and slightly sinking in the now softening gelatine. They are, in fact, beginning to liquefy the latter. About this period, also, the superficial colonies are usually beginning to show the violet pigment — though sometimes it occurs much earlier and at others is delayed — the colour appearing first as a mere trace of purplish cast on a dirty- white matrix, gradually spreading over the colony, till the latter is deep violet or almost blue black. The pigment does not spread into the gelatine, but is strictly confined to these superficial colonies, at least for a long time. The totally submerged colonies may remain dirty-yellow, or buff-coloured, even for weeks, so that oxygen may have to do with the development of the pigment ; in some cases, however, I have found them faintly purple in four or five days (cf. Figs. 4 -6). It is interesting to note that slow growth, with the formation of the pigment, occurs at temperatures even as low as 50 or 6° C. on gelatine-plates kept cool by ice. If there are very few colonies on the plate, they may eventually dry up with the non-liquefied gelatine, and I have plates in the laboratory more than two years old where this has occurred. The hard, smooth, deep purple discoid colonies on these plates look just like dried ink-drops, each about 10-12 mm. in diameter, with irregular margins, and each surrounded with a pale zone, 2-6 mm. broad, scooped out of the yellower dry gelatine, and representing where liquefaction 6 2 Ward. — A Violet Bacillus from the Thames. occurred. The water having evaporated, the colonies are flattened on to the glass. In stab-cultures in gelatine a white, button-like mass forms at the infection-point in three days at 150 C., while minute points develop along the line of inoculation. In ten days a thistle-head funnel of liquefaction is developed, still all quite white. In eighteen days the liquefaction at the top has reached the walls of the tube, or nearly so, and the liquefied gelatine above -is found to have a dense, tough membrane on its surface, the white matrix of which is turning purple. This membrane (zoogloea) lines the sides of the funnel-shaped depression, as the surface of the gelatine sinks owing to evaporation (Fig. 7 c). In a month, at this temperature, the gelatine was liquefied about one-eighth of the distance down, and a deep violet zoogloea-membrane floated on the top, while violet flecks had fallen to the bottom of the liquid, and rested on the still solid gelatine below (Fig. 7 d). All the gelatine-tube cultures show liquefaction of the gelatine. The rate at which this progresses at 20° C. may be estimated from the following : in six days the upper one-third of the gelatine of a stab-culture was liquefied, and with a beautiful deep violet, funnel-shaped mass of the Schizomy- cete tailing off below into the solid gelatine. Even after three months the gelatine is only liquefied about two-thirds down the tube, and six months’ tubes may still have some gelatine solid below. The liquefied gelatine is very viscid. The thin streak in the solid gelatine was still buff in colour, apparently from lack of oxygen. Streak-cultures on gelatine, at i5°-20° C., grow fairly rapidly. In twenty-four hours at 1 50 C. a white opaque streak was developed, which in three days began to show traces of the violet pigment, and the gelatine began to soften below, so that the culture sank in. In nine days the gelatine was excavated by a large scoop-like cavity, filled with viscid liquid (the surface being kept flat by inclining the tube till nearly horizontal) on which floated a deep violet membrane, while Ward— A Violet Bacillus from the Thames . 63 in the liquid itself flecks of white and violet colonies were floating or sunk to the bottom (Fig. 8). In three or four months all the gelatine is liquefied, and dirty-violet flecks are floating in the liquid Closer examination shows that the violet- mass is composed of a complex folded and wrinkled zoogloea-form, and that the pigment is confined entirely to this, and does not diffuse into the liquefied gelatine in which the trembling jelly-like mass floats. These zoogloea-masses are found scooping out cavities in the otherwise solid gelatine, and lining them with the folded membrane. Later on, the liquefaction slowly spreads away from the zoogloea, but the pigment is confined to the latter. The same occurs, somewhat more rapidly, at 200 C. On removing the zoogloea-membrane, these cavities are left apparently devoid of Bacilli ; but the microscope shows their presence in abundance, and in a few days those remaining have multiplied and again covered the concave surfaces with the membrane, which soon thickens, wrinkles, and obtains the violet colour as before. When the gelatine is completely liquefied, the violet colour is confined to the membrane at the surface, or to pieces of it which sink in : the non-oxygenated submerged Bacilli forming a precipitate at the bottom are buff or greyish-yellow, or even dirty-white. On agar, at 20° C., a thick white or slightly buff opaque streak forms in twenty-four hours or so, and in three days has a waxy appearance, often with raised margins. On the fourth day the purple colour begins to show here and there in the white, and in ten days the whole surface of the agar may be covered with an intensely deep purple, corrugated membrane, which is very tough and may be lifted off bodily from the medium. Underneath, the growth is still white, and the colour is confined to the zoogloea-mass — it never diffuses into the agar (Fig. 9). In feeble cultures the growth may be so slow as scarcely to extend beyond the path traced by the inoculating needle ; but here too, the deep, ink-like, purple- black drops soon appear as before. On potatoes, at 20°C.; a thickish, dirty-white streak, tinged 64 Ward . — A Violet Bacillus from the Thames. with grey or yellowish-brown, forms in from two to five days, and slowly spreads ; it varies considerably as to thickness, wet or dry appearance, and rate of growth (Fig. io). Generally, but not always, the violet colour begins to appear about the fourth day, and then invades the mass, and a tough, deep violet, corrugated and mamillated membrane is formed as on agar, but not so extensively. In some cases the cultures on potatoes fail altogether to produce the pigment, and I am unable to say why. Very often the violet colour does not extend to the edges, but remains towards the centre. In broth, at 20° C., turbidity is just perceptible on the second day, and this increases. Later on a thick membrane forms above, the liquid below being very turbid. The mem- brane is at first white, and may remain so for ten days or more. In strong cultures the violet hue appears sooner or later, and in twenty days there is a tough, much corrugated, violet membrane floating on the still turbid liquid, at the bottom of which a white deposit collects. The colour of the membrane deepens as the culture ages, and it clings so tightly to the walls of the tube that the liquid does not escape on upturning. Broth at 350 gives no growth in four or five days — the liquid is perfectly clear. Carrot at 2 50 C. Spreads as a white film in forty-eight hours, and in three days is a purple and white paste. In fifteen days much spread as a white, wet, pasty layer, but only small patches of purple here and there. Artichoke at 25° C. Much as on carrot, but perhaps more waxy. In fifteen days the white wet paste is smeared with blue-violet patches. Milk undergoes slight coagulation, and the separated casein is slowly peptonized, or peptonization may occur without evident precipitation. The reaction of the liquid is alkaline. After some time — it may be three or four weeks — the violet pellicle forms on the surface, deepening in colour from day to day. In some cases, at 250 C., traces of violet occur above in a week, and in the deposit in a fortnight, and still no change is evident in the milk, except a grey colouration and Ward . — A Violet Bacillus from the Thames . 65 marked alkalinity. The grey colour is due to the spreading of the purple through the milk. In a month nearly all is peptonized, and the liquid purple and very strongly alkaline. The colouring pigment cannot be detected in the Bacilli themselves under the microscope, but seems to be in the mem- brane— the swollen cell-walls forming the zoogloea-matrix — or even external to this. It is hardly, if at all, soluble in water, as is shown by the fact that on filtering water in which large quantities of the purple flecks &c. have been crushed, the liquid comes off quite or nearly clear ; but it is readily dissolved out by absolute alcohol. This alcoholic solution is, moreover, extremely stable, and I have kept a tube half full of it for more than six months unaltered in the dark at ordinary (spring and summer) temperatures. The beautiful blue-violet colour has a slight reddish cast in it, reminding one exactly of a solution of gentian-violet. The alcoholic solution shows one broad absorption-band, extending from the red to the green-blue ; but even a layer half-an-inch thick lets some of all the rays through. Acetic acid slowly renders the alcoholic solution paler. NaHO turns it bluish-green, the violet colour returning, but paler, with slight excess of HC1.1 On evaporating the alco- holic solution to dryness over a water-bath, the purple sedi- ment dissolves up again in alcohol apparently unaltered.. Stable as it thus is, however, the solution exposed to the bright sunlight of an August day is completely bleached in from one to two hours. Old milk-cultures yielded Bacilli, the gelatine-colonies from which grew well, but were quite colourless, and very like those of B. Coli communis . It required several passages through broth and gelatine to get the colour up again. I have made numerous experiments with this Schizomycete to determine its relations to light, and find it one of the most sensitive as yet tried. It is quite easy to obtain sun-prints with agar-plates over which a stencil- letter is placed, with 1 Practically the same reactions were observed by Mace with the violet pigment from his B. violaceus {Traite pratique de Bacteriologies 1892, p. 541). F 66 Ward. — A Violet Bacillus from the Thames. a couple of hours’ exposure on a spring day. I have also attained fairly satisfactory results in bright sunshine by placing a plate behind an ordinary photographic negative : the print was slightly wanting in sharpness, because of course I had to have a plate of sterilized glass between the agar-film and the negative ; but in cases where the glass was thin enough the picture was very good indeed. Tubes of water containing large quantities of this Bacillus, can be almost completely sterilized by a few hours’ exposure to sunshine. No doubt the question of temperature comes in in all these cases, but it must be noted (i) that I have kept them alive for four or five hours in a hanging-drop at 30 to 35° C., and (2) that in the experiments with the solar and electric spectra, where the exposures are made over ice, the maximum light-effect is not towards the red end at all, but in the blue- violet. I therefore regard these bactericidal effects as due to the blue-violet light-rays, and not to the high temperature. I have made numerous attempts to watch and measure the growth and division of the rodlets, but although I have been able to convince myself that the filaments segment up into the short rodlets, the fact that both filaments and rodlets are moving actively in all the available media, as soon as the light is turned on them1, though in some cases at any rate they seem to become quiescent again in the dark, has completely baffled all my endeavours. These movements seem to depend on a fairly high temperature — 20 to 250 C. or so — and I should say the divisions are completed about every twenty minutes or so2. The shorter segments (Bacilli) often move actively, each for itself, before separating from the other segments — rarely more than four in all — of the filament, which has a slow, undulating movement as a whole. I have not investigated the question of cilia, but the character of the movements of 1 This fact had already been observed by Engelmann (Unters. aus dem Physiol. Lab. zu Utrecht, 1882, p. 252). 2 Brefeld (Bot. Unters. IV, p. 46) found that B. subtilis, at 240 R. and under favourable conditions, divided once every thirty minutes. Ward • — A Violet Bacillus from the Thames. 67 the individual Bacilli is such as to lead one to expect that each has one or more. As already stated, I have entirely failed, after many attempts, to make this Schizomycete develop recognizable spores. In old cultures on agar or gelatine the rods and filaments, both long and short, are usually found to contain small stainable bodies, which, owing to their brightness and general appear- ance in the fresh preparations, are often very like spores ; especially when, as frequently occurs, they lie singly in the rodlets or at regular intervals in the filaments. These are not spores, however, for they stain as readily with the ordinary bacillar dyes — -such as methylene-blue and methyl- violet — as they do with carbolized fuchsin, and aniline, methyl-violet, & c. Moreover, these stainable masses are of all sizes, and in all positions ; there is nothing definite or spore-like about them. In old cultures they are also often met with in the interior of giant rods and filaments of irregular shapes, which are obviously of the nature of the so-called involution forms. All attempts to get them to germinate have failed, and they do not withstand high temperatures : moreover, the rodlets and filaments containing them are still capable of further growth and division when placed in hanging-drops. On the evidence, therefore, I conclude that no spores are developed under ordinary conditions. The following table exhibits a summary of the characters. Violet Bacillus. Habitat. In the Thames , especially in winter , and not a common for?n. Morphological Rodlets or filaments, from 2-3 /x to 60 /x and upwards long by characters. about 0.75 to o-8 fx broad. Often quiescent, but may be actively motile. Involution-forms in old cultures. No spores found. In old cultures the rodlets are so short as to be almost cocci. Gelatine-plates. Visible in about three days at 15-20° C. At first small, milk- white, opaque, circular colonies, growing very slowly. Under the microscope they are yellowish, granular, and. sometimes faintly zoned ; on emerging to the surface they spread irregularly in very thin sheets, with notched margins. May be rugose and contoured like Typhoid. Violet pigment begins to show faintly about the tenth day, and the gelatine is softening and liquefying immediately 68 Ward. — A Violet Bacilhis fro7n the Thames. around by this time. Liquefaction may be complete, or soon arrested. Submerged colonies may remain yellowish, or only become faintly coloured : the exposed ones become deep blue- violet, and all stages to white may occur on the same plate. Streak-culture At 150 C. forms a white, opaque, milky streak in 24-48 hours. on gelatine. In three days may show signs of sinking as gelatine softens, and violet hue may begin to appear along the axis. In nine or ten days the gelatine is scooped, and a violet membrane floats on the liquefied mass, in which violet and dirty-white flecks are dis- tributed. More or less complete liquefaction of the gelatine follows, but it often requires many weeks, even at 20° C., for completion throughout the tube. Stab-culture in In three days at 150 C. a small white button, like a drop of gelatine. milk, is formed at the point of inoculation, and minute white points in the tunnel. In ten days or so the button has sunk in a depression of liquefied gelatine, producing a thistle-head funnel. Culture still white. In eighteen days the liquefaction reaches nearly to the walls of the tube, the funnel being lined by a violet membrane. In about four weeks the gelatine is liquefied to'* one-eighth of its depth or so, a deep violet, folded membrane floating on the top and through the liquid. , At 20° C. the same phenomena are observed, but proceed more rapidly at first. It takes many weeks or even months to completely liquefy all the gelatine. Submerged in gelatine : no growth. Agar. Forms a thick white streak in 24-48 hours at 20° C. which turns violet along the axis on the fourth day or later. In about ten days a magnificent corrugated, deep violet membrane is formed, which can be lifted, and shows white growth below. When water of condensation is collected, the submerged growth is white. Potato. In from two to five days, at 20° C., a more or less copious, dirty- white to yellowish patch. Later on the violet colour spreads from the centre, and a corrugated, deep violet membrane is found by the tenth day. But this often does not extend to the edges, and in some cases fails altogether. Broth. Traces of turbidity in two days at 20° C., and by the sixth day very turbid, especially towards the top, where a thick white mem- brane forms. After ten or twelve days, or even earlier, the thicker membrane begins to turn violet and a white precipitate falls in the still turbid liquid, in which are violet and white flecks. Milk. The casein is precipitated as a coagulum, and slowly dissolved. The liquid is alkaline. Later on a violet membrane forms above. Glucose. No fermentation, and no perceptible change at all beyond a very faint turbidity during the first few days. Air require- ments. Aerobic. Ward. — A Violet Bacillus from the Thames . 69 Temperature. Liquefaction. Rapidity of growth. Pigment. Reactions to light. Pathogenicity. The Bacilli are still alive after several days at 350 C., but no growth is maintained; growth occurs at 50 C. The best temperature seems to be near 20° C. Old cultures withstand 50° to 6o° C. for some hours, though no true spores are known. Liquefies the gelatine, but only slowly after the upper parts are fluid, and the deeper parts may be solid months afterwards. Slow. In many cases colonies are found only 10 or 12 mm. in diameter six months after making a plate, the gelatine around having dried up ; they are usually about half that diameter when the gelatine liquefies, and they float in it. Deep violet. Not in the cells, and does not diffuse into the solid media, except perhaps a little on potatoes. Insoluble in water ; very soluble in alcohol, and looks like gentian-violet ; very stable except in sunlight. Turns bluish green on adding NaHO, the original colour almost restored by excess of acid. Acetic acid to original solution makes it paler. Absorbs yellow-orange to green-blue. Easily killed on exposure to direct sunlight. At moderately high temperatures the Bacilli, quiescent in darkness, move actively when illuminated. Prof. Kanthack finds it not pathogenic for guinea-pigs. I have to thank Prof. Kanthack for examining pathologically a large number of these Thames Schizomycetes. Interesting and important results were obtained on reviving No. 2 from an agar-tube which had remained untouched from August 15 to June 9, i. e. ten months. The plate-colonies at 20-22° C. eventually came up white and quite normal, except that they were more tough and membranous than expected, and the needle lifted each colony whole from its liquefying disc. At first they were pure white, as usual ; but in ten days the purple colour appeared in the middle or at the margins, and rapidly spread. Until the purple hue appeared, I was inclined to suspect the plates were not pure, and that the colonies were those of a capsuled Bacillus. The tube-cultures were also normal, and the purple mem- branes of the broth- and milk-cultures appeared on the tenth to eleventh days at 20-25° C., so that no doubt need exist as to the species. At the same time it should be noted that some of the plate-colonies remained white or yellowish-white to the end, and that great variations were exhibited as to the degree of liquefaction and coherence of the colonies. I was 70 Ward . — A Violet Bacillus from the Thames. unable to determine the causes here at work, but neither temperature nor impurity of culture were among them. It is evident that this is a well-marked Pigment-Bacterium, and at first sight one might suppose it not difficult to identify. On looking further into the matter, however, it turns out that several violet or deep blue Pigment-Bacteria are described, some of them very vaguely. So far as I have been able to discover, the following are the forms hitherto introduced into the literature. Passing over Schroter’s Micrococcus violaceus 1, a form described as not liquefying the gelatine, and non-motile — though in other respects it would seem to be very similar to the elliptical cocci got in old gelatine-cultures of the Thames species, and in any case it must be regarded as imperfectly described — there are at least nine or ten alleged specific forms in the handbooks, viz. B. violaceus of Plagge and Proskauer, usually regarded as identical with their B. lividus 2 ; B. janthinum of Zopf3; B. membrauaceus amethystinus of Eisenberg4 ; B. violaceus Laurentius of Jordan5 ; B. violaceus of Frankland 6 ; B. coeruleus , Smith 7 ; B. coeruleus of Voges 8 ; B. violaceus of Mace 9 ; B. berolinsis indicus of Claessen 10 ; and B. indigoferus of Voges n. Assuming these to be all autonomous forms — which, how- ever, is extremely doubtful — I now proceed to examine their principal characters with reference to the Thames form. Frank- land’s species — B. violaceus — naturally suggests itself, seeing that it was also reported from the Fondon as well as the Berlin waters ; but it is noticeable that it was first observed in the Spree. Apart from small differences as regards size and manner 1 Schroter in Cohn’s Beitrage, Bd. i, p. 109 ; and Adametz, Die Bakterien der Nutz- und Trinkwasser, Wien, 1888. 2 Zeitschr. f. Hygiene, Bd. ii, p. 463. 3 Die Spaltpilze, 1885, p. 68. 4 Bakteriol. Diagnostik, 1891, p. 421. 5 Report of State Board of Health, Massachusetts, 1890, p. 838. 6 Zeitschr. f. Hygiene, Bd. vi, p. 394. 7 Medical News, 1887, p. 758. 8 Centralbl. f. Bakt., Bd. xiv, 1893, p. 303. 9 Annales d’hygiene publ. &c., t. xvii, 1887. 10 Centralbl. f. Bakt., Bd. vii, p. 13. 11 Ibid., Bd. xiv, 1893, p. 307. Ward. — A Violet Bacilhis from the Thames, 71 of movement, the morphological characters of Frankland’s species agree very well with mine, except that I have not observed the spores described as occurring in the agar- cultures. It is true that Frankland says they only occur here and there ; but since they bulge out the Bacillus, they ought to be easily seen. However, the authors give no further par- ticulars about these spores, and so it is impossible to form any opinion concerning them. In many respects, also, there are resemblances between the Franklands’ form and mine in the characters of the plate- and other colonies ; but they do not refer to the zoogloea-membrane, and describe the agar- cultures as forming a smooth, bright layer over the surface, and not a corrugated membrane. Perhaps the sum of the differences gives us sufficient char- acters to separate the two forms, and we may say that mine differs from the Franklands’ form in certain morphological characters : in forming no spores, in the marked development of zoogloea-membranes, and in the cultures on agar and potato. It might possibly be that the smooth, bright layer referred to by the Franklands is a young stage, or that they were working with a weaker form. I often find that after being some months in culture this form grows feebly on agar and potato. The B. violaceus Laurentins of Jordan agrees fairly well in size in forming no spores, and in some other morphological characters. The plate-colonies are very different, as are also the stab-cultures — especially as the author insists that no membrane forms above, and the gelatine rapidly liquefies in the tunnel. The cultures in milk and in broth are also markedly different. The B. violaceus of Mace is believed by Mace himself1 2 to be the same as the Micrococcus violaceus of Schroter, and the one found by Bujwid in hail3. 1 I assume that the breadth is 0.7 [i and not 7^', as stated in the original (1. c. p. 838), obviously by a printer’s error. 2 Mace, Traite pratique de Baeteriologie, 1892, p. 541. 8 Bujwid in Ann. Pasteur Inst., 1887, p. 592. 72 Ward. — A Violet Bacillus from the Thames. Zopf’s Bacterium janthinum differs in some characters — size, milk-cultures, &c. — but might very well pass for a feeble form of mine in other respects. Lustig 1 regards Zopf’s form as identical with the one found by Bujwid in hail, and agrees with Jordan2 that it is the same as one previously found by Hueppe. Jordan also points out the resemblances between Zopf’s form and Rosenberg’s Bacterium h.3, as well as the doubts as to what the latter is — possibly identical with Frank- land’s B. violaceus or with Jordan’s B. violaceus Laurentius. Zopf’s B. janthinum has been found often in water, e. g. by Plagge and Proskauer 4, Roszahegyi 6, Jordan 6, &c. The latter also discusses the difference between the three forms last mentioned. As regards Smith’s B. coeruleus , from the Schuylkill, if the colouring-matter is really formed in the cells, as stated, and is insoluble in alcohol, there is no doubt a sufficient difference ; moreover, the pigment is a blue one, and we may safely regard this as a distinct form on the evidence to hand. Claessens’ B. berolinsis indicus , from the Spree, also pro- duces a blue (indigo) pigment, insoluble in alcohol, and many other differences exist, in addition to its not liquefying the gelatine and not rendering broth turbid 7, which suffice to distinguish it. Voges’ B. coeruleus is too stout and short for my form, and again the pigment is a different colour — blue, not violet — and does not form in broth. The agar- and milk-cultures are also decidedly different. Moreover, the colour is soluble in water. There can be no doubt this form is quite different from the one I have isolated from the Thames. Voges’ B. indigoferus is still more different, and especially in its minute size. There remains Eisenberg’s B. membranaceus amethystinus. The differences between this form and mine are so few and 1 Diagnostik del* Bakterien des Wassers, 1893, p. 76. 2 1. c., p. 841. 3 Ueber die Bakterien des Main-wassers, Arch. f. Hyg., Bd. v, p. 458. 4 Zeitschr. f. Hyg., Bd. ii, p. 458. 5 Lustig, 1. c., p. 76. 6 1. c., p. 840. 7 According to Voges, 1. c., p. 302. Ward . — A Violet Bacillus from the Thames. 73 so trivial that I am strongly inclined to regard them as acci- dental. Eisenberg’s description of the plate- and other gelatine-cultures, of the development of the membrane on agar and broth, absence of spores, and so on, are quite like mine ; but the rods are extremely short, and non-motile, though their thickness agrees very well. His description of the potato-cultures also agree with some of mine, though in other cases I find the pigment developed on these also. On the whole, it seems pretty certain that this violet Bacterium from the Thames is, then, the one found by Jolles in water, and named B. membranaceus amethystinus by Eisenberg. Other Pigment-Bacteria, with blue or violet hues, are referred to in the literature, but the descriptions are so incomplete that I cannot compare them. Thus Schroter’s B. Lacmus1, Ehrenberg’s B. syncyanus 2, Beyerinck’s B. cyaneo - fuscus 3, forms a blue pigment in one of its stages, but does not properly come under the head of violet Bacteria. What Jordan’s B. cyanogenus 4 is, and whether it is the same as Hueppe’s5 Bacillus of blue milk, I cannot decide, but both may safely be regarded as distinct from any of the forms referred to, and the pigment is quite different. The same is true of Alvarez’ B. indigogenus* , and so far as I can discover these exhaust the list of blue and violet Pigment- Bacteria. It not unfrequently happens that plate-colonies refuse to colour. I have made numerous attempts to determine the cause of this. Exposure to light, and cultivation at too high a temperature, may bring it about, and in one series I had it in culture for nearly a year as a white or slightly yellowish form which refused to form the violet pigment. Eventually, however, the colour appeared in a milk-culture, and was 1 Pilz. Schles., p. 158, referred to by De Toni and Trevisan in Saccardo’s Syll., Vol. viii, p. 978. 2 Verhandl. d. Berl. Akad., 1840, p. 202, and Saccardo, 1. c., p. 979. 3 Bot. Zeit., 1891, No. 43. 4 Jordan, 1. c., p. 832. 5 Mitth. aus dem Kaiserl. Ges., Bd. ii, p. 355 (see also Heim, ibid., Bd. v, p. 518). 6 Sternberg, Bacteriology, 1893, p. 476. 74 Ward . — ^4 Violet Bacillus from the Thames . regained more and more on transference. In some cases, however, the non-pigmented variety appeared on normal plates, and I was quite unable to refer it to any cause. This white variety would probably repay further study. In any case I think the evidence is against the multiplicity of species of violet Bacteria which now exist in the literature. At the same time, the difficulties of microscopic cultures of the present form, and the lack of information regarding them in other alleged ‘ species ’ or forms, should make us hesitate before we decide as to the autonomy of any, since it may be taken as certain that we do not know the whole life-cycle of even a single member of this type. EXPLANATION OF FIGURES IN PLATE VI. Illustrating Professor Ward’s paper on a violet Bacillus. Fig. i. Rodlets from fresh gelatine ( a ) not stained ; (b) similar rodlets stained with methylene-blue — the ends often stain more deeply than the centre. Fig. 2. Rodlets, &c., from an old Agar culture (a) stained with methyl-violet ; (b) similar preparation from old gelatine-cultures showing ‘ involution-forms.’ Fig. 3. Rods and filaments from a twenty-four hours’ broth-culture. Fig. 4. Plate-colonies on gelatine after ten days’ growth at 20°C., from a culture a year old (nat. size). Fig. 5. Plate-colonies in various stages of development on gelatine : (a) after twenty-four hours at i8° C. ; (b) the same on third day; ( c ) submerged colonies on fourth day at i8° C. All under i obj. Fig. 6. Plate-colonies: (a) after ten days at i8°C., nat. size, showing develop- ment of pigment, and liquefaction ; (b) the non-pigmented variety after three weeks at i8° C. ; the plate is liquefied, and cream-coloured or yellowish colonies are floating in the liquid. Fig. 7. Stab-cultures in gelatine : (a) after three days; (b) after ten days, lique- faction is beginning, but the colonies are still white ; ( c ) after eighteen days, the violet pigment appearing in the funnel ; (d) after a month, the gelatine liquefied some way down. All at ordinary temperatures. Fig. 8. Gelatine-streak after ten days at 15° C. ; the purple growth lies in a groove of liquefaction. Fig. 9. Agar-culture at 20° C. ; (a) after four days, the purple colour appearing in the white, and ( b ) on the tenth day, the white nearly all gone. Fig. 10. Potato-culture, ten days at 20° C., the purple hue has only invaded part of the growth — the rest remains dirty brown in colour. \ /InruxZs of Bo£ajty Fig. 6. WARD. VIOLET BACILLUS. VoL.XlI,Pl.VI. Fwf.7. The Polymorphy of Cutleria multifida (Grev.). BY ARTHUR H. CHURCH, B.A., B.Sc., Jesus College , Oxford: Demonstrator of Botany in the University . With Plates VII, VIII, and IX. SINCE the classical researches of Falkenberg1, carried out at Naples in 1878, in which he showed the necessity of fertilization for the oospheres of Cutleria multifida (Grev.), and established the identity of the product of germination of such sexually-produced spores with Aglaozonia reptans (Kutz.), but little has been done to clear up the mystery underlying his conclusion that Cutleria must therefore present an anti- thetic alternation of generations, in which Cutleria multifida represents the gametophyte, while Aglaozonia reptans is to be regarded as a true sporophyte-generation. Conclusive evidence of such an alternation of generations would be of greatest theoretical interest ; not so much from its being the only example described in the Phaeophyceae of an antithetic alternation at all comparable with the phe- nomena occurring in the Coleochaetaceae and the Florideae, 1 Die Befruchtung und der Generationswechsel von Cutleria: Mittheilungen aus der Zoolog. Stat. zu Neapel, vol. i. 1879. [Annals of Botany, Vol. XII. No. XLV. March, 1898.] 7 6 Church. — The Polymorphy of which are regarded as possibly representing evolutions of alternation, in existing groups of Green and Red Algae, parallel to that which has reached such a degree of com- plexity in the Archegoniatae ; as from its presenting, within the limits of a very narrow alliance, a sudden transition to a true alternation from the simpler life-history of the homotype genus Zonardinia . Falkenberg showed, beyond all doubt, that Aglaozonia plants were normally the ultimate product of sexual repro- duction in Ctitleria midtifida ; but, owing to the death of his plants, the question as to the subsequent relations of the Aglaozonia to Cutleria — whether the latter arose from spores, or was the result of merely vegetative reproductive processes — was left open ; as was also that of the strictness of the alternation on the side of the perennial and more widely distributed Aglaozonia. Although the following observations, made at the Marine Biological Laboratory at Plymouth, cannot be regarded as finally solving the problem, it is hoped that they may con- tribute to a more complete acquaintance with the life-history of these plants. Both Citleria and Aglaozonia grow in the estuary of the river Yealm, near Plymouth, at 2-3 fathoms below low-water mark, and may easily be obtained, at all states of the tide, by dredging. Cutleria grows as a summer annual, reaching its maximum development in July and the beginning of August. It rapidly diminishes in quantity in September, and has completely disappeared by October. Aglaozonia , on the other hand, is perennial, growing on stones and shells, especially oyster-shells, in the same locality and depth, reaching its finest development in October and November. Poor in quality during the winter-months, possibly owing to its being eaten by Mollusca & c. in the absence of other vegetation, it recovers in the spring, and bears reproductive sori in March and April. It is also usually in poor condition throughout the summer. From these data, it would appear, therefore, that in English waters Cutleria is a rapidly- 77 Cutler ia multifida (Grevl). developing summer sexual plant, Aglaozonia a slow-growing perennating winter form ; and that these two growth-forms had become complementary in structure and habit, as well as in reproductive functions. A consideration of the conditions under which these two plants live in the Bay of Naples shows, however, that this does not represent the whole truth. The majority of the most interesting of the summer annuals growing in shallow water on the southern shores of England are, in the Bay of Naples, early spring- and even winter-plants ; while others, on the other hand, retire to deeper water. A few examples will make this clear : thus, Cutleria at Plymouth grows in company with abundant Spor'ochnus, A rthrocladia^ Stilophora , Asperococcus bullosus , Dictyota , and such Florideae as Du- dresnaya coccinea and Scinaia furcellata ; all these forms reproducing freely in July and the beginning of August, and growing in 2-3 fathoms of water at a temperature of i8°C. From data given by Falkenberg for the Bay of Naples, it appears that Sporochnus and Arthrocladia flourish also there in July and August, but at a depth of 20 fathoms ; Stilophora follows the Cystoseira zone at a slightly less depth and grows in early summer ; Asperococcus bullosus also in spring and summer, varying from 1-8 fathoms ; Dudresnaya at 2 fathoms in March and May ; and Scinaia at the same depth from February to June, although the latter has been dredged at Messina in July at 1 5 fathoms. Of the plants with which Cutleria grows in English waters, therefore, some retain at Naples the same annual period, but live in far deeper water ; but more generally the depth of water remains fairly constant, while the season of the year is changed. There can be little doubt that the two determining factors of external environment are temperature and intensity of sunlight, with the latter being associated a greater degree of purity in the water, which in the Bay of Naples allows vegetation to flourish as far out as 40 fathoms, while, in the immediate vicinity of Plymouth Sound, only scanty traces are met with at even 10 fathoms. Cutleria 78 Church. — The Polymorphy of at Naples follows the rule of the majority of its English associates and vegetates in shallow water from December to April, vanishing, like Dictyota in the Mediterranean, on the approach of summer. It is important to note that while Aglaozonia is also perennial in the Bay of Naples, Cutleria is the winter-form, completely disappearing by April, its existence being apparently terminated by a rise of temperature, instead of by a fall as on the English coasts. It is clear, therefore, that the vital capacities of the sexual plant towards temperature are much more limited than those of the asexual Aglaozonia , which is perennial, not only in the warmer waters of the Mediterranean summer, but in the cold waters of the North Atlantic and North Sea winter. If now we compare the geographical distribution of the known species of the Cutleriaceae l, we find that the order belongs naturally to the warmer seas. Thus, omitting the doubtful C. Laminaria , Kutz. of the Mediterranean, the group consists of two little known species, C. pacifica from Samoa, and C. compressa from La Guayra ; of C. adspersa of the Mediterranean district only2 (Cadiz to Suez); of Zanar- dinia collaris , Mediterranean, West Indies, Polynesia, the Atlantic shores of Europe as far as Brest (Crouan). drifted specimens at Jersey (Harvey); and of C. multifida , also a Mediterranean and Atlantic type3. But this last, alone of the group, extends northwards to England, Shetland Islands, and the coast of Norway to Nordland (Kjellman) ; on the other hand, it is poorly represented in the North Sea district and absent in the Baltic. That is to say, the north- ward distribution of the sexual form appears to be limited 1 De Toni, Sylloge Algarum, vol. iii. p. 300. 2 Sauvageau (Journ. de Bot. 1897, p. 177) since gives C. adspersa and an undetermined Aglazonia, but neither C. multifida nor Zanardinia , as being abundant in winter, at low-tide mark, in the Gulf of Gascony ; and Mr. Batters informs me that C. adspersa is found at Brest. 3 A doubtful Polynesian form of Aglaozonia described as Zonaria parvula var. duplex (Heydr. Beitr. Algenfl. v. Kais. Wilh. Land), placed by De Toni under C. multifida , might more possibly belong to C. pacifica . 79 Cutleria multifida ( Grev .). by the temperature of the northern summer, and in the English Channel we are already beyond the natural home of the Cutleria family. Parthenogenesis of Cutleria. The first recorded specimen of C. multifida was picked up after a storm on Yarmouth beach by Dawson Turner on August 3r, 18041. It was a female plant, covered with oogonia as the date would suggest, and was described in Smith and Sowerby’s English Botany in 1805, under- the name of Ulva multifida (No. 1913)- It appears as Zonaria multifida in Agardh’s Sp. Alg. 1824, and as Sporochnus multifidus in Sprengel’s Systema Vegetabilium of Linnaeus, in 1825 ; it received its modern title in 1830, from Greville2, who formed for it a new genus, named in honour of Miss Cutler of Sidmouth. Greville, also, knew only the female, or as it was considered, the sporangiate plant. Antheridial plants were described later by Dickie, but these were very rare, and Harvey in his Phycologia Britannica (1846) mentions that he had never seen more than one such plant, which had been sent him from Sidmouth. At this time no sexual significance had been attributed to the reproductive cells of Algae, or these antheridia might have been a source of difficulty ; but they were commonly regarded as imperfectly formed swarming cells which were consequently destined to remain sterile 3. The first definite statements with regard to the emission and germination of the spores were made by Thuret4 in 1850. He observed the discharge of the oospheres in the early hours of the morning, as also their active movement and strong positive heliotropism by means of which they rose to the surface of the water. In all cases germination was direct ; the 1 Harvey, Phycologia Britannica, i. 33. 2 Algae Britannicae, p. 60. 3 Cp. Nageli, Bot. Zeit. 1849, p. 569. * Ann. Sci. Nat. iii. 14, p. 32. 80 Church. — The Polymorphy of oospheres came to rest, the pointed end grew out to form a rhizoid, the body of the spore giving a brown filament of a few cells. A number of female plants, kept in a vessel of sea-water, continued to give off oospheres for several successive days, which in all cases germinated perfectly without the admission of antherozoids. Thus, although Thuret was fully satisfied as to the necessity of antherozoids for the fertilization of the oospheres of Fucus , he concluded that no act of fertilization took place in the Cutleria spores he had under observation. It is also of special interest to note that he found antheridial specimens to be extremely rare at Saint Vaast-la-Hogue, where these researches were con- ducted ; he states that he often collected from the oyster-beds there, where Cutleria grew in profusion, over a hundred female specimens before finding one male ; and he points out that this rarity of antheridial specimens not only agrees with what Harvey had stated to be the case in English waters, but would to a certain extent militate against the view that antherozoids possessed sexual functions of such importance to the plant. In the summer of 1855, the brothers Crouan1 repeated these observations at Brest, and came to identical conclusions with regard to the perfect parthenogenesis of the oospheres. At the same time, they noted a peculiar phenomenon in connexion with the fate of the antherozoids. These at first rose to the surface of the water, forming an orange film on the side nearest the light, in the manner typical for all swarming cells of the Brown Seaweeds ; but on coming to rest, they became agglutinated by their gelatinous membranes into a pseudo-tissue mass of a brown colour, which was even capable of being sectioned. They therefore concluded that the antherozoids were non-sexual, but still possessed a certain degree of germinative capacity. It is so far clear that to the older observers who worked on the French shores of the Channel, the constancy of the germination of the oospheres Bull. Soc. Bot. France, ii. p. 644. 8 1 C tit l evict multifida ( Grev .). was so apparent that the question of the non-sexuality of Cutleria was never in doubt. The converse was however asserted by Reinke 1 in carrying out his researches at the Naples Station in 1875-76. He confirmed Crouan’s observations on the peculiar pseudo-tissue formation of the antherozoids, but attributed to it no real germinative significance, since in all his experiments, anthero- zoids and oospheres, isolated from each other, constantly underwent no further development. On the other hand, in vessels containing both male and female plants, germination took place freely, and actual fertilization by the antherozoids was observed. From these facts he deduced the perfect sexuality of C. multifida and the essential importance of the antherozoids ; as also that Thuret’s observations must have been due to an accidental parthenogensis. It is inter- esting to note that he gives male and female plants as occurring in the Bay of Naples in the ratio of three male to two female. Similarly Falkenberg2, in 1878, described male and female plants as being about equally abundant in the Bay of Naples, and carrying out his experiments with great care in obtaining pure cultures of emitted oospheres and antherozoids, he fully confirmed Reinke’s results. Moreover, as his cultures were free from extraneous growths of Diatoms, &c., which had ultimately induced pathological conditions in Reinke’s cul- tures, Falkenberg succeeded in developing the germinated embryos to a considerable size. In all his experiments, antherozoids became immotile and useless in twenty-four hours, and then died ; oospheres retained the capacity for fertilization for four or five days, but never commenced segmentation; fertilized oospheres germinated directly and rapidly; while unfertilized oospheres never got beyond the formation of a thin cell-membrane. Further, Janczewski 3, at Antibes in 1883, showed in the case of C. adspersa , which is also a Mediterranean spring-plant, 1 Nova Acta der K. L. C. Deutsch. Akad. xl. 1878. 2 Loc. cit. 3 Ann. Sci. Nat. vi. 16. p. 210. G 82 Church. — The Polymorphy of that neither male nor female sexual Cells present any of these curious suggestions of direct development, but that both oospheres and antherozoids die the same day if copulation does not take place. During the summer of 1896, male plants were extremely rare at Plymouth, only two or three being seen, although female plants were dredged in considerable quantity. In 1897 the same proportion obtained : thus, in a dredging taken on August 11, a score or so of very fine female plants were collected, but only one male. Some of these were placed in a vessel of filtered water with the object of obtaining embryos, but owing to the heat or some other cause, the plants all died ; nor was it until the end of the month, when cold and wet weather set in, that freshly dredged material could be kept alive more than a day or two. At this time and onwards, all the plants obtained were female, no more male plants being seen for the year. With the object of testing Thuret’s observations on direct germination, a number of female plants were on August 20 placed in filtered water (temp. 1 8° C.) standing in a north window. In the course of three weeks, the water having only once been changed, the fronds were found to be sprinkled all over with innumerable young plants, which by September 16 presented unmistak- able Aglaozonia characters (Figs. 14-21). Although the immediate proximity of such numbers of these young plants to the sori of oogonia suggested at once the direct germina- tion of oospheres which had lost their motility soon after discharge, it was quite possible that fertilization might have taken place before collection. More plants were accordingly collected in September, and washed and placed in filtered water. In a week the surface of the vessel was covered with thousands of germinating oospheres which had risen to the surface in virtue of their strong positive heliotropism ; of these, the majority at least must have been parthenogenetic, as it is evident that any few antherozoids, which might have survived collection and washing on the female plants, would not have sufficed for such a multitude of oospheres. 83 Cutler ia multi ft da ( Grev. ). On adding a fresh supply of filtered water, another week gave a second similar crop of germinating oospheres, and a third week yet another, thus confirming Thuret’s original observations. Several distinct cultures, some containing fragments of female thallus, still producing oospheres, which had been growing in filtered water in the laboratory since August, others containing freshly discharged oospheres only, were made towards the end of September. In all cases germination proceeded directly and quite normally, although slower than in the case of the first crop obtained in August, and far slower than in the experiments of Falkenberg, who states that his fertilized oospheres produced a plant of 3-4 cells in the first twenty-four hours, whereas the Plymouth cultures in September did not do more than this in the first week. It is probable, however, that this rate of growth varies directly with the temperature. Finally, separate cultures, from small pieces of female plants collected on September 21, were made on October 12, and brought to Oxford and kept in a sunny window. In all cases germination again took place normally ; in three days sufficient oospheres had collected on the side nearest the light to form a visible film. The majority of the oospheres were covered with a well-marked membrane and many had already put out the first rhizoid. The temperature was low (14°) and the weather dull, but after two days of bright sunshine the plants increased to about five cells and a long rhizoid, and by October 25 they formed well-grown embryos in which segmentation was rapidly proceeding (Fig. 13). At the end of three weeks (November 1), the culture- vessels having latterly been standing in bright sun for a few hours every day, an immense number of young plants in all stages of development were to be seen, the small piece of thallus in the culture continuing to give off oospheres. The oldest plants showed the * foot-embryo ’ now at its maximum size, but with so far no formation of dorsiventral lobes (Fig. 14). That is to say, the germination of these unmistakably 84 Church . — The Polymorphy of parthenogenetic spores had proceeded at a rate equal to and with results in a given time identical with those observed in the first culture of August 11, and in which the possibility had not been eliminated that fertilized oospheres might have already become attached to the plants before they were gathered. That oospheres did this in the natural state was observed on specimens dredged in September, but it is clear that continued crops of free-swimming oospheres, germinating at the surface, were beyond suspicion. It is also of interest to note that the old plants which continued to give these crops of germinating oospheres had been, since the middle of September, in a rapid state of disintegration, and by November 1 were but partial skeletons compared with the perfect summer-plants ; nor, at this time, would they have been found by dredging. No Cutleria was dredged at Plymouth in 1896-97 after the middle of September, the plants then evidently decaying and easily losing their point of attachment. The general result of these observations, therefore, is not only to confirm the original observations of Thuret and Crouan, made on the opposite shores of the Channel, as to the absolute constancy of parthenogenetic development of the oospheres at the end of the summer ; but, bearing in mind the equal constancy of fertilization observed by Reinke and Falkenberg in early spring at Naples, it further leads us to correlate the apparent contradiction of these observations with the fact that the conditions of external environment are so widely different in the case of plants growing in the Channel and in the Bay of Naples respectively ; and further to suggest that the parthenogenesis of the Channel plants may be due to the fall of the temperature of the sea at the end of the northern summer, which, by diminishing the sexuality of the oospheres, causes the plant to become an asexual form by degeneracy, although morphologically retaining the distinction of sex. Cutleria multifida ( Grev .). 85 Germination of the Oospheres. In all cases, whether in later undoubtedly parthenogenetic cultures or in the earlier ones only doubtfully so, germina- tion proceeded along lines absolutely identical with those described by Falkenberg for the sexually-produced spore. The spore secretes a cell-membrane, becomes pear-shaped, and divides into a shoot-cell and a first rhizoid (Fig. 11) ; the latter elongates and reaches a considerable size if germination takes place at the surface of the water, but remains short on contact with any foreign body. The shoot-portion of the plant gives rise to a filament of a few cells only (6-10) by intercalary rather than apical segmentation (Fig. 12), and then definitely ceases to elongate ; this being, according both to Falkenberg’s and the Plymouth experiments, all that remains in this type of germination to mark the primitive filamentous condition of the Cutleria (Fig. 12, one week old). Irregular segmentation commences immediately through- out the young plant ; any and ultimately every cell dividing repeatedly by walls in different quadrant-planes, until the embryo becomes a more or less club-shaped multicellular mass of tissue, attached by one extremity and still exhibiting radial symmetry (Figs. 12, 13, 14). To this stage Falkenberg has given the name of the c Foot,’ and it is probably representative, both phylogenetically and ontogenetically, of a primitive thalloid condition in which the main axis of the plant was radially symmetrical and segmented behind the growing region in the regular manner seen in such a form as Stypocaulon. When well-developed, the foot may form a well-marked tissue-mass (Fig. 14) ; but it is often, and this was more general in some cultures than others, to a great extent abbreviated in development, ultimately giving rise to an embryo which was practically dorsiventral throughout (Fig. 16), and identical with the oldest embryos obtained by Janczewski1 in C. adspersa . 1 Loc. cit. p. 220. 86 Church . — The Polymorphy of At one or more points in the ‘ Foot,’ any single super- ficial cell may initiate a new growth (Fig. 15), which, by successive T-shaped walls, gives a lobed outgrowth which exhibits dorsiventral symmetry, and by laying down the marginal segment-walls preferentially in a radial vertical plane, assumes a fan-shaped outline, the commencement of an Aglaozonia disc (Figs. 15, 19, 20). The formation of these lobes appears to be mainly due to the stimulus of contact, and thus a majority form discs at the point of attachment (Figs. 19, 22); but if the apex of the ‘Foot’ bends over, a symmetrical outgrowth may take place there, either alone or in addition to another at the base (Figs. 17, 18, 20). In the case of the foot lying more or less prostrate, several (6-8) distinct lobes may be produced, which develop rhizoids on the side towards the substratum (Figs. 20, 21). In plants which have become detached, the dorsiventral lobes con- tinue to be formed and exhibit a tendency to curl up, indicating a return to radial symmetry comparable to that of the proliferating ‘ cups ’ of Z anardinia. Although many distinct cultures were made, and hundreds of embryos observed, in no single case was any further development noticed in the ‘Foot’; the dorsiventral lobes slowly but steadily increasing along definite Aglaozonia lines. Following Falkenberg, this type of plant may suitably be distinguished as the Foot- Embryo. It will therefore be noted not only that these observa- tions on the development of the Aglaozoma-ths.\lus from oospheres of Cutleria absolutely confirm those of Falkenberg, but that such confirmation was necessary, since the embryos observed by Thuret1 at Saint Vaast-la- Hogue in 1850 were unmistakably different : it was in fact the figure given by Thuret of a free-growing filament of thirty-six cells with branches towards the base, which appeared, as being a vegetative growth homologous with an adventitious branch, to confirm his assertion of the non-sexuality of the oospheres. 1 Etudes Phycologiques, and Ann. Sci. Nat. iii. 14. 8 7 Cutleria multifida ( Grev .). Not only were the Plymouth plants truly parthenogenetic, as opposed to Falkenberg’s truly fertilized ones, but they were grown in the autumn months, whereas Falkenberg’s were grown in the spring ; the only factor in common therefore appears to be this, that in either case the spores were obtained from mature plants about to die, from summer- heat in the latter case, but from winter-cold in the former. Germination of Zoospores of Aglaozonia. Aglaozonia plants were first described by Greville1 in 1828, from specimens found, appropriately enough, by Miss Cutler at Sidmouth, growing at low-tide mark on exposed sand- stone rocks ; these sterile plants being placed as a new species in the genus Zonaria of C. Agardh under the name Zonaria parvula . Later Greville2 founded a new genus, and changed the name to Padina parvula ; and in 1833 similar sterile plants found by Crouan3 at Brest were dis- tinguished as Padina rep tans. Reproductive organs were first found on Skagerack specimens by Areschoug in 1843, and the genus refounded as Padinella. Areschoug’s plants were very small, and possibly dead before examination, as his figures 4 are quite misleading 5. The genus Aglaozonia was ultimately established by 1 Crypt. Flora, t. 360. 2 Alg. Brit. 1830, p. 63. 3 Florule du Finisterre, p. 169. 4 Linnaea, 1843, p. 260. 5 Areschoug obtained his plants on oyster-shells at Koster, and was satisfied that they were identical with Sidmouth specimens described by Greville. His drawing appears to have been made from a squeezed-out sorus, rather than from a section ; and the appearance which it presented induced Reinke to revive the old name of Zonaria parvula for a plant he obtained at Naples in 1875 (Nova Acta, xl, No. 1. p. 34), which was of distinctly Dictyotacean nature. Reinke’s plant differs fundamentally from Aglaozonia in the structure of the thallus, the well-marked ‘ tetraspore/ and, above all, in the embryology, which is again that of the Dictyotaceae. It is quite obvious that Greville’s Sidmouth plants were Aglaozonia , as they still grow there abundantly, and Areschoug had received specimens from that locality ; but it is not clear why Reinke’s distinctly Dictyotoid plants should be classed as Cutleriacean by De Toni (Sylloge Algarum, Fucoideae, p. 234). 88 Church. — The Polymorphy of Zanardini1, and Kiitzing, in his Species Algarum (1849). gives both Aglaozonia parvula for the English and Mediterranean plants, and Aglaozonia reptans for Crouan’s specimens. The sporangia and the emission and asexuality of the zoospores were correctly described by the brothers Crouan 2, at Brest, in 1856, from the large quantities of material they found thrown up by a storm on April 5 of that year, while the first correct drawings were given by Zanardini 3 in 1 860. Since then the plants have been known as Aglaozonia reptans , it being clear that Crouan’s specimens were not only identical with those found elsewhere, but were the first on which the reproductive organs were definitely observed. The discs of Aglaozonia are perennial, and are distributed from the Mediterranean along the Atlantic shores of Western Europe to the coast of Norway, being much more abundant and more widely distributed than is Cutleria along the Norwegian coast4. Again, they are more general than Cutleria in the North Sea, and are found abundantly in the more northern portion (Berwick) where Cutleria is unknown ; and finally, they penetrate into the milder climate at the entrance of the Baltic, and are moderately common in the Skagerack5, where Ctitleria is very rare, or only found as very young specimens. Aglaozonia reproduces in the Mediterranean in late autumn, in the Channel in early spring, and it would appear that Areschoug found his Swedish specimens in reproduction during the summer months. On March 29, 1897, shells bearing fine plants of Aglaozonia with reproductive sori were dredged in the river Yealm, Plymouth. One or two sori were carefully removed, placed in a glass dish of filtered water, and allowed to stand in a window exposed to a north light. Zoospores were set free in great numbers, and rising to the surface, swam towards the side nearest the light, forming in a day or two 1 Saggio di classificazione nat. delle Ficee, 1843. 2 Bull. Soc. Bot. de France, 1857. 8 Icon. Phycolog. Adriatica. 4 Kjellmann, Handbok Skand. Hafsalgflora, 1890, p. 17. 5 Gran, Algenvegetationen i Tonsbergfjorden. 89 Cutleria multifida ( Grev .). a distinct brown film. Germination, as already described by the brothers Crouan, took place immediately, and with considerable rapidity (Fig. i). As in the case of the oospheres of Cutleria , the zoospores came to rest, the anterior end became attached to the sides of the vessel or to another plant, and grew out into the first rhizoid. In the case of free-floating spores the rhizoid elongated considerably if it did not come into contact with anything, but ceased to elongate further on contact (Fig. 2). A simple filament of 3-6 cells was formed in a few days, and this agrees with the rate of germination observed for the oospheres of Cutleria grown at an approximately equal temperature. As growth proceeded, the film stretched over the surface of the water in the vessel, forming a pure culture of ger- minating spores, from which portions could be readily trans- ferred to other vessels of water similarly filtered by a Berkefeldt filter. Beyond keeping the vessels covered, to prevent evaporation and the entry of dust, no further change was made ; the best results, in fact, being obtained from the original culture in which the water remained unchanged for over a month. The filaments continued to elongate, by intercalary rather than apical growth, but the characteristic cessation of growth observed in the foot-embryo of Cutleria did not set in ; steady intercalary growth enabling the filaments to double their length each week (Figs. 5, 6). In the second or third week, differentiation in the cells of the filament became marked. The cells in the basal region of the plant increased in bulk and commenced segmenting irregularly by walls in different planes, thus rendering a lower region of the embryo multicellular by the same quadrant- walls, and at about the same age, as in the segmentation of the foot-embryo (Figs. 5-9). This basal multicellular region is therefore homologous with the foot itself, but the embryo differs in that the filamentous terminal portion goes on growing by a definite intercalary zone (Figs. 7- 10). 90 Church.— The Polymorphy of A further exaggeration of the basal segmentation resulted in the formation of a small irregular attachment disc (Fig. io); but the latter exhibited no immediate tendency to extend into dorsiventral lobes, the main energy of growth being, in this embryo, clearly localized . in the filamentous portion. This continued to grow, throwing out branches above and rhizoid attachment-hairs below. In the case of plants growing on the sides of the vessel, the filaments showed a tendency to attach again at any point in their length, sending out rhizoids, and initiating a new intercalary zone of growth above each such attachment : but it is possible that this may be an abnormal result of cultivation, as the same tendency can be observed in cultures of old Cutleria plants, where the reproductive filaments elongate and attach themselves to the sides of the vessel by rhizoids and bear gametangia at irregular intervals ; the filamentous portions of the adult Cutleria are, in fact, still in the condition of these filamentous embryos. It will be seen that even these young plants present the majority of the essential characters of the Cutleria- thallus : there is, for example, the same intercalary growth of a filamentous apex, with irregular segmentation behind the growing-point leading to a multicellular condition, and the same throwing-out of branches of similar growth and of attachment-rhizoids to supplement the primitive holdfast. The only point lacking is the aggregation and fusion of the branches behind the growing-points to the peculiar fasciated thallus of the adult Cutleria. At the beginning of May, a culture of these young plants, now a month old and forming tufts of actively assimilating filaments, was taken to Oxford and kept under observation in a shaded situation in a south window of the Botanical Laboratory. The plants continued to live and assimilate vigorously, forming a bright brown woolly growth of Ecto- carpus- like filaments in the unchanged water, and maintaining their position at the surface in virtue of the gas-bubbles evolved. In still water, this phenomenon affords the surest 9i Cutleria multiftda ( Grev .). test of the health of a culture, death rapidly ensuing if the plants once sink and are unable to again raise themselves. In no case was any further advance made in the formation of the adult Cutleria^ thallus ; the filaments in some cases showed the rope-like aggregation characteristic of the main branches of many Ectocarpus- forms, but no fusions to a pseudo-tissue took place. The filamentous mass increased in bulk for over a month, but after that the plants began to be sickly, and by the end of June the whole culture was undoubtedly so, and portions of it commenced to die off. Before dying however, in July, the plants produced multi- locular reproductive organs in great abundance throughout the culture, which, on maturity, proved to be unmistakable antheridia of Cutleria (Fig. 3). In the same culture (the only one which reached this stage, for all the plants left at Plymouth died at an early date) many of the young plants had, in addition, thrown out Aglaozonia- lobes from their attachment discs, and some of these fully equalled in extent a two months’ old Aglaozonia grown from Cutleria- oospheres (Fig. 4). Although abnormal conditions may have led to patho- logical results, it was undoubtedly shown that Aglaozonia- zoospores, under certain conditions, not only give rise to a Protonematoid stage of Cutleria , which on impoverishment and exposure in a sunny window in summer became pre- cociously antheridial, but that they may, on the other hand, produce the Aglaozon ia- fo r m again, and thus the antithetic character of the alternation would fail to be established. As already indicated, these observations are still incomplete, since the observation of the development of the mature assimilating thallus of Cutleria has yet to be made ; but this is not absolutely essential, since a filamentous plant bearing oogonia, but presenting even fewer of the vegetative characters of a Cutleria , in that it was almost a constantly uniseriate filament throughout, has already been described by Kuckuck under the name C. multifida var. confervoides1 . 1 Wissenschaftliche Meeresuntersuchungen, Biolog. Anst. Helgoland, 1894, i. p. 251. 9 2 Church. — The Polymorphy of Kuckuck’s plants came up spontaneously in the tanks of the Heligoland Laboratory in the summer of 1893, and grew as short filaments attached to stones which had been collected in the North Haven in fairly shallow water (1-3 fathoms). Similar plants were found in reproduction in July, forming brown Elachista- like tufts on Plocamium , and sterile plants also as late as December. Normal Cutler ia is said to have been gathered at Heligoland by Wollny, but has not been known to occur there since ; and although Kuckuck appears to infer that his plants reproduced their like, it is quite probable that they had all sprung from Aglaozonia-spores, and owed their late and feeble development to the cold spring and early summer of the North Sea; and that thus unfavourable conditions had led to a vegetative degeneration similar to that observed in the Plymouth cultures. But if they had been reproduced from oospheres similar to those they bore, the confirmation of the development of a protonematoid embryo from a Cutleria- form would be of still greater theoretical interest, as confirming Thuret’s original observation, and thereby assisting in the demolition of the theory of inherent necessity of an alternation of growth- forms. Seasonal Dimorphism. From the preceding considerations it is obvious that the polymorphy of Cutleria presents little in common with the antithetic alternation of primitive gametophyte and nursed sporophyte of the Archegoniatae ; and still less with the case of Coleochaete and the Florideae, in which the origin of what in these forms is generally regarded as a sporophyte may be sought in polyembryony. From the homology of the Aglaozonia-thaMus with other asexual Algae such as Battersia or even Laminaria , Aglao- zonia has as much claim to be regarded as theoretically a gametophyte as any other Alga. It might even be urged, 93 Cutler ia multi ft da ( Grev .). that, as It has not only a wider geographical, but higher tide-mark distribution, in virtue of its greater power of resistance to extremes of temperature and wave-action, and is moreover perennial and capable of reproducing its like, Aglaozonia has even a better claim to be regarded as the most important of the two forms, and therefore more entitled to be regarded as the phylogenetic and theoretical gameto- phyte than the delicate, sexual Cutleria- shoot itself. To this view, however, there are serious morphological objections. In comparing the development of the two stages, it becomes evident that not only is there no special evidence of alter- nation which could be included under a theoretical alterna- tion of generations, but that even the polymorphism is less evident than at first sight appears. It is important to note that this polymorphy originates only in the embryonic history, leading to the formation of the embryos designated the Foot-Embryo and the Protone - matoid Embryo respectively; and it is in this that its impor- tance lies. Thus, as far as present data go, the dorsiventral Aglaozonia cannot recreate the erect plant vegetatively, nor can the adult Cutleria reproduce from the base of its main axis, clothed with rhizoids to form a secondary holdfast, the dorsiventral basal lobes ; although a longitudinal section of the attachment-disc shows that at the extreme base it never gets beyond the simple segmentation of the foot (Fig. 10). While the protonematoid embryo is clearly on the way to a true Cutleria- form, it is the foot-embryo which is aberrant in development, in that it presents an anomalous cessation of terminal growth at an early period ; and thus the cause of polymorphy may possibly be sought in the solution of the problem as to what induces this arrest of a free-growing axis : that is to say, — Have we to do with the influence of environment on the germinating spore itself, or does it act upon the parent organism? Now, in com- paring the April cultures of Aglaozonia with the September cultures of Cutleria , at Plymouth, it is difficult to see what 94 Church. — The Polymorphy of marked differences of light or food-supply there can be at the two equinoxes. Thus, the two stages of germination were carried out in similar vessels, in the same window, at a laboratory temperature varying from i2°-20°, although the average was nearer 140 for the spring and 180 for autumn, and in similar water, in which, as it was not changed, similar gas-conditions must have obtained. Apparently the only factor which varied was that the Aglaozonia-p\a.nts had been resting throughout the winter at a temperature of 8°~9°, and were in April reproducing on the spring rise to 12°, while the Ctitleria- plants were vegetating and reproducing throughout July and August at 180. But although they continued to discharge oospheres on into the autumn at 1 4°, and possibly at less than 12° in November, the par- thenogenetic oospheres all gave true foot-embryos, without exception. As far as present data go, therefore, it would appear that inherited characters may play a certain part; but it has not yet been established that heredity has attained such a degree of importance in the life-history that any alternation is inevitable, much less sexually beneficial. Sum- ming up these data, it remains shown that : — ] . Cutleria oospheres, whether fertilized in the Mediterranean (Falkenberg), presumably or actually parthenogenetic in the autumn in the English Channel, developed a foot-embryo, which resulted in definite Aglaozonia- thallus and nothing beyond. 2. Aglaozonia zoospores produced a recognizable Cutler ia- form, presenting all the essential characters of a Cutleria- thallus with the exception of the fasciation of the branches (protonematoid embryo —C. multifida var. confervoides Kuckuck), — and under adverse conditions (since the plants died) producing antheridia in great numbers ; but also true Aglaozonia- discs. 3. Cutleria oospheres, germinated parthenogenetically by Thuret, under conditions not described, gave a true protonematoid embryo, which if it had lived would have undoubtedly given a Oitleria plant. 95 Cutler ia multifida ( Grev. ). And added to these, that : — 4. Janczewski, at Antibes in 1883, germinated a true foot- embryo from fertilized oospores of C. adspersa , and this in subsequent development became a dorsiventral structure, closely comparable with many of the Plymouth embryos (Fig. 16). (A special point of interest attaches to Janczewski’s cultures, since in them the arrest of terminal growth which forms the special characteristic of the foot-embryo is not complete ; the young foot-stages being figured as bearing a terminal filament with an intercalary zone of growth.) 5. Reinke, at Naples in 1876, obtained protonematoid em- bryos, identical with Thuret’s Cutleria-z mbryo, from both zoospores and fertilized oospores of Zanardinia collaris. (These stages, which were the most advanced obtained by Reinke for members of the Cutleriaceae, were also distinguished by presenting no trace of the tissue-fusion necessary to form the true adult thallus. They further differ from the Plymouth cultures very considerably in their rate of growth ; a definite protonematoid embryo being only obtained by Reinke in three months, while one month at Plymouth gave the furthest vege- tative stage observed. Nor was any trace whatever noticed of the peculiar phenomenon, suggested to be pathological, which Reinke describes and figures under the name of ‘secondary spores,’ in the germination of Zanardinia , Cutleria , and Aglaozonia). 6. Reinke also found the protonematoid embryo of Zanar- dinia growing in its natural habitat, and, as already indi- cated, the protonematoid embryo of C. mtdtifida has been described as reproducing naturally at Heligoland (Kuckuck). 96 Church. — The Polymorphy of The Relation of Cutleria to Physical Environment. i . Means of Dispersal. From evidence derived from floating bottles, it is clear that anything that will float will be carried indefinitely along the lines of currents and prevailing winds. When it is borne in mind that the spores of Cutleria and Aglaozonia will germinate freely on the surface and float for at least the first month of their existence, and that the film of germinating spores may be at least equal in area to the plant producing them, it is clear that these plants must have practically unlimited powers of dispersal, and that their presence or absence at any given spot must be solely deter- mined by the conditions of external environment. 2. Relation to Temperature. Of the factors of external environment which influence the growth of marine Algae, temperature, light-intensity, the transparency of the water and velocity of current, the first- mentioned is the one most easily measured ; and beyond small daily and local variations on the grand annual curve this is so remarkably constant that it is probable that the sensitiveness of marine vegetation to temperature will be found to lie within far narrower limits than in the case of subaerial vegetation. Thus, although seasonal changes are strongly marked, the extreme annual range which at Naples is 20°, is only 12° at Plymouth, less than 8° at Shetland, and as little as 6° at points along the east coast of Scotland (Isle of May). The maximum temperature is found at the end of August, the minimum in February ; the sea thus undergoing a steady and rapid rise in early summer, and a rapid fall in late autumn. The former is accompanied by a great amount of light-supply during the summer solstice, the latter by a great diminution in light-intensity towards co W P^ £ H < p4 w PL< s w H ■ W u 2 P4 o CO fe O W PQ < H Literature Berthold. Berthold. Admiralty Chart. Berthold. Admiralty Chart. Whitley. Shore waters in summer approximate higher numbers, the lesser being for open sea. Sea temperatures rather than shore water. Monthly averages. Dec. 12 to 9 *>. ib Nov. rOO *-• 00 lb S' 3 2 M CO 00 Oct. Tj- o N l8— 22 CO Sept. VO O -'+* Cl Cl VO CO «b Aug. rj- o 00 o Cl 4-» Cl vO ci 0* VO Cl do Cl VO 00 * to O N lO Cl Cl Cl M VO « ON to ON 7 »o 'g 7 o April 00 O O ON lb Mar. 2 on o 7 oo 00 Tt- $ 6 to 8 lO CO CO 7 cq 8 7 to 8 to n to O 3*4 Plymouth . . Naples. . . . Adriatic . . . Yarmouth . . TABLE OF SURFACE-TEMPERATURES (C°). Jan. Feb. Mar. April May June July Aug. Sept. Oct. Nov. Dec. Literature Plymouth . . 7 to 8 to 11 6 to 8 7 to 9 8 to 10 IO to 13 12 to 16 13 to 17 14 to 18 16 to 14 14 to 13 to 12 to 9 Shore waters in summer approximate higher numbers, the lesser being for open sea. Naples. . . . 8-10 15-19 20-25 25 to 27 18-22 Berthold. 15 13 15-19 24-26 18 17 Berthold. Admiralty Chart. Adriatic . . . 10-15 12-13 u-14 16 12-17 25 23-26 16-22 10-17-5 12-19 Sea temperatures rather than shore water. Berthold. Admiralty Chart. Yarmouth . . 3-4 4 4.8 7-9 IO *4 l6.2 18.6 15-3 n 8-3 5-1 Monthly averages. Whitley. Bell Rock . . 4.8 4-7 4-8 5-9 s„ io-6 12.4 13-3 12-6 II-2 8.4 6-5 Surface. Fishery Board for Scotland, 1895. Abertay . . . 4-7 4.8 4.8 6-4 8-4 io-8 12-8 13-6 12.7 10.7 7.6 6.2 ; Three fathoms. Monthly averages. Ditto. Heligoland . . 3-7 2.7 3-i 5-2 8.5 12-5 15.6 16.8 i6-i i3-i 9-3 6-2 Karsten. Kiel .... 1-8 i-5 2-5 6-i 10.7 .5-8 18 18 15.8 1 1.9 7-3 3-7 Surface. Karsten. 2-97 2-55 3-08 5-3 8-6 12-8 15-2 16.1 I5'4 12.4 7-9 4-2 Five fathoms. Orkney . . . > 6 7 5-7 8-9 9-1 1 1 1— 1 2 12-13 12-14 II « 9 Whitley and Ad- miralty Chart. Shetland . . . 6 4-7 5 5-7 8-9 9-12 11-13 n-13 1 1— 1 2 8-1 1 8-9 6-8 Whitley and Ad- miralty Chart. Skagerack . . 5 to 7 I to 4 O to 3 4 to 5 6 to 8 IO to 15 to 17 15 II to 13 9 to 10 9 to 6 Open water. Pettersson. Christiania Fjord. 2-1 17.2 12 to 5-7 Cold winter (Baltic water). Hjort. 4-5 8-11 Warm winter (North Sea water). Hardanger Fjord. (Bergen) 3 (cold) 4 15 to 9 8 to 10 Hjort and Nor- weg. N. At- lantic Exped. 6 (warm) Vigten Fjord 4 2 12 to 13 Ditto. Lofoten Is. . . !-5 1-3 to 12-7 6 Ditto and Ad- miralty Chart. To face p. 96.] 97 Cut l evict multifida (Grevi). the winter solstice. It is clear that in the case of free assimilating plants the light-intensity must be of supreme importance for food-supply and increase in bulk, and there can be no doubt that it is the abundance of light-supply on the ascending part of the temperature-curve which brings forward so rapidly the summer vegetation of the British seas ; but on the other hand, these same two factors in the Bay of Naples appear to lead to the death of the greater number of the same plants, and the optimum vegetative period occurs there on the descending part of the temperature-curve with a diminishing light-supply. Berthold1, from his observations at Naples, came to the conclusion that optimum light-intensity and velocity of the current were the main factors, and temperature, although important, subsidiary to these. He bases his view chiefly on the manner in which the winter and spring annuals last on in deeper water or shaded situations ; but it is clear that these conditions would also imply a lower degree of tempera- ture by affording a protection from the direct heating effect of the sun’s rays ; while rapid movements of water, by carrying off the heated surface-layers, would produce a similar result. In the case of a plant which requires light for assimi- lation, it is difficult to isolate the heating effect from the light-intensity of the sun’s rays ; and as temperatures are easily recorded, it has been thought worth while to collect a few of the available data for different localities. As Cutleria grows in fairly shallow water, i. e. less than five fathoms, the temperatures of the surface-water will be sufficient ; but it is to be noted that records of surface-temperatures along the shore vary much more than in the open sea, and those alone would be absolutely reliable which were taken on the spot where the plant was growing : for the influence of the land and winds on shallow water in enclosed areas leads to a source of error of possibly 3° C. from the average at Plymouth, while local variations are still greater in the Mediterranean. 1 Mittheil. Zool. Stat. Neapel, iii. 1882, p. 293. H 98 Church. — The Polymorphy of The observations made at Plymouth suggested that Aglao- zonia vegetated at an optimum of io°-i2°, but was perennial within the annual range of 6°-i 8° ; the zoospores germinated in the spring at 120, and the optimum range of temperature for Cutleria was from I2°-i6°5 this being accompanied by the great light-intensity of May and June. Under these con- ditions the thallus was mature and in full reproduction in four months, of which one month may be included in the protonematoid embryo stage. It died (in the Laboratory) at 20°, this temperature being reached at the end of August in shallow enclosed water, and also on the autumn fall to 130. The autumn fall to T2°, although accompanied by great diminution in light in October and November, led to re- newed growth of the Aglaozonia which perennated through the winter, growing slowly at its optimum temperature, but stopping if the light-supply was slight, as in dull weather in November and December. The same conditions were also fatal to the whole of the foot-embryos in all stages, which had not yet thrown out Aglaozonia expansions. Complete data are not available for Naples , but the same annual period of maximum and minimum obtains, the range being from 8° in January and March to 27° in July and August. As before noted, temperatures to have more than an approximate value require to be taken where the Cutleria is growing, and as Berthold gives temperatures of 15° and 170 for February and December, it would seem that Cutleria lives in these waters under similar temperature-conditions to those which obtain in the Channel in summer; i.e. it commences growth on the autumn fall in December and matures in four months, completely disappearing in shallow water in April, as the temperature rises to near 20°. At the same time the Aglaozonia is perennial over the summer heat of 270, or at any rate may exist in deeper and colder water. According to Berthold, a stunted growth of Cutleria also occurs in deep water in July and August to a certain extent, the temperature at the depth of growth, forty fathoms, being I4°-I7°, thus approximating the English temperature. This growth appears 99 Cutleria multi fida ( Grev .). however not to have been investigated, the researches of Reinke and Falkenberg having been carried out on the winter-plants (cf. p. 109). At Antibes the isothermals for surface-temperature very closely approximate those of Naples, and Cutleria has here the same annual period. North Sea , Yarmouth , Heligoland. With a cold winter, late cold spring, and a rapid summer rise, the North Sea presents a variation of from o° to 180; the Heligoland curve averaging a degree lower on its rise than that of the western shore. The fall to zero is exceptional, and the mid-winter average for Heligoland is well over 2°, and possibly higher in sheltered localities. More complete data for the occurrence of Ctitleria in the North Sea would be of great interest, as from the preceding it would appear that here the high degree of temperature necessary to form the mature plant did not obtain, as a rule, throughout a sufficient length of time ; and this may possibly be the explanation of the fact that Cutleria has been found at Heligoland, but in recent years has only occurred in the protonematoid form as C. confervoides. Similarly, Cutleria is found at Yarmouth, but is not known to occur along the east coast until the sea at Orkney feels the influence of the warm Atlantic current, although Aglao- zonia is often common (e.g. Berwick). It is therefore probable that we here reach the minimum heat-supply for the develop- ment of the typical Cutleria-XhdMus, owing to the fact that the optimum degree of temperature does not obtain over a sufficient period in the brightest months. Similarly, Kiel has a still greater range, with a high summer temperature but very low mid-winter average, being below 30 during the winter months, and often below zero. This appears to limit the Aglaozonia , and neither Aglaozonia nor Cutleria occur at Kiel or inside the Baltic. Orkney and Shetland , owing to the presence of warm Atlantic water, show the smallest amount of annual variation, the characteristic feature being the warm winter ; while the summer maximum is only 140, the winter minimum is 4°-5°. H 2 IOO Church. — The Polymorphy of Orkney has a noticeably milder winter than that of Shetland, the February temperature being 6°. It is fully in agreement with previous statements that Aglaozonia should here peren- nate safely and Cutleria vegetate in the summer without reaching any great bulk ; and it would appear that Pollexfen found all the Orkney specimens to be of the delicate ‘ peni- cillata , variety, which suggests but small amount of growth beyond the protonematoid condition. Skagerack and N orwegian Coast. From the extensive researches of Pettersson and Hjort, it is known that the temperature of the Skagerack and West Norwegian shores varies from year to year according to the manner in which the summer-heated waters of the Baltic find an outlet into the Atlantic, giving rise in summer to a superficial Norwegian coast-current (the Baltic current) running close along the shore, and thus forming a strip of water from Christiania to Nordland warmer than that of the North Atlantic summer. The Skagerack average temperatures given by Pettersson, especially the higher ones which represent those of warmer seasons, compare very favourably with those of Yarmouth with the exception of the late spring-rise. Christiania Fjord is the warmest portion of the Norwegian seas, and the best Norwegian stations for Cutleria occur in this Fjord. In warm winters the surface-temperature does not fall below 5°, while the summer maximum is 17-2°; these numbers falling well within the suggested temperature-limits for Cutleria and Aglaozonia . On the other hand, as in the North Sea, it is clear that the critical temperatures are reached in passing up the West Norwegian coast, where Cutleria is found sparingly, Aglaozonia more commonly, extending as far as Nordland but not to Lofoten. Winter-temperatures along the coast again vary in different years according to the relative strength of the Baltic current, now cold at 2° or less, and the open Atlantic at 6°-j°. In warm winters the surface-temperature may be as high as 6° at Hardanger Fjord and as much as 40 at Vikten. On the other hand, the surface-layers lose heat in contact with the extreme low temperature of the air at TO I Ctttkria multifida ( Grev .). Lofoten, and here the surface-temperature of the open sea falls to i*5° in February; this being again the Kiel average for the same month. There is therefore a certain amount of evidence in favour of the view that temperature rather than light-intensity is the determining factor as far as the actual existence of these plants is concerned, and that while Cutleria vegetates at a mean of 160, with a range of four degrees above and below, Aglaozonia prefers a mean of io°, with a maximum consider- ably over 20°, and a minimum below 3° : further, that a continuance of this low temperature limits the existence of both these plant-forms, in that it destroys the perennating thallus, both in the Baltic and along the Norwegian coast. In order to test these data, observations were attempted at Plymouth in January, 1898, on perennating plants of Aglao- zonia, both the adult thallus and also the young perennating plants of the first winter germinated the previous summer, large numbers of which were now from *5 to 2 mm. long, and had been growing for months at an average temperature of 14°. It is clear that in this case the action of a constant degree of temperature over a longer period of time than was available would be preferable, and the observations were not so successful as might have been wished owing to the difficulty of maintaining a constant temperature over a long period and at the same time maintaining general health-conditions by frequent change of water; and it is probable that the data derived from actual distribution are more likely to be correct than those derived from cultures in the laboratory so long as the difficulty of accurately imitating the natural environment remains. Thus, at 25°-26°, the summer maximum for Naples, young perennating forms remained healthy for ten days, and although a few died, many were alive and well after sixteen days. At 27°-29°, similarly, both young and old plants remained healthy after six days, and there seemed reason to believe that at temperatures below 30° Aglaozonia might perennate successfully. 102 Church. — The Polymorphy of At temperatures above 30°, on the other hand (30°-33°), death occurred sooner or later ; young plants dying in 2-4 days ; older ones in 4-6 days, dying irregularly in patches. This is of interest as showing the unlikelihood of Cutleria crossing the Tropics where the maximum surface-temperature is above 30°. Experiments at low temperatures were not conclusive, young plants remaining perfectly healthy after being surrounded by melting ice for six days. Theory of Sexuality. The theory of the sexuality of the Phaeosporeae, which in point of fact still remains based on the classical researches of Reinke and Falkenberg on Cutleria , and those of Berthold on Ectocarpus siliculosus , has more recently been called in question by such accurate observers as Kuckuck and Sauvageau 1, who have repeatedly failed in obtaining union of gametes in various species of Ectocarpus and allied genera. Thus Kuckuck maintains that Ectocarpus siliculosus is constantly parthenogenetic at Kiel, and it may be noted that Reinhardt has observed both copulation and direct germina- tion of gametes in this species at Sevastopol ; while Sauvageau in 1895 obtained direct germination in the case of the gametes of seven species of Ectocarpus and wholly negative results as regards a sexual process. No one, again, has ever observed sexual fusion in any of the plants of the Giffordia section of Ectocarpus which possess apparent antheridia, nor again with certainty in any of the Tilo- pterideae. The facts in the case of Cutleria , however, appear to point to the narrow range of external conditions within which the sexual process can be effected : if these conditions do not obtain, the plant may fall back on parthenogenesis, which in Cf. Ann. de Sci. Nat. 1896, p. 223. C titter ia multi fid a ( Grev .). 103 northern waters is associated with a correlative diminution of the now useless male organs ; so that, under extreme conditions, the admittedly asexual mode of reproduction alone remains on the perennating form. There can thus be little doubt that in the case of Gijfordict and the Tilopterideae, purely morphological considerations may be a better guide to the theoretical degree of sexual specialization than the physiological observation of the act of fusion of the gametes ; and further, that until more complete physical data are forthcoming as to the exact conditions of the experiment, a single positive result must far outweigh many negative ones, and the evidence that the so-called plurilocular sporangia of the Phaeosporeae are not potentially gametangia remains inconclusive. Nor, on the other hand, do the data for Cutleria point so much to an imperfectly differentiated or incipient sexuality, as to an actual and progressive loss of that function ; and thus, by analogy, the conception that the primitive Ectocarpus-Ytke ancestor of the Phaeosporeae was a sexual plant with iso- gamous gametes would be strengthened rather than under- mined. At any rate it is clear that the actual data for any given plant can only be obtained by actual observations taken at different times of the year at different points of distribution. Phylogeny of Cutleria. All generalizations as to the phylogeny of existing Algae must, in the present state of our knowledge, be necessarily more or less founded on the very hypotheses the scientific botanist most desires to prove. At the same time the only proof of such hypotheses at present attainable consists in their complete agreement with ascertained facts ; and thus so long as the tentative character of the proceeding is clearly borne in mind, it may become of interest to construct a phylo- genetic scheme for the life-history of the genera Cutleria and 104 Church. — The Polymorphy of Z ancirdinia which will not only include the existing data, but may present some suggestions towards the solution of other algological problems. The evolutionary specialization of the Cutleriaceae, with which we are here concerned, takes into account the vegetative structure only. A comparison of Cutleria multifida (including Aglaozonia reptans ), Cutleria adspersa (and its suggested Aglaozonia c kilos a), and Z anardinia collaris shows that in the structure of the reproductive organs the three types are identical. In both genera the asexual sporangia give rise to a few (6-1 o) large biciliated zoospores, and thus present an intermediate reduction-specialization as opposed, on the one hand, to the numerous spores from the unilocular sporan- gium of Ectocarpus siliculosus , &c., and to the immotile monospore of the Tilopterideae on the other. The antheridia show a slight advance on the primitive Ectocarpoid multilocular sporangium, in the more complete delimitation of the antherozoid-tissue, best seen in the skeleton framework remaining after emission of the antherozoids, but they have not attained such a high degree of specialization as that exhibited by the bottle-like antheridium of the Tilopterideae. In the same manner, a further degree of reproductive con- centration has, in correlation with the increase in bulk of the female gametes, reduced the segmentation of a multilocular gametangium to an oogonium of sixteen loculi, each producing a single oosphere ; but this again is a lesser degree of reduc- tion than that obtaining in the Tilopterideae with huge solitary oosphere. Hence, in all three forms of reproductive organ, the Cutleri- aceae offer a condition intermediate between the isogamous Ectocarpaceae and the completely heterogamous Tilopterideae ; and they may, in view of the present vegetative condition of these two groups, be regarded as descended from a filamentous form which had attained the present comparatively high degree of sexual differentiation before passing beyond the branched filamentous condition in its vegetative structure. Cutleria multi fida ( Grev. ). 105 Again, in the Cutleriaceae, the presence of the filamentous stage is clearly marked (1) The mature plant itself is but a fasciated structure of which the growing regions are still in the purely filamentous form ; the assumption of growth by intercalary division, as opposed to the primary apical growth, being general through- out the whole group of the Phaeosporeae. The reproductive portions of the thallus are still wholly in the filamentous condition,, and filaments bearing reproductive organs can be induced to grow and attach by rhizoids ; while the attach- ment-disc is also a mere felted mass of rhizoids ; the only portion of the thallus, in fact, which is not filamentous being the highly specialized assimilative region. (2) The filamentous condition is characteristic of the em- bryogeny for a short period in the foot-embryo, but persisting to the adult condition in the protonematoid embryos which produced antheridia in the Plymouth cultures, as also in the female form C. confervoides found at Heligoland by Kuckuck. In the evolution of the vegetative thallus from such a simple filamentous form, in which intercalary growth of the branches supersedes the original apical development, the first step in advance is marked by the regular segmentation of the cells of the main axes, by successive divisions by walls in planes at right angles to one another, leading to the more or less regular formation of a multicellular condition such as exists iq. many Phaeosporeae, e. g. simpler Sphacelarias, Myriotrichia , Des- motrichum. This massive type of thallus with purely radial symmetry is represented (1) by the foot-embryo, (2) by the basal region of the plant in the protonematoid embryo. In the same way this method of segmentation is ontogenetically repeated in the formation of sterile hair-like branches on the adult thallus. As an example of such a plant-form in which the cortical cells send out basal lobes forming a dorsiventral disc around the point of attachment of the plant, Sphacelaria cirrhosa may be instanced ; and it is clear that in the Sphacelariaceae differentiation has proceeded in two lines from such a simple io6 Church. — The Polymorphy of form, giving (i) a further specialization of the shoot-system in Stypocaulon and Cladostephus ; (2) a suppression of the shoot system, and reduction to the creeping dorsiventral disc alone in Battersia. Such a reduction to a dorsiventral creeping thallus is again in Algae, from a vegetative point of view, a distinctly down-grade specialization. The plant, by adopt- ing a prostrate habit, exposes far less assimilatory surface to the action of light and free-flowing currents which bring both food and oxygen supply ; on the other hand, it gains in the struggle to resist the tensions and tractions of wave motion, and will thus exist safely, not only throughout more stormy seasons, but farther up towards the tide-mark. The relation of Cutleria to Aglaozonia , from a vegetative point of view, is simply that the two growth-forms represent- ing the extreme cases of specialization of such types as Clado- stephus and Battersia are here combined in a single species, and become fixed, one way or the other, at a very early stage. Thus it is interesting to note that the delicate Cutleria- thallus is confined to comparatively quiet waters and depths at least two fathoms below low-tide mark, and possesses a very slender point of attachment in relation to the bulk of the full-grown thallus ; while Aglaozonia rises, from equal depths, to rocks even above the tide-mark in many localities (Sidmouth). Further, accompanying feeble powers of growth and nutri- tion, and possibly correlated with them, Aglaozonia possesses an increased power of withstanding extremes of temperature. Beyond the ‘ massive ’ stage, the Cutleriaceae make one more advance, which forms the unique characteristic of the order. This consists in the fusions which take place between the axes produced from the independent filaments of the apex of the thallus, leading to a fasciated growth which further presents the complication of dorsiventrality. That this is not only the essential feature of the Cutleriacean type, but is the last and most recently acquired and there- fore most mobile character, is suggested by the following considerations. (1) The specific characters are essentially based on the Cutleria multi fida (Grev.). 107 characters of the fasciated shoot ; variations in this point being still considerable, giving rise to growth-forms which have been regarded as varieties ; those with a lesser degree of fasciation being extremely common, especially in localities where external conditions are unfavourable. (Orkney ; cp. C. penicillata , Lamour ; C. penicillata , Kutzing). (2) It is now the only character left which has not been observed in cultures, and it would therefore appear to demand the most delicate adjustment of external assimilative conditions. Considering the three types from the standard of attainment of this assimilative growth-form, it would appear that C. mul- ti fida, which covers the widest range of temperature in dis- tribution, presents these specializations in a high degree in its sexual shoot, but maintains existence under conditions unfavourable to its development by reduction to a degenerate creeping Battersia- like form, which alone persists at the extreme northern limit of distribution. C. adspersa , with a considerably narrower range of distribu- tion, exhibits a thallus, meagre by comparison with C. multifida , but more dorsiventral. Its Aglaozonia- stage, suggested for A. chilosa by Falkenberg, has not been more definitely isolated ; but Janczewski showed that the foot-embryo passed directly into a dorsiventral disc on the approach of summer at Antibes. Finally, in Zanardinia collar is the fasciated shoot, with extreme dorsiventral development, becomes itself prostrate, and by vegetating in the manner of an Aglaozonia- disc in the hot summer does away with the necessity for such a basal formation from the embryo ; and thus, being itself homotypic and obtaining the perennating advantages of the procumbent growth-form, presents the paradox of becoming degenerate by carrying to extremes the last variation of the family. During a portion of the time in which these observations have been made, the writer has occupied the Oxford University Table at the Laboratory of the Marine Biological Association at Plymouth ; and in acknowledging the goodwill and unfailing io8 Church. — The Polymorphy of courtesy with which the resources of the Station have always been placed at his disposal by the Director, Mr. E. J. Allen, he would wish to draw the attention of English algologists to the facilities afforded by the geographical position of the Plymouth Laboratory for the study of our native Algae. Grateful acknowledgments are also due to Mr. E. A. Batters for kind assistance on many out-of-the-way points not easily obtained from the literature. Postscript. — Since writing the above, Dr. P. Mayer kindly informs me that sea-temperatures taken on different days at Naples during the months December, 1897, and January, 1898, for localities in which Cutler ia is known to occur, ranged between 13-5° and 13*9°. This adds confirmation, therefore, to the data obtained at Plymouth, which tended to show that the young Cutleria-thdMus vegetates normally between o o 12 -14 . EXPLANATION OF FIGURES IN PLATES VII, VIII, AND IX. Illustrating Mr. Church’s paper on Ctitleria multifida. PLATE VII. All figures drawn with Zeiss D. Oc. 3, and slightly reduced in reproduction. Fig. 1. Germination of zoospores of Aglaozonia (2-3 days). Fig. 2. Older stage (one week old). Fig. 3. Protonematoid Cutleria producing antheridia (August). Fig. 4. Aglaozonia- disc produced at the base of protonematoid Cutleria (August). Fig- 5- Germination of zoospores of Aglaozonia (two weeks old). Cutleria multifida (Grev.). 109 PLATE VIII. All figures drawn with Zeiss D. Oc. 3, and slightly reduced. Figs. 6-9. Protonematoid Cutleria, three weeks old. Fig. 9. Embryo with multicellular ‘ foot ’-like base. Fig. 10. Protonematoid Ctitleria , four weeks old, with attachment disc. Fig. 11. Parthenogenetic germination of oospheres of Cutleria. Fig. 12. Embryos, one week old. Fig. 13. Embryos, becoming multicellular, two weeks old. Fig. 14. Foot-embryos, three weeks old. PLATE IX. Figs. 15-21, Zeiss D. Oc. 5 ; Figs. 22, 23, Zeiss D. Oc. 1 ; and all slightly reduced. Fig. 15. Development of Aglaozonia- disc from a single superficial cell by successive T-shaped walls. Fig. 16. Small foot-embryo, wholly growing into Aglaozonia form. Fig. 17. Terminal development of disc. Fig. 18. Late: al development of disc. Fig. 19. General case of basal development of Aglaozonia expansion. Figs. 20, 21. Three to four weeks old embryos, giving disc-growths at various points. 20, dorsal; 21, ventral surface. Fig. 22. Older foot- embryo with basal disc. Fig. 23. Older embryo with Aglaozonia expansion well developed (Nov.). Annals of Botany Vol XII, PI VII ft ft-* Cm/R.CH.bCL.9^ CHU RCH, — CUTLERIA. . • ■ > ; Annals of Botany Vol. XII, PI. VIII A H- Church . Je.vh/'jy C H U R C H, CU T LE R I A. Annals of Botany Vol XII, PL IX CHURCH C U T L E R I A. On the Structure of an Ancient Paper1. BY M. DAWSON, B.Sc. Botanical Laboratory , Cambridge. IN November last, Professor Marshall Ward handed to me a specimen of ancient paper, with the request that I should attempt to determine the botanical nature of the materials of which it was made. Examined macroscopically, the paper may be described as a light-brown felt-like substance, consisting of layers — easily separable one from the other — of closely interwoven fibres. 1 The paper here referred to was one of four pieces of ancient MSS. which were sent to me by Mrs. Gibson of Castle Brae, Cambridge, for examination as to the materials of which they were composed. These MSS. are parts of a series discovered by Mrs. Gibson and Mrs. Lewis at Cairo in 1897, and which proved of some historical interest. They almost certainly came from the Genizeh or lumber-room of the Synagogue in Old Cairo, whose contents Mr. Schechter has brought to Cambridge. The writing on them is Hebrew, and refers to legal matters. Mrs. Gibson informs me that on one fragment there is conclusive evidence of the date, 1038, and further inquiry leads to the conviction that this is one of the oldest fragments of such writing in England. I subsequently received by Mr. Schechter’s kindness another fragment of a similar paper. On testing these five papers I found them to be made of flax or some similar fibre, and the subject seemed so interesting that I asked Miss Dawson to go more fully into the matter, which she has successfully done. — H. Marshall Ward. [Annals of Botany, Vol. XII. No. XLV. March, 1898.] I I 2 Dawson. — On the Structure of As is well known, many ancient writings are on vellum or papyrus : this, however, was evidently neither a prepared animal skin nor so complicated a tissue as that of the Cyperus Papyrus. It clearly was composed of fibres which had by some process been extracted from the tissues of some plant. After teasing in water and examining with the microscope, the paper was seen to consist of fibres, often occurring in strands, accompanied by elements showing spiral or simple- pitted markings, and here and there by cells of wood-paren- chyma or medullary rays. The layers of the paper appear to have been stuck, not woven together ; an examination of the dust, shattered from the paper, showed, after treatment with iodine, yellow and bluish-black particles — the latter some- times lying along the fibres, and giving to them a blue colour. This suggested that some form of starch may have been used in the manufacture, but all attempts to determine its nature were unsuccessful. This alteration of the starch-grains, which had evidently taken place, was probably due to the action of numerous Fungus-hyphae, and groups of Bacteria-like bodies, which could be seen amongst the fibres. The individual fibres are somewhat cylindrical in shape, with pointed ends, and very narrow lumen. Their walls are thick, and show regular longitudinal striations, but no visible pits. Their breadth is practically constant (-015 mm.), but the length varies considerably: the extreme measurements taken were 6*4 mm. and 3*1 mm. ; an average of ten measure- ments gave 4*25 mm. As, however, so many of the fibres were broken, accurate measurements of length were not possible, and these figures cannot be relied upon. It can, nevertheless, be affirmed that the fibres are shorter and narrower than those of cotton, and are not, like them, flattened and twisted. They show a marked tendency to tear into fibrillae, often in spirals, and show traces of swellings or kinks. With Schulze’s solution, they give a purple colour, with iodine pink, and with iodine and sulphuric acid blue : but no trace of an Ancient Paper . 113 lignin-reaction can be obtained either with phloroglucin and HC1 or with aniline sulphate. When treated with Cu. Amm., the fibres behave in a characteristic manner : the walls swell greatly — causing the striations to become beautifully clear — and the lumen persists for some time as a very narrow, dark, wavy line, running down the middle. The above results lead to the conclusion that they are some sort of non-lignified bast-fibres. Among the fibres which seemed likely to be met with as ancient paper-materials are Flax, Hemp, Boehmeria , Brous- sonetia , Cotton, and Nettle1. We may exclude, by the re- actions, such lignified fibres as jute, straw, and wood, met with in modern papers : and, probably, certain Indian and out-of-the-way fibres, as Rice, Bamboo, Daphne, New Zealand Flax, &c., may also be neglected. A careful examination of the above list gave the following results 2 : — 1. Cotton fibres are flat and often twisted ; they are broader, longer, and show a wider lumen than those under investigation ; and in addition, when treated with Cu. Amm. they coil into spirals before complete solution. 2. Nettle ( Urtica dioica ) fibres, like those of cotton, are flat, and somewhat twisted ; they are even broader than those of Cotton, and show as wide a lumen. Their walls have distinct longitudinal and transverse striations, and under the action of Cu. Amm. they become swollen and much crumpled, leaving the broad lumen very clearly visible. 3. Boehmeria nivea fibres resemble the fibres of the paper in their breadth, narrow lumen, and non-lignified walls, but are clearly distinguished from them by their club-shaped ends, and their characteristic behaviour with Cu. Amm., viz. rapidly coiling into loose spirals. 1 J. Wiesner, Die Rohstoffe des Pflanzenreiches, p. 447. 2 See also Cross and Bevan, A Text-book of Paper-making, pp. 30-61. I 1 14 Dawson. — On the Structure of 4. Hemp ( Cannabis sativa) fibres are long and narrow, with a very narrow lumen. Their ends, however, are slightly flattened, and their micro-chemical reactions distinguish them from the fibres of which the paper is made : thus they give a yellowish colour with aniline sulphate, green with iodine and H2 So4, and in Cu. Amm. they swell greatly, showing longitudinal stria- tions, with a comparatively broad central wavy line. 5. Flax ( Linum perenne ) fibres are long and narrow, the walls are thick and non-lignified, and show charac- teristic swellings or kinks, where fibres have crossed one another. Their behaviour, under the action of Cu. Amm., is strictly comparable to that of the fibres of the paper. The chemical reaction of the walls also agrees closely in the two cases. A comparison of the above results shows at once that we must exclude the fibres of Cotton, Nettle, Boehmeria , and Hemp. I was unable to examine those of Broussonetia , but Wiesner’s 1 description of them puts them also on one side. On the contrary, the shape, narrow lumen, and pointed ends, the chemical properties of the walls, their reaction with cupric ammonia, longitudinal striations, tendency to fray into fibrillae, and the traces of slight swellings or kinks, all point to the conclusion that we are dealing with bast-fibres of Flax Considering the subject from a historical standpoint, we learn from De Candolle2 that Flax was familiar to the Chaldeans, Egyptians, and Hebrews. The plant figures in Egyptian drawings, and, as the microscope has revealed, the bandages, used as mummy-wrappings, were made of linen. Moreover, the annual Flax was cultivated for thousands of years in Mesopotamia, Assyria, and Egypt, and it was, and still is, found wild in the districts lying between the Persian Gulf and the Caspian and Black Seas. 1 J. Wiesner, Die Rohstoffe des Pflanzenreiches, p. 459. 2 De Candolle, Origin of Cultivated Plants, 1884, pp. 1 23-130. an Ancient Paper. ii5 It is interesting to note also, that Karabacek, in his preface to the descriptive catalogue of the Archduke Rainer’s Collec- tion of MSS.1, gives reasons for concluding that the Arabs learnt the art of paper-making from plant-fibres, from the Chinese, about A. D. 751. We are not, therefore, guilty of any anachronism in assuming that, by the year 1038 — the date which has been assigned to this paper — the process of manufacturing paper from the fibres of the Flax-plant was both known and employed. 1 Fiihrer durch die Ausstellung Papyrus des Herzog Rainer, 1892, Th. I, pp. xvii-xxiv. NOTES. CORRELATION OP GROWTH UNDER THE INFLUENCE OF INJURIES. — In the paper on this subject which appeared in the Annals of Botany, Vol. xi, No. XLIV, December, 1897, reference was made on p. 513 to Laurent’s valuable paper, Etudes sur la Turgescence chez le Phycomyces, but by an oversight the name of the author was omitted. C. O. TOWNSEND. GELATINE AS A FIXATIVE. — Microtome-sections passing through embryonic and parenchymatous tissues embedded in paraffin are sufficiently fixed to the microscope-slide, for staining purposes, by their own simple adhesion to the glass. This, however, is not the case when the section comprises a large proportion of woody tissue. For such preparations, collodion, agar-agar, and albumen have been recommended as fixatives. The first of these, so far as my experience extends, is the most certain. But it has the disadvantage that with its use the paraffin-section cannot be floated out on water on the slip and caused to flatten out by gentle warmth. The same objection applies to albumen \ and in addition, I have found it to be very easily coloured by stains (especially the blue dyes) which are often essential to use in microscopic work. I have no experience with agar-agar, but Zimmermann states that it becomes dyed with haematoxylin — one of the most important stains, and that the sections often come loose from the glass during the staining and washing manipulations. This latter objection, perhaps the most vexatious of all, applies, to some extent, to albumen also. Recently I have used as a fixative a dilute solution of gelatine in a watery solution of bichromate of potash. The solution should be quite fluid at io°C. In use the ribbon of paraffin-sections is laid on a drop of this solution on the slide. Wrinkles in the sections may 1 I used the preparation given by Zimmermann, Bot. Mikrotech. 1 18 Notes. be removed by gently warming the slide over a flame. Then the superfluous fluid is drawn off by blotting paper, and the gelatine is allowed to dry and harden. During this process it should be exposed to a bright light. The action of the light on the bichromated gelatine renders it quite insoluble even in warm water, and so re- moves all danger of the sections becoming detached from the slide. The bichromate of potash in the gelatine has this additional ad- vantage, that after exposure to light it prevents the latter from taking up the dyes used as stains. So far as I have at present tested it, this preparation of gelatine is unaffected by saffranin, fuchsin, acid fuchsin, haematoxylin, iodine green, gentian violet, and aniline blue. With aniline blue, however, a precipitate is sometimes formed along the line which formed the edge of the paraffin ribbon and in cracks in the ribbon ; but in no case is the substance of the gelatine itself stained, and so it offers a marked advantage over albumen and agar-agar. HENRY H. DIXON. Trinity College, Dublin. LATHRAEA SQUAMARIA. — I find that in my paper on this subject, which appeared in Annals of Botany, Vol. xi, 1897, P- 3^5, I omitted to include in my references to the literature the observations of C. Darwin and F. Darwin, of which an account is given in ‘ The Power of Movement in Plants ’ (footnote, p. 85). The authors show that Lathraea can excrete large quantities of water underground, and state that the water is secreted by glands lining the cavities of the scales. PERCY GROOM. On the Development of the Leaf and Sporocarp in Marsilia quadrifolia1, L. BY DUNCAN S. JOHNSON. With Plates X, XI, and XII. LTHOUGH Marsilia and the related Pilularia have l\. been frequently studied during the present century, the exact origin and morphological significance of the sporocarp has never been satisfactorily made out in either genus, the chief reasons for this being apparently the com- plexity of the apical bud and the dense covering of trichomes over all the younger parts. The present work was under- taken at the suggestion of Dr. J. P. Lotsy, then of the Johns Hopkins University, in the hope that a detailed study of the development of the leaf and the sporocarp of Marsilia would give some indication of the morphological nature of the latter. The work has been carried on during the winters ’95-’96, ’96-97, in the biological laboratory of the Johns Hopkins 1 Accepted as a thesis for the Degree of Doctor of Philosophy by the Board of University Studies of the Johns Hopkins University, Baltimore, U.S.A., June, 1897. [Annals of Botany, Vol. XII. No. XL VI. June, 1898.] K 120 Johnson. — On the Development of the University, under the stimulating direction of the late Prof. J. E. Humphrey, to which any value the work may have must be largely attributed. The material was collected, partly at New Haven, Conn., through the kindness of Prof. W. A. Setchell, and partly at Cromwell, Conn. It was fixed in 95°/0 alcohol, i°/o chro- mic acid or a 5°/o sublimate-acetic mixture. For staining, gentian-violet or Mayer’s haemalum, the latter either alone or in combination with Bismarck brown, were found most satisfactory. The Development of the Leaf. Our knowledge of the development of the leaf in Marsilia is due almost entirely to the work of J. Hanstein on the embryo, and frequent reference to this will be necessary in the following pages. The leaves arise in two rows, one on each side of the median line on the dorsal surface of the stem. Each leaf is developed from a typical two-sided apical cell formed from part of one of the dorso-lateral segments of the tetra- hedral apical cell of the stem. This apical cell of the leaf is recognizable when the stem-segment in which it is formed is only the third or fourth in its series from the apical cell of the stem. It is larger in size and projects more than the neighbouring cells ( L , Fig. i), and its position is such that its edges are directed toward the base and the apex of the stem. Hanstein (’65) has already shown the shape and position of the apical cell to be as above described in all but the very earliest leaves of the embryo of Marsilia , which agrees thus with most other Leptosporangiates that have been studied. They have been thus described by Hofmeister (’62) in Aspidinm , by Kny (’75) in Ceratopteris , by Klein (’87) in Polypodium , by Meunier (’87) and Bower (’89) in Pihdaria , by Campbell (’87) in Onoclea , and by Bower ( 89) in Tricho - manes. Pteris is apparently the only known case where, Leaf and Sporocarp in Marsilia qnadrifolia , L. 121 according to Hofmeister (’62) and Klein (’84), the position of the two-sided apical cell is transverse to the stem. This apical cell of the leaf in Marsilia continues its growth and activity, cutting off segments alternately toward the right and left of the young leaf, which has its ventral . side facing toward the stem-apex. When about fifteen or sixteen pairs of segments have been formed, the activity of the apical cell is ended, probably by a periclinal wall like that seen by Sadebeck (’73), Kny (’75), and Bower (’84) in other Lepto- sporangiates, and that to be described later in the sporocarp of Marsilia. The great regularity of the segments of the leaf in Marsilia , as well as the fact that certain of the cells remain of the full length of the segment (Fig. 3), make it possible to determine quite definitely the number of segments formed (Fig. 2). The only doubt is in regard to the first segments, some of which fuse with the tissue of the stem, and it is thus not certain that the segment numbered 1 in Fig. 2 is not really the second one. The young leaf is about 1 mm. long and -15 mm. in diameter at the base when the activity of the apical cell ceases, and is a slightly tapering conical organ curved upward (Fig. 2) and ventrally over the stem-apex. It is almost exactly circular in cross-section until the formation of the pinnae begins, and is not at all spatulate as described by Hanstein (’65) in M. Drummondii , and by Campbell (’96). The segments of the apical cell of the leaf, or ‘ primary marginal cells ’ of Hanstein (’65) and Sadebeck (’74), are nearly semicircular blocks with the upper and lower surfaces slightly concave toward the apex (Figs. 4-6). The first division-wall appearing in these segments is a longitudinal and radial anticline (/, Fig. 6), cutting off about one-third of the segment toward the dorsal side to form what we may call a section , and leaving on the ventral side a secondary marginal cell. Wall I is apparently the ‘ tangential wall ’ of Sadebeck (’74), and section I is the ‘ Schichtzelle * of Hanstein (’65), but this terminology does not seem appropriate when the real position of this and later section-walls is taken into K 2 122 Johnson. — On the Development of the account, since it refers to the position of these walls at the intersection with the surface, when the thing of importance is, as we shall see, their position in the interior. The second wall formed in the segment is also a longi- tudinal anticline in the secondary marginal cell, and nearly parallel to the inner or median border of the segment (II, Figs. 4, 6) ; and thus is formed a tertiary marginal cell (me.3, Fig. 6). This latter is then divided by a transverse anticline (/. a.1, Figs. 4, 5), the radial wall of Sadebeck, into an upper and a lower tertiary marginal cell. Then in each of these further section-walls, to the number of three, are formed near and parallel to the dorsal and ventral sides alternately (III, IV, V, Figs. 4, 5, 7, 8), and there are thus formed two marginal cells of the sixth grade in each segment (mcl, Fig. 8). In a less frequent type of division only four section-walls are formed, and the ultimate marginal cells are thus of the fifth grade. The Petiole. The first nine or ten pairs of segments of the leaf go to form the petiole, and the six primary divisions of the segments (taking the type where the ultimate marginal cell is of the sixth grade) break up into cells, as will now be described. About the same time that wall II is formed, there appears in section I a pericline (pi. w., Fig. 6), cutting off at the inner end a part of the plerome contributed by this section to the longitudinal bundle of the petiole. This is followed by a longitudinal and radial anticline, the halving anticline, cutting the outer cell into two ( h . a., Figs. 5, 7, 8). Each of the other sections and the marginal cell in turn cuts off plerome at the inner end (pi. w., Figs. 7, 8), but no halving anticline is formed in any of them. Then there appears a pericline in the outer end of the halves of section I, in the outer ends of each of the other sections and of the marginal cell, separating a layer of outer cells which give rise to the epidermal structures of the petiole from an inner one of cells forming the mesophyll. We may for the sake of brevity, Leaf and Sporocarp in Mar si Ha quadrifolia , Z. 123 though not with strict propriety, call these layers dermatogen and periblem, and the pericline itself a dermatogen-wall (d. w., Fig. 7, d.,pb., Fig. 22). The dermatogen soon splits by another pericline into hypodermal and epidermal layers (Figs. 8-10, 22), each of which remains of one cell in thickness even at maturity, though numerous anticlines, both longitudinal and transverse, are formed in each. On the line of the median wall, of each section-wall and of the halving anticline of section I, there are formed intercellular spaces (a.} c.} Figs. 8, 22) between the periblem and hypodermis, which are the beginnings of the fourteen (primary) longitudinal air-canals of the petiole (a. c., Fig. 9). The single periblem-cell of each half of section I cuts off by a pericline at the inner end a second portion of plerome (pi., Figs. 8, 22). Then each of the remaining periblem-cells of section I, the single periblem-cell of each of the other sections and that of the marginal cell, divides by a pericline into an inner and outer cell (Figs. 8, 22). Of these the inner cell divides by anticlines and periclines to form the loose mesophyll-tissue of the mature petiole (mp., Figs. 8-10), while the outer cell gives rise to both the longitudinal and transverse partitions between the adjacent air-canals (p. c., Fig. 8). These latter cells swell in the middle (as seen in cross-section of the petiole) and grow out at the ends into papilla-like tips (Figs. 8, 11, 12), touching their fellows of the adjacent sections, but leaving an intercellular space sur- rounding each tip. The tips thus formed are soon cut off by longitudinal anticlines (Figs. 8, 11), forming a pair of nearly isodimentional cells (c. p. c., Figs. 8, 11) in each air-canal, opposite each primary partition-cell, of which there are usually eight in the length of each segment. From these eight pair of cells are developed the eight transverse partitions of the air-canal in each segment. These remain one cell thick even at maturity, but during their later development many intercellular openings or pores are formed, allowing the passage of air through them ( c. p. p ., Figs. 9, 10, 13). 124 Johnson. — O71 the Development of the The portion remaining of each primary partition-cell (l.p. c., Fig. 11) grows in a radial direction and splits by periclines, while it at the same time grows in the direction of the length of the petiole and divides by transverse anticlines. Thus are formed the longitudinal partitions between the adjacent air- canals, which are also one cell in thickness (Figs. 9, 10, 13). As each of these primary longitudinal partitions elongates with the lengthening of the petiole, it is seen (/. p. c., Fig. 12) that the primary cross-partition cell at one end is nearer the upper wall and that at the other is nearer the lower wall. Then when the first transverse anticline is formed it is some- what oblique and forms thus two wedge-shaped cells, each with a cross-partition cell at the broad end and none at the narrow one (/. p. c ., Fig. 12). The cross-partitions in adjacent canals are thus alternate. These wedge-shaped cells continue to elongate and divide by transverse anticlines (Fig. 13) till in the mature petiole the cross-partitions are far apart. Here again in the longitudinal partitions we find at maturity many small intercellular openings or pores, the ‘ meats ’ of Meunier (l-P-P; FigS. 10, 14). When the epidermal surface of a section is two cells broad and four cells long (Fig. 15), there is cut out of each cell, by a semicircular anticline at the upper end, a small cell which gives rise to one of the numerous trichomes that clothe the young leaf. The rest of the epidermal cell then divides further by anticlines (Fig. 15), and more trichomes arise in the cells thus formed, while the epidermal cells at maturity become much elongated (Fig. 16). Each trichome-cell grows out beyond the surface of the epidermis, and swells to a knob at the outer end ( tc ., Figs. 2, 3, 9), which soon elongates in the direction of the length of the petiole. On the lower or basi- scopic side it projects but little (Fig. 3), while it grows out toward the apex of the leaf to the long multicellular hair ( tc ., Fig. 17) that is supported by the basal or stalk-cell which remains wedged in between the epidermal cells ( b . c. tc ., Figs. 9, 16, 17); later in the development most or all of these trichomes are cast off, and the petiole thus becomes naked at maturity. Leaf and Sporocarp in Marsilia quadrifolia, L, 125 Stomata also occur on the petiole, but apparently not until quite a late stage, and their development was not studied. When the longitudinal partitions are about three or four cells in width (radially), longitudinal rows of mesophyll-cells, usually one opposite each partition, have become specialized to form the so-called tannin-sacs (t. Figs. 9, 10). The cells composing these are, like the surrounding mesophyll-cells, about twice as long as broad at maturity and rounded off laterally, forming many small intercellular spaces connecting with the large air-canals. While the dermatogen and periblem have been developing as described above, the plerome of sections I-IV has given rise to the axial vascular bundle of the petiole. The plerome of section II divides by two longitudinal anticlines into quarters, of which the one in the angle between I and III never divides further in any direction, but forms the large trachea of its side of the bundle (tr., Figs. 3, 8-10, 22). The nucleus of this cell may divide many times, so that in a trachea of half a millimeter in length (Fig. 3) we may find twenty-five or thirty nuclei, but these and all other proto- plasmic contents disappear later and the end-walls assume the characteristic oblique position, always with the dorsal edge directed towards the base of the leaf. The remaining three quarters of this section and all the plerome of sections I, III, and IV break up by numerous longitudinal walls (Figs. 8-10), and later by fewer transverse walls (Fig. 3), to form the remaining tissues of the bundle which later still develops a bundle-sheath (b. s., Figs. 9, 10). The portion of the marginal cell within what we have called the plerome-wall never forms any part of the vascular bundle, and the same is usually true of the same portion of section V, though it does occasionally form a small portion of it (Fig. 9). The Lamina. Just before the activity of the apical cell ceases, the tenth and eleventh (or eleventh and twelfth) segments on each side begin to grow out laterally and ventrally to form the first 126 Johnson . — On the Development of the pair of pinnae (p1., Fig. 18). Each pinna is formed from the whole length of one segment (usually the lower one) and most, but not all, of a second (Fig. 19). In this respect the pinnae resemble those of Ceratopteris (Kny 75) and differ from those of Asptenium serpentini (Sadebeck ’73) and Onoclea (Campbell ’87), in which the pinnae are equal in extent to the segments. Soon after the apical growth ceases, the segments beyond the first pair of pinnae, except part or all of the one next the lower pinna on each side, begin to swell out in a similar manner to form the terminal pair of leaflets.' In a transverse section of the leaf through the pinnae, which is practically the same for both pairs, we see that the swelling mentioned is due to the continued activity of the marginal cells (Fig. 20). No pericline is formed in these, as in the marginal cells of the petiole, but anticlines parallel alternately to walls IV, V are formed continually until the pinna is two millimeters broad or more (Fig. 2i). The additional sections thus formed divide like the earlier ones by periclines to form the three meristem-layers of the lamina. The marginal cells also divide frequently by anticlines perpendicular to the edge of the pinna, thus constantly increasing in number and giving the pinna a fan-like shape with rounded outer edge formed by the actively dividing marginal cells (Fig. 19), as was shown by Hanstein. The pinnae are directed more ventrally than laterally from the petiole, and the upper pair soon come to have their upper or ventral surfaces nearly in contact, while the lower and older pair fold together (Fig. 21) to enclose the younger ones between them in the bud (Fig. 25). A branch of the axial bundle is given off to each pinna, which branches to form the anastomosing veins characteristic of Marsitia ; but the exact development of these bundles of the pinna was not studied, and I cannot state whether they arise, as Sadebeck (74) has shown them to, in Aspteiiium. The epidermal cells of the leaf give rise to stomata on the upper or both sides, and to deciduous trichomes like those of the petiole. Leaf and Sporocarp in Marsilia quadrifolia , L. 127 The Sporocarp. The bean-shaped sporocarps of Marsilia quadrifolia are usually borne in pairs, the stalks of the two uniting below, as shown by A. Braun (TO), to form a common stalk joining the petiole of the fertile leaf on its inner side near the base. Occasionally but a single sporocarp is found, or two with stalks separately inserted on the petiole, or more rarely three or four, usually with a common stalk. In the half-grown sporocarp we find the smaller or younger one of the pair is borne on the side of the stalk toward the petiole. If sporo- carps occur on any leaves of a given branch they are usually found on all. Plants of Marsilia which were left out of water in September by the drying up of a pond, matured many more sporocarps than plants growing where the water-level was constant. Although the latter had an equal number of young sporo- carps in July, nothing but small and often shrunken rudiments were found on most of the plants in September ; these might, however, be borne on large and well-developed petioles, so that there is no regularity in the retardation in development of the fertile leaves. Bischoff (’28) says the sporocarp of Marsilia arises as a slight prominence on the anterior side of the petiole, while Mettenius (’46) states that it originates endogenously, and later breaks through the epidermis of the petiole to form a solid mass of tissue, in the interior of which later the sori and canals are developed. The youngest sporocarp studied by Russow (72) had a two-sided apical cell, but was already differentiated into stalk and capsule (probably about the stage of that shown in Fig. 42). He thought the soral canals arose by the splitting apart of certain cells in the interior of the capsule and the formation of pits on the ventral surface into which these slits opened, to close again later by the growth of the cells on the ventral surface. On the walls of these canals arose the ‘ soral cells,’ in each of which later a tetra- 128 Johnson. — On the Development of the hedral apical cell was formed, which cut off a number of segments that, according to Russow, gave rise to the placenta with its vascular bundle and to the microsporangia, while the apical cell itself finally became the macrosporangium. Goebel (’82) states that the soral canals of Marsilia are external in origin, and that the sporangia arise from super- ficial cells ; Blisgen (’92) describes the first rudiment of the sporocarp as ‘ eine scheinbare grosse Liicke ’ in the tissue of the young leaf, and he thinks it probable that all the soral cells of each sorus are derived from a single superficial cell of the ventral surface. The placenta, microsporangia and macrosporangia, he states, are formed as Russow has described from these soral cells. According to my own observations on M. quadrifolia , the sporocarp makes its appearance when the young fertile leaf consists of about six or seven pairs of segments, and thus long before the appearance of the lamina. It is developed from an apical cell exactly like that of the leaf, formed in one of the ultimate marginal cells of the inner side of the petiole (F. m. c., Fig. 22) and placed transversely to the latter. The marginal cell involved may be either the upper or lower of (apparently always) the second segment of the inner side of the petiole, though, because of the crowding together of the various rudiments of the bud, this could not be made out with certainty (F., Fig. 23). The sporocarp is thus not, strictly speaking, epidermal in origin, but resembles closely in its origin the single sporangium of Lygodium from a mar- ginal cell of the fertile pinnule, as described by Prantl (’81). The apical cell of the sporocarp thus formed goes on cutting off segments, alternately toward the base and apex of the leaf, or to the right and left of the sporocarp itself, until about twenty-three pairs of segments are formed. It thus gives rise to a papilla, much like the very young leaf, which bends laterally to grow up beside the petiole with its ventral side facing in the same direction (Fig. 24), and then bends ventrally upon itself at the point where the stalk joins the capsule (Fig. 25). Finally, at about the time that the Leaf and Sporocarp in Marsilia quadrifolia , L, 129 activity of the apical cell is ended, by the appearance in it of a periclinal wall, the capsule or upper part of the sporocarp lies with its ventral side nearly in contact with that of the stalk (Fig. 31 a). The capsule is at this time about 1 mm. long, and the sori at the base about as far developed as that shown in Fig. 36. There is never any curling in of the extreme tip of the capsule, suggesting the circinate coiling of the leaf, and the sharp bending mentioned above is partially straightened out later, as Russow has shown, by the more rapid growth of the capsule at the base. The shape and size of the segments of the apical cell are very nearly like those formed in the leaf, and their earlier divisions are exactly the same. Walls I and II (Fig. 27) are followed by the transverse anticline dividing the marginal cell (t. a.1, Fig. 34), and wall III is formed in the same position as in the leaf. This is followed, however, by another anticline parallel to III (IV, Fig. 27), and then the regular alternation is resumed, wall V being on the ventral side, and VI, the last wall, on the dorsal side of the marginal cell (Fig. 28). We thus have one more section dorsal to the marginal cell than in the leaf, and the ultimate marginal cell is of the seventh grade instead of the sixth. The position of the section-walls given above is in general that found in all of the segments of the sporocarp, but certain exceptions are worthy of note. Thus wall IV, instead of running through to wall II as usual, often bends down to meet wall III at some distance from II (dotted line, Fig. 27). This type of division, however, was never seen in the lower or basiscopic marginal cells of the soral segments of the capsule. Again, section V is usually narrower in the basiscopic marginal cells of the soral segments of the capsule ( V, Fig. 34), and hence the basiscopic ultimate marginal cells, which are evidently ‘ the sorus mother-cells ’ of Biisgen, are the largest ones of the ventral side of the capsule. Finally, any marginal cell of the sporocarp, except the basiscopic one of the soral segments, may form a pericline instead of wall VI, and thus make the ultimate marginal cell of the sixth grade. This 13° Johnson . — On the Development of the behaviour is apparently analogous to that of certain marginal cells of the fifth grade in the leaf. When the second sporocarp of a pair is formed, it arises usually from a marginal cell of the second or third segment of the first sporocarp on the side of the latter toward the petiole on which it is borne ( F 2, Fig. 26), and a third probably arises in the same way from the second. The position of the apical cell of this second sporocarp, with reference to the first, is transverse, like that of the apical cell of the first with reference to the petiole. This mode of origin of the younger sporocarp from older ones shows that the common stalk of the pair is simply the portion of the stalk of the older one below the point of origin of the second, and the same is true of the stalk common to the second and third sporocarps of a trio. It is probable that where two or more sporocarps are inserted on the petiole by separate stalks, as happens occasionally in M. qtiadrifolia> and constantly in forms like M ’. polycarpci, we should find them to originate from marginal cells of successive segments of the petiole, but the early stages of this type were never seen. The Stalk. The further development of the seven divisions of each segment mentioned above differs in the different regions of the sporocarp. In the four or five oldest pairs of segments that form the stalk, their later history is very like that of the segments of the petiole, except that the axial bundle is here formed entirely from the plerome of sections I and II and a part only of that from section III. All the remaining portions of these segments, and all of the other segments, give rise to mesophyll and to hypodermal and epidermal structures, but no tannin-sacs are formed among the mesophyll-cells, and only very small air-canals between these and the hypodermis. A structure is thus formed of smaller diameter than the petiole, and of much firmer tissue, which swells out at the upper end (/. t., Fig. 44) to form the lower tooth of the capsule. Leaf and Sporocarp in Marsilia quadrifolia , L. 131 The Capsule. In the seventeen or eighteen segments forming this part of the sporocarp, we find that plerome- and dermatogen- walls are formed in each section, as in those of the petiole (Figs. 27, 28), and the halving anticline of section I is followed by periclines cutting off another portion of plerome from each half. The dorsal bundle, which is a continuation of the axial bundle of the stalk, is made up entirely from the plerome of section I (d.b., Fig. 31), some of the cells of which differ from all others of the capsule by remaining of the full length of the segment. The bundle is thus much more restricted in origin than that of the petiole. The dermatogen-layer in the capsule splits, as in the leaf, into epidermal and hypodermal layers, of which the former remains one cell in thickness and gives rise to stomata and deciduous trichomes, while the latter divides {hy.> Figs. 39,31) to form the two layers of thickened cells of the wall of the mature capsule, differing thus from the hypodermis of the leaf (Fig. 10). The periblem of the capsule gives rise to the several layers of loosely packed cells between the vascular bundle and the hypodermis, and between these cells and the latter are developed the numerous but small air-canals con- fined mostly to the dorsal side (a. c ., Fig. 29) and separated by partitions arising like those in the petiole. In the first three segments at the base of the capsule no sori are formed, but there is formed in the youngest pair from the plerome of sections III and IV, in a way to be described in treating the soral segments, a forked lateral branch of the dorsal bundle (Fig. 44). The plerome of sections II, V, and VI of this pair of segments, and all but section I of the next older pair, is apparently devoted to the formation of the basal portion of the gelatinous ring on which the sori are borne when the capsule bursts. In the oldest pairs of seg- ments there is formed a two-layered wall of thickened cells, like those of the hypodermis, stretching completely across the 1 32 Johnson . — Cbz the Development of the base of the capsule (b.w., Figs. 31 a, 42, 44). In the periblem above the dorsal bundle in these segments and in several of the older soral ones is developed a wall of thickened cells, enclosing between it and the dorsal hypodermal wall a lens- shaped mass of looser cells, the ‘ linsenformige Raum ’ of Russow (/. c., Fig. 44). There is an opening into this space just above where the dorsal bundle pierces the basal wall, and another at the anterior end, out of which there projects a rod of brown-walled cells (br., Fig. 44). The epidermal cells above this cavity swell out later to form the upper tooth of the capsule {it. /., Fig. 44). Each segment of the next eight or nine pairs give rise to a sorus. Section I in these segments develops much as in the segments of the stalk, as we have seen above. But the other sections have a peculiar history. Sections III, IV, and VI, dorsal to the marginal cell, widen rapidly at their outer ends, while sections II and V do not. The ultimate marginal cell is thus pushed around to a ventral position, the inter- polated section IV contributing largely to this end (Figs. 27, 29 ; cf. Figs. 8, 9). Finally, all other cells in this region grow out beyond the cells formed from the basiscopic ultimate marginal cell, and grow together over the outer ends of them, completely enclosing them (Figs. 29-33). The Vascular Bundle-System. The dorsal bundle, arising from the plerome of section I in all the segments of the capsule, gives rise in each soral segment to a lateral branch that runs down back of each sorus, between this and the lateral wall of the capsule. At a point about opposite the middle of the sorus the lateral bundle splits to three branches, as Russow has shown. Two of these (/. b . Fig. 33) continue on in the course of the single part of the bundle, while the third turns abruptly inward to connect with the placental bundle of the sorus (/. br.> Figs. 33, 44). The dorsal and single portion of the lateral bundle arises from the basiscopic half of the plerome Leaf and Sporocarp in Marsilia quadrifolia , L. 133 of section III, and from a part of the same region of section IV (/. b., Figs. 29, 31, 32). Of the two outer forks of this bundle one arises in the basiscopic quarter of the acroscopic half of the plerome of the same segment (/. b.f, Figs. 35, 39, 40), and the other from the acroscopic quarter of the acroscopic half of the next older segment. These forks are formed very early, but grow in length with the sorus, and finally on beyond it to the ventral edge of the capsule (Figs. 33-44), where their ends become connected in a more or less regular way with those of their fellows of the same side of the capsule (/. b.f., Fig. 44). The third or placental branch of the lateral bundle arises from a part of the plerome of the basiscopic half of section VI (pa., br., Figs. 31, 32, 33, 42), and the placental bundle with which this connects is developed from the plerome of the same part of this section (pa., b., Figs. 30, 32, 36, 41, 42). The Sori. Of the sections on the ventral side of the marginal cell, the plerome of section II develops ultimately into the large-celled tissue of the dorsal portion of the gelatinous ring (pi.2, Figs. 29, 31 ,g. r., Figs. 32, 33), described by Hanstein (’62) and Russow. The plerome of section V grows around under the inner end of the marginal cell (Figs. 28-32), and probably takes part ultimately in the formation either of the gelatinous ring or perhaps of the stalks by which the indusia are attached to the latter, but this was not determined with certainty. The periblem of both these sections apparently develops very slightly, and seems to form a part of the stalk of the indusium (pb., Figs. 28-31), but the boundary between this and the dermatogen soon becomes indistinguishable. The dermatogen of both sections grows rapidly in a radial direction (d., Figs. 28-31, o. ind., Figs. 32, 33), and gives rise to that portion of the indusium on the median side of the sorus. The outer or ventral cells of these sections soon grow over laterally to meet section VI, and thus enclose the cells of the sorus, while certain cells of these just below the ventral wall give 134 Johnson. — On the Development of the rise to part, if not all of the ventral portion of the gelatinous ring. The inner portion however, in the basiscopic half of the segments at least, remains of a single cell in thickness in each, even at maturity. We come now to the most important division of the soral segment, the basiscopic ultimate marginal cell (Figs. 34-38), from which are derived all the sporangia of the sorus. This is the ‘ Sorusmutterzelle ’ of Biisgen, but this name seems inappropriate as there is no single mother-cell of the sporangia of the sorus alone, nor of the whole sorus including the indusium after the single marginal cell of the third grade. No dermatogen-wall is formed in these marginal cells, and the sori being derived thus, like the young sporocarp itself, from a cell capable of forming at least two of the meristem- layers, are not of strictly epidermal origin. As the young sporocarp increases in size, we find that soon after the forma- tion of section VI the basiscopic marginal cell elongates in the direction of the length of the organ, and divides by a transverse anticline into halves, of which the acroscopic one soon comes to be the larger (Figs. 34-38). Then each of these divides by another anticline (Fig. 38), forming thus four cells, of which the basiscopic one of the acroscopic pair soon becomes the largest ( p . ma-sp. m. c ., Fig. 38), while its sister-cell on the acroscopic side splits by still another anti- cline. We have formed thus a series of five cells, of which the middle and larger one (p. ma-sp. m. c., Figs. 34, 38) is the primary macrosporangium mother-cell giving rise to all the macrosporangia of the sorus. The adjacent cells on either side of the latter (p. mi-sp ., m. c., Figs. 34, 35, 38) are the primary microsporangium mother-cells, while the outer cells of the five (i. ind ., Figs. 34, 38) give rise to the inner layer of the indusium on each side. The outer layer of the indusium on each side is formed by the splitting in two of the acroscopic marginal cell by a transverse anticline ( 0 . ind., Figs. 35-38), one half helping to form the indusium of the sorus of its own segment, and the other of the sorus of the next younger segment. Leaf and Sporocarp in Mar si li a quadrifolia , L. 135 In horizontal section it is seen that the three middle or sporangial cells become more densely filled with protoplasm than the indusial cells (Fig. 35), and also become separated at the ventral surface from the cells of section V (s. c Figs. 29, 35), thus forming the beginning of the soral canal. Otherwise the development of all of the five cells is much alike at first, and if we take transverse sections in the plane of the sporangial cells, we find that each elongates considerably in a radial direction (Fig. 39), and later divides into two by a pericline (Fig. 30). Then by the further growth and division of both of these cells ( ma.-sp . m. c Fig. 31), there is formed a row of seven or eight cells reaching from about the centre of the capsule nearly to the ventral surface (Fig. 32), all of them separated by the soral cavity or canal from that part of the indOsium formed from section V. In sagittal section (Fig. 43) it is seen that the microsporangial and indusial cells have divided in a similar manner. From the occurrence in them of nuclear spindles and their relation to the surrounding cells;, there can be no doubt that all the sporangial cells of the sorus are derived from the marginal cell, and none from cells dorsal to this in the interior of the capsule, as Biisgen thought possible. In the bending of the soral canal that takes place as it increases in length, the sporangial cells may come in contact with the inner layer of the indusium, but there is certainly no growing together, and the phenomenon has no significance. It is during the development of this row of soral cells that they are outgrown and finally enclosed by the surrounding cells, forming at first a ‘funnel-like pit’ at the ventral end of the soral canal (Figs. 30, 31), but finally closing entirely, though leaving traces of the fusion for a long time (.y. c.} Figs. 32, 33). While the sporangial and indusial cells have increased in numbers by the radial growth and division, there have been other important changes. The- macrosporangium mother-cells ( ma-sp . m. c ., Figs. 35, 36) are pushed by the growth of the plerome of section VI ( pa . b., Figs. 36, 37) out into the soral cavity, far beyond the microsporangium mother-cells, swell L i 36 Johnson. — On the Development of the laterally to several times their former size, and in so doing push the microsporangial cells around (Figs. 36, 37) to a position nearly at right angles to their former one. The macrosporangium mother-cell finally divides by three inclined walls to form the tetrahedral apical cell of the macro- sporangium (ma. sp., Figs. 32, 37). This apical cell cuts off two more segments on each of the three sides below (Fig. 41), which form the stalk and basal wall of the sporangium ; then a pericline is formed near the outer end of the apical cell, cutting off the archesporium (arc., Fig. 41) and completing the sporangium wall. The archesporium, as Russow has shown, then gives rise to the tapetum and spores. While the microsporangial cells are being pushed aside as described above, each has divided by anticlines approximately parallel to the segment wall, first to two (Fig. 36) and then to four (Fig. 41). These come to lie parallel to the segments of the apical cell of the macrosporangium, and are evidently the cells which Russow supposed to be segments of this, but there can be no doubt that they are really derived as described above. Of the four cells formed from each of the microsporangial cells as just described, the lower three go to form sterile tissue of the placenta (pa., Fig. 41), while only the upper one, next to the macrosporangium, actually forms microsporangia. Each of these upper cells divides by walls transverse to the axis of the sorus to form four cells on each side of each macrosporangium (which are well seen in a sagittal section of a capsule somewhat older than that shown in Fig. 43). Then each of these four cells swells out from the placenta, and divides into a basal cell (st. c., Fig. 41) and an outer cell, in which is formed later the tetrahedral apical cell giving rise to the stalk, walls, and archesporium. This basal cell of the microsporangium may be considered as homologous with the stalk-cell found in other Leptosporangiates, but nothing was seen in the development of the macrosporangium that could be regarded as such. In this latter respect Marsilia appears to differ from Pilularia, where Campbell (*93j states that such a cell is formed, at least occasionally. Leaf and Sporocarp in Marsilia quadrifolia , L. 137 In the development of the sporangial cells just described, the plerome of the acroscopic part of the basiscopic half of section VI has played an important part. The cells derived from this ( pa. b., Figs. 35, 36) push in back of the swelling macrosporangium mother-cell and between the placental cells (pa., Figs. 37, 41) derived from the primary microsporangium mother-cells. Most of these cells derived from section VI form the middle portion or axis of the placenta, but a row of them next to the base of the macrosporangium (pa. b., Fig. 36), and running the whole length of the sorus (pa. b., Figs. 32, 33, 44), develop the vascular bundle of the placenta, while at a point about opposite the middle of the sorus these same cells become modified across the whole width of sec- tion VI (pa. br., Figs. 31-33, 42, 44), to form the placental branch connecting the placental bundle with the lateral bundle. During this activity of the other cells of the sorus the indusial cells have been developing also. The acroscopic part of section V and the acroscopic marginal cell have each split by a transverse anticline (Figs. 35, 36) to form, in connexion with the cells of section II, the complete outer layer of the indusium (o. ind ., Figs. 35-37, 41, 43). The inner indusial cells derived from the basiscopic marginal cell and the basi- scopic portion of section V (i. ind., Figs. 32-37, 41, 43) complete the inner layer also. Each of these layers remains one cell in thickness throughout ; but by growth of the cells in a direction parallel to, and division by walls perpendicular to the surface of the indusium, the latter pushes out so as to accommodate the growing sporangia. During the growth of the indusium intercellular spaces appear at many points between the two layers (Fig. 41), and other larger ones between the outer layers of the indusia of adjacent sori, both laterally and along the median wall (i-s.c , Figs. 41, 42). By the increase in size of the latter spaces the indusia of adjacent segments become entirely separated, and the sori of each side of the capsule push into the furrows between the sori of the opposite side (Fig. 42). At a time a little i 38 Johns on. — On the Development of the before this happens the sori may appear opposite each other, though they are really alternate in origin, as we have seen above (Fig. 34). Finally we come to speak briefly of the last six or seven pairs of segments of the capsule, beyond the youngest soral segments. In these there is no single dorsal bundle, as this divides to two just beyond the origin of the lateral bundles of the last pair of sori. These two divisions run along nearly parallel to each other near the dorsal wall of the capsule (Fig. 44}, and each gives off three or four branches which arise, like those in the soral segments, from the plerome of sections III and IV, and are joined like those also with their fellows of the same side near the ventral margin. The exact region of origin of the two divisions of the dorsal bundle was not made out satisfactorily. All the plerome of this region, except the little devoted to the dorsal and lateral bundles, is apparently devoted to the formation of the gelatinizing tissue of this part of the capsule. Summary and Conclusions. The leaves of Marsilia arise in two rows on the stem each from a cell quite near the growing-point. The two-sided apical cell formed in this leaf mother-cell cuts off fifteen pairs of segments, and these are divided by radial anticlines into six main divisions, five sections, and an ultimate marginal cell of the sixth grade. Four of these divisions on each side take part in the formation of the axial bundle of the petiole, while all of them help to form the mesophyll and epidermal tissues. One quarter of the vascular tissue contributed by section II develops without further division to the large trachea of its side of the bundle, which has its oblique end-walls always inclined in the same direction. Fourteen air-canals are formed between the 'mesophyll and hypodermis of the petiole, and a single longitudinal row of the mesophy 11-cells gives rise to both the longitudinal and transverse partitions between Leaf and Sporocarp in Marsilia qnadrifolia , L. 139 these. Another longitudinal row of the same cells gives rise to each of the tannin-sacs. The pinnae or divisions of the lamina are formed by the continued activity of the marginal cells of certain segments, but their limits do not correspond exactly with those of the segments, the lower pair being nearly two segments in length and the upper pair about three. In its mode of origin, then, the leaf of Marsilia agrees closely with that of other leptosporangiate Ferns, as it does also in its further growth by the segmentation of a two-sided apical cell. But the position of the first division-walls in these segments, while very like that described for Asplenium serpentini by Sadebeck (’74), is apparently quite unlike that described for Ceratopteids by Kny (’75), for Onoclea by Campbell (’87), and that given by Campbell (’95, p. 325) for the Leptosporangiates in general. In the development of the lamina also Marsilia is unlike other described forms except Ceratopteris , since the pinnae are not co-extensive with the segments as in Onoclea and Asplenium , though all agree in having the pinnae formed by the activity of a series of marginal cells. There is however great need of more detailed work on the origin of the leaf and the differentiation of this into petiole and lamina. The sporocarp of M. qnadrifolia is developed from a trans- versely placed apical cell, arising in a marginal cell on the inner side of the young leaf. The second sporocarp when present (usually) arises in the same way from a marginal cell of the first. The two are thus respectively primary and secondary branches of the leaf. More rarely we may find two or more sporocarps inserted separately on the petiole, both on the same side. Then the suggestion is a tempting one, more especially in cases like M. polycarpa , where ten or more sporocarps may be borne in the same way, that the sporocarps represent pinnae homo- logous with those at the tip of the petiole, and the study of abnormal pinnae by Biisgen may perhaps seem to favour this. But before accepting this we have to account for the 140 Johnson . — On the Development of the occurrence of the sporocarps on one side of the petiole only, and also for their origin by a single apical cell instead of a series of marginal cells like the pinnae. Growth by the apical cell continues till more than twenty pairs of segments are formed. In the primary division of the segments one more section is formed dorsal to the marginal cell than in the leaf. The epidermis is formed much as in the leaf, but the mesophyll and its air-canals are less developed, while the hypodermis is of two much-thickened layers. The longitudinal bundle (axial in the stalk and dorsal in the capsule) is derived from section I only ; the lateral branches of this in the capsule are formed in sections III and IV ; and the placental bundle and branch from section VI. The sporangia of each sorus are all derived from one macro- sporangial cell, and two microsporangial cells are formed in the basiscopic marginal cell of each soral segment. The microsporangia and the macrosporangia are thus derived from sister-cells, and the former do not come from segments of the apical cell of the latter as described by Russow and Biisgen ; neither is the view of these authors as to the origin of the placental bundle from these same segments the correct one, as was stated a few lines above. A stalk-cell, homologous perhaps with that of the other Leptosporangiates, is formed in the development of the microsporangium, but nothing that could be so interpreted was seen in the macrosporangium. The soral canals arise by the separation of the primary sporangial cells from the outer cells of section V, and are entirely external in origin. The indusium surrounding each sorus arises by the more rapid growth of the superficial cells of the ventral side of the capsule which grow out and close together over the ends of the sporangial cells. Its develop- ment thus seems to warrant the statement that it is a true indusium morphologically as well as physiologically. The gelatinizing tissue of the dorsal part of the capsule is apparently the equivalent of a part of the vascular tissue of the petiole, while that at the ventral edge probably comes Leaf and Sporocarp in Marsilia quadrifolia , L. 141 from the outer, and that at either end of the capsule from all three meristem- layers. The walls of the capsule, including the vascular bundle- system. are developed entirely, or practically so, from the four sections in each segment dorsal to the marginal cell. Hence the two valves into which the capsule splits at bursting cannot be homologized with the divisions of the lamina, since these are developed from the numerous sections formed on both sides by the continued activity of the marginal cells. For this reason also any seeming similarity in the branching of the vascular bundle-systems of the two organs can have no meaning in the direction of homology. We have here then another reason, in addition to the one mentioned above in speaking of the mode of origin of the sporocarp from the petiole, for not believing with Goebel that it represents a single leaflet or pinna with its edges folded in to meet at the ventral margin of the capsule. And the same objections hold against other views involving a belief in the laminar nature of the valves, such as that of Russow and Btisgen, who regard the capsule as made up of two leaflets with ventral surfaces facing each other, or that of Campbell and Meunier, who compare it to a folded pinnate leaf with a sorus for each pinna. As far as developmental history gives any clue, the sporo- carp of Marsilia is homologous with the petiole only of the sterile branch of the leaf. But before adopting this unre- servedly we have to explain why there should be the marked difference in the development of the longitudinal vascular bundle in the two, especially in such very similar structures as the petiole and stalk. So far as we have light at present, then, we may consider the capsule as the swollen end of a petiole in which the marginal cells are devoted to the formation of the sporangia instead of a lamina. 142 Johnson . — On the Development of the List of Writings Referred to. Bischoff, G. W., ’28: Die kryptogamischen Gewachse. Nuremberg, 1828. Bower, F. O., ’89 : The Comparative Examination of the Meristems of Ferns. Ann. Bot., Vol. iii, 1889. Braun, A., ’70 : Ueber Marsilia und Pilularia. Monatsb. Berl. Akad., 1870. Busgen, M., ’90: Untersuch ungen iiber normal e und abnormale Marsilienfriichte. Flora., Bd. lxxiii, 1890. Campbell, D. H., ’87 : The Development of Onoclea Struthiopteris. Mem. Boston Soc. Nat. Hist., 1887. — » — ’93 : The Development of the Sporocarp of Pilularia Americana. Bull. Torrey Bot. Club, Vol. xx, 1893. ■ ’95 : The Structure and Development of the Mosses and Ferns. New York, 1895. Goebel, K.^’82 : Ueber die ‘ Frucht ’ von Pilularia globulifera. Bot. Zeit., Bd. xl, 1882. Hanstein, J., ’62 : Ueber eine neuhollandische Marsilia. Monatsb. Berl. Akad., 1862. ’65 : Die Befruchtung und Entwickelung der Gattung Marsilia. Pringsh. Jahrbiicher, IV, 1865. Hofmeister, W., ’62 : On the germination, development, and fructification of the Higher Crypt ogamia, &c. Ray Soc., London, 1862. Klein, L., ’84 : Vergleichende Untersuchungen liber Vegetationspunkte dorsi- ventraler Fame. Bot. Zeit., Bd. xlii, 1884. Kny, L., ’75 : Entwickelung d. Parkeriaceen dargestellt an Ceratopteris thalic- troides. Nova Acta, &c., Bd. xxxvii, 1875. Mettenius, G., ’46 : Beitrage zur Kenntniss der Rhizocarpeen. Frankfurt a. M., 1846. Meunier, A., ’87 : La Pilulaire. Etude anatomico-genetique du Sporocarpe chez la Pilularia globulifera. La Cellule, tome iv, 1887. PoiRAULT, G., ’90 : Recherches d’histogenie vegetale. Mem. de l’Acad. Imp. de St.-Petersbourg, tome xxxvii, 1890 L. Prantl, K., ’81 : Untersuchungen zur Morphologie der Gefasskryptogamen, Heft ii, Die Schizaeaceen. Leipzig, 1881. Russow, E., ’72: Vergleichende Untersuchungen d. Leitbiindelkryptogamen. Mem. de l’Acad. Imp. de St.-Petersbourg, tome xix, 1872. Sadebeck, R., ’73 : Wachsthum der Farnwedel. Verhandl. d. Bot. Vereins der Prov. Brandenburg. Bd. xv, 1873. — . — ’74 ; Ueber die Entwickelung des Farnblattes. Berlin, 1874. 1 I regret that Poirault’s work did not come to my notice before the present paper went to press, so that I might have mentioned certain points in his work concerning the origin of tissues in the leaf of Marsilia which my own work confirms. Leaf and Sporocarp in Marsilia quadrifolia , L. H3 EXPLANATION OF FIGURES IN PLATES X, XI, AND XII. Illustrating Dr. Johnson’s paper on the Development of Marsilia quadrifolia. Abbreviations used : — A., direction of the apex ; a. b., axial bundle of the petiole ; a. c., air-canal ; arc., archesporium ; a. s. w., acroscopic segment wall ; B., direction of base ; b. c. tc ., basal cell of the trichome ; b. m. c., basiscopic marginal cell ; br ., branch (axial bud) ; b. r. , rod of brown-walled cells ; b.s., bundle-sheath ; b.s.w., basiscopic segment-wall ; b. w., basal wall of the capsule ; c., capsule ; c.p., trans- verse partition; c.p.c ., transverse partition-cell; c.p.p ., spores in transverse partition; /)., dorsal side; d., dermatogen; d1., d2., &c. dermatogen of sections I, II, &c. ; d.b., dorsal bundle of sporocarp ; d. w., dermatogen-wall ; ep ., epidermis ; F., sporocarp ; F K, first sporocarp ; F2., second sporocarp ; F. m. c., mother-cell of sporocarp ; g. r ., gelatinous ring ; h. a ., halving anticline of section I ; hy., hypo- dermis ; ind ., indusium ; i. ind ., inner layer of indusium ; i. s., intercellular space ; i.s. c., inter-soral cavity ; L., leaf ; l.a., first longitudinal anticline in plerome of section V ; /. b., lateral branch of dorsal bundle ; l. b.f, fork of the lateral branch ; l.p., longitudinal partition; l.p.c., longitudinal partition cell; l.p.p., pores in longitudinal partition ; /. t ., lower tooth ; ma-sp., macrosporangium ; ma-sp. m. c., macrosporangium m other- cel 1 ; m.c., marginal cell; m. c1., in. c.2, &c., marginal cell of the first, second, &c. grade; mi-sp ., microsporangium ; mi-sp. m. c., micro- sporangium mother-cell ; mp., mesophyll ; m. w., median wall ; 0 , hy., outer layer of hypodermis of capsule; o.ind., outer layer of indusium; p1., lower pinna; p2., upper pinna ; pa., placenta ; pa. b., placental bundle ; pa. br., placental branch ; pb1., periblem of section I; pb2., periblem of section II; pb., periblem ; p.c ., partition-cell ; pi., plerome ; pi1., plerome of section I ; pi2., plerome of section II ; pi. w., plerome-wall ; p. ma-sp. me., primary macrosporangium mother- cell; S ., stem ; sc., soral cavity ; st., stalk ; st. b., stalk-bundle ; st. c„ stalk-cell of microsporangium ; s. w., segment-wall; ta1., to2., &c. first, second, &c. transverse anticlines of marginal cell ; tc., trichome ; tp., tapetum ; tr., trachea ; t. s., tannin- reservoir; u. t ., upper tooth; V., ventral side; X., apical cell; I. II. Ill, & c. first, second, &c. section-walls ; 1. 2. 3., &c. first, second, &c. segments of the apical cell on one side. All figures are camera drawings, and all are from microtome-sections, except Figs. 25, 31% and 44. PLATE X. Fig. 1. Transverse section of stem through apical cell of young leaf, x 300. Fig. 2. Sagittal section of a leaf nearly at the end of apical growth, x 200. Fig. 3. Part of a sagittal section of the petiole of an older leaf, x 300. Fig. 4. Ventral surface of tip of a young leaf, x 400. 144 Johnson. — On the Development of the Fig. 5. Dorsal surface of the tip of a similar leaf, x 400. Fig. 6. Half of a nearly transverse section of a young leaf showing the shape of a segment and the position of the first two section- walls, x 700. Fig. 7. The same still older, x 750. Fig. 8. Transverse section of petiole in which epidermal and hypodermal layers are completed and the partition-cells are nearly ready to cut off the cross-partition - cells, x 750. Fig. 9. The same section of a still older petiole, x 400. Fig. 10. Transverse section of a nearly mature petiole, x 60. Fig. 11. Tangential section of petiole showing the air-canals and partitions, x 400. Fig. 12. The same in a slightly older petiole, x 400. Fig. 13. The same still later showing the lengthening of the longitudinal partition, x 400. Fig. 14. Surface view of a longitudinal partition showing the pores in a nearly mature petiole, x 125. Fig. 15. Surface view of petiole showing the arrangement of the trichomes. x 750. Fig. 16. Nearly mature stage of same, x 300. Fig. 17. Nearly mature trichome. x 60. Fig. 18. Horizontal section of the tip of a leaf, showing the beginning of the first pinnae, x 400. Fig. 19. Sagittal section of a leaf through one of the well-developed lower pinnae, x 400. Fig. 20. Transverse section of a leaf through pinnae, a little later than Fig. 18. x 300. Fig. 21. A similar section still later, x 125. Fig. 22. A transverse section of petiole showing origin of sporocarp. x 750. Fig 23. Part of an approximately horizontal section of base of a leaf showing the apical cell of sporocarp. x 75°* Fig. 24. Transverse section of stem, and a young leaf with two sporocarps, all three nearly parallel to the stem, x 150. Fig. 25. Inner side of a young leaf with a sporocarp in which the segmentation of the apical cell is nearly finished, x 25. PLATE XI. Fig. 26. Nearly horizontal section of a stem through two leaves, an axillary branch and two sporocarp-rudiments, showing the under surface of part of older leaf, and cross-section of the first sporocarp arising on this ; also the origin of a second sporocarp from the first, x 400. Fig. 27. Transverse section of a young sporocarp. x 75°* Fig. 28. Transverse section through the basiscopic ultimate marginal cell of a sporocarp after all six sections are formed, x 75°- Fig. 29. The same section of a capsule at the time of beginning of soral canal, showing relative thickness of older and younger walls, x 750. Fig- 3°* Part of transverse section of older capsule. x 75°- Fig. 31. The same still older, x 400. Leaf and Sporocarp in Marsilia quadrifolia , L. 145 Fig. 31®. A slightly older sporocarp than Fig. 25, on a petiole, showing capsule bent against the stalk. Fig. 32. The same as Fig. 31 at the time of closing of soral canal, x 500. Fig. 33. The same still later, x 45. Fig. 34. Ventral surface of capsule a little older than Fig. 25. x 750. Fig* 35* Part °f horizontal section of about the age of Fig. 30 near the ventral surface, x 750. Fig. 36. Horizontal section of capsule about the age of Fig. 31 near the ventral surface, x 750. Fig* 3 7. The same of about the age of Fig. 32. x 750. Fig. 38. Sagittal section through the marginal cells of capsule about the same age as Fig. 29. x 750. Fig. 39. Sagittal section through sections III and IV of capsule, the same age as the last, x 750. PLATE XII. Fig. 40. Sagittal section through sections III and IV of a capsule of about the age of Fig. 31. x 750. Fig. 41. Part of horizontal section of capsule of the age of Fig. 33, x 750. Fig. 42. The same of whole capsule of same age as Fig. 41. x 50. Fig. 43. Sagittal section through sori of same capsule as Fig. 40. x 750. Fig. 44. View of inner side of one of the valves of a nearly mature capsule, x 8. TKW. rhW- D.S. Johnson del. XI - f ~L -L i=^ '1 >*■ 1 1 ] ! J , .1 i TT !rOJ yry ^JcT M r4i I mi -J - ytruzals ofBo&xny JOHNSON. — MARSIL1A. Voi.i/tn.x. T71W. University Press, Oxford '4r Annuls of Bo tuny V0IJ/4MX. D.S.Johnson del. JOHNSON. MARSILIA. University Press. Oxford tAnsuzZ# ojFBo£an$? D.S. Johnson del. Vol.M,PLI/. University Press, Oxford. t.&mals of Botany Vol.P//,PLAY. o irui rru.sp.m.c. V m,w s.c. T/i _J_n Y :~i,.iruL- mi-sp.TM:- -b.svr. Lind. V ; p.mi-spi D.S.Johnsoii del. University Press, Oxford. IVI A R S 1 L 1 A Vol.X/lPl.111. ^Annals of Botany oinch. pcc.br m..w. S.TTUL-Sp. b.ijzcb. D.S. Johnson del. University Press, Oxford. JOHNSON. MARS1L1A. On some points in the Histology of Monocotyledons. BY JOHN PARKIN, M.A., Trinity College , Cambridge. With Plate XIII. I. Observations on the Raphides (Figs. 1-12). HEN recently examining microscopically a number V V of Monocotyledonous leaves and reserve-organs for the purpose of investigating the occurrence and distribution of carbohydrates in them, I also paid some attention to the raphides so often present, and accumulated certain facts con- cerning them, some of which are apparently new. It was my intention to make a more complete investigation of this class of calcium-oxalate crystals ; but at present, not having sufficient time at my disposal, it seems worth while to notify the few features that have come under observation. The word ‘ raphides,’ introduced first by De Candolle, is used to denote bundles of needle-shaped crystals which are arranged generally, but not necessarily, in a parallel manner. Each bundle arises in and occupies a single cell. Such cells I term raphide-cells. Although raphides occur in some Dicotyledons, they are pre-eminently characteristic of Mono- [Annals of Botany, Vol. XII. No. XL VI. June, 1898.] 148 Parkin. — On some points in the cotyledons, often existing in great abundance in the leaves, stems, and especially the reserve-organs. Raphide-cells are very constant in character, only differing from one another in size and in the length of the individual crystal. The bundle of raphides does not fill the cell-cavity, but is suspended in a mass of mucilage. The nucleus and protoplasm originally present in the developing cell disappear when it is fully formed, as Hilgers1 showed in the case of Polygonatnm. In tissue preserved in alcohol, the bundle is seen lying in the centre of the cell surrounded by a sheath composed of the precipitated mucilage. In the genus Iris the typical raphide-cells are absent, but crystal-sacs of a different kind are present. These contain each a large acicular crystal, and possess nuclei and proto- plasm, but no mucilage. They have been known for a con- siderable time, and have been investigated by Hilgers 1 in two species of the genus. De Bary 2 also refers to them. What I desire to point out is that intermediate stages occur in certain petaloid Monocotyledons between these two types of crystal-sacs ; although in using the term ‘ intermediate,’ I do not want it to be inferred that there is necessarily any genetic relationship between the two. The five cases which I have observed are found in Fnnkia ovata , Convallai ia majalis , Phormium tenax var. atropurpurea , Tritoma Uvaria , and Polianthes tuberosa. Funkict ovata. Raphide-cells of the usual large type exist in fair abundance in the leaf-lamina and petiole (Figs. 1 and 2). Besides these, however, there is another kind of crystal-cell, which is more widely distributed in the plant, occurring in the leaves and root-stock. They might, without special attention, be passed over as ordinary raphides. The petiole is a gcod region in which to observe these two forms. By carefully examining transverse and longitudinal sections the differences between the two become very manifest. 1 Hilgers, Pringsheim’s Jahrbiicher, Vol. vi, 1867-68, p. 285. 2 De Bary, Comparative Anatomy of Phanerogams and Ferns, p. 138: he refers to Unger’s Anatomic und Physiologie, 1855. Histology of Monocotyledons. 149 In a transverse section certain of the crystal- cel Is are seen to be of smaller diameter, to contain fewer individual crystals of greater sectional area, and to be without the envelope of mucilage (Fig. 3). They are arranged rather differently, being situated near the xylem, one to three being seen in a section to each bundle ; odd ones also occur in the ground-tissue along with the ordinary raphide-cells. In longitudinal section, the differences are more apparent. The new form of crystal-sac is narrower, and contains from ten to twenty needles of a larger size than the numerous ones composing the raphide-bundles. This type of crystal-sac also possesses both nucleus and protoplasm (Fig. 4). The raphide-cell averages about 250 //. in length, and the individual raphide about 60 \x. The crystal-sac is very little longer than the crystals themselves, which average 100 /uu, about double the length of the raphide. Convallaria majalis resembles Funkia closely as regards its crystals, possessing in its leaves both ordinary raphide- cells, and sacs containing larger and fewer needles without mucilage. The latter are the more numerous. The roots and rhizomes seem only to have the mucilaginous raphide- cells. Hilgers 1 examined the mucilage of the raphide-cells, but makes no mention of the other kind of crystal-sac. In Polygonatum multiflorum , however, I have not succeeded in finding any type of crystal-cell other than the mucilaginous raphide-cells. Phormium tenax var. atropurpurea. This plant furnishes a more interesting case. In the leaf the crystals which attract attention are very similar to those found in Iris , being solitary prisms in cells without mucilage arranged in longi- tudinal rows (Figs. 5 and 6). Careful observation, however, also shows here and there ordinary raphide-cells. These are very sparsely scattered. The others are comparatively common, and odd ones contain two crystals instead of the usual solitary one (Fig. 7). Hilgers, loc. cit. 150 Parkin. — On some points in the Tritoma Uvaria . In this plant none of the ordinary mucilaginous raphide-cells could be found either in the root, stem, or leaf; but the other kind are numerous, varying considerably in size, from 60 to 160 ju. in length. The crystals, which almost fill each sac, are fairly abundant. Nuclei and protoplasm can also be made out in them (Figs. 1 1 and 1 2). Polianthes tuber osa. The tuber only has been examined, and in it are found both kinds of crystal- cells. • The muci- laginous ones are more numerous and larger (Fig. xo). The others occur chiefly near the vascular bundles (Figs. 8 and 9). From the fact that these two kinds of crystal-sacs occur in the same plant and even in the same tissue, it looks as if they may have arisen independently, and not have been derived the one from the other. Nevertheless the above instances suggest that as the raphide-cells decrease in number their place is taken by the other type. In the order Iridaceae raphides seem wholly wanting, only the solitary crystals being present. Besides species of Iris , I have observed these prisms in the genera Crocus , Spar axis, Schizostylis , Xiphium , Freesia , Babiana , Gladiolus , and Mont- bretia ; and thus they appear characteristic of the order. It may be that the non-mucilaginous crystal-sacs discovered in Funkia , Convallaria , Phormium , Tritoma , and Polianthes are genetically connected with one another and with those of the Iridaceae. The distribution of special crystal-cells in the large petaloid family, the Liliiflorae, is instructive. The tribes Tulipeae and Allieae are exceptional in containing none1. They seem very scarce in the Colchicaceae ; Veratrum possesses a few raphide-cells, but I failed to find any in Colchicum or Uvular ia. The Iridaceae, as previously shown, possess the large solitary needles. The other groups, as far as I know, contain the ordinary mucilaginous raphide-cells, such as the Hyacintheae, Anthericeae, Yuccoideae, Hemerocallideae, 1 De Bary, loc. cit. p. 142. Histology of Monocotyledons. 1 5 1 Convallarieae, Asparageae, Dracaeneae, Pontederiaceae, and Amaryllidaceae. Then in addition to the ordinary raphide- cells, Funkia , Convallaria , Phormium , and Polianthes possess , crystal-sacs without mucilage ; while Tritoma appears to have the latter only. Funkia , Phormium , and Tritoma belong to the tribe Hemerocallideae ; Hemerocatlis fulva and H. flava, both of which I have examined, contain ordinary raphide-cells, but not the other type of crystal-sac. The Convallarieae come very near the Hemerocallideae in habit, differing chiefly in the baccate fruit. Thus the four genera, Funkia , Phormium , Tritoma , and Convallaria seem fairly closely allied. Polianthes is placed among the Agaveae, a sub-order of the Amaryl- lidaceae. It is generally considered that the Amaryllidaceae have been derived from the Liliaceae by the ovary becoming inferior, and sometimes it is inferred that the Iridaceae have arisen from the Amaryllidaceae. It may be that the Iridaceous forms began to appear just about the time the Amaryllideae were evolving, both having a common origin in some Liliaceous type. The similarity of the crystal-sacs would support a relationship between the Iridaceae and the Liliaceous tribe Hemerocallideae. A special study of these crystals, which are formed in the growing organs, and hence belong to Schimper’s class of primary calcium oxalates, might be of value from a phylo- genetic point of view, as well as a means of throwing light on their function, of which at present we seem very ignorant. II. An absciss-layer in THE LEAVES of Narcissus , Galanthus , AND Leucojum (Figs. 13, 14). Having had occasion to examine species of Narcissus , Galanthus , and Leucojum at various stages in their annual growth, I observed that the foliage does not simply die down and wither away, but that each leaf is detached from its tunicate base (bulb-scale) by means of a layer of cells M 152 Parkin . — On some points in the becoming merismatic. Not finding in botanical literature any description of such an absciss-layer in these plants, it seems worth while making a note of it, and pointing out a few details connected with the occurrence. Von Mohl 1 in his researches on ‘ leaf-fail ’ in reference to Monocotyledons, mentions merely the falling of their perianth- segments and immature capsules. Bretfeld 2 has investigated the mode of detachment of leaves in Dracaena , many of the Orchidaceae and Aroideae, and generalizes from the study of these plants, that, whereas the leaf-fall in Dicotyledons is brought about by a new tissue formed a short time before the shedding of the foliage, in Monocotyledons it results from the action of a special mechanism, produced along with the other tissues in the developing leaf, similar to the contrivance which brings about the dehiscence of dry pericarps. However, the method here observed resembles that of Dicotyledons, although it is perhaps simpler in detail. Some time before the leaves turn yellow, certain of the parenchymatous cells situated a little way above the tunicate base of the foliage-leaf become merismatic, and divide to form a zone of narrow cells with conspicuous nuclei and abundant protoplasm ; this region is visible to the unaided eye as an opaque line on holding the leaf up to the light (Figs. 13, 14). In Galanthns nivalis the cell-divisions were just commencing when examined on April 1 6, and were well advanced on May 1 ; by the end of May the leaves, having turned completely yellow, are easily detached by means of their absciss-layers from their swollen bases, which now become the scales of the bulb, full of reserve material. The epidermal and mesophyll-cells and nucleated cells belonging to the vascular bundles take part in the divisions ; the raphide-cells and, of course, the vessels remain passive, their lumina becoming obliterated by the pressure exerted on them by the adjacent dividing cells. As a rule, about four or 1 Mohl, Botanische Zeitung, i860. 2 Bretfeld, Pringsheim’s Jahrbiicher, xii, 1880. Histology of Monocotyledons. 1 5 3 five irregular layers of small narrow cells result from the activity. In these plants, besides the perfect foliage-leaves, there are sheathing phyllome-structures external to them, the lower parts of which also swell to form reserve bulb-scales, while the upper parts remain membranous and afford a covering to the active region of growth in the young foliage leaves. These upper sheathing parts are likewise divided off from their lower reserve-storing portions by the formation of absciss- layers. These in the case of Galanthus nivalis are well formed by the end of March, that is some time before those of the foliage-leaves. The plane of detachment is through the middle of the absciss-layer, and previously to the separation the walls of the newly-formed cells become suberised, giving a deep yellow coloration with iodine and sulphuric acid, whereas the other cell-walls stain blue. The plants in which these absciss-layers have been noticed by me are Narcissus P seudo-narcissus and N. poeticus , Galan- thus nivalis , Leucojum vernum and L. aestivum. No doubt other species of these genera exhibit them, and most likely other bulbous Amaryllideae. Such an absciss-layer, by means of the corky walls formed and by the closure of the vessels, may possibly be a protection against the entrance of Bacteria or fungus-hyphae into the scales, and also a check to the passage of water out of them. Botanical Laboratory, Cambridge. M % 3 1 54 Parkin . — Histology of Monocotyledons. EXPLANATION OF FIGURES IN PLATE XIII. Illustrating Mr. Parkin’s paper on the histology of Monocotyledons. Figs, i— 1 2 taken from sections, cut from spirit-material and mounted in alcohol, so as to retain the mucilage in the cells. Drawn under a Reichert No. 5, objective. Funkia ovata — petiole. Fig. 1. Transverse section of a raphide-cell. Fig. 2. Longitudinal section of the same : r. bundle of raphides ; m. mucilage precipitated by alcohol. Fig. 3. Transverse section of a crystal-sac with mucilage. Fig. 4. Longitudinal section of the same : c. bundle of acicular crystals ; n. nucleus ; p. protoplasm. Phormium tenax var. atropurpurea — leaf. Fig. 5. Transverse section of a crystal-sac without mucilage. Fig. 6. Longitudinal section of the same. Fig. 7. Transverse section of a crystal-sac with two crystals : c. solitary crystal. Polianthes tuber osa — tuber. Fig. 8. Transverse section of a crystal-sac without mucilage. Fig. 9. Longitudinal section of the same. Fig. 10. Longitudinal section of a raphide-cell : r. bundle of raphides ; c. bundle of acicular crystals ; n. nucleus ; p. protoplasm. Tritoma Uvaria — root-stock. Fig. 11. Transverse section of a crystal-sac. Fig. 12. Longitudinal section of the same: c. bundle of crystals; n. nucleus; p. protoplasm. Narcissus poeticus — foliage leaf- base. Fig. 13 (natural size). Line a-b shows the position of the absciss-layer; l. upper leaf ; s. bulb scale. Fig. 14. Semi-diagrammatic longitudinal section through region marked c-d in Fig. 13. The absciss-layer in process of formation is shown by the zone of narrow cells recently formed with conspicuous nuclei : v. vessel ; r, raphide-cell ; the lumens of both being obliterated by the pressure of the dividing cells ; oe. outer or lower, ie. inner or upper epidermis. , /l/t ru:/Is o f Botany VoU//,PLXtU On Aecidium graveolens (Shuttlew.)1. BY P. MAGNUS, Berlin. With Plate XIV. IT was stated by Jacob Eriksson in his work, Studien liber den Hexenbesenrost der Berberitze (. Puccinia Arrhejia- theri , Kleb.2), that the mycelium of the Rust of the Witches’ Brooms of Berber is vulgaris (which had been named Aecidium magellanicum , Berk., in consequence of my remarks on the subject in Hedwigia, 1876) grows in the interior of the cells of the cambium of the witches’ broom. In a paper published in the Berichte d. Deutsch. Bot. Ges., Vol. xv, 1897, pp. 148- 152, I questioned these observations, and showed that an intercellular mycelium occurs in the branches of this plant in the pith, in the cortex and in the phloem, which sends out knot-shaped haustoria into the neighbouring cells ; but I was unable to observe the intracellular mycelium described by Eriksson. In a subsequent paper3 this author criticized my results, stating that he had observed only the cambium, while I had given most attention to the pith and the tissues of the cortex, and therefore my observations counted for little as regards 1 Read before Section K of the British Association, Toronto, 1897. 2 Cohn, Beitrage zur Biologie der Pflanzen, Vol. viii, Heft I. 3 Berichte der Deut. Botan. Ges., Vol. xv, 1897, pp. 228-231. [Annals of Botany, Vol. XII. No. XL VI. June, 1898.] 156 Magnus . — On Aecidhim graveolens ( Shuttlew .). the cambium. As I had, however, looked for the mycelium in all the tissues of the stem, as is evident from my paper, ‘ Ueber das Mycelium des Aecidium magellanicum , Berk1,’ this criticism is of little weight. Eriksson in fact admits that I had investigated the cambium when he quotes my statement that I did not find the mycelium in it. He further states that while he had worked with living material, I had only examined specimens preserved in alcohol. I do not regard such an objection as a reasonable one ; nevertheless this year I have expressly used living material, and have obtained the same results. I have also extended my observations to ascertain how and where the mycelium develops in the branches of the witches’ brooms, how it arises in the buds, and how it develops in the leaves which, even in the bud, are covered with the fructification of the Fungus. For these observations I was able to use material rich in the fungus which was sent to me by Hofgartner Reuter from the Pfaueninsel near Potsdam, on April 24, 189 7, and by Herr J. A. Baumler, from Pressburg in Hungary, on May 8, 1897. I owe my best thanks to these two gentlemen for this material. The latter was especially valuable to me, inasmuch as it consisted of a large number of young buds growing into elongated branches. Of course I was not able to decide, in response to Eriksson’s wish, whether this mycelium is derived from aecidiospores or from sporidia of germinated teleutospores. Up to the present time we do not know of any observations supporting the idea that such mycelia are different in any respect except in their extension in the host-plant. I have shown in earlier papers that the mycelium developed from the sporidia of the hibernated teleutospores often extends further into the tissue of the host-plant than that derived from the germinating uredospores or aecidiospores, which form in most cases a mycelium restricted to the region of infection (e. g. Puccinia Oreoselini , Strauss., and P. Cyani , Pass.) : but no other 1 Berichte der Dent. Bot. Ges., Vol. xv, 1897. Magnus,— On A ecidium graveolens (Skuttleiv.). 157 difference in the relation of these mycelia to the tissues of the host-plant has ever been observed. Therefore I cannot accept Eriksson’s hypothesis that the mycelium of the aecidiospores can bear any relation to the tissues of the branches of the witches’ broom, differing from that of the mycelium developed out of the germinating sporidia of the teleutospores. The investigation of the fresh material, as I have already mentioned, has confirmed my earlier observations. The hyphae are always intercellular, and put out knot-shaped haustoria into the cells between which they grow. They occur in the pith, in the cortical parenchyma, in the phloem, and in the medullary rays (I had forgotten to mention the last in my communication of February, 1897). In the me- dullary rays the walls of the cells on which the hyphae grow are swollen (Figs. 7-9), and the hyphae send out numerous knot-like haustoria into the parenchymatous cells. I could not be certain that haustoria are also sent out into the wood- cells bordering on the medullary rays. I believe that this does take place, but I am not satisfied that the cells in which it seems to occur are not medullary ray-cells. In the elon- gated cells of the phloem the branches of the haustoria are sometimes not coiled up into a knot, but lie free in the cell (Fig. 6). The branches are not straight, but are curved or crumpled, and constricted here and there. As I have de- scribed earlier (loc. cit.), the mycelium is formed in the young branches of the witches’ broom in the pith, in the medullary rays, and in the primary cortical parenchyma. In the latter it is often cut off from the surrounding tissue by the formation of cork in the infected tissue, so as to form island-like masses in the cortex h Out of the primary cortical parenchyma in the older twigs the mycelium grows into the phloem and gradually penetrates into it as development proceeds. In the phloem also the infected parts of the tissue are often enclosed as islands by the formation of cork. These ring-like formations of cork Vide Berichte der Deut. Bot. Ges., Bd. xv, 1827, Taf. iv, Fig. 5. 1 5 8 Magnus. — On A ecidium gravcolens ( S hut t lew'). give a very characteristic appearance to the transverse sec- tions of the older stems of the witches’ broom. In these mycelial hyphae, which have a very small lumen, I have never observed the yellow colouring-matter of the Uredineae described by Eriksson (loc. cit.) in the mycelial filaments of the cambium, which he states are intracellular. I will at once remark that I have generally observed the yellow colouring- matter of Uredineae only in the mycelial filaments which are exposed to the light, never in those portions of the mycelium which are embedded so deeply in the tissue of the host-plant as not to be exposed to the light. As I have already stated elsewhere1, the spermogonia appear on the whole surface of the first leaves, which are developed in April and the beginning of May, and the aecidia appear between the spermogonia ; on the later leaves of the infected buds are found single larger or smaller groups of aecidia only, while the latest formed leaves are altogether free from the Fungus. At the end of April or the beginning of May a large number of these buds had already put out branches with long internodes. The leaves of these branches are free from the Fungus, as I have said. If longitudinal sections of these long branches be examined at the end of April or beginning of May, the hyphae in the pith will be seen growing in a longitudinal direction into the region of the merismatic tissue (Fig. j). These mycelial strands are inter- cellular, and occur in the actually dividing cells of the parenchyma of the pith. The cells around these mycelial strands are sometimes more elongated than the others, and remain for some time in this condition, while the neighbouring cells are undergoing transverse division (Figs. 2, 4). This intercellular mycelium often sends into these neighbouring parenchymatous cells haustoria of the same kind as those already described (Figs. 2, 4). From these longitudinal mycelial threads horizontal threads grow out laterally (Fig. 2). These are the threads which grow into the medullary rays, 1 Verhandl. des Botan. Vereins der Provinz Brandenburg, Sitzungsberichte, 1875, pp. 87-89 (which was also published in Hedwigia-, 1876, No. I). Magnus. — On A ecidium graveolens (Shull lew.). 159 and especially into the spaces between the vascular bundles where the leaves come off. These mycelial strands grow next spring into the buds which are formed at this place, and form spermogonia and aecidia over the whole surface of their leaves. I will here shortly recapitulate these observations. The mycelium always grows between the cells, and gives off haustoria into them. In the first spring the hibernating mycelium grows into the developing buds, and forms sper- mogonia and aecidia on the whole surface of the first leaves. In the case of those short shoots which grow out into branches with long internodes, the mycelium grows directly into the pith and continues to grow with the merismatic tissue. This also takes place in the spring. From these medullary mycelial strands branches grow outwards, but these do not penetrate into the leaves. They pass through the medullary rays into the primary cortical parenchyma, and especially through the openings in the vascular cylinder, where the young leaves are given off to the axillary buds, from whence, in the following spring, the mycelium enters into the first developing leaves. As the branches of the witches’ broom increase in thickness, the mycelium spreads from the primary cortical parenchyma into the phloem. Both in the primary cortical parenchyma and in the phloem the rows of cells affected by the mycelium are enclosed more or less com- pletely by a cylindrical cork-formation, and are thus separated from the less affected tissue. What, then, is the tubular mycelium in the cambium-cells described and figured by Eriksson (loc. cit Fig. 5, PI. II), containing, contrary to what is observed in the mycelial strands of Uredineae growing in tissues not exposed to light, the yellow colouring-matter of the Uredineae? Leaving aside for the present the yellow colouring-matter, his figure re- sembles more than anything else the young cells of the cambium with horizontal transverse walls, the contents of which are contracted by plasmolysis. It is often the case in elongated cells that the plasmolyzed contents are for the 160 Magnus. — On Aecidium graveolens (Skuttlew.). most part retracted only from the longitudinal walls, while remaining more or less in contact with the short horizontal or oblique walls (Fig. 10). These contracted cell-contents then resemble a tube placed longitudinally in the empty cells, and if the plasmolyzed contents are not retracted from the common transverse wall of neighbouring cells, they present the appear- ance of a continuous tube passing through the cell-cavities. Eriksson’s figure appears to me to strongly suggest such plasmolytically contracted cell-contents, although it must be borne in mind that I have not examined Eriksson’s pre- parations. It is well known that the wood of the young Barberry is yellow, and that this is caused by the yellow cell-sap which is found even in the young wood-cells, thus indicating that the colouring-matter exists in a state of solution, and not in the form of yellow granules. The yellow colouring-matter of the granules observed by Eriksson in the tubes, which appear to me to be nothing more than plasmolyzed cell- contents, may be connected in some fashion with the colouring- matter of the young wood-cells. In making preparations, if the cells are cut across, the yellow sap flows out into the water, and then they appear colourless. The granules of the young wood-cells do not appear yellow. I cannot decide whether the yellow tint of the granules figured by Eriksson may not be due to the colouring-matter of the wood-cells seen through the cambium-cells (especially as Eriksson mentions in a foot-note to his paper in the Ber. d. Deutsch. Bot. Ges., Vol. xv, 1897, p. 229, that all the granules appeared yellow, and not as the lithographer has by mistake indicated, a few of them only), or whether they represent the first appearance of the yellow colouring-matter of the wood-cells. In no case have I been able to observe in the mycelium of any of the Uredineae not exposed to the light any trace of yellow colouring-matter. Even in the mycelium of the cortical parenchyma and the phloem no colouring- matter is found. It seems to me that these results are of special interest, because Eriksson has lately maintained the theory of the Magnus. — On A ecidium graveolens (. Shuttlew .). 1 6 1 ‘mycoplasmastadium’ of Uredineae1. In the case under discussion we can observe in succession the mycelium itself entering the buds, wintering there, and the following spring penetrating into the young leaves of the sprouting shoots, and there developing organs of fructification. The annual development of this parasitical Fungus does not exhibit the 4 mycoplasmastadium ’ of Eriksson. There is no ground here for such a theory. In conclusion I will give a short account of the identi- fication of the European species producing the witches’ broom on Barberry. In 1875 and 1876, when I had dis- tinguished the aecidium producing the witches’ broom of Berberis vulgaris from the aecidium of Puccinia graminis , I at first thought I had discovered a new species ; but I found afterwards, to my great surprise, that Berkeley had described in Hooker’s Flora Antarctica, II, pp. 450-451, a similar aecidium on Berberis ilicifolia , Forst., from the Straits of Magellan, and then I identified it as Aecidium magellanicum , Berk. On the ground of the similarity of appearance of the parasite, I believed that the aecidium of the witches’ broom in Europe was identified with this aecidium of Berkeley, and accordingly I named the fungus causing the witches’ broom of Berberis vulgaris, Aecidium magellanicum , Berk. All the later authors have followed me. Subsequently this identi- fication appeared to me doubtful, as the cultures of Peyritsch and Eriksson had demonstrated that Puccinia Arrhenatheri (Kleb.), Eriks., ( = P- magellanica , Peyr.) on A rrhenatherum elatius belongs to the aecidium of the witches’ broom of Berberis vulgaris , whilst no Arrhenatherum or Avena occurs in Patagonia. In the Berichte der Deutsch. Bot. Ges., Vol. xv, 1897, pp. 270-276, I have shown that in Patagonia and Chile another aecidium occurs on Berberis buxifolia , Lam., which causes witches’ broom, and which is well distinguished from the European species by the remarkable nestlike formation of 1 See ‘ Vie latente et plasmatique de certaines Uredinees,’ Comptes Rendus, 1897, Mars, and * Der heutige Stand der Getreiderostfrage,’ Berichte der Deut. Bot. Ges., Bd. xv, 1897, pp. 192-194. 1 62 Magnus . — On Aecidium graveolens (Shuttlew i). its growth on single nodes, by the withering of the branches, by the swelling of the nodes which bear the witches’ brooms, and by the absence of spermogonia. I have called this species Aecidium Jacobsthalii Henrici , P. Magn. It differs also from the Aecidium magellanicum , described by Berkeley on Berberis ilicifolia , Forst., which may be a third species attack- ing Berberis. The European Fungus forming the large witches’ broom on the Barberry, with a great many elongated erect branches, can therefore no longer be called Aecidium magel- lanicum. It may be designated either as Aecidium graveolens, Shuttlew., which name was found by Cooke in the Paris Herbarium, or as the aecidium of Puccinia Arrhenatheri (Kleb.), Eriks. I am deeply indebted to Mr. Harold Wager for having translated my German manuscript into English. EXPLANATION OF FIGURES IN PLATE XIV. Illustrating Prof. Magnus’ paper on Aecidium graveolens. The figures which accompany this paper were drawn from nature in my presence by Dr. Paul Roeseler. Fig. i. Longitudinal section of apex of spring shoot (in May, 1897) developed from a bud of witches’ broom. At M. M. the mycelium is seen growing in the pith up to the merismatic region of the apex, x 68. Figs. 2 and 3. Portions of Fig. 1 more enlarged. Fig. 2 from part marked A , and Fig. 3 from that marked B. Showing longitudinal tracts of intercellular mycelium with some haustoria : at A. a horizontal tract is given off. x 420. Fig. 4. Similar longitudinal section of pith at apex of spring shoot, with inter- cellular mycelium and haustoria. The parenchymatous cells near the mycelium are more elongated than the others, and less divided by horizontal walls. x 420. Magnus. — On A ecidium graveolens ( Shu ttlew. ) . 163 Fig. 5. Transverse section of young pith, with transverse sections of intercellular mycelium-threads. The walls in which the mycelium-threads occur are swollen, x 420. Fig. 6. Longitudinal section of the phloem of an infected branch of witches’ broom. In the elongated cells the haustoria are not knot-like. Fig. 7. Medullary ray in transverse section of an infected stem, with mycelium and haustoria. x 420. Figs 8 and 9. Medullary rays in tangential section of an infected stem, with mycelium and haustoria. The walls in which the mycelium-threads occur are much swollen, x 420. Fig. 10. Tangential section of the very young wood in the neighbourhood of the cambium. The cell-contents are plasmolytically contracted, and withdrawn from the longitudinal walls, but not from the oblique transverse walls, x 420. ftrmxxls of Botany P. Roseler del. MAGNUS.- M YCE LI U M OF AEC1D1UM GRAVE OLE N S. Vol.XIlPl.IIV Fig. 6. University Press, Oxford. 'A/utcds of Botany Fiy.3. -v Ficf.1. I P. Roseler del. B Fig.Z. a. MAGNUS.- MYCELIUM OF AECID1UM GRAVEOLENS. Vol.XIl, Flint University Press, Oxford. The Coagulation of Latex. BY R. H. BXFFEN, Frank Smart Student , Gonville and Caius College , Cambridge. WHILE engaged during the latter part of 1896 in studying the functions of latex, my attention was frequently called to its spontaneous coagulation when in contact with the air. De Bary describes the phenomenon as follows 1 : — ‘ As soon as latex comes in contact with the air, and still more quickly on treatment with water, alcohol, ether, or adds, coagula appear in the hitherto apparently homogeneous clear fluid itself, and independently of the aggregation of the insoluble bodies described by Mohl (Bot. Zeit. 1843, No. 33). The coagula collect together and separate with the insoluble bodies from the clear fluid. These phenomena of coagulation which appear under the action of so various agencies point especially to a complicated composition of the fluid, and deserve further investigation.’ An examination of the subject was therefore commenced with the small quantities of latex obtainable from plants grown for the purpose in the Cambridge Botanical Gardens. The results obtained were of some interest, and accordingly the experiments were continued, together with other researches 1 De Bary, Comp, Anat. of Phanerogams and Ferns, p. 184. [Annals of Botany, Vol. XII. No. XL VI. June, 1898. 1 66 Biffen. — The Coagulation of Latex. on a larger scale, in Mexico, Brazil, and the West Indian Islands. Rubber-yielding plants which always have laticiferous cells, were for the most part chosen on account of the ease with which large quantities of latex could be obtained, and because the various processes used in the preparation of crude rubber seemed likely to throw some light upon the subject. A microscopic examination of any one of these latices shows that its milky appearance is due to the presence of innumerable small granules of caoutchouc, which in themselves are soft and sticky, for they readily cohere to form a small mass of rubber if the cover-glass is lightly rubbed on the slide. Some of the processes employed to prepare this rubber may be described here. In the preparation of Para rubber, a thin layer of the latex of Hevea brasiliensis (Muell. Arg.) or other species of Hevea , is exposed to the action of the smoke of burning ‘ urucuri ’ nuts ( Attalea excelsa, Mart.); coagulation is immediately brought about, resulting in the formation of a soft, curdy mass of rubber, which on drying becomes tough and elastic. The same process is now being applied with good results to the preparation of Ceara rubber from the latex of Manihot Glaziovii (Muell. Arg.). The usually accepted explanation of this is that the water contained in the latex is simply evaporated off1 ; but as the coagulation is brought about in so short a time, and moreover as there is no loss of weight on its occurrence, this is obviously incorrect. On passing the smoke of the burning Attalea nuts through a condenser, condensation occurs and two layers of liquid are found in the receiver, one colourless and limpid, the other dark brown and oily. If these are separated by means of a pipette, or with a moistened filter paper, and analyzed, the former is found to consist mainly of acetic acid, and the latter of creosote and traces of pyridine derivatives. 1 Ernst, Trinidad Bulletin, vol. iii. p. 235. Biffen . — The Coagulation of Latex. 167 On adding acetic acid to the crude latex of Hevea coagulation occurs immediately. This process of smoking the latex may then be classed with those mentioned by De Bary under the heading of treatment with acids. As other examples, the preparation of Lagos rubber from the latex of Ficus Vogelii (Miq.), in which case lime-juice is added L, and Heifer’s process of adding acetic acid to the latex of Artocarpus Chaplasha (Roxb.) 1 2, may be quoted. It is worthy of note that the latex of Hevea brasiliensis is in itself alkaline, and that the addition of a solution of ammonia preserves it indefinitely from spontaneous coagula- tion. The addition of alkalies bring about coagulation, however, in the latex of Castilloa elastica. In Mexico and Nicaragua, where this tree abounds, a decoction is made of the stems of the Moonflower, Ipomoea bona-nox ( Calonictyon speciosum [Choisy] ), and added to the latex 3. The alkaline properties of this extract are well known to the native Indians, who frequently employ it in the manufacture of soap. The latex has an acid reaction towards litmus-paper, and the addition of acids does not cause coagulation. Another method of clotting latex is to add an excess of common salt. This method is almost invariably applied in the case of Hancornia speciosa (Gomez) to produce the ‘ mangabeira ’ rubber. It is also reported to have been employed at times to coagulate the latex of species of Hevea and Manihot Glaziovii (Muell. Arg.). Coagulation may also be brought about by boiling the latex, as, for example, in the preparation of ‘ balata ’ from Mimusops globos a (Gaertn.) in Venezuela and Trinidad. There are several other methods in general use besides the few that have been quoted, and many others have been suggested from time to time 4. 1 Kew Bulletin, 1890, Art. 142, p. 89. - Watt’s Diet. Economic Products of India, vol. iv. p. 343. 8 Belt, Naturalist in Nicaragua, p. 33. 4 For a complete account see Le Caoutchouc et la Gutta-percha , Seeligman, Lamy, and Falconnet, Paris 1896. N 1 68 Biffen. — The Coagulation of Latex . As the rubber exists in particles in the latex, it seemed possible that the centrifugal method of separation might be adopted in examining the phenomena of coagulation. A modified form of the ordinary centrifugal milk-tester was therefore designed capable of being rotated some 6,000 times per minute. The latex was taken directly from the trees, strained through wire-gauze to remove any pieces of bark, and then, if very thick, diluted to about the consistency of thin cream. The first experiments were made with the latex of Castilloa elastica. After centrifugalizing for from three to four minutes, the rubber- particles completely separated as a thick, creamy, white layer, from the deep brown solution containing tannic acid in which they had been suspended. This layer was taken off, shaken with an excess of water to thoroughly wash it, and again separated. The separated particles were then shaken with water so as to form an emulsion, and alkalies were added. No coagulation now occurred, even though the mixture was allowed to stand for several days. The particles could however be brought into a solid mass by pressure, by gently heating, or by drying off the water with a porous tile. So prepared, the rubber formed a pure white mass, without any trace of its usually characteristic smell. On exposure to the air for several days the surface gradually became brown, probably owing to oxidation. The percentage of rubber in the latex was estimated at the same time by separating 50 c. c. The weight of the dry substance was 12-5 grammes, which, as the specific gravity of Castilloa elastica latex is practically i-o, gives a yield of 25 per cent. On treating the latex of Hevea brasiliensis in the same way for a slightly longer time a similar separation occurred. The same purely physical means as those employed in the case of the separated Castilloa rubber-particles caused them to coalesce to form a solid mass, while the addition of acetic acid and the action of the smoke of burning urucuri nuts had no effect. 169 Biffen. — The Coagulation of Latex. The yield of rubber, estimated as before, was from 28 to 30 °/o. The latex of Manihot Glaziovii also separated readily and gave results completely parallel with those mentioned above. This latex is interesting, as it is readily clo ted by churning. A soft spongy clot is formed in a few minutes containing in its meshes the greater part of the solution in which the rubber-particles were suspended. If this clot is cut into slices while still soft, and pressed between sugar-cane crushers, or in a heavy press, the bulk of the solution is extracted and a fairly pure rubber is found. On drying it does not give off the putrid smell characteristic of the ordinary Ceara ‘ scrap.’ Other latices can also be clotted by churning, but the process is a long one. The latex of Hancornia speciosa and of Mimusops globosa gave similar results on centrifugalizing. In the case of the latter the pink colouring-matter which characterizes ‘ balata ’ was found to have separated as a thin layer at the bottom of the tubes. Artocarpus incisa (Linn.) contains a very viscous latex employed by the Brazilians as a bird-lime or as a substitute for glue. When diluted and centrifugalized it separates readily, giving a creamy white layer which dries to a resinous mass somewhat resembling gutta-percha. At the ordinary temperature this is quite hard and brittle, but if the tempera- ture is raised slightly it becomes plastic, and at the temperature of boiling water it is soft and excessively sticky. The substance is soluble in carbon bi-sulphide, and insoluble in alcohol and water. Urostigma gamelleira (Miq.1) yields a similar substance of a chocolate-brown colour. We thus see that the mere action of centrifugal force effects the separation of rubber ; and from the failure of the processes usually employed, involving the use of chemical reagents, to bring about the clotting of the separated and washed rubber- particles, we must infer that no chemical change occurs in the 1 Mart. FI. Bras. 4. 1. 93, Ficus doliarum of Mart. Sys. Mat. Med. Bras. p. 88. N 2 170 Biffen . — The Coagulation of Latex . rubber itself, and that the cause of coagulation must be looked for in the medium in which they are suspended. From our knowledge of the constitution of latex it is evident that the proteids are the most likely substances to cause this when treated with acids, alkalies, excess of salt, & c., and when boiled. Unfortunately few latices have as yet been examined for their proteid constituents, chiefly on account of the difficulty of obtaining them in their natural condition in European laboratories, owing to their coagulating and undergoing decomposition during the journey from the tropics 1. The investigations so far made prove the presence of albumin, globulin, albumose, and peptone in several rubber-yielding latices 2. In the clear solution left after separation of the rubber-particles the xanthoproteic reaction always showed the presence of proteid matters, but under the circumstances it was impossible to identify them. Now albumins are characterized by the coagulation of their solutions on heating, especially in the presence of dilute acids, and globulins by their ready precipitation with the salt- solution and their coagulation on heating. Thus when the latex of Hevea brasiliensis is held in the smoke of the burning urucuri nuts, the albumin it contains 3 is clotted by the action of heat in the presence of dilute acetic acid. The globulin of Manihot Glaziovii latex coagulates on heating when the temperature rises to 74-76° C.4 The acid latex of Castilloa elastica contains an acid albumin, which on neutralization forms a gelatinous precipitate. These coagula on forming gather up the rubber-particles (and probably starch-grains also, in the case of starch- containing latices) in the same way as the white-of-egg gathers 1 This does not apply to the latex of Mimusops Globosa, or Hancornia speciosa , both of which may be kept for months without undergoing any change. 2 J. R. Green, Proc. Roy. Soc. 1886, p. 28. 3 Faraday — see Le Caoutchouc et la Gutta-percha. 4 J. R. Green, ibid. Biffen . — The Coagulation of Latex. 171 up particles in suspension when clotted for the purpose of clearing jellies. We may even push the old analogy of blood and latex further, and compare the formation of a rubber-clot, in many cases, to the formation of a blood- clot, the rubber- particles being bound together by coagulated proteids in the same way as the blood-corpuscles are bound together by fibrin. In this case, however, we must remember that the rubber- particles, owing to their being sticky bodies unprotected by any external film, as e.g. the fat-particles of milk are, are capable of aggregating together of their own accord to form a solid mass. Rubber then, as now prepared, contains among other substances proteid matters. To these must be ascribed the well-known * fermentative change ’ which causes a considerable loss by converting the solid blocks of rubber into a foul- smelling spongy substance. In the Para rubber the creosote, absorbed from the smoke of the burning nuts, acts as an antiseptic and prevents this proteid decomposition k To test for the coagulated proteids is not an easy matter; continued boiling with a concentrated solution of caustic potash will however extract small quantities of alkali- albumin. £ Balata ’ gives good results most readily. On extraction with caustic potash a flocculent precipitate is obtained, which is readily soluble in dilute nitric acid, and is reprecipitated on the addition of alkalies. Boiling precipitates it either in acid or alkaline solutions, and it gives no precipi- tate with acetic acid and potassium ferro- cyanide. The proteid is thus identical with the albumose described by Green from the latex of Mimtisops globosa. 1 Cf. the smoking of fish &c. for preserving purposes. Botanical Laboratory. Cambridge, February , 1898. The Development of the Cystocarp in Rhodymeniales : II. Delesseriaceae. BY REGINALD W. PHILLIPS, M.A., B.Sc., Professor of Botany in the University College of North Wales , Bangor. With Plates XV and XVI. IN a former paper in the Annals of Botany (’97) I published the results of observations on the development of the cystocarp in certain species of the families Bonnemaisoniaceae, Rhodymeniaceae, Sphaerococcaceae, and Ceramiaceae. In still earlier papers (’95 and ’96) I had already described the structure of the ceramidia of several species of Rhodome- laceae. But the cohort Rhodymeniales as constituted by Schmitz (’89) contains yet a sixth family, the Delesseriaceae ; and in this paper I propose to give the results of an investi- gation of the structure of the cystocarp in the following species belonging to it : — Delesseria sanguinea, Lamx. Delesseria alata, Lamx. Delesseria Hypoglossum, Lamx. Delesseria ruscifolia, Lamx. Delesseria sinuosa, Lamx. Nitophyllum laceratum, Grev. Nitophyllum Hilliae, Grev. L Annals of Botany, Vol. XII. No. XLVI. June, 1898.] 174 Phillips . — The Development of the Cystocarp Delesseria sanguinea, Lamx. This plant is one of the most conspicuously beautiful of all the Red Seaweeds, and must have been known from early times. It was the Fucus sanguineus of Linnaeus (1767), and on the disintegration of that comprehensive genus became the Deles- seria sanguinea of Lamouroux (’13). J. G. Agardh (’51) could not, however, find that its characters harmonized with those of 'other species of Delesseria , and he therefore adopted for it the generic name W ’ormskioldia, proposed by Sprengel (’27), but unlike that author, made it the single species of the genus. Later (’76), finding that this name had already been appropriated, Agardh utilized Stackhouse’s (’01) generic name Hydrolapathmn. More recent writers have assigned it to Delesseria or to Hydrolapathum as they recognized or denied its near relationship to such other typical species of Deles- seria as the D. Hypoglossum , D. alata, and D. ruscifolia of Lamouroux. Kiitzing, by transferring such species as those last named to a genus Hypoglossum , and retaining the designa- tion Delesseria only for D. sanguinea and two South Atlantic plants, has shown a certain agreement with Agardh’s view of the generic distinction of the Delesseria sanguinea of Lamouroux from the other species of that author. Schmitz (’89), in his c Uebersicht,’ has reverted to the older arrange- ment of Lamouroux. This course has, as late as last year (’97), called forth a protest from Agardh, who re-asserts his belief in the generic isolation of D. sanguinea. There is, moreover, a wider question at issue among algo- logists in connexion with this species. While Kiitzing has separated D. sanguinea from other species of Delesseria , he has retained it in the same family with them. Agardh (’76), however, and in this he has been followed by Hauck (’85) and others, has removed it altogether from among Delesseriaceae, and placed it among Rhodymeniaceae. This course he still defends in his latest publication (’97). It thus appears that there is a difference of opinion, not in Rhodymeniales : IL Delesseriaceae. 175 only upon the generic position of D. sanguined , but even upon its ordinal place. I propose to recur to this question later. The segments of the thallus of D. sanguinea simulate the appearance of leaves to a remarkable degree. As, however, Goodenough and Woodward (1795) long ago remarked, ‘when attentively considered and compared with others, they appear to be branches growing up into, or dilated into, a thin mem- brane.’ The rudimentary plant consists of a leaf-like lamina, attached by a holdfast. In course of time the winged portion of the thallus disappears, and further growth is provided for by proliferation from the persistent, more or less cylindrical midrib. I believe that in this species a new series of pro- liferations occurs every year, and that therefore the age of the plant can be accurately measured by counting the joints of the sympodium, of which the plant, from the holdfast to the still growing apex, consists. Among the first prolifera- tions of each recurring period of growth the fertile shoots arise in great numbers (Fig. 1). Occasionally the winged portion of an old shoot may persist for some time after the proliferations of a new period have begun to grow, in which case the fertile shoots appear in two rows, right and left of the midrib, on each surface of the leaf-like shoot. These fertile shoots are at first precisely similar in appearance to the corresponding stages of the sterile shoots, but remain small in comparison to the later stages of the latter. The pro- duction of reproductive organs, whether tetrasporangia or antheridia or cystocarps, seems to drain the resources of the shoots and to dwarf their vegetative growth. Whilst the sterile shoots of one period may reach many inches or even a foot in length, the tetrasporiferous shoots are hardly more than half an inch long, and the antheridiferous and cysto- carpic shoots are even less. I believe, however, that when fertilization fails, procarp-bearing shoots may, and often do, take on again the vigorous apical growth of the sterile shoots, and later become indistinguishable from them. When, however, fertilization does take place, the female 176 Phillips . — The Development of the Cystocarp shoots gradually become transformed into the so-called ‘ pedi- cellate ’ cystocarps. These usually occur in considerable numbers along the lateral margins of the midribs, and have much the appearance, when mature, of aggregates of minute berries. By a careful comparison of a series of these struc- tures of varying degrees of maturity, the transformation of the young proliferations into the ripe cystocarps may be made out. With the aid only of such magnification as is afforded by a simple lens the following changes may be observed to occur. The young shoot in its primary condition is about 1 mm. long, and is lanceolate in outline, its width at its widest part being about twice that at its base (Fig. 2). It is perceptibly thicker along the mid-line, as if traversed by a rudimentary midrib. The first appearance of a cystocarp is a slight swelling at the mid-line on one of the surfaces. No dorsi- ventrality can be detected in these proliferations, and the swelling may arise on either surface. The circular base of this swelling gradually extends to the margins of the lamina, so that the outline is changed from lanceolate to ovate (Fig. 3). Widening still more, the outline might be described as rotund, were it not that a small triangular apical region of the originally lanceolate thallus persists as such (Fig. 4). In elevation the swelling rises so that the vertical diameter is soon as great as the horizontal diameter (Figs. 5, 7), and the swelling is roughly dome-shaped. With the growth of the cystocarp there becomes apparent near the summit of the dome a pore, the margins of which, later, somewhat protrude, trans- forming the dome into a broad-based urn. The lamina upon which the urn is situated does not remain flat, but, with the progress of growth, becomes somewhat depressed and convex below. The cystocarp, however, never becomes globular, the one-sidedness of the swelling and the existence of the triangular apical flap rendering such a term inappropriate. With the appearance of the cystocarp in the substance of the thallus its apical growth is arrested, and the urn-shaped swelling soon occupies almost the whole of the surface. When in Rhodymeniales : II. Delesseriaceae. 177 mature the fructification assumes a dark-red colour, from the dense ‘ nucleus ’ of highly-coloured carpospores. It would appear, therefore, that the so-called stalked cysto- carp is more accurately described as a minute flattened branch, upon one of the surfaces of which an urn-shaped cystocarp has arisen. It is, however, with the minute structure of this cystocarp, and the details of its process of maturation, that I am chiefly concerned. In order to make the following description in- telligible, it is necessary to recall the histological structure of the sterile fronds of Delesseria sanguinea and its congeners. This has been the subject of detailed investigations by several observers, and it appears that these species conform rigidly to the law of growth of the Floridean thallus first clearly enunciated by Schmitz (’83). That is to say, the growth is exclusively apical, no transverse division ever occurring in a segment cut off from the apical cell, and no longitudinal division passing through the organic axis of the segment. It is true that Schmitz, returning to this subject later (’92), somewhat limited the application of this law, excluding in particular from its operation the tribe Nitophylleae of the order Delesseriaceae. The tribe Delesserieae, which, standing as it does next to the excluded tribe, it may be assumed that Schmitz examined afresh, was, with the large majority of Florideae, still regarded as falling under the law of exclusive apical growth. With this view I agree, as I have seen nothing in my observations on Delesseriaceae which could be regarded as evidence of intercalary cell-division. Wille (’87) has, it is true, both described and figured intercalary cell- formation as occurring in D. sanguinea in the axial row of cells during the course of the development of the midrib. As, however, he makes the statement incidentally and without comment, although Schmitz’s work had been published some years before, it is probable that the statement was based on a too superficial observation of the thallus. The cells shaded in Figs. 1 and 3 of Taf. I of his work, are the products of the pericentral cell lying immediately above the central 178 Phillips. — The Development of the Cystocarp cell. By appropriate means, the much longer, somewhat attenuated axial cell may be seen lying below these cells, and it may be traced, even in the vegetative thallus, in an un- divided condition for great distances behind the situation described by Wille. In the young proliferations in which the cystocarps arise, there is no difficulty in tracing the undivided cells of the axial row from base to apex. I now proceed to describe the structure and arrangement of the procarps in these phylloid branches. For some few cells behind the apical cell, the axial cells, while they give off laterally pericentral cells, which grow out to form the lamellar wing, right and left of the mid-line, do not cut off cells parallel to the flat surface of the thallus. This, however, soon occurs, and immediately following upon the appearance of these pericentral cells above and below, is the appearance of the carpogonial branch. The pericentral cell cuts off, obliquely, right or left posteriorly, a cell, which is the first cell of a 4-celled carpogonial branch. The branch curves round right or left of the pericentral in a plane roughly parallel to the surface, and in such a way that the carpo- gonium itself is brought forward to a level with the apical part of the pericentral cell which bears the branch, and the trichogyne there passes out obliquely to the surface. The trichogyne is inflated at the extremity, and extends but little beyond the surface. Of the four cells of the carpogonial branch, the first two are each larger than the third and fourth, and the second is considerably larger than the first. As this cell takes up Hoffmann’s blue readily, the position of the carpogonial branch can easily be determined even with a low power by its means. The third and fourth cells are small, and are with difficulty distinguishable from one another for some time after the cell-division which gives rise to the carpogonium. Such a carpogonial branch is borne by the vertically situate pericentral cell of every joint along a considerable length of the fertile branch. They lean however to the right or left of the mid-line in regular alternation. Further, when in Rhodymeniales : II. Delesseriaceae. 1 79 the corresponding pericentral cells on the opposite surface are examined, it is found that they also each give rise to a carpogonial branch which curves round the pericentral cell in the same direction as the carpogonial branch corresponding to it above. These relationships will be most readily realized by means of figures. Figs. 8 and 10 show the arrangement of the carpogonial branches as seen from the surface. Fig. 9 gives the appearance of the pairs of carpogonial branches when viewed from the side. The section is supposed to be taken a little to one side of the median line, and the difference in the depth of the shading of the alternate pairs is intended to indicate the slight difference of level. A single thallus- segment thus often gives rise to from thirty to forty procarps, the position of all of which can be distinguished in material appropriately stained, and swollen in glycerine. No procarp ever arises elsewhere than in these situations along the midrib. It does not follow, however, that all these procarps are functional at the same time. They are produced in acropetal succession, and those whose trichogynes protrude at any time are a few in the apical region on each surface. Further back, the midrib becomes stouter by the peripheral growth of the sterile filaments derived from the vertical pair of peri- central cells. Right and left of the middle line, moreover, the lateral pair of pericentral cells and their derivatives give off cells to each surface. By this process the midrib soon becomes six or eight cells thick, and the procarps, which are at the level of the cells nearest to the axis, tend to become more and more immersed. A furrow may at first be detected on the surface on each side of the midrib, joining the points of exit of the trichogynes, where minute pit-like depressions for some time remain. It may well be that the convection of spermatia to the trichogynes is facilitated by the existence of this groove, where they might more readily lodge than on the even surface. As I have already said, I am inclined to think that some of the procarp-bearing segments grow out into the ordinary 180 Phillips . — The Development of the Cystocarp vegetative leaf-like shoots when fertilization fails. They resume the vigorous apical growth, and, ceasing to give rise to procarps, attain great lengths. This I infer from the fact that in young branches which are considerably longer than those in the receptive stage, I could still distinguish the remains of procarps in the basal region. Vegetative shoots certainly arise here and there in the fringe of fertile shoots. Such a reversion cannot, however, occur on a large scale ; for the young procarp-bearing proliferations may be counted by scores, while the number of vegetative branches on a plant hardly reaches a dozen. The first indication of the appearance of a cystocarp in these fertile branches is best afforded by staining. One cell of the axial row, the four pericentral cells connected with it, and the adjacent axial cells before and behind, seven cells in all, become so deeply stained by Hoffmann’s blue, that this region can then be readily distinguished with the aid of even a hand-lens. I have always found that this stain is taken up with greatest avidity by those cells in which there is great metabolic activity, or a relatively large quantity of proto- plasm. Thus, the apical cell, the cells of the carpogonial branch, the auxiliary cell, and carpospores and tetraspores in Florideae, all stain deeply. In the ordinary vegetative cells, it is only the chromatophores and the nucleus that stain readily, the ordinary cytoplasm being hardly tinged in glycerine-material. The deep staining of the cells referred to above indicates that one of the procarps connected with the axial cell, which is the centre of the group, has in all probability become fertilized, and that physiological changes ensue in neighbouring cells, analogous to those which occur in an ovary when the seeds are fertilized. When this stage is reached, however, it may be inferred that a considerable time has already elapsed since the attachment of the spermatium to the trichogyne. In the cases examined, the trichogyne was already so much immersed and obliterated, that it was useless to look for evidence of the presence of the spermatium. Such deeply stained groups of cells show also a consider- in Rhodymeniales : II. Delesseriaceae . 1 8 1 able modification of parts, though each is still capable of identification with the earlier condition already described. First, the filaments, other than the carpogonial branch, derived from the same pericentral cell with it, which, when no fertiliza- tion takes place, help to form the thickened midrib, now take on a characteristic appearance. These filaments are two in number ; one, the larger, springing laterally, and the other, smaller, posteriorly (Figs. 12 and 13). In all, the tuft which these filaments constitute consists of a score or so of cells. When fertilization of a procarp takes place, these cease to grow further, although all the adjacent filaments in a similar situation take on a more active growth. The cell-walls become greatly thickened and highly refractive, and sharply contrast on this account with the rest of the tissue. The neighbouring filaments growing more vigorously, soon arch over and bury them, without however completely closing the aperture above. The gap thus left is the apical pore of the future cystocarp (Fig. 16). A small portion of the external surface is thus covered in, and may still be distinguished by the foreign substances adhering to it. Both these changes, that in the tuft of filaments which cease to grow, and that in the adjacent filaments which grow the more vigorously, indicate beyond doubt on which side of the thallus fertilization has taken place. I have never found these changes taking place on both sides the thallus, or at more than one spot on the thallus. As it is unlikely that only one procarp becomes fertilized on a branch, it is probable that the demand for nutrition consequent upon the occurrence of the first act of fertilization prevents the formation of a second cystocarp. The case is analogous with that of the ovule of Pinus , for example, where of many possible embryos only one normally matures. To turn however to the carpogonial branch and the peri- central cell from which it is derived. At the stage above described it is still possible to distinguish the cells of the carpogonial branch, especially since they form a characteristic filament owing to the inequality in the size of the successive cells. The cells, however, have by this time greatly altered 1 82 Phillips . — The Development of the Cystocarp in appearance. Instead of readily taking up the stain, they are now the least stained in the whole section. No longer full of dense protoplasm, they are now vacuolated and granular. It is noteworthy, however, that the outline of the carpogonium is larger than before fertilization. The peri- central cell itself, at the earliest stage which I could obtain, had already divided as in Fig. 13, cutting off a large cell towards the apex of the branch. This derived cell is much larger and more conspicuous than the pericentral cell, which is greatly reduced in size by its formation. It is from this cell, undoubtedly, that the gonimoblast-filaments afterwards arise (Fig. 14) ; and it is highly probable that it is this cell, and not the pericentral cell, which constitutes the auxiliary cell and is fertilized by means of an ooblastema-tube from the carpogonium. Were the pericentral cell itself the auxiliary, it might be fairly argued that it would directly give rise to many gonimoblast-filaments, which it does not. An almost precisely similar case is that of Polysiphonia , in the Rhodomelaceae, where a cell derived from the pericentral cell is now considered to be the auxiliary, since from it, and not from the pericentral cell, the gonimoblast-filaments arise. Other Rhodomelaceae, like Chondria , occur, in which the pericentral cell seems to be the auxiliary. In all Ceramiaceae, it is a cell derived from the cell bearing the carpogonial branch ; and this I believe to be the case here. I did not succeed in finding any trace of the ooblastema-tube. It may be noticed, however, that the close contiguity of the carpo- gonium to the auxiliary cell is favourable for the process of fertilization. Fig. 14 represents an early stage in the development of gonimoblast-filaments from the auxiliary. The early cells of these filaments have an appearance which I have repeatedly observed. They are disc-like in shape, and seem to be separated by concave walls, fitting one into another like a series of cups. They probably arise in quick succession, pushing forward into the dense mucilage derived from the decadent sterile filaments. The pressure thus produced in Rhodymeniales : II. Delesseriaceae. 183 reacts on the pericentral cell, which is pushed back against the central cell, and thus lost sight of. This may possibly account for the statement that the pericentral cell gives rise to the carpospores. After the stage represented in Fig. 14, I have not been able to find any vestige of the carpogonial branch. It probably atrophies and disappears. The rest of the development of the cystocarp consists chiefly in the luxuriant branching of the gonimoblast-filaments, by which the sporogenous tissue attains considerable bulk in this species. The sterile derivatives of the pericentral cell are pushed off, and may often be seen lying at the peripheral part of the fertile tissue. In the mature condition the contents of the cystocarp exhibit a lobed appearance (Fig. 10), owing probably to the partial separation from one another of the products of a few main branches. Adventitious filaments seem to arise along the larger branches comparatively late in their development (F‘g- J5)- The mature cystocarp-bearing branches are considerably longer than the procarp-bearing branches. This is probably due to a general elongation of the cells already formed rather than to continued apical growth. Wille (’87) has shown that in this species there arise from the internal cells of the thallus in the older parts numerous hypha-like cells whose function he considers to be storage. These arise in the basal parts of the cystocarpic branches, and to some extent may account for the greater length. Delesseria alata, Lamx. This plant presents several striking differences of habit from D. sanguined. While it possesses an equally well-marked midrib, the laminar portions are so reduced that the appear- ance is more that of a winged stem, which is the true morphological equivalence in both cases. In D. sanguinea , however, the proliferations of one season do not branch again O 184 Phillips . — The Development of the Cystocarp in the same season ; and when the new proliferations arise in the next season, it is exclusively from the persistent midrib. In D. alata the plant bifurcates repeatedly in one plane by marginal growth near the apex ; and as the apical growth is apparently continuous from one season to another, the plants come to consist of a dichotomously branched thallus of considerable length. In the neighbourhood of the axils between the branches, it also gives rise to dense tufts of adventitious shoots, similar to those which come off the mid- rib in D. sangumea. It is on these structures for the most part that the reproductive organs occur ; but as far as the production of tetrasporangia and cystocarps is concerned, they also occur, but less commonly, on the surface of the ultimate forkings of the ordinary thallus. Hence it would seem that D. sanguined is a more highly specialized plant than D. alata. The apical growth of the thallus is as pronounced in this species as in the other, and no true intercalary growth occurs throughout its structure. D. alata has been selected by Kny (’86) in his well-known ‘ Wandtafeln , for illustration of apical growth. Wille (’87), in his figures of the apex, seems to consider that cells of the axial row divide by means of c horizontal ’ walls, i. e. by transverse divisions, which is not the case. The so-called ‘ hyphal * cells which arise from the inner cells some distance behind the apex in D. sanguined occur also in this species. Wille regards their function here as that of conduction, not of storage. The young axillary proliferations of D. alata serve well for the study of the development of the cystocarp, as in the same tuft there may be found varying stages of growth. The cystocarp arises on the midrib some distance behind the apex, and gradually enlarges as a papillar elevation until it can be seen in profile by means of a hand-lens. The mature cystocarp, however, never so completely transforms the appearance of the branchlet as it does in D. sanguined. This is partly because the cystocarp is not so bulky, and partly because the proliferation is in the end larger than those of D. sanguined. in Rhodymenia ies : 11. Delesseriaceae. 185 When the cystocarp occurs on the midrib of one of the ultimate forkings of the thallus, it is still smaller in proportion to the size of the thallus. It is only in this condition that it is figured by Harvey (’51). I have found the arrangement of the procarps to correspond closely to that already described as occurring in D. sanguined. The carpogonial branches are 4-celled, and arise on the pericentral cells above and below. They lean to the right and left alternately as in D. sanguined (Fig. 17). Of the four cells of the branch, the second is here too' by far the largest, exceeding in bulk the other three put together. The tricho- gyne is inflated where it reaches the surface, and protrudes but little. The next recognizable stage is elucidated by the same selective staining of the axial cell of the fertile joint, and of the six adjacent cells. At this stage the enlarged peripheral sterile derivatives of the pericentral cell are a conspicuous feature, although their appearance is dissimilar from the equivalent structures in D. sanguined. They consist similarly of two branches, but the posterior branch consists of only two cells, and the other branch of four (Fig. 1 8). These cells are relatively much larger than in D. sanguinea , and form a loose aggregate of cells, whose pit-connexion it is not easy to follow. The carpostome is formed by the over-arching of the surrounding vegetative filaments, though their growth seems to take place in a common mucilage and without any such invagination of the external surface as occurs in D. sanguinea. This is doubtless associated with the circumstance that the cystocarp does not attain the large size of that of D. sanguinea. The gonimoblast-filaments arise exclusively from an anterior derivative of the pericentral cell, and are directed forward in the early stage (Figs. 18, 19). There is the same probability that this derivative of the pericentral cell, rather than the pericentral cell itself, is here also the true auxiliary ; but the disorganization of the carpogonial branch at this stage 1 86 Phillips . — The Development of the Cystocarp renders it difficult to find any evidence of conjugation of the carpogonium with either cell. Delesseria Hypoglossum, Lamx. The ordinary vegetative thallus of D. Hypoglossum pro- liferates regularly from the midrib, and there is never any such forking by marginal growth as is found in D. alata. The lateral veins which are so marked a character of the vegetative thallus of D. sangninea , and which occur more obscurely in D. alata , are absent from this species. The segments which bear the reproductive organs are otherwise indistinguishable from the ordinary vegetative segments : hence it would appear that this species is a still less specialized form than D. alata. The apical growth has long since been accurately described by Naegeli (’47). Owing to the great obliquity of the cell- divisions in the lateral pericentral cells, the apical region presents a beautiful appearance which does not occur in any other British species of the genus. I have found this character useful in distinguishing this species from D. ruscifolia with which it is sometimes confounded in herbaria. The hyphal filaments which occur along the midrib in D. sangninea and D. alata , and which in the older parts greatly obscure the primitive arrangement of the cells, do not seem to occur in D. Hypoglossum , at any rate at the corre- sponding stages. The cystocarp-bearing plants occur only very rarely on the coast of Anglesey. Goodenough and Woodward (1795) contrast the east and west coast of England in this respect. According to these authors, it was only cystocarpic plants that had in their time been found on the coast of Norfolk. I have already ('96) referred to the case of Plumaria elegans. While I could never find female plants of this species on the coast of Anglesey or Carnarvonshire, I found them fre- quent at Sidmouth. Again, it is well known that Laurencia obtusa and L . pinnatifida rarely occur as cystocarpic plants in British waters. Mr. A. H. Church, who was good enough to send me cystocarpic material of D. Hypoglossum and other in Rhodymeniales : //. Delesseriaceae . 187 species from Plymouth, suggests that temperature probably affects the frequency or infrequency of the female plants. Considering that, for the most part, male and female plants among Florideae are, in comparison with the tetrasporic plants, small and apparently depauperized, it may be that the cystocarpic plants are few where the conditions of tem- perature and illumination are favourable for vegetative development. It is, however, hazardous to generalize in the present state of our knowledge of the natural history of our marine Algae. Not more than one cystocarp usually arises on D. Hypo - glossum in the course of a single branch, though one cystocarp may often be found on each of the many proliferations of a branch which itself bears a cystocarp. The cystocarp is relatively small, and the branch persists after the cystocarp has discharged its spores. The procarps occur along the thallus in the same regular way that has been described for D. sanguinea and D. alata. On the female plants, at the proper season, every leaf-like branch seems equally to bear carpogonial branches throughout its course, until a cystocarp arises, when the production of procarps generally ceases (Fig. 21). Since the midrib does not thicken to the extent that it does in the species previously described, it is possible to detect the unfertilized carpogonial branches for great distances along the thallus (Fig. 22). It is moreover possible to follow more readily the sequence of events in the young cystocarp. Fig. 22 represents a surface-view of the tuft of gonimoblast- filaments, derived from the auxiliary cell, and the carpogonial branch may still be seen - in an attenuated condition lying alongside the tuft. No signs of a conjugation of the car- pogonium and the auxiliary can be seen at this stage ; though the incapacity of the cells of the carpogonial branch to any longer absorb the blue stain, which they so readily take up at an earlier stage, suggests that the protoplasmic contents have undergone change. It is fair to add that all the cells of the carpogonial branch behave alike in this respect. 1 88 Phillips , — The Development of the Cystocarp Delesseria ruscifolia, Lamx. D. ruscifolia is undoubtedly closely allied to D. Hypoglossum , with which indeed it seems to have been confounded until Turner (’02) pointed out the distinguishing characters. In its dark-red colour D. ruscifolia resembles D. sang'uinea rather than D. Hypoglossum. Its segments are oval or oblong rather than lanceolate, as in D . Hypoglossum ; a lateral venation is obscurely traceable also in D. ruscifolia , and the cells are much smaller than are those of D. Hypoglossum. There is also a marked difference in the shape of the cysto- carp, which in D. Hypoglossum is somewhat flattened, but in D. ruscifolia has an elongated neck, and a carpostome with everted rim. As in D. Hypoglossum , each branch of the female plant usually bears a cystocarp, though occasionally, as in Harvey’s figure (’51), two may occur on the same midrib. The procarps occur on the midrib only, but the arrangement is by no means so regular as in the three preceding species. Occasionally no procarp will occur on an axial cell, or two successive joint-cells bear procarps inclined towards the same side, or a procarp will occur on one surface but not at the corresponding situation on the opposite surface. A more important deviation is the rare occurrence of two carpogonial branches arising, one on the right, the other on the left of the same pericentral cell. The cells of the carpogonial branch are also more uniform in this species, though in this respect it resembles D. Hypoglossum. The condition of the young cystocarp figured in Fig. 20 may often be found. It is probable that it represents a pause in the sequence of events between the fertilization of the trichogyne and the production of the gonimoblast-filaments. If so, this lends additional support to the idea that the anterior cell cut off from the pericentral cell is the auxiliary cell, as the halt may be accounted for by the fact that at this stage the conjugation of the carpogonium with the auxiliary cell would occur. As in the other species, it is from this anterior cell alone that the gonimoblast-filaments arise. in Rhodymeniales : II. Delesseriaceae. 189 Delesserta sinuosa, Lamx. This plant presents so great a general similarity in ap- pearance to D. sanguined , that the collector would readily acquiesce in its inclusion in the same genus with it. As D. sanguined has been appropriately called the f Dock-leaved ’ Delesseria^ so might this plant be called the ‘ Oak-leaved ’ Delesseria . In the remarkable similation of the veined ap- pearance of the leaf of Flowering-plants these two species stand out conspicuously among British Seaweeds. Kiitzing (’49), however, constituted for D. sinuosa the separate monotypic genus P hy codry s, and although Schmitz in his list of Floridean genera (’89) included the species in Delesseria , he seems later (’92) to have contemplated the possibility of its restoration to the position assigned to it by Kiitzing. A brief description of the macroscopic characters will explain this disinclination to include the species in the genus Delesseria. Its branches are traversed by a midrib from which diverge veins into the substance of the distended lamina. This midrib persists when the winged portion disappears, and gives rise by proliferation to the new phylloid branches. More commonly, however, a thallus-segment forks by the more vigorous growth of one of the lobes. Even when a lobe does not develop so as to form a distinct segment, its vein may become a strong secondary rib, from which prolifera- tions may arise as they do from the midrib. Again, D . sinuosa produces its tetrasporangia in marginal stichidia, like those of some species of Nitophyllum , and not along the sides of the midrib of the vegetative branches, or on special proliferating branches as in species of Delesseria. Further, the cystocarps are scattered in considerable numbers over the marginal region of the thallus-segment away from the midrib, while in Delesseria they usually occur one for each segment, and on the midrib. It is true that in the end the cystocarp of D. sinuosa is found to be seated on a prominent 190 Phillips . — The Development of the Cystocarp vein, as shown in Harvey’s figure, but this vein arises only after the establishment of the cystocarp. When the manner of growth of the thallus comes to be considered, D. simiosa presents a marked contrast to the species already dealt with. In these there is present at the geometrical apex a single conspicuous cell, which, by its repeated transverse divisions, gives rise to the axial row, the cells of which never divide transversely again, and from which later cells are cut off longitudinally by divisions which do not pass through the organic axis. The pericentral cells thus cut off repeat, in a modified form, the behaviour of the apical cell, and thus the thallus arises. Were it possible to isolate the pericentral cells with their respective products, the whole thallus would resolve itself into a system of branched filaments like a Callitkamnion . The growth is apical, in the sense that multiplication of cells takes place at innumerable apices, of which the most important coincides with the geometrical apex, the others lying at the margin and surface, or imbedded in the substance of the thallus. This method of growth Schmitz at first regarded as characteristic of all the Red Seaweeds, exclusive of the Bangiaceae. Naegeli and Schwen- dener (’67), in their work on the Microscope, had already selected D. simtosa and Nitophyllum laceratum as typical cases of growth by intercalation, which they illustrated by figures. Returning to this subject in his later writings, Schmitz (’92) conceded the whole tribe Nitophylleae (including D. sinuosa ) as affording evidence in the structure of the thallus, sooner or later, of intercalary growth. He demurs, however, to the figures of Naegeli and Schwendener, which, he said, left much to be desired. He seemed still to deny that the growth in thickness of the thallus of Nitophyllum is ever due to intercalation, and in particular he refused to acquiesce in Johnson’s (’92) suggestion that the callosities of Nitophyllum versicolor afforded an instance of growth by intercalation. With regard to the growth in thickness of the thallus in the neighbourhood of the cystocarps, in both Nitophyllum and D. sinuosa , my own observations convince in Rhodymeniales : II. Delesseriaceae. 19 1 me that it is effected in precisely the way in which it takes place in other families of Florideae, that is to say, by ex- clusively apical growth. As to the growth in area, where, by inference, it is to be concluded that Schmitz believed growth by intercalation to take place, I have seen no evidence of this either, but do not claim to have given, as yet, adequate attention to the phenomena figured by Naegeli and Schwendener. From all this, however, it is manifest that in the way in which the thallus of D. sinuosa arises, it is more akin to Nito- phyllum than to the species of Delesseria already described. When the arrangement of the procarps in D. sinuosa comes to be considered, the divergence from Delesseria is equally striking. They are found to be distributed in great numbers, without any regularity, in the marginal portions of the thallus. The ultimate ramifications of the veins are obscure lines, traceable only with the aid of the microscope, where a row of axial cells gives off a pericentral cell above and below. Between these veins the thallus is only one cell thick. It does not appear that the procarps in their inception are related in any way. to the veins, although a strong vein always arises in connexion with a fertilized procarp. When the procarps arise, an axial cell cuts off a pericentral cell above and below, and from each pericentral cell there springs a 4-celled carpogonial branch, which curves in a characteristic manner before the trichogyne emerges slightly at the corresponding surface (Fig. 29). By the time the carpogonial branches are formed, the pericentral cells from which they originate divide again and give off externally other cells, and a minute swelling in the thallus is the consequence. The trichogynes emerge on the slopes of this swelling. When fertilization fails there is no further development, and great numbers of such unfertilized procarps may be found among the few which are fertilized. The first indication of the development of a procarp into a cystocarp is afforded by the staining properties of the central cell concerned. This extends gradually to the neighbouring axial cells. At this time too 192 Phillips. — The Development of the Cystocarp the whole area round the spot where the fertilized procarp lies increases in thickness by the cutting off of pericentral cells, in which divisions occur parallel to the surface, giving rise to vertical rows of cells. The peripheral cells, derived from the pericentral cell which bears the fertilized procarp, do not take part in this vigorous growth, but remain in number as at the moment of fertilization. They however enlarge considerably, and assume a characteristic appearance, which is the first indication as to which procarp has been fertilized. The rows of cells surrounding these sterile deri- vatives arch over them, leaving a central depression which is the carpostome. Pressed by the convergence of the adjacent filaments, the sterile cells, which are five in number, become pyriform, with their pointed ends outwards, and their walls become at the same time thickened and highly refractive. Fig. 2 7, which represents this stage in Nitophyllmn Hilliae , might also serve for D. sinuosa. The pericentral cell cuts off a segment, which is the auxiliary cell, and from which later the gonimoblast-filaments arise. The sterile cells are then pushed off, and eventually disappear, supplying in their decadence a copious mucilage. As will be shown, in these particulars D. sinuosa more nearly resembles N itophyllum than the typical species of Delesseria . The systematic position of the species will be discussed later. Nitophyllum laceratum, Grev. N itophyllum laceratum may be regarded as the typical species for the genus as established by Greville. There is no percurrent midrib, and the cystocarps are embedded in the substance of the thallus. As it possesses, however, a distinct anastomosing venation in its older parts, and a more obscure venation throughout, it was placed by Kiitzing (’49) with other species like it in these respects in a genus Crypto- pleura, an arrangement, however, which has not found acceptance. in Rhodymeniales : //. Delesseriaccae . 193 The procarps are scattered along the margin of the thallus. Fig. 25 represents a procarp in surface-view ; its similarity to that of D. sinuosa , as shown in Fig. 29, will be apparent at once. Fig. 24 represents a section through the thallus at a point where the pair of procarps arises. In the later stages the sterile cells derived from the peri- central form a compact group similar to that shown in Fig. 27. Nitophyllum Hilliae, Grev. Nitophyllum Hilliae is the largest and firmest of all the British species of Nitophyllum . The thallus is for the most part more than one cell thick. Like N. laceratum , it has the obscure venation of Ktitzing’s genus Cryptopleura. I am indebted to Mr. A. H. Church for fine specimens of this plant from Plymouth. The procarps are scattered over the thallus as in N. lace- ratum and D. sinuosa. On closer examination they differ in one important particular from both these species. While each pericentral cell in D. sinuosa and N. laceratum gives rise to a single carpogonial branch, in N. Hilliae I have found that each pericentral cell very regularly gives rise to two such branches. These curve in a crescent on opposite sides of the pericentral cell, and the trichogynes emerge on opposite declivities of the papillar elevation which marks the position of the procarps. As two carpogonial branches emerge on each surface, the swelling marks the site of four carpogonial branches. In this respect N. Hilliae is alone among the Delesseriaceae here described, though isolated instances of the same phenomenon occur in D. ruscifolia. In the genus Ceramium two carpogonial branches also arise regularly from one cell. Fig. 26 represents a surface view of a pair of carpogonial branches. Fig. 27 represents an early stage of a cystocarp. The group of pyriform sterile cells, which belong to two filaments, have already been referred to. The auxiliary cell has been formed, but has not yet given rise to gonimoblast- 194 Phillips . — The Development of the Cystocarp filaments. The two unfertilized carpogonial branches of the opposite side of the axis can still be traced. Fig. 28 is an enlarged view of the essential parts of Fig. 27. When the cystocarp has matured, the papillar outgrowth becomes tuberculated in N. Hilliae , owing to uneven growth of the vertical filaments. I desire now to discuss the mutual relationships of the seven species, the structure of whose cystocarps has been described, and then to consider the affinities of the family to which they belong. To consider first the four species, D. sanguined , D. alata , D. Hypoglossum , and D. ruscifolia. It is clear that the two plants most closely related are the two last. The leaf-like branches are equally vegetative and reproductive in both, and the branching is exclusively adventitious. In D. alata , while the ordinary forkings may give rise to the reproductive organs, there is a marked tendency to produce them ex- clusively on somewhat specialized adventitious branches. What is only a tendency in D. alata has become a fixed condition in D. sanguinea. The four species thus form a natural series of which D. ruscifolia is perhaps the least specialized, and D. sanguinea certainly the most specialized. As has been shown, the structure and the arrangement of the procarps and the development of the cystocarp in D. san- guinea present so many features in common with the other three, that there can bg. no doubt about its inclusion in the same family with them. On account of considerations arising chiefly from a study of the tetrasporiferous segments, Agardh (76) placed the species as Hydrolap a thu m sanguineum in the family Rhodymeniaceae. This decision he has recently (’97) discussed in some detail and confirmed. He regards the stichidia or tetrasporiferous branches as unlike those of any other Florideae, finds the closest analogy in the origin and in Rhodymeniciles : II. Delesseriaceae. 195 arrangement of the spores in the genus Chylocladia , and is convinced from all the characters of the wide separation of Hydrolapathum from Delesseria. On the other hand, it is now clear that in the young condition of the procarp-bearing segment, in the remarkable arrangement of the procarps along the midrib in this segment, in the structure of the carpogonial branch, as well as in the general course of development of the cystocarp, there is the closest agreement between this species and the typical species of Delesseria . And as, from all analogy, the sexual reproductive structure affords the safer criterion in the search for affinities, the similarities here disclosed must be taken to outweigh any dissimilarities appearing in the origin and arrangement of the tetrasporangia. On these grounds, therefore, I cannot but think that Schmitz’s contention that Hydrolapathum san- guineum should be restored to the Delesseriaceae, and even to the genus Delesseria , is fully justified. D. sinuosa has had a similarly uncertain position. Schmitz (’92) was latterly inclined to divide the Delesseriaceae into two tribes, the Delesserieae and Nitophylleae, on account of differences in the mode of growth of the thallus. But this division would involve the inclusion of D. sinuosa among the Nitophylleae, and he therefore suggested the adoption of Kiitzing’s genus Phycodrys for its reception. This is a course which receives strong support from the study of the development of its cystocarp. In the true Delesserieae the procarps are borne in pairs along the midribs, whereas in D. sinuosa they are scattered as in Nitophyllum . In the more compact texture of the thallus, moreover, it resembles Nitophyllum . In the existence of a well-marked midrib with diverging veins it resembles species of Delesseria . These characters seem to mark for it a position intermediate between Nitophyllum and Delesseria , and the adoption of Kiitzing’s proposal would meet the case. With regard to the two species of Nitophyllum , N. laceratum comes nearest to D. sinuosa. It seems to be premature to suggest the occurrence of two carpogonial branches on each 196 Phillips . — The Development of the Cystocarp pericentral cell in N. Hilliae as a ground for the generic separation of this species from Nitophyllum , though, from all analogy, it would seem to indicate a deep-seated difference. Turning now to the diagnosis of the family Delesseriaceae as given by Schmitz and Hauptfleisch (’97) in Engler and Prantl’s s Pflanzenfamilien,’ it would seem to require modifica- tion in several particulars. 1. The carpogonial branches are described as 3- or 4-celled. In all the species I examined the carpogonial branch was invariably 4-celled. I never found the number to vary from four cells in the family Rhodomelaceae either. In a recent paper on Grinnellia americana , Harv., a monotypic genus of Delesseriaceae, which seems to stand near to Delesseria , Brannon (’97), has, it is true, described the carpogonial branch as 3-celled, but as he has also stated that it arises directly from a central cell, and not from a pericentral cell, it is just possible that he has missed the real pit-connexion of these cells. Otherwise, Grinnellia differs from all known Deles- seriaceae in the origin of the carpogonial branch, as well as from those here described in the number of the cells which constitute it. 2. The carpogonial branches are said to arise singly on an inner cell of the cortex. This does not now cover the case of N. Hilliae , where they regularly arise in pairs, nor exceptional cases of D. ruscifolia , where the same thing occurs. 3. The external pericarpial wall is described as formed by a tearing away of the cortical filaments from the middle layer of the thallus. While such a tearing may occur in certain species of Nitophyllum with somewhat flattened cystocarps, it is certainly not general. In Delesseria , the filaments sur- rounding the cystocarpic cavity become strongly curved, being pressed back at first by the copious mucilage derived from the walls of the sterile derivatives of the pericentral cell, and later by the tuft of spore-producing gonimoblast-fila- ments. Under this pressure, the cells of the filaments elongate, so as to resemble rows of cylindrical cells, yet the correlation of part to part in the course of growth is in Rhodymeni ales : II. Delesseriaceae . 197 so gradual, that no rupture of their continuity can be detected. 4. The tearing away of the pericarp is further said to be commonly omitted immediately above the auxiliary cell, where a strand of filaments remains connecting that cell with the pore of the cystocarp above. The strand of cells referred to is doubtless the group of sterile derivatives of the peri- central cell so often described in the foregoing accounts, and if so, it is not a strand of cells but a bushy tuft in D. san- guined, reaching only a part of the distance to the pore, and pushed aside when the gonimoblast-filaments subsequently arise. In the other species of Delesseria the cells are fewer in number, and lie loosely imbedded in a mucilage above the auxiliary cell, still less resembling a strand of cells connecting it with the pore. In D. sinuosa and the species of N itophyllum the cells are more compact, and the gonimoblast-filaments grow round and over them, justifying the description ‘ nabel- formig’ to this stage in the appearance of the cystocarp. In all cases the pore is the gap left above these cells by the over-arching converging filaments. It is not accurate to describe the sterile group as in any way attached to the pericarp at the pore. 5. As to the formation of a second chamber below, sepa- rated from the spore-containing cavity by the middle layer as a kind of diaphragm, a condition figured by Schmitz and Hauptfleisch for N '. punctatum, I have not been able to find it in the species examined. In the maturer stages of the growth of the cystocarp, the site of the auxiliary cell is the apex of a papilla projecting into the cavity, and while the luxuriant gonimoblast-filaments depress the base of the cavity round about it, they do not, as far as I have been able to see, enter a second cavity on the opposite side of the middle layer. 6. A general fusion of the auxiliary cell with neighbouring cells is described as taking place at the ‘placenta.’ While such a confluence apparently occurs in some species of Nito- phyllum , and is very general in Florideae, it is strikingly 198 Phillips. — The Development of the Cystocarp absent in the genus Delesseria. Brannon (’97) found moreover that no such fusion occurs in Grinnellia. 7. The pericentral cell is described as playing the part of the auxiliary. I have already given reasons for believing that the auxiliary is an anterior cell cut off from the pericentral cell. An auxiliary cell so derived occurs in most Rhodome- laceae and in all Ceramiaceae. An interesting feature in the development of the cystocarp in the species here under consideration is the fact that, when the cells adjacent to the central cell in a fertilized procarp become charged with nutriment prior to the formation of the gonimoblast-filaments, they also become multinucleate, as many as eight or ten nuclei at times occurring in one cell. I have found that elsewhere in the thallus of D. sangninea the greatly elongated axial cells contain more than one nucleus. With regard, finally, to the systematic position of the Delesseriaceae, I have no hesitation, on the ground of the remarkable correspondence in the process of development of the cystocarp, in placing them close to the Rhodomelaceae. There is the same invariably 4-celled carpogonial branch, the auxiliary cell is derived anteriorly from the pericentral cell, and there are always found, in the early cystocarp, two sterile filaments which degenerate later into mucilage. Indeed, the observation of the way in which these sterile filaments arise in Delesseriaceae affords a clue as to their origin in Rhodomelaceae which would otherwise be wanting. In Rhodomelaceae they would seem to be vestigial structures, and the cylindrical Rhodomelaceae would seem to have been derived from forms with a flattened thallus like Delesseriaceae. The two families form one alliance ; the simplest forms being represented by Nitophyllum , and the most complex by the polysiphonous Rhodomelaceae. in Rhodymeniales : II. Delesseriaceae. 199 Bibliography. 1. Agardh, ’51 : Sp. Gen. et Ord. Algarum : Vol. ii, pars 1 ; J. G. Agardh. Lundae, 1851. 2. Agardh, ’76 : Sp. Gen. et Ord. Algarum ; Vol. iii, Epicrisis ; J. G. Agardh. Lipsiae, 1876. 3. Agardh, ’97: Analecta Algologica ; Continuatioiv; J. G. Agardh. Lundae, i897. 4. Brannon, ’97 : The Structure and Development of Grinnellia americana , Harv. ; M. A. Brannon; Annals of Botany, vol. xi. Oxford, 1897. 5. Goodenough and Woodward, 1795 : Observations on British Fuci ; S. Goodenough and T. J. Woodward ; Trans. Linn. Soc., vol. iii. London, I795' 6. Harvey, ’51: Phycologia Britannica ; W. H. Harvey. London, 1845-51. 7. Johnson, ’92: Callosities on Nitophyllum versicolor ; T. Johnson; Proc. of the R. Dublin Soc., N. S., vol. vii. Dublin, 1892. 8. Kny, ’86 : Botanische Wandtafeln ; L. Kny ; Tafel 77. Berlin, 1886. 9. Kutzing, ’49 : Species Algarum; F. T. Kiitzing. Lipsiae, 1849. 10. Lamouroux, ’13 : Essai sur les genres de la Famille Thallassiophytes non- articulees. J. V. F. Lamouroux ; Mus. Hist. Nat. Paris, 1813. 11. Linnaeus, 1767: Mantissa Plantarum ; C. Linnaeus; Iiolmiae, 1767* 12. Naegeli, ’45 : Wachsthumsgeschichte von Delesseria Hypoglossum ; Carl Naegeli ; Schleiden und Naegeli’s Zeitschrift fur wissensch. Botanik, Heft 2. Zurich, 1845. 13. Naegeli und Schwendener, ’67 : Das Mikroskop ; C. Naegeli und S. Schwendener. Leipzig, 1867. 14. Phillips, ’95 : On the Development of the Cystocarp in Rhodomelaceae ; R. W. Phillips; Annals of Botany, vol. ix. Oxford, 1895. 15. Phillips, ’96 : On the Development of the Cystocarp in Rhodomelaceae (II). R. W. Phillips; Annals of Botany, vol. x. Oxford, 1896. 16. Phillips, ’97 : On the Development of the Cystocarp in Rhodymeniales ; R. W. Phillips ; Annals of Botany, vol. xi. Oxford, 1897. 17. Schmitz, ’83 : Untersuchungen iiber die Befruchtung der Florideen ; Fr. Schmitz; Sitzungsber. der Berl. Akad. der Wissensch. Berlin, 1883. 18. Schmitz, ’89 : Systematische Uebersicht der bisher bekannten Gattungen der Florideen. Fr. Schmitz; Flora, Heft 5. Marburg, 1889. 19. Schmitz, ’92 : Kleinere Beitrage zur Kenntniss der Florideen, I, II ; Fr. Schmitz ; La Nuova Notarisia, Series III. Padova, 1892. 20. Schmitz und Hauptfleisch, ’97 : Delesseriaceae ; Fr. Schmitz und P. Haupt- fleisch : Die Natiirlichen Pflanzenfamilien ; A. Engler und K. Prantl ; Lief. 1 49-1 50. Leipzig, 1897. 21. Sprengel, ’27 : Linnaei Systema Vegetabilium, edit, xvi curante C. Sprengel. Gottingen, 1827. 22. Stackhouse, ’01 : Nereis Britannica ; J. Stackhouse. Bathoniae, 1801. 23. Turner, ’02: Description of four new species of Fucus; Dawson Turner; Trans. Linn. Soc., vol. vi. London, 1802. 24. Wille, ’87 : Beitrage zur Entwickelungsgesch. der physiol. Gewebesysteme bei einigen Florideen; N. Wille. Halle, 1887. P 200 Phillips. — The Development of the Cystoccirp EXPLANATION OF FIGURES IN PLATES XV AND XVI. Illustrating Prof. Phillips’ paper on the Development of the Cystocarp in Rhodymeniales. The figures are to be regarded as of a semi-diagrammatic character. They were sketched for the most part by means of the camera lucida, but cells lying at different levels have often been figured together, and cells are sometimes left out for the sake of clearness. The cells hatched by means of oblique lines are in all cases those of the carpogonial branch, those hatched by means of horizontal lines in Figs. 20 and 25 are the central cell and its adjacent readily stained cells. The cells with red outlines are those of the sterile filaments derived from the pericentral cell of a fertilized procarp. In Figs. 21, 27, 28, 30, 31, and 32, the red colour is also employed for these cells, although the procarps are not yet fertilized. The auxiliary cell and its derivatives are shown with somewhat thickened outlines. In the figures disposed horizontally, the same orientation of parts is preserved throughout, that is to say, the anterior part of the thallus is to the right. Abbreviations : aux. c., auxiliary cell ; c. c., central cell ; carp ., carpogonium ; carp.br ., carpogonial branch; cpst. , carpostome ; gonbl gonimoblast-filament ; in. st. br ., inferior sterile branch ; /. st. br., lateral sterile branch ; peric. c., peri- central cell ; tr., trichogyne. PLATE XV. Figs. 1-16. Delesseria sanguinea , Lamx. Fig. 1. An old midrib, giving rise to three sterile branches, and numerous cystocarp-bearing branches in various stages of development. Natural size. Figs. 2, 3, 4. Views of the flat surface of a thallus- segment upon which a cystocarp is developing, x 15. Figs. 5 and 7. Lateral or profile views of two stages in the development of a cystocarp upon such a thallus-segment. X 15. Fig. 8. Front view of a procarp-bearing branch, showing the row of axial cells, the pericentral cells derived from them, and the carpogonial branches turned alternately to the right and left, x 75. Fig. 9. Lateral view of a procarp-bearing branch at the same stage. The axial cells are shown, and the paired pericentral cells of the upper and lower surface, each bearing a carpogonial branch emerging at the corresponding surface, x 75. Fig. 10. Mature cystocarp-bearing segment, showing the apical flap remaining, and the lobing of the contents of the cystocarp. x 10. Fig. 11. Magnified view of a segment at the same stage, and of the same view as in Fig. 8. The eighth segment behind the apical cell has an immature procarp, afterwards the carpogonial branches are fully developed, x 600. in Rhodymeniales : II. Deles seriaceae. 201 Fig. 12. A single procarp (unfertilized) showing the relation of the sterile filaments (red) to the carpogonial branch, x 900. Fig. 13. A single procarp at a stage subsequent to fertilization. The carpogonial branch is still present but by no means so conspicuous relatively as in the figure. The auxiliary cell has been cut off from the anterior part of the pericentral cell. The sterile filaments begin to undergo mucilaginous degeneration, x 500. Fig. 14. Median vertical section of a segment about the stage shown in Fig. 7. The carpogonial branch is still visible, but only with difficulty traceable in a section. The auxiliary cell has developed a minute tuft of gonimoblast-filaments. The section is drawn through the carpostome, and the invaginated portion of the external surface is shown below, x 300. Fig. 15. A strong gonimoblast-filament giving rise to an apparently adventitious lateral branch, x 300. Fig. 16. Lateral view of the apical region of a procarp-bearing segment. Magnification inferior to that of Fig. 11. x 350. PLATE XVI. Delesseria alata , Lamx. Figs. 17, 18, 19. Fig. 17. Vertical section through a procarp-bearing segment close behind the apex. The pair of procarps corresponding to every axial cell is shown. The first and third pair of carpogonial branches are on the hither side of their pericentral cells, and the second on the farther side, x 600. Fig. 18. A fertilized procarp in vertical section. The auxiliary cell has already divided once. The carpogonial branch is not shown, x 400. Fig. 19. A stage later. The auxiliary has given rise to a tuft of filaments by repeated branching, x 400. Delesseria ruscifolia , Lamx. Fig. 20. Fig. 20. Surface view of a fertilized procarp. The cells hatched by horizontal lines are three cells of the axial row and two lateral pericentral cells. These five cells stain readily. The carpogonial branch is distinctly visible. The pericentral cell has divided off (anteriorly) an auxiliary cell. The sterile derivatives of the pericentral cell coloured red. The other cells are superficial, x 750. Delesseria Hypoglossum , Lamx. Figs. 21, 22, 23. Fig. 21. Two cells of the axial row close to the apex with its derivative peri- central cells. From each of these are derived sterile derivatives and a carpogonial branch, x 1 200. Fig. 22. Three axial cells far behind the apex. The carpogonial branches immersed in the tissue of the midrib and atrophying, x 300. Fig. 23. A view from the surface, corresponding to that shown in Fig. 20, for D. ruscifolia . The auxiliary has grown out into gonimoblast-filaments. The sterile cells are omitted, x 750. • Nitophyllum laceratum , Grev. Figs. 24, 25. Fig. 24. Vertical section of the thallus, showing a pair of carpogonial branches, arising from the pericentral cells of one axial cell, x 75°* Fig. 25. Surface view of a procarp. The axial cell is not shown, x 75°* 202 Phillips. — The Cystocarp in Rhodymeniales. Nitophyllum Hilliae , Grev. Figs. 26, 27, 28. Fig. 26. An axial cell with its pericentral cell. Surface view. Two carpogonial branches arise from the one pericentral cell. The sterile derivatives are not shown. Fig. 27. Vertical section of a stage later, when the auxiliary cell has been cutoff. The sterile filaments are a group of cells with the pointed apices directed towards the carpostome. The carpogonial branch may still be traced. Two others on the opposite side the axis are faintly traceable, x 180. Fig. 28. Enlarged view of the central region of Fig. 27. x 500. Fig. 29. Delesseria sinuosa , Lamx. Surface view of a procarp, to show its similarity to N. laceratum (Fig. 26). Fig. 30. Rhodomela subfusca. — Products of the pericentral cell at the stage of fertilization diagrammatically represented. Fig. 31. Similar diagram for D.alata , D. ruscifoliatD . Hypoglossum, D .sinuosa , N. laceratum , N. Hilliae. Fig. 32. Similar diagram for D. sanguinea, to which Dasya coccinea approximates. fZ/maZs of Botany PHILLIPS.- CYSTOCARP IN R H 0 D Y M E N I A L E S , II . Vol.X/I, Ph.XV. University Press, Oxford. fZnsuUs of Botany Vol.Xff.Pl.XV PHILLIPS.- CYSTOCARP IN R H 0 DY M E N I A LE S , II . cpst. University Press, Oxford.. Fig 10 X 10 ptric.c . Fig. 14 ZnnaZs of BoZx/?^ c.c peroc.c. - ecu'/). Fi£.17 X 600 percc.c .R.W.P. del. carp. br. PHILLIPS.- CYSTOCARP IN R H 0 D Y M E N I A L E S 7 II. VoLXIl;PLXVI. University Press, Oxford. 'JbwxzZs of Botany peric.c Fig.Zl Fig. 27 .R.W.P. del. carp.br. PHILLIPS.- CYSTOCARP IN RH 0 DYM E N I A LE S , II. Vol.M,Pl.m. University Press, Oxford. The Vascular Structure of the Sporophylls of the Cycadaceae1. BY W. C. WORSDELL, F.L.S. With Plates XVII and XVIII. NE of the most marked characters of the plants com- prising the order Cycadaceae is the sympodial nature of the development of the stem, whereby the main axis usually terminates in a peduncle bearing a cone with sporo- phylls, while the vegetative axis is continued as a lateral branch. An exception, however, is found in the female plant of Cycas, where sympodial branching does not obtain, the vegetative axis growing continuously through ; here the sporophylls, instead of being borne on differentiated portions of the axis known as cones, occur in whorls intercalated amongst the ordinary foliage-leaves of the stem. The structure of the peduncle in this order has recently been described 2, and has been found to possess certain structures not known in the vegetative axis, such as the primary concentric strands in the cortex and in the central 1 From the Jodrell Laboratory, Royal Gardens, Kew. 2 Scott, Ann. Bot., Vol. xi, 1897. [Annals of Botany, Vol. XII. No. XL VI. June, 1898. 1 204 WorsdelL — The Vascular Structure of cylinder, and the centripetally-formed xylem of the bundles composing the latter. These are, I think, ancestral characters which, in the vegetative part of the stem, have been lost owing to special modification of its tissues for functional purposes, and which in the peduncle, as having undergone much less modification in structure, have been retained. This more primitive structure of the peduncle has been emphasized by Solms-Laubach 1 as a result of his observations on the course of the vascular bundles in this organ, and also by Scott 2, who first discovered the presence of centripetal xylem here. Precisely as in the case of the peduncle, we shall find that the sporophyll , the normal foliar appendage of the former, has also retained certain primitive characters which are not found in the foliage-leaf. But in the following pages two factors have to be considered in the structure of the former — the presence of these primitive characters, and the modification of its general structure in connexion with the sporangiferous function. In order to throw light upon the structure of the sporophyll I shall proceed to compare it with that of the foliage-leaf. In the base of this latter organ, which is the first -formed portion, the vascular bundles have a purely endarch structure ; higher up in the petiole, however, a mesarch structure of the bundle obtains, the centripetal xylem gradually increasing, and the centrifugal xylem gradually decreasing, in quantity from below upwards, until the former is almost entirely pre- dominant. The number of the bundles is almost uniform in every part of the petiole. The difference in structure between the bundles in the leaf-base and those in the remaining portion of the petiole is explicable from the fact that a transition occurs in the bundles of the lower region of the petiole from the structure peculiar to the stem to that which is peculiar to the leaf. When the structure of the bundle-system in the sporophyll 1 Die Sprossfolge der Stangeria und deriibrigen Cycadeen, Bot. Zeit. 1890. 2 Loc. cit. 205 the Sporophylls of the Cycadaceae . is examined, a similar phenomenon, as might be expected, prevails. Here also in the base of the organ, near the point of its attachment to the axis of the cone, the bundles are, or tend to be, endarch in structure, while in proportion as they pass higher up towards the lamina the endarch is gradually replaced by the mesarch character, until, in the lamina of the organ, the latter predominates. Thus the same general structure obtains here as in the foliage-leaf. Now the sporophyll has a function to perform which is entirely different from that of the foliage-leaf. Whereas the function of the latter is to elaborate food-material for the supply of the various parts of the plant, to aid in the con- duction of water through the stem, and to discharge the surplus quantity of this water from its expanded surfaces, the function of the sporophyll is to bear the sporangia, and to act as channels for the conduction of nutriment of all kinds to these latter ; it also, in some cases, takes a small part in assimilation. Though both foliage-leaf and sporophyll are morpho- logically of foliar nature, and possess essentially the same structure, their physiological relations are seen to be widely divergent It is the special sporangiferous function of the sporophyll which has induced those modifications in the structure and arrangement of the vascular bundles of this organ which are considered worthy to form the subject of a separate paper. There exists a considerable range of variation as regards the external development of the sporophyll in the different genera, and sometimes also within the same species between the two sexes. They may be either large and leaf-like or very much reduced and modified in shape, so as no longer to bear much resemblance to ordinary foliar organs. In the male sporophyll , as is known, the sporangia are thickly arranged in sori on the lower surface of a large flattened area situated between the extremely short stalk and the terminal, swollen, sterile portion ; the function of the latter is to afford protection to the sporangia within ; the apex 206 WorsdelL — The Vascular Structure of of this part is often prolonged into a short acuminate blade, recalling its foliar nature, or it may be produced on each side into two reflexed horns. The numerous bundles which are found in the stalk of this organ usually have their origin in a single bundle which is given off as a lateral branch from one of the members of the central cylinder of the axis of the cone ; this bundle, while traversing the cortex, divides into two, and these in the base of the sporophyll split up at first into three and very soon after into a larger number. Occasionally the two bundles may arise separately and independently from the cylinder of the axis. The bundles of the stalk are usually of small size, their xylem being mainly of centrifugal origin, sharply defined and deeply staining with safranin, and often, but not always, accompanied by a (usually) much smaller quantity of centri- petal xylem, which, however, is not so sharply defined nor so brightly stained, the tracheides of the centrifugal portion being evidently better lignified and chiefly functional in conduction. In proportion as the bundles pass up into the flattened portion of the sporophyll bearing the sporangia, the centripetal increases and the centrifugal xylem decreases in significance and quantity, until, in the sterile portion above, the former predominates. In this latter portion the structure of the bundles is essentially the same as that in the foliage-leaf. The centripetal xylem is as sharply defined as, and often even of a brighter colour, than the centrifugal, and is usually present in greater quantity. The bundles are of very varying sizes ; they have no definite orientation, lying in all directions. They arise from successive branching of those in the stalk. They have no longer any important conducting function to perform, the xylem being probably chiefly serviceable as a strengthening framework, so that the phloem is usually reduced in quantity and definiteness. The change of function of the bundles in this region and their relative unimportance will probably have much to do with the inconstancy of structure displayed, for the two kinds of xylem are very unequally developed in different bundles, the one often the Sporophylls of the Cycadaceae. 207 entirely supplanting the other, and vice versa. The very frequent presence here of the bundles with an endarch structure probably indicates the far-reaching influence of the special physiological function of the sporophyll, which causes, even in the region where it can no longer have any meaning, the suppression of the centripetal in favour of the centrifugal xylem. In the extreme portions the bundles are often reduced to a few irregular large tracheides. Transfusion- tissue is of very frequent occurrence in connexion with the bundles, and is sometimes of considerable development. In the female sporophyll there are usually about four, some- times a larger number, or only two bundles to be seen in the stalk. Two bundles arise independently, as a rule, from two distinct members of the central cylinder of the axis ; these, where a larger number supplies the sporophyll, divide up in the cortex before entering the latter. The structure and behaviour of the bundles in the female sporophyll are essentially the same as has been described for those in the male organ ; but the difference in sex of the sporangia, necessarily involving a difference in their function, a great reduction in their number, a considerable increase in their individual size, and an altered position on the sporophyll, goes hand in hand with a certain difference in the develop- ment of the bundles. The bundles which are told off to supply the sporangia are, in most cases, two in number, one at either side of the much elongated stalk ; each of these branches in the base of the expanded portion, sending off one or two bundles for the sporangium, which in nearly all genera is situated at the margin of the sporophyll in the sinus formed by the reflexed lateral portion of this part of the sporophyll, or on the tip of the reflexed portion, so that its apex is directed towards the axis of the cone. There is considerable variation in the number of the bundles which enter the sporangium and in the amount of branching which takes place after the one or two strands have been given off from the bundle proceeding from the stalk. The two lateral bundles in the stalk which supply the sporangia are always 208 Worsdell.— The Vascular Structure of much larger than, often extremely large as compared with, the other intermediate bundles which run right through into the sterile portion without taking any part whatever in the conduction of substances to the sporangia. These latter bundles are always quite small, their xylem and phloem being but poorly developed. The lateral bundles, on the other hand, have a great development of centrifugal xylem and phloem, as would, from their function, be expected. For each one has to supply a large megasporangium, and this over a protracted period, during which the processes of sporangial development, fertilization, and embryonic growth must supervene, demanding a constant and adequate stream of nutritive substances through the tissues of these bundles. Hence is the difference in structure and development between the bundles of one and the same female sporophyll, and between these and the bundles of the male sporophyll, easily explained. These larger bundles very often have an element or two of centripetal xylem, while this tissue is usually absent in the smaller bundles, the large development of the bundle seeming to go hand in hand with the presence of centripetal xylem, as will be seen more clearly when the individual genera are treated of. The structure of the bundles in the thickened, expanded portion of the female sporophyll agrees essentially with that of the same region in the male organ. An exception to the general structure prevailing here is found in the bundles, or a certain number of them, which enter the megasporangium, and occur immediately below the base of the latter ; these bundles assume, in many cases, a perfectly concentric structure, which may or may not enclose a central pith, and may be large or quite minute in size ; others are not completely but only partially concentric, while others again are quite collateral in structure. All these kinds of structure may occur in the same group of bundles entering the sporangia. These con- centric strands I regard not as purely adaptive structures connected with the radial symmetry of the sporangium they supply, for though this idea might explain the structure of 209 the Sporophylls of the Cycadaceae . the single large concentric strand entering the central chalazal portion of the sporangium, and which breaks up into a group of bundles radiating from the common centre to the different parts of the organ, it will hardly entirely account for the extremely small concentric bundles occurring in no very definite position in the group, and which must necessarily take a subsidiary part in supplying the sporangium. I am inclined, on the other hand, to regard these concentric structures and the constancy with which they occur in the region concerned, as relics of the original primitive structure appearing in what must be considered as the most primitive tissue of the sporophyll. It is the function of the individual bundles of the sporophyll, both on the male and female side, which readily explains their various degrees of development and varying types of structure. The bundles of the male sporophyll, at the level of the insertion of the sporangia, are small and insignificant in appearance because each one is told off to supply one of the numerous small sori scattered over the surface of the sporo- phyll, the microsporangia composing each having but a very temporary existence, insomuch that the function of the bundle supplying them ceases with the dispersion of the spores. Even in the lowest region of the stalk, before much branching of the bundles has occurred, and where they are therefore somewhat larger and fewer in number, the latter have quite a small development of centrifugal xylem as compared with that of the bundles supplying the megasporangia, this being clearly correlated with the respective functions of these bundles. But why in some genera the bundles in the basal region of the stalk of either male or female sporophyll should possess centripetal xylem, while in the case of other genera this tissue should be quite absent, and why this variation should exist both between the bundles of sporophylls of opposite sexes and between the bundles of one and the same sporo- phyll, is no more obvious than the reason why the same variation in structure should prevail in the bundles of the peduncle, not only of different genera, but also of different 2io Worsdell. — The Vascular Structure of species, as has recently been clearly observed 1. The greater or less development of the centrifugal portion of the bundle does not appear to be the sole or even the principal con- ditioning factor in determining the amount of centripetal xylem present, for where the latter is fairly well represented, the centrifugal portion may be better developed than in those cases where it is absent. There is evidently in these organs, as in the peduncle, the axis of the cone, and the foliage-leaf, a tendency, in spite of other prevailing influences, for the primitive mesarch characters to appear in the bundles, this tendency being stronger in some genera than in others, notably in those in which the sporophylls are of large size, this latter being doubtless indicative of their more primitive character as compared with the smaller and more highly modified sporophylls of other genera. In the cones of both sexes of all genera there are always, at the base of the cone, a number of sterile sporophylls, which may either have the form of the fertile organs and be crowded together like these, or may be elongated structures more like bracts in shape and position ; frequently a gradual transition occurs between these latter and the normal fertile sporophylls. It is interesting to note that the vascular structure of these barren sporophylls, while partaking in a general way of that of the fertile organs, exhibits a more primitive character, owing chiefly to the fact of the absence of the sporangiferous function in these organs, whereby these primitive structures, which consist of the more frequent presence of concentric structures and the occasional abortion of parts of the vascular system, have not been so much interfered with and altered as in the case of the fertile sporophylls. The small bundles in the cortex of the upper portion of the peduncle of both sexes, which are usually observed in pairs, and which supply the sterile sporophylls, possess very often a small quantity of centripetal xylem 2, which is frequently developed as trans- fusion-tissue both in a ventral and lateral position. Scott, loc. cit. 2 Cf. Scott, loc. cit., p. 406. 2 I i the Sporophylls of the Cycadaceae. I will now proceed to describe the structure of the vascular bundle-system of the sporophylls of the various genera of the order. Cycas revoluta, Thunb. Male Sporophyll. The few sporophylls available for investigation were supplied from a male cone preserved in spirit in Museum No. I in the Royal Gardens. On this account the course of the bundles from the central cylinder of the axis of the cone to their entry into the sporophyll was not observed ; but this has been adequately followed in C. siamensis Miq. and C. circinalis L. by Thibout \ who finds that a single bundle enters the cylinder from the sporophyll. The sporophyll does not assume such a perpendicular position with regard to the axis of the cone as in most genera, but is more or less curved. It is also larger, not being quite so reduced in size as in other cases. There are a number of bundles in a row whose size is in correspondence with that of the sporophyll. They possess sharply defined centrifugal xylem in fair quantity. Most of them have an almost equal quantity of centripetal xylem ; but the latter, in one or two cases, far exceeds the centrifugal, consisting of a great number of elements extending some distance from the protoxylem. The tracheides of the centripetal are often far larger than those of the centrifugal xylem. But in every case the centripetal is less brightly and sharply stained than the centrifugal portion, this fact indicating that the latter is more strongly lignified and thus chiefly functional in conduction. The chief feature of the structure of these male sporophylls of Cycas is the appearance, in the bundles of the lower portion of the stalk, of centripetal xylem. In no other genus is this tissue so well-developed in this region of the sporophyll, and it affords an instance of the way in which the centripetal tracheides may form the chief part of the xylem quite low down in the stalk. In the sterile portion the centrifugal 1 Recherches snr l’Appareil Male des Gymnospermes, 1896. 2 1 2 Worsdell. — The Vascular Structure cf xylem of the bundle is greatly reduced in quantity, its elements less sharply defined, and less brightly stained. The centripetal xylem usually exceeds it in quantity. Female Sporophyll. This organ is situated on the main axis of the vegetative stem, and not on a strobilus as in all the other genera. It occurs in whorls alternating with the foliage- leaves. It is larger in size than the sporophyll of any other genus, and more nearly resembles a foliage-leaf in external conformation ; herein lies probably a more primitive and less modified character ; the sporophylls of other genera, from their crowded arrangement on a cone, having undergone a much greater modification in form. These sporophylls differ also from all others in possessing a very long stalk, and in bearing a larger number of megasporangia than two, these being situated, not on the terminal expanded portion of the sporophyll, as in other genera, but at regular intervals on short projections from the stalk. The size of the sporangia is in correspondence with that of the sporophyll, being the largest of the whole order. The structure of the vascular bundles of the sporophylls is remarkable and interesting ; it is precisely that which is found in the bundles of the peduncle of Stangeria. The strands, as seen in transverse section, are arranged laterally in groups which form a single row (Fig. i). Some bundles or groups of bundles are entirely or partially surrounded by a thick zone of cells, probably pericyclic, filled with dense protoplasmic contents and conspicuous nuclei. This, again, is surrounded by a belt of stone-cells, possibly representing an endodermis, which have conspicuous reticulate thickenings on all their walls. But these two tissues are of inconstant appearance. The centrifugal xylem and the phloem are of very great development, which is quite equal to that of the bundles in the peduncle of other genera, and is in evident correlation with the size and number of the sporangia which they have to supply. The centripetal xylem, either in contact with the protoxylem, or as scattered tracheides or groups of 213 the Sporophylls of the Cycadaceae. tracheides lying some distance away, is always present ; in position and development it is the exact counterpart of that in the peduncle of Stangeria. It sometimes occurs opposite a gap between the bundles. Transfusion-tissue , of quite small tracheides with bordered pits or close reticulations, occurs in proximity to the centri- petal xylem both in a ventral and lateral position, and its derivation from this tissue is obvious. It is present in con- nexion with the cells containing dense protoplasmic contents surrounding the bundle ; where these cells are absent the transfusion-tissue is also no longer to be found. Some of the bundles are curved into a semi-concentric shape. On the dorsal side of nearly all the bundles are small strands with inverted orientation, their phloem being directly in contact with that of the main bundles. Each has an active cambium with (usually) a fair amount of xylem and phloem developed. Some have protoxylem ; other smaller ones have none, so that they are not all secondary. In this respect they resemble the similar strands observed in the peduncle of Ceratozamia . In one case, on the dorsal side of the protoxylem, i.e. on the side towards the periphery of the sporophyll, two tracheides occur which have evidently been formed by a cambium centrifugally, for regular radial rows of cells are seen on the dorsal side of these two tracheides. Also, on the ventral side of some of the tracheides of the centripetal xylem of the main strand, evidence of cambial divisions occurs such as I have also seen in the peduncle of Stangeria ; these were always observed in sister-cells of such tracheides lying on the inner (ventral) side of the latter towards the pith. Such division-walls were noticed nowhere else in the surrounding parenchyma-cells. Two strands are observed to pass off to a sporangium and enter it without previous branching. One of these is almost completely concentric in structure ; it has a large amount of secondary centrifugal xylem, and about three groups of protoxylem are observed around the inner margin of the latter. In the greater part of its contour the strand is 214 W or s dell. — The Vascular Structure of evidently of primary origin, but a small segment of the cylinder is probably formed of purely secondary tissue, though this is not perfectly certain ; but no crushed phloem- elements are observed at the periphery of this part. In the central parenchymatous tissue, or pith, of the strand occur a few tracheides isolated or in groups, some of which are attached to the protoxylem ; they do not extend to the centre, but quite a small central free space is left ; they evidently belong to the primary centripetal system (Fig. 4). This concentric strand, for about three-fourths of its con- tour, has small inverted strands attached to it, whose tissues appear to be mostly secondary ; there is no sign of primary phloem anywhere, no crushed elements of the latter being visible, most of the elements in this region occurring in the same radial rows with cambial cells. But the xylem of these strands, besides the numerous layers of secondary tracheides, has in all cases some primary xylem, some of the elements being small, others large and developed as transfusion-tissue with reticulate thickenings on the transverse wall, which occur both in a lateral and a ventral position, often in great quantity. In only one of these small bundles could I discern any sign of protoxylem, in most this tissue appears to be absent. In one or two cases there seems to be a con- siderable quantity of primary centrifugal xylem. In the tissue below the place of insertion of other megasporangia the strands are quite collateral, with a few centripetal tracheides, of which, in one instance, a radial row of three was attached to the protoxylem ; a sheath of stone-cells extends round close to this row of tracheides. In some of the bundles a large amount of transfusion-tissue runs out amongst the cells with conspicuous nuclei and dense pro- toplasmic contents occupying the region between the strands and the sheath. In the sterile region of the sporophyll, above the insertion of the sporangia, the bundles are much reduced as regards their centrifugal xylem and their phloem, so that the former approximates more to the dimensions of the centripetal xylem, 215 the Sporophylls of the Cycadaceae. while in the pinnae of the organ the bundles undergo a still greater reduction, and resemble there in their structure those of the ordinary foliage-leaf. Cycas circinalis, L. Female Sporophyll. In this plant the same structure of the strands prevails, but the latter are usually much elongated tangentially (Fig. 2). Some of these tend to become revoluted towards the dorsal side, a phenomenon similar to that oc- curring in the root of C. Seemanni , Al. Br.1 A single strand passes off to the sporangia, and in so doing divides into 3-5 strands which enter the latter ; some of those have a perfectly concentric structure. Fig. 3 repre- sents a bundle from the lamina in transverse section. The sporophylls are much more slender and the sporangia smaller than in the last species. Stangeria paradoxa, Th. Moore. Male Sporophyll. In the stalk are a number of bundles in a row (except at the very base where there are only two). In the well-developed centrifugal xylem the tracheides are clear and well defined, taking the stain brightly. Centripetal xylem is quite absent. In the sterile portion the centrifugal xylem is much reduced, its elements being inconspicuous and not so brightly stained as in those of the stalk ; in many bundles one, three, or four elements of centripetal xylem occur, in other bundles there are none seen at all. Female Sporophyll. Two bundles leave the cylinder of ' the axis of the female cone, each from a separate strand, and may either enter the sporophyll without dividing further, furnishing to that organ two large bundles of equal develop- ment, or they may previously divide so as to form three or four bundles, of which the outer one on each side supplying the sporangium is considerably larger than the one or two 1 Gregg, Ann. Bot., Vol. i, p. 4, Fig. i, 1887. Q t 216 WorsdelL — The Vascular Structure of intermediate ones (Figs. 6 and 7). Only in one or two bundles are two or three very small centripetal tracheides seen. The bundle at either side which supplies the sporangium forks just on entering the base of the lamina, one branch passing upwards, the other bending off towards the reflexed lateral lobe, between which and the stalk the sporangium is seated. Before this bundle assumes its downward course into the lobe, however, and opposite the place of insertion of the sporangium, it gives off a branch which forks : of the two bundles thus formed, one enters the sporangium without further change, the other, immediately before entering the latter, divides into two (Fig. 5). The bundles of the lamina are characterized by a great reduction in the centrifugal and a corresponding increase in the centripetal xylem, so that the two parts are about equal in development (Fig. 8). The lamina of a female sporophyll at a very young stage was examined ; in this is a number of very small bundles, most of which have an element or two of centripetal and one or two of centrifugal xylem. A few of them have no centripetal xylem besides the protoxylem. It is difficult to make out which is primary and which secondary centrifugal xylem in some bundles ; the smaller elements in connexion with the protoxylem are probably primary. The centrifugal and centripetal xylem appear to be developed simultaneously. Many of the ele- ments of the latter are very small and closely united to the protoxylem, two or three together, and would most likely have the same spiral thickenings as the latter ; others usually single, lie some distance away. Scale Leaves and Barren Sporophylls . At the base of the peduncle of the male cone are large, narrow, elongated scales, consisting of a fleshy central portion, with a wing on either side. They contain at the base three bundles which are collateral in shape, and possess secondary centrifugal xylem ; a tracheide belonging to the centripetal xylem is here and there seen. In the upper part of the scale the bundles increase to four or six, becoming very much reduced in size, 217 the Sporophylls of the Cycadaceae. and exceedingly small. They are here seen to possess one or two centrifugal and one or two centripetal tracheides ; there is a very small quantity of phloem ; no definite cambium can be distinguished, so that it is possible that the centrifugal tracheides are primary. The bundles are arranged in an arc. Sclerotic cells are scattered about in the ground-tissue. The uppermost scales are more leaf-like and much broader than the lower scales, and represent sterile sporophylls, re- sembling in shape the fertile organs. One of these, whose structure was examined, has an extremely long foliar base, running ridge-like, a considerable distance down the peduncle. About half way up this foliar base a minute concentric bundle becomes differentiated from the ground-tissue, arising quite independently of any other strand (Fig. io). Higher up it gradually increases in size, and in the upper free part of the organ is observed to give rise to an imperfect branching system forming the nervation of one side of the sporophyll. About three or four branches occur, all very imperfect, interrupted for considerable spaces in different parts of their course, and dying out before reaching the margin of the sporophyll ; they appear to have a very reduced structure, and their tracheides are much twisted and contorted. From their evident progress towards extinction and the concentric structure of the bundle from which they spring, they may be said to represent part of the original primitive bundle-system of the sporophyll which has long since ceased its connexion with that of the peduncle. A portion of the system, however, has still retained this connexion, and after a large part of the sporophyll became fused with the peduncular tissues, the bundle constituting the base of this system assumed a more direct course into the axis, passing in in the upper part of the foliar base, instead of, as in the case of the small con- centric bundle, running down in the latter towards its lower extremity. This bundle, which enters the barren sporophyll from the peduncle, possesses, whilst still in the latter, two distinct and well-marked groups of centripetal xylem. The subsequent course and branching of this bundle Q 2 2i8 W or s dell. — The Vascular Structure of and the nervation to which it gives rise, though imperfect, are not nearly so rudimentary and reduced in character as the one above described, for the bundles composing it are perfectly functional and nearly everywhere continuous ; the tip of the organ is occupied by 3-5 bundles formed by the division of one of the two main branches of the bundle coming from the axis ; these bundles have the structure characteristic of those in the sterile portion of the sporophyll, while the large bundle in the lower part has an endarch structure. The presence of these bundles in the cortex supplying the barren sporophylls was observed in two or three female and also in several male peduncles ; they were most abundant in the former, where they often occur in longer* or shorter and very straight tangential rows in the outer part of the cortex, but may also occur quite isolated. Most of these bundles possess, besides the usual endarch portion, two, often very large, groups of centripetal xylem, lying each slightly to one side of the median ventral line, and with tracheides of which those farthest towards the ventral side are three or four times the size of those composing the centrifugal xylem. Protoxylem is seen to be attached to each group (Fig. 9). But there is considerable variation in the development of these centripetal xylem groups, for they may be very much more reduced than those shown in Fig. 9, or even almost entirely absent ; they may be also in much closer connexion with the centrifugal xylem, or even, as in the case of one such group belonging to a bundle in a male peduncle, perfectly continuous therewith. The reason why I attach special importance to these bundles arises from the fact that I regard each such bundle, with its three distinct groups of xylem, as the vestige of a primitive concentric leaf-trace bundle which was probably characteristic of the ancestors of these plants. Now in the cortex of the stem of certain species of Medullosa just such concentric leaf-traces are known to occur, which during their course outward become split up into three portions, the concentric structure becoming thereby lost. My Figure 9 the Sporophylls of the Cycadaceae. 2 1 9 may be aptly compared with Fig. 8 (in the text) of Weber and Sterzel’s work1, and with Fig. 9, Plate V of Solms’ paper. It will be seen that in the bundles of the fossil plants secondary xylem is present in all three parts in considerable quantity, while in those of modern plants this tissue occurs in small quantity in the complete dorsal portion only. In a male peduncle of Stangeria one of the bundles with its two groups of centripetal xylem was traced into the elongated barren sporophyll, as above described, when the centripetal xylem - groups were found to gradually die out as the bundle passes into the sporophyll, so that a purely endarch structure of the strand remained. I regard the two ventral groups of xylem each with its inner protoxylem-strand, as the remnants of two distinct bundles, whose phloem has disappeared and which are now more or less separated from each other and from the dorsally-placed complete bundle with which, in the ancestors of the plant, they formed a compact whole as a concentric leaf-trace bundle. In the same female peduncles groups of three bundles some little distance apart, with their xylems converging, are often seen close to the central cylinder. These appear to me to represent, in a rather different way, the same phenomenon of a vestige of an ancestral concentric strand ; in this case, however, the three parts have not become in any way reduced, but they have, however, undergone much more complete separation from each other than in the previous case. Bowenia spectabilis, Hook. Male Sporophyll. A single bundle leaves the vascular cylinder of the axis of the male cone and very soon, almost immediately, divides into two. These run through the cortex and enter a sporophyll, where they forthwith begin to divide up into a number of bundles. In the stalk of the sporophyll the bundles are nearly all 1 Weber u. Sterzel, Beitrage zur Kenntniss der Medulloseae, Chemnitz, 1896 ; pp. 17 and 18, Figs. 7 and 8 of text, and Plate III. Solms-Laubach, Ueber Medullosa Leuckharti, Bot. Zeit., 1897. 220 WorsdelL — The Vascular Structure of extremely minute, with centrifugal xylem only, the tracheides of which are well defined and thick-walled. The bundles are very irregular in position and orientated in different directions. In the sterile portion the centrifugal xylem is much reduced and quite insignificant. The centripetal xylem is fairly well developed and its tracheides much scattered and wandering. In some of the bundles the phloem is scarcely developed, and the parts of the xylem cannot be properly orientated. Female Sporophyll. Two bundles leave the central cylinder of the axis of the female cone ; in the outer part of the cortex each of these divides into two, so that four bundles are produced. In those cases where other bundles, besides these four, occur, in the stalk of the sporophyll, these are cut off by the two original bundles just before entering the latter, and passing off towards the ventral side of the sporophyll, they, in the majority of cases, twist on their axes, assuming thereby a more or less abnormal orientation (Fig. n). The number, orientation, and structure of these ventral bundles in the stalk varies greatly ; the following are the principal cases met with : — 1. One large bundle with inverted orientation. 2. Three bundles, two with inverted orientation, and the median one concentric (Figs, n and 12). 3. Two bundles, closely contiguous, lying sideways with their xylems facing each other. 4. Two bundles, one of which lies sideways, the other obliquely. 5. One bundle, lying close on the ventral side of one of the normal bundles, and with normal orientation, but placed rather obliquely. 6. One bundle, lying sideways. 7. One bundle, with normal orientation. All the ventral bundles pass up into the lamina, without taking any part in supplying the sporangia ; those which in the stalk were inversely orientated, in the lamina twist on their axes once more to assume approximately the normal orientation. Here the same change as regards the relative 22 I the Sporophylls of the Cycadaceae. development of the centrifugal and centripetal xylem occurs as in previous cases. Of the four normally orientated bundles in the stalk, the two lateral ones which supply the sporangia are considerably the largest. One or two small elements of centripetal xylem may occur in any of the four bundles, but are nearly always present in the large lateral ones. From each of these latter, as they enter the lamina, two bundles branch off to the sporangium ; and each of these, immediately before entering the latter, may divide up into two or three. In transverse section of the part immediately below the insertion of the sporangium, one of the two bundles is seen to be divided into three. Scale Leaves and Barren Sporophylls. At the base of the male cone are a number of scale leaves similar to those in Stangeria. They contain a row of many bundles with endarch structure and very small in size ; a few bundles also occur on the ventral side of this row and appear to be normally orientated. Immediately below the compactly- arranged sporophylls of the male, are two or three elongated, barren sporophylls (Fig. 13). In the case of one cone, each barren sporophyll contains, in its lower portion or stalk, a row of four or five bundles, of which the two end ones of the row are concentric in structure (Fig. 14), one of these being very small, the other of normal size. The intermediate bundles are rather larger in size than those of the fertile organ, this being due, possibly, to the smaller number of the former ; their structure is similar, the centrifugal xylem being sharply defined and the phloem well developed ; in the upper part, or lamina, the smallest concentric bundle dies out, while the larger one assumes a collateral structure ; while, as regards the other bundles, a similar change in their structure takes place as in the fertile sporophyll. DlOON EDULE, Lindl. Male Sporophyll. In the narrow stalk is a single row of bundles. All are of quite small size and of nearly equal 222 Worsdell. — The Vascular Structure of development as regards the centripetal xylem, which is quite small in amount. Nearly all the bundles have three or four (one bundle had seven or eight) tracheides of centripetal xylem ; these are, however, always faint, rounded, and weakly stained compared with those of the centrifugal xylem. In the sterile portion the usual structure of this part prevails ; some of the centripetal xylem is developed as transfusion- tissue which is both lateral and ventral in position. Female Sporophyll. Two bundles leave the cylinder of the axis of the female cone, and on their way through the cortex divide up into a considerable number of bundles, so that in the stalk of the sporophyll a row of about nine bundles occurs. Though the sporophyll is one of the largest in the order (equal in size to that of Encephalartos'), the mega- sporangia are not nearly so large as in that genus ; con- sequently, the bundles have also a considerably less development, and are all regularly orientated. In those sporophylls which bear two well-developed sporangia, the one or two bundles at either end of the row are larger and have better-developed phloem and centrifugal xylem than the intermediate ones (Fig. 15). In those with one abortive sporangium, the most external bundle of the row on the same side of the stalk is scarcely better developed than the rdst. In those with both sporangia abortive, the same applies to the external bundle on both sides. The bundles of the lamina exhibit the ordinary structure (Fig. 1 6). The external bundle of the row on one side of the stalk and a branch from the bundle next to it bend off in the lamina and pass down to supply the sporangium ; each of these divides into three, so that six bundles enter the latter (Fig. 17). Of these two or three of the smaller ones are perfectly concentric in structure. Some of the bundles of the inner portion of the integument of the sporangium have also a concentric structure, the protoxylem occupying the centre of the bundle and enclosed by primary tracheides ; the phloem, however, on the ventral side is rudimentary in character; the occurrence of these bundles in the integument seems to show that the 223 the Sporophylls of the Cycadaceae . concentric structure of those occurring in the sporophyll below the insertion of the sporangium and in the base of the latter is not entirely due to the radial symmetry of the sporangium which they supply. Encephalartos villosus, Lehm. Male Sporophyll. In the winged stalk is a row of bundles, of which the one on either side situated between the wing and the central region is the largest, and has a large amount of centrifugal xylem ; a certain small amount of centripetal xylem occurs at the sides of the bundle and is continuous with the centrifugal portion. The smaller bundles in the rest of the stalk have also very conspicuous centrifugal xylem of fairly large elements. Female Sporophyll. The axis of the cone to which these sporophylls were attached was not available for investigation, so that the course of the bundles therein could not be traced. The sporophylls are the largest of any in the order except those of Cycas. Their vascular system is, consequently, very well developed, and much resembles that of Cycas. As in that genus, the bundles have a large amount of centrifugal xylem, and many possess besides a considerable quantity of centripetal xylem, though others have less of the latter or even none at all. As in Cycas also, there are inverted strands on the dorsal side of many of the bundles, some of which have a small amount of centripetal xylem, while others have none ; most, if not all, appear to have protoxylem ; two strands placed back to back may be of equal size. But the bundles have a very irregular arrangement, being orientated and grouped in all kinds of ways, resembling somewhat the arrangement of the bundles in the basal region of the peduncle in Stangeria. At each side of the stalk there is an aggregation of bundles in close proximity and with variously orientated parts ; these are the strands which supply the sporangium ; the other bundles are smaller. In shape the bundles of the sporophyll are rounded and often very much curved, so as to 224 WorsdelL — The Vascular Structure of form an almost concentric structure (Fig. 18); indeed, in another species, E. horridus , Lehm., a primitive concentric structure for these bundles appears to be indicated by the peculiarity presented by some of them, as, for instance, one bundle which is curved into a horse-shoe shape so that the xylem from opposite sides almost meets at the open part of the strand so as to enclose a pith with protoxylem round its periphery ; in another larger, tangentially-elongated strand which is but slightly curved, there occur on the ventral side of two groups of centripetal xylem, and intimately connected therewith by radial rows of cells, two distinct groups of crushed phloem-cells, with a slight indication of these latter having at one time been united by similar tissue ; thus the combination of this phloem with the groups of centripetal xylem affords the phenomenon of two small strands with inverted orientation on the ventral side of the main strand ; other groups of centripetal xylem occur belonging to the same strand, but isolated and without any sign of phloem near them (Fig. 19). In the lamina the strands have about the same general structure ; but here the centrifugal xylem is extremely reduced as compared with that in the stalk, its tracheides also being usually much smaller ; it is about equal in de- velopment to the centripetal which occurs in considerable quantity, and in some bundles extends as transfusion-tissue, with scalariform thickenings on the transverse walls of the tracheides, both laterally, and even quite on to the dorsal side of the phloem. In E. horridus , Lehm., protoxylem was only seen on the side of the centripetal xylem. The megasporangium receives 4-5 bundles, some of which have a very distinct concentric structure (Fig. 20). Macrozamia Fraseri, Miq. Male Sporophyll. There are about a dozen bundles in the stalk lying in a regular row and somewhat unequal in size. They all have sharply-defined and brightly-stained centri- 225 the Sporophylls of the Cycadaceae. fugal and an entire absence of centripetal xylem. They arise from the branching of three bundles (which really represent two, of which one has undergone premature division) which enter the sporophyll from the axis of the cone, these three originating in the cortex from the single bundle which leaves the central cylinder. Owing to this premature division of one of the bundles entering the sporo- phyll, there are at first a larger number of bundles on one side of the organ than on the other. In a rather young male sporophyll of M. spiralis , Miq., there are about nine bundles in the stalk. Female Sporophyll. This is one of the large types of sporophyll. Owing to the fact that the material at my disposal was dead and withered, the arrangement of the bundles in the stalk could not be ascertained, as the sections broke up into small fragments. The strands appear to be of the same type, in form and structure, as those in the female sporophyll of Encephalartos , as they possess a large amount of centrifugal and also a small but well-marked quantity of centripetal xylem. In the lamina some of the bundles are a good size, the centrifugal and centripetal parts of the xylem being of about equal development, though the former seems to stain rather more sharply than the latter. Zamia latifolia, Lodd. Male Sporophyll. In the lower part of the stalk are three bundles with the typical centrifugal and no centripetal xylem. A little higher these divide up into a rather larger number, of which one from each side is seen to pass off to the dorsal side of the sporophyll. One of the bundles in this region show a good example of primary centrifugal xylem ; it has a single, minute, flattened tracheide of secondary centrifugal xylem separated from the other portion of this tissue by a layer or two of roundish parenchyma-cells; next, towards the ventral side succeeds a row of four or five tracheides, very well-defined, and with brightly-coloured walls, and then 226 WorsdelL — The Vascular Structure of a second row of as many, but rather smaller and less clearly- defined tracheides, these two constituting the primary centri- fugal xylem ; these tracheides do not lie in the same radial rows as the cambium-cells further out, but are separated therefrom by the parenchymatous cells afore-mentioned, some of which overlap in position two of the cambium-cells. In the sterile portion the mesarch structure of the bundles is very pronounced. Female Sporophyll. In the stalk, besides the four bundles which exhibit the usual structure and no centripetal xylem, there is a very small one on the ventral side with inverted orientation and almost concentric structure, which has evi- dently sprung from one of the four bundles opposite to which it lies. Zamia pumila, Linn. Male Sporophyll. In the stalk is a row of about six bundles, quite small in size, one or two having only a single tracheide of centrifugal xylem. Of these small bundles, two occur on the dorsal side of the row, one of them lying sideways and being semi-concentric in shape. That there is a tendency for a concentric structure to appear in the bundles is shown also by one in which a rudimentary phloem occurs on the ventral side of the protoxylem. In one of the bundles two elements of centripetal xylem occur. In the sterile portion the bundles have the usual mesarch structure. Zamia Skinneri, Warsz. Male Sporophyll. In the stalk are three bundles with well- developed secondary tissue, of which the tracheides, especially in the median bundle, are of considerable size and very brightly coloured. To each bundle a single tracheide of centripetal xylem is seen to be attached. In the region just above where the bundles pass off to the sporangia, occur two large bundles with well-developed, typical mesarch 227 the Sporophylls of the Cycadaceae. structure, both the centrifugal and centripetal xylem having each its distinct group of protoxylem. Higher up in the sterile portion these bundles split up into a larger number, some of which are partially concentric in structure like those in the female sporophyll of Encephalartos. Zamia muricata, Willd. Male Sporophyll. There are three very small bundles in the stalk, with the usual structure, the two outer ones of which have a sidelong position. But at a certain level there occur in close proximity to two of these the rudiments of a bundle lying obliquely towards the ventral side of the normal bundle ; in one case the rudimentary bundle appears to have a mesarch structure, a minute protoxylem -group intervening between the reduced phloem and two rather scattered tracheides. In the sterile portion many of the bundles exhibit a structure intermediate between mesarch and endarch, tracheides occurring at the sides of the proto- xylem of which it is not easy to say whether they belong to the centrifugal or centripetal portion of the xylem ; in the same region are other bundles with a definite mesarch structure, some of which have no centrifugal xylem at all. Zamia Lindeni, Regel. Male Sporophyll. In the stalk is a short row of four bundles which are normal in every respect (Fig. 21). In the sterile portion are a great number of bundles orientated in every direction and with the usual structure. Barren Sporophyll. The smaller compactly-arranged barren sporophylls at the base of the male cone, which are similar in conformation to the fertile organs, present a very interest- ing structure. The bundles in the stalk, sometimes one, sometimes several in a row, are all more or less to be inter- preted as having a partially concentric structure ; in some 228 WorsdelL — The Vascular Structure of cases the appearance is as if two distinct bundles were closely united, but not completely fused, by their ventral faces ; in others the bundle has a curved, or horse-shoe shape. The bundles are all irregularly orientated, mostly assuming a sidelong position (Fig. 22). But these structures are of importance from the fact that they exhibit more primitive characters than the bundles of the fertile sporophyll, the latter differing entirely therefrom, both in structure and arrangement, being perfectly collateral and normally orien- tated. It is owing to the absence of the sporangiferous function in these barren sporophylls that this primitive structure has been retained. ZAMIA LODDIGESII, Miq. Female Sporophyll. Two bundles leave the cylinder of the axis of the female cone, each from a distinct strand thereof. On their way through the cortex each divides into two, thus forming four bundles. The small bundles which occur in pairs in the cortex of the peduncle and which supply the barren sporophylls at the base of the cone have, many of them, centripetal xylem, of the normal kind and in the form of very distinct transfusion-tissue, both in a ventral and a lateral position (Fig. 24). Of the four bundles in the stalk of the sporophyll, the smaller, newly-formed ones move somewhat forward towards what becomes the ventral side of the organ. The two lateral bundles are somewhat larger, though not so strikingly so as in other genera. All these bundles have several elements of centripetal xylem which are often larger than the tracheides of the centrifugal portion ; in this respect this plant affords quite an exception to the general rule. One bundle turns off to the megasporangium on each side, dividing into two main branches ; one of these, the proximal one, before entering the sporangium, divides in different directions into three branches ; the distal bundle, which 229 the Sporophylls of the Cycadaceae. passes obliquely away to the far side of the sporangium, remains undivided. As seen in transverse section, a bundle about to enter the sporangium has a curved contour, with very well-developed centrifugal and often a large amount also of centripetal xylem ; the occurrence of this latter is perhaps correlated with the evident tendency of these bundles to revert to a concentric structure. In many of the bundles in the most distant portions of the lamina, transfusion-tissue is very markedly developed, resulting from the extension, even on to the dorsal side of the phloem, of the centripetal xylem (Fig. 23). As in nearly all other cases, the tracheides of the centrifugal are always much smaller than those of the centripetal xylem. Zamia furfuracea, Ait. Female SporophylL In the stalk are four bundles with the usual structure ; no centripetal xylem occurs here ; the two lateral bundles are rather better developed than the others and have larger centrifugal tracheides. A little higher up in the sporophyll one of the middle bundles gives off a very small branch which gradually, in passing upwards and towards the ventral side, turns on its axis, assuming directly afterwards a perfectly concentric structure, which again, higher up, opens out and becomes collateral ; there may be more than one bundle on the ventral side of the normal row ; these bundles pass up into the lamina without again assuming a normal orientation. In this latter region of the sporophyll transfusion-tissue is well developed in many of the bundles. Two bundles enter each megasporangium ; one of these, the largest, being the lateral, unbranched bundle of the stalk on the same side which passes up round the axil of insertion of the sporangium ; the other, which is very much smaller, as a descending branch from a bundle which has passed a considerable distance up into the lamina ; there is thus 230 Worsdell. — The Vascular Structure of a difference in this respect from what occurs in the case of if. Loddigesii. The smaller bundle is collateral in structure and somewhat curved in shape, with two or three elements of centripetal xylem, the other larger one has well-developed centrifugal and no centripetal xylem, and is curved inwards so as to assume an almost concentric structure ; in the base of the sporangium this bundle begins to separate into two, as a result of which two almost distinct bundles come to lie in close contact, with their xylems facing each other. Zamia Fischeri, Miq. Female Sporophyll. The arrangement and structure of the bundles here is very similar to what obtains in the last plant ; a small ventral, inverted bundle also being present. Zamia Leiboldii, Miq. Female Sporophyll. The sporophyll examined bore three megasporangia (Fig. 25); as a consequence of this the arrangement and structure of the bundles in the stalk are somewhat modified to suit this abnormality. On the side on which the two sporangia occur is a bundle much larger than is usually the case, and close to it another bundle almost or quite as large, but lying somewhat obliquely and with inverted orientation. The bundle at the other side of the sporophyll is very much smaller, corresponding more in size to the bundles supplying the sporangia in previous species. At a slightly higher level the inverted bundle on the side of the two sporangia turns on its axis and becomes united laterally, in the normal position, with the adjacent bundle, so as to form a single exceptionally large bundle, considerably elongated in the tangential direction, and with very well developed centrifugal but no centripetal xylem. The unusual size of this bundle is obviously correlated with the extra sporangium attached to that side of the sporophyll. 231 the Sporophylls of the Cycadaceae . At a somewhat higher level still, this large strand divides up into a row of four, of which the two end-bundles are larger, and have their centrifugal xylem developed in such a way as to indicate that they are the bundles which are to supply the sporangia. Hence, probably on account of the lop-sided character of the sporophyll, owing to the extra sporangium being present on one side, the vascular system, to counterbalance this, is concentrated on this side, the normal row of four bundles, instead of being evenly spread over the tissues of the sporophyll, as in all other cases, is limited to the side on which this extra sporangium occurs. The smaller bundle on the other side of the stalk has nothing to do with any of the sporangia, but passes up into the lamina. In the upper part of the stalk about three smaller ventral bundles are seen, one of which is concentric in structure, the others irregularly orientated ; tracing these downwards, one of them is seen to spring from one of the four bundles of the row, the other two, of which the concentric one opens out and becomes collateral, end blindly each close up against the rim of a mucilage- canal. Higher up, the two end-bundles of the normal row are seen to pass off respectively right and left to the sporangia (the one supplying the two sporangia much earlier than the other). The extra sporangium may be fed by subsequent branching of the bundle passing off to the normal sporangium on that side ; but the exact course of the bundles to each sporangium was not followed. The bundle passing off to supply the single sporangium on the other side appears to give off a small branch while doing so ; this does not happen in the case of the bundle supplying the two sporangia. The large bundle found entering the sporangium is almost concentric, as in other cases ; on entering the sporangium it splits up into a number of parts which radiate out from the common centre. Fig. 25 a and b represent respectively ventral and lateral- views of the sporophyll. R 232 Worsdell. — The Vascular Structure of Ceratozamia latifolia, Miq. Male Sporophyll. The course of the bundles from the central cylinder of the axis of the male cone to the sporophyll has been thoroughly described by Thibout in C. mexicana, Brongn.1 The occurrence of medullary bundles in the axis has also been mentioned and figured by him 2, and by Scott 3 in that species. In the present species I observed in one case three small bundles lying in the pith immediately within the normal ring which, one after the other, end blindly upwards. In another case a single large bundle in a similar position was traced upwards for a long distance ; in one part of its course it split into two bundles which directly after- wards again fused together into one. It twice partially fused with a bundle of the ring during its course, but was not seen to completely fuse therewith. A similar large bundle, which at first was collateral, became higher up concentric , and at a still higher level again collateral ; it eventually ended blindly in the pith. These characters are confined, in this species, to the male cone ; they do not, however, occur in the male cone of C. Miqueliana , H. Wendl. In the lowest part of the stalk of the sporophyll three bundles occur, of which the median one is much smaller than the others ; the centrifugal xylem is clearly defined and brightly coloured, and there is no sign of any centripetal xylem. For the rest, the structure is similar to that of the bundles in the male sporophyll of other genera. Ceratozamia mexicana, Brongn. Male Sporophyll. In the lowest part of the stalk are three bundles, the median one as in the last species, being much smaller than the others, and lying out of the row or a short distance towards the dorsal side (Fig. 26). In the sterile 1 Loc. cit. p. 23. 2 Loc. cit. p. 24. 3 Loc. cit. p. 412. 233 the Sporophylls of the Cycadaceae. portion of a young sporophyll in which the parts of the bundle are not as yet fully developed, the xylem consists of protoxylem, on the ventral side of which are two or three elements of centripetal, and occasionally, on the dorsal side, one or two elements of primary centrifugal xylem, and in the lateral region are primary elements intermediate between the two kinds. In one or two bundles one tracheide was seen which appears to be secondary in origin, and which was deeper in colour and larger than the others, and isolated, lying quite apart from the group of primary tracheides. Female Sporophyll. Two bundles arise independently from a distinct strand of the central cylinder of the axis of the female cone ; on their way outward through the cortex they each divide up into a number of bundles, some of which pass up on to the ventral side of the others, turning on their axes as they do so, so that in the stalk of the sporophyll there is a ring or double row of bundles with their xylems all pointing towards the centre. The ventral bundles of the stalk with inverted orientation nearly all, in the lamina, turn on their axes and become normally orientated, so that two rows of bundles pass upwards through this region ; one or two bundles, however, lying somewhat out of the ventral row, towards its inner side, retain their inverted orientation. As the sporophyll was very young, the bundles had not yet attained their full development ; but in the stalk, centrifugal and no centripetal xylem was developed. C. MlQUELIANA, H. Wendl. Female Sporophyll. In the lowest part of the stalk there is a row of three or four bundles, of which the two outer lateral ones are of great size, with a very well-developed mass of centrifugal and two or three quite small elements of centri- petal xylem. The one or two intermediate bundles are quite small, elongated radially, and may have an element or two of centripetal xylem. In some sporophylls there occur on the ventral side of and lying usually some distance away from R 2 234 Worsdell. — The Vascular Structure of the normal row, several small normally-orientated bundles whose origin, owing to the cone being in a state of decay and the tissue therefore connecting the sporophylls with the axis being partially destroyed, could not be ascertained (Fig. 27). In the case of a small bundle 'with inverted orientation, lying on the ventral side of one of the large lateral bundles, this was observed to be cut off from this latter, whose phloem gradually extended round the xylem, so as to render the bundle partially concentric in structure, when part of it became severed towards the ventral side, which part, twisting on its axis, formed the small inverted bundle. In the lamina, the bundles have usually an equal quantity of centrifugal xylem, but in some the latter is absent. In the young female sporophyll the developing bundles of the lamina appear to form centrifugal and centripetal xylem at the same time ; in some bundles both kinds are seen together (Fig. 28), in others centripetal alone ; in others, again, centrifugal alone ; in some there is an intermediate stage where the tracheides lie at the side of the protoxylem. This variation in the respective development of the centrifugal and centripetal xylem coincides with what is found in the mature sporophyll. In the barren sporophyll of this genus the bundles are all of uniform development, and none are so markedly developed as those supplying the sporangia in the fertile sporophyll. Summary. The result of my investigations into the structure of the sporophyll for each genus may be thus summarized: — Cycas, $ : in lower part of stalk (though probably not at extreme base) centripetal xylem of bundles is equal to, or greater in amount than, centrifugal xylem. ? : borne directly on vegetative axis ; larger and more leaf-like than in any other genus ; bundles of large size, with large development of centrifugal and the Sporophylls of the Cycadaceae. 235 considerable and constant quantity of centripetal xylem, with small inverted strands on their dorsal side, structure exactly resembling that of bundles in peduncle of Stangeria ; large strand supplying mega- sporangium has almost completely concentric structure with few primary tracheides in central pith. Stangeria, d : two bundles in stalk with purely endarch structure. : — $ : two to four bundles in stalk, of which two lateral ones supplying sporangia are much larger in size ; centripetal xylem of two or three tracheides only in one Or two bundles. Bowenia, d : number of very small bundles in stalk, irregu- larly orientated, and with purely endarch structure. ? : four principal bundles in stalk, of which two lateral are largest ; few tracheides of centripetal xylem always present in latter and may also occur in smaller bundles ; besides these, occur small bundles on ventral side very differently orientated in different sporophylls, and sometimes quite concentric in structure. Dioon, d : number of small bundles in stalk ; most have three or four or more centripetal tracheides. ? : one of the largest in order, but megasporangia relatively small in size ; number of bundles in stalk, but extreme base not represented. Lateral bundles of row more strongly developed than rest only when sporangia they supply are not abortive ; small quantity of centripetal xylem present in some bundles. Eneephalartos, d • r°w of bundles in stalk, of which some are larger than others, and have small amount of centripetal xylem. $ : one of the largest in order ; bundles in stalk resembling those of $ of Cycas in structure, but very irregularly orientated and grouped, and more curved in shape, some having almost concentric 236 W or s dell. — The Vascular Structure of structure. Centrifugal xylem nearly always present, often in large quantity. Small inverted strands attached to dorsal side of bundles, as in Cycas. Macrozamia, $ : two to three bundles in lowest part of stalk, which rapidly divide up into larger number with usual endarch structure ; no centripetal xylem. ? : one of the largest in order ; bundles have very well-developed centrifugal and well-marked quantity of centripetal xylem ; bundles resemble very much those in ¥ of Encephalartos. Zamia, $ : three to four bundles in stalk with rather irregular arrangement and orientation ; centripetal xylem some- times present ; some bundles are partially concentric in structure. Z. Loddigesii , Miq., ? : four bundles in stalk, of which two lateral are largest. All bundles have centripetal xylem, whose elements are often larger than those of centrifugal portion. Z . furfur acea, Ait. ; Z. latifolia , Lodd. ; Z. Fischeri , Miq., ? : the four bundles have no centripetal xylem. A small bundle, either concentric or collateral in structure, occurs on ventral side of row as a branch from one of normal bundles. Z. Leiboldii , Miq., ? : abnormal, bearing three sporangia ; in consequence, vascular system of stalk is concentrated on side on which two sporangia are borne. Ceratozamia latifolia , Miq., <3 : medullary bundles, either con- centric or collateral in structure, in axis of cone ; three bundles in stalk of sporophyll ; no centripetal xylem. C. mexicana , Brongn., $ : of three bundles in stalk, median one is much smaller than other two. $ : ring or double row of bundles in stalk, whose xylem is in all cases directed inwards ; these arise from division of two original bundles arising from cylinder of axis. Owing to young state of organ no difference in respective development exists. 237 the Sporophylls of the Cycadaceae. C. Miqueliana , H. Wendl., ? : row of three to four bundles in stalk, of which two lateral are of great size compared with the others ; along with largely developed centri- fugal, few small centripetal tracheides usually present, as is also the case with smaller bundles ; other smaller bundles, quite apart from normal row, with either normal or inverted orientation. The more general characters of the sporophyll are the following : — Male. A single bundle leaves the cylinder of the axis of the cone, which, on entering the stalk of the sporophyll, divides into three The bundles supplying the sporangia are much smaller in size than the similar ones on the female side, owing to the comparatively brief period of attachment to the sporophyll of the microsporangia and the short functional activity of the latter. They also diverge less from the mesarch structure of those of the foliage-leaf than do the bundles of the female sporophyll. Female. Two bundles leave the cylinder of the axis of the cone, usually dividing up in the cortex into a larger number, so that, as a rule, four bundles occur in the stalk of the sporophyll, of which the two lateral ones are much larger than the rest, this being correlated with their function of supplying the megasporangia during the long period of development and attachment of the latter to the sporophyll. The divergence from the ordinary mesarch structure of the foliage-leaf is much more marked here. In the sterile portion of the sporophyll of both sexes, i. e. the part above the insertion of the sporangia, the mesarch structure of the bundles prevails, showing how, the sporangial element being eliminated, the ordinary and typical structure of foliar bundles reappears. As compared with the bundles of the stalk, the centripetal is as a rule much more developed than the centrifugal xylem, while the phloem, which shares the development of the centrifugal xylem in the stalk, becomes reduced and insignificant. 238 Worsdell. — The Vascular S true here of As a general theory resulting from the preceding investi- gations, it appears to me a quite probable view that the sporophyll in Cycadaceae is a more primitive foliar organ than the foliage-leaf, and for the following reasons : — a. The theory of Bower with regard to the development and morphology of spore-producing members in the Lyco- podineous series of plant-forms and the phylogenetically late appearance of assimilating foliar appendages, which arise by progressive sterilization of the tissues of the former, appears to me to be a highly natural hypothesis and one containing a large element of truth. Now, it seems to me highly probable that the Fern- sporophyte had a very similar origin, that it also arose from a fertilized oospore which, instead of producing motile re- productive organs, gradually replaced these latter by asexual, non-motile spores contained in sporangia which eventually became raised into an aerial position by progressive steriliza- tion of the tissues of the young developing sporophyte : a theory which will not be accepted by those who incline to the view that the Fern-sporophyte had an entirely different origin from that of the Bryophytes 1, with which group it is here my purpose to emphasize its similarity of origin. There is, in fact, no reason to think that the order of development of the two kinds of foliar appendages, sporophyll and foliage- leaf, should, in the Fern-series, have been fundamentally different from that in the Lycopodineous series : the onto- genetic developmental history of the two types is to-day so similar as to point to the probability of their phylogenetic developmental history having been also similar. This being so, we may assume that evolution may have proceeded as follows. First in order of time, sporophytes which bore sporophylls only would exist ; next to these would succeed those in which both sporophylls and assimi- lating foliar organs were present. But it is quite conceivable that subsequent modification of all the sporophylls might 1 Scott, Address to Botanical Section of Brit. Assoc. Meeting, Liverpool, 1896, pp. 9-10. the Sporophylts of the Cycadaceae . 239 have taken place in certain forms, producing types like the Marattiaceae at an early period, so that the sporangia came to be borne on assimilating fronds, as in the case of most modern Leptosporangiate Ferns. On the other hand, the original type, viz., that in which the sporangiferous organs were distinct from the assimilating leaves, probably persisted right through, even down to the present day. From such forms as these latter I imagine it to be quite conceivable, and even probable, that modern Cycads took their origin. b. From the investigations of Solms-Laubach, Scott, and myself, it appears that the peduncle of these plants has in several ways, notably in the simplified vascular bundle-system, the presence of a mesarch structure in the bundles of the central cylinder, and of concentric strands in various parts of the organ, a more primitive structure than the vegetative stem, and probably therefore more nearly represents the original typical stem-structure. This being the case, it might naturally be expected that the foliar appendages of the peduncle would possess a more primitive structure than the foliar appendages of the vegetative stem. c. This primitive structure is represented by the concentric bundles which occur in both the fertile and barren sporophylls of several genera, especially in the latter organs, where the special physiological function of the former has not interfered with the original structure. These concentric bundles are absent from the foliage-leaves, the structure, number, and orientation of whose bundles are extremely regular, constant, and well-defined, whereas in the sporophylls the reverse is the case, a fact which probably points to their possessing a more primitive structure, viz., one not so perfectly adapted and stereotyped to subserve a special physiological function, as is the case with the foliage-leaves. In conclusion, the special point upon which I desire to lay stress in this paper is this : that though the sporophyll, according to my view, is a more primitive organ than the foliage-leaf, for the reasons above adduced, the main and, physiologically, most important part of its vascular structure 240 Worsdell. — The Vascular Structure of has become, as a result of the sporangiferous function, much more highly modified from the primitive type than that of the foliage-leaves, this special function influencing more or less all parts of the vascular system. But though this modifica- tion in structure has taken place, it has not obscured, to such an extent as in the foliage-leaves, the primitive character of the organ which, either in the form of concentric or partially concentric strands, of irregular orientation and arrangement of the bundles, of great variability in size of the latter, or of a tendency to abortion of certain of the more primitive parts of the vascular system, appears over and over again, in one place or another, throughout all the genera investigated. I desire to express my obligation to Dr. D. H. Scott for all the assistance and the numerous criticisms which he has, as usual, so kindly afforded me. EXPLANATION OF FIGURES IN PLATES XVII AND XVIII. Illustrating Mr. Worsdell’s paper on the Cycadaceae. The following are the abbreviations used: — ph. phloem; ph}. ventral phloem; x1. centrifugal xylem ; xl. centripetal xylem ; px. protoxylem ; tf transfusion-tissue ; pb. bundle from peduncle ; fb. foliar bundle ; ped. peduncle ; bs. barren sporophyll ; sph. sporophyll (fertile) ; sp. sporangium ; Im. lamina ; st. stalk. Fig. 1. Cycas circinalis , L. Diagrammatic transverse view of the arrangement of the vascular bundles in the stalk of the 9 sporophyll. x 35. Fig. 2. Cycas circinalis , L. Transverse section of a strand from the stalk of the 9 sporophyll. x 65. Fig* 3* Cycas circinalis , L. Transverse section of a bundle from the lamina of the ? sporophyll. x 100. Fig. 4. Cycas revoluta , Thunb. Transverse section of a concentric strand about to enter the megasporangium, x 65. Fig. 5. Stangeria paradoxa , Th. Moore. Diagram of the vascular bundle-system of the 9 sporophyll. Fig. 6. Stangeria paradoxa , Th. Moore. Diagrammatic transverse view of the arrangement of the vascular bundles in the stalk of the 9 sporophyll. x 35. the Sporophylls of the Cycadaceae . 241 Fig. 7. Stangeria paradoxa , Th. Moore. Transverse section of one of the bundles in the stalk of the 9 sporophyll which supply the megasporangia, x 145. Fig. 8. Stangeria paradoxa , Th. Moore. Transverse section of a bundle from the lamina of the $ sporophyll. x 145. Fig. 9. Stangeria paradoxa , Th. Moore. Diagram of the vascular bundle- system of a barren sporophyll at base of unknown from homotaxial deposits N = Norway ^ in Britain, s w = Sweden 1 Preglacial. Early Glacial. Interglacial. Late Glacial. Neolithic. Clematis Vital da, L E ? Thalictrum minus , L E S ,, jlavum, L E E . . . Sw Ranunculus aquatilis, L E E E S E S E S I M „ sceleratus, L E E E „ Flammula , L E E S M „ Lingua , L E S , , repens , L E E E S E S M „ bulbosus, L E „ Sardous , Crantz . . . E „ parvijlorus, L E Caltha palustris, L E E E S Nuphar luteum , Sm E E E Nymphaea alba , L * G Sw Papaver somniferum, L S Fumaria officinalis , L S Cakile maritima , Scop Sw Viola palustris, L E E S E S S M W Silene maritima, L E Lychnis alba, Mill. ... S „ diurna, Sibth S S „ Floscuculi, L s S Stellaria aquatica , Scop E ,, media, Cyr E E E S s „ uliginosa, L .. s Arenaria trinervia, L G Spergula arvensis , L ... s Montia fontana, L E E s Hypericum quadrangulum, L. . . . s ,, elodes, L . . . s Tilia platyphyllos , Scop G Sw Linum, sp . . . s Geranium columbinum, L G Oxalis Acetosella, L ... S E S Llex Aquifolium, L G .. . E Rha?nnus Frangula, L ... E ... Sw Geological History of the British Flora . 247 Preglacial. Early Glacial. Interglacial. Late Glacial. Neolithic. Acer campestre , L E G „ monspessulanum , L ... E Prunus communis , Huds E E E S „ domestical , L E S „ Avium , L E E S ,, Padus , L. E S E S Spiraea Ulmaria , L E ... S Rubus Idaeus , L • • . E E E S E S „ fruticosus , L E E E S M ,, caesius, L ... ... Sw ,, saxatilis , L Sw Dryas octopetala , L • . . S Potentilla Tormentilla , Neck. . . . ... E S ? S M „ Comarum , Nestl S E S M Alchemilla arvensis , L ... S Poterium officinale , Hook E E E E S S M canina, L . . . E S Pyrus iorminalis, Ehrh ... • . . E ,, Aucuparia , Gaert ... ... w „ communis, L ... ... E Crataegus Oxyacantha , L E E • . . E W S Saxifraga oppositifolia , L G „ Hir cuius, L . . . G „ aizoides, L ' • • • G Hippuris vulgaris , L E E E S E S E S M Myriophyllum spicatum , L E E E S E S S I M „ alternifolium , L. . . ... . . . Sw Trapa natans, L E • • . G ... Sw F Hydrocotyle vulgaris, L E • • . E S M Apium nodijlorum , Reich ... • . . S Cicuia virosa, L ... Sw .SVwzw latifolium, L . . . • • . Sw Oenanthe Lachenalii, Gmel. . . . E S ,, croc at a, L , , , ... e"? ,, Phellandrium , Lam. • • . E E E Aethusa Cynapium, L S Angelica sylvestris , L E Sw Peucedanum palustre , Moench. . ... Sw Heracleum Sphondylium , L. ... E E Hedera Helix , L ... E E W Cornus suecica, L • . . Sw ,, sanguinea, L E E ... E W Sambucus nigra, L E E E S Viburnum Opulus, L E E Galium boreale, L E? „ palustre , L • • • G S ? ,, uliginosum, L G ,, Aparine, L E ? Valeriana officinalis, L E W S Scabiosa succisa, L E Eupatorium cannabinum, L. . . . ... E E E S s 248 Reid. — Further Contributions to the Bidens cernua , L ,, tripartita , L Chrysanthemum segetum, L. MatricaHa inodor a, L. ... Tanacetum vulgare , L. . . . Tussilago Farfara , L Senecio sylvaticus, L Carduus crispus, L ,, lanceolatus , L. . . . „ palustris, L Centaurea Cyanus, L Lapsana communis , L. . . . Picris hieracioides , L Crepis virens , L Leontodon autumnalis, L. Taraxacum officinale, Web. . . Sonchus arvensis , L Vaccinium Oxycoccos, L. . . . „ Viiis-fdaea, L. . uliginosum, ,, Myrtillus, L. . . . Arctostaphylos alpina , Spreng. . „ Uva-ursi , Spreng. Andromeda Polifolia, L. . . . Loiseleuria procumbens , Desv. . Fraxinus excelsior , L Menyanthes trifoliata , L. . . . Myosotis sylvatica , Hoffm. . Solanum Dulcamara , L. . . . Bartsia Odontites , Huds. . . . Pedicularis palustris, L. . . . Mentha aquatica, L Lycopus europaeus, L Thymus Serpyllum, L. . . . Prunella vulgaris , L Stachys palustris, L „ sylvatica , I, Galeopsis Tetrahit , L Ajuga reptans , L Littorella lacustris, L. . . . A trip lex patula , L Polygonum aviculare, L. . . . ,, Hydropiper, L. . . ,, Persicaria, L. . . ,, lapathifolium, L. „ amphibium , L. . . ,, viviparum, L. . . Oxyria digyna, Hill Rumex conglomerate , Murr. „ maritime , L „ obtusifolius , L. . . . G Sw G Sw E S S E S S Sw Sw Sw Sw E W S M Sw Sw S E S S Sw E E S M S S E G Sw S S s s Sw S S S I E S S E S S ? Sw E S M Neolithic. Geological History of the British Flora . 249 Preglac’ ’ Early Glacial. Interglacial. . Late Glacial. Neolithic. Rumex crispus , L E E E S E W S „ Hydrolapathum , Huds, . . • • . • • • Sw ,, Acetosella , L E E Hippophae rhamnoides , L ... ... ... Sw Viscum album L ... ... Sw Euphorbia Helioscopia , L • • • S ,, amygdaloides, L. . . E Mercurialis perennis , L ... E E S Ulmus montana ?, Sm E ... E E Urtica dioica, L • . . E Myrica Gale , L . . . G Sw Betula alba , L E ... E S E W S I „ nana, L E E S E S Alnus glutinosa, L E E E S E S E S Carpinus Betulus, L E E Corylus Avellana , L E E S • . . E W S Quercus Robur , L E E . • . E W S Castanea saliva , Mill ... E ? Fagus sylvatica , L. ...... . G E Salix pentandra, L ... ... G ? Sw ,, cinerea , L E G E Sw „ aurita , L G Sw „ Caprea , L . G E W „ phylicifolia , L Sw Sw ,, nigricans , Sm Sw' ,, repens , L G S E S „ lanata , L Sw Arbuscula , L G ’?” Sw ,, Myrsinites, L ... E „ herbacea , L . S E S M ,, polaris, Wahlb • . . E E S ,, reticulata , L S Populus canescens , Sm E ? ,, tremula , L G W Empetrum nigrum , L S S M Ceratophyllum demersum , L. . . . E E E E E Juniperus communis , L G Sw Sw Taxus baccata , L E E E S Picea excelsa, Link E G Sw Pinus sylvestris, L E E E S I Stratiotes aloides , L E E Iris Pseudacorus, L E S Sparganium ramosum , Curtis . . . E E E S E W S ,, simplex , Huds. . . . G „ minimum , Fr • . . G Alisma Plant ago, L E E E S Sagittaria sagittifolia , L. .... Sw Scheuchzeria palustris , L Sw Sw Potamogeton natans, L E <• . W ,, rufescens , Schrad . . . S E „ heterophyllus , Schreb E E S w s s 2 250 Reid. — Geological History of the British Flora. IC Potamogeton lucens , L. . . „ praelongus, Wulf. „ perfoliatus, L. ,, crispus, L. . „ obtusifolius , M. „ pusillus , L. . ,, trichoides , Cham. „ pectinatus , L. ,, filiformis , Nolte. Ruppia maritima , L. . . Zannichellia palustris, L. Zostera marina , L. . . Najas Jlexilis , Rostkov. . „ marina , L. . . . Eleocharis acicularis , Sm. ,, palustris , Br. . Scirpus pauciflorus , Lightf. ,, caespitosus , L. . . ,, jluitans , L. . . ,, seta ecus, L.- . . „ lacustris , L. . maritimus , L. . . „ sylvaticus , L. . . Blysmus rufus, Wahlb. . Eriophorum vaginatum , L. „ angustifolium , Roth. Cladium Mariscus , Br. Carex dioica , L. . . ,, echinata , Murr. ,, remota, L. . . ,, alpina , Sw. . „ canescens , L. ,, panicea, L. . . „ distans, L. . . „ jlava , L. . . . ,, Jiliformis , L. ,, Pseudo-cyperus , L. „ paludosa , Good. ,, riparia, Curtis . ,, rostrata , Stokes vesicaria , L. . . Phragmites communis , Trin Pteris aquilina , L. . . . Athyrium Filix-foemina, Roth Scolopendrium vulgare , Sm Lastraea Thelypteris , Presl Osmunda regalis, L. . Equisetum palustre , L. ,, limosum , Sm „ hyemale, L. Isoetcs lacustris , L. . . Preglacial. Early Glacial. Interglacial. Late Glacial. Neolithic. E S G Sw I E E S E E E S I M E E E S E E E E E S ... Sw Sw E Sw E E E S E G Sw ? Sw E E W E E E E S S I M E E E S S E E E E S ... E E S E E E E S E S ... E W Sw E E G Sw E E Sw G ... Sw E S S S G ... S E M S S E S E S Sw G Sw ^ E E E Sw E S s ? Sw E E E W Sw S w E ? G Sw E ... E ... G G Sw E ? E E S S S NOTES. ON APOGAMY AND THE DEVELOPMENT OP SPOR- ANGIA UPON PERN-PROTHALLI. By William H. Lang, M.B., B.Sc.1 — The two most important deviations from the normal life-history of Ferns, apogamy and apospory, are of interest in them- selves, but acquire a more general importance from the possibility that their study may throw light on the nature of alternation of generations in archegoniate plants. They have been considered from this point of view by Pringsheim, and by those who, following him, regard the two generations as homologous with one another in the sense that the sporophyte arose by the gradual modification of individuals originally resembling the sexual plant. Celakovsky and Bower, on the other hand, maintain the view that the sporophyte, as an inter- polated stage in the life-history arising by elaboration of the zygote, is not the homologue of the gametophyte, and is only represented in a few Thallophytes. In the light of the theory of antithetic alternation no weight is attached to apogamy and apospory for phylogenetic purposes. In the paper of which this is an abstract, the results obtained by cultivating the prothalli of a number of species of Ferns under conditions slightly different from the natural ones are described, and their bearing on the problem of the nature of alternation considered. The behaviour of Scolopendrium vulgar e, Sm., and Nephr odium dila- tum , Desv., in which sporangia were borne upon the prothallus, has already been described in a preliminary statement2. It is therefore sufficient to express the results of prolonged cultivation of these and the remaining species in a tabular form. 1 Abstract of a paper read before the Royal Society, March 3, 1898. a Roy. Soc. Proc., Vol. lx, p. 250. [Annals of Botany, Vol. XII. No. XLV I. June, 1898.] 252 Notes. Table of the Results of cultivating Prothalli for a period of Two Years and a-half. [Note. — In every species normal embryos were produced when conditions permitted fertilization.] Names. Scolopendrium vulgare , Sm., var. ramulosissimum. var. marginale. Nephrodium dilatatum, Desv., var. cristatum gracile. Nephrodium Oreopteris , Desv. var. coronans. Aspidium aculeatum , Sw., var. multifidum. Results. Gametophytic budding. Development of archegonial projections. Development of cylindrical process usually from the apical region of the prothallus. ,Tracheides in cylindrical process. Leaves, roots, and ramenta on process. Apogamy. J Sporangia on the process. Vegetative buds from tip of cylindrical process, or in place ^ of an archegonial projection. Similar to var. ramulosissimum , but no spor- angia, isolated ramenta, or leaves found. Gametophytic budding. Development of archegonial projections. Development of cylindrical process, usually from the under-surface just behind the apex, which formed a ‘ middle lobe.’ ITracheides in middle lobe and cylindrical process. Sporangia, sometimes associated with ramenta, on middle lobe and process. No vegetative buds. Gametophytic budding. Development of archegonial projections. Development of cylindrical process from apex of prothallus. STracheides in cylindrical process. Ramenta on cylindrical process. Vegetative buds (rare). Gametophytic budding. Development of archegonial projections. i Tracheides in prothallus. pogamy. j yegetatjve bU(js (rarey Notes. 253 Names. Aspidium angular e, Willd., var. folio sum multifidum. var. acutifolium multifidum. Results. Gametophytic budding. Development of archegonial projections. ( Ramenta on prothallus. pogamy. j yegetatjve buds (frequent). Gametophytic budding. Development of archegonial projections. No apogamy seen. Athyrium niponicum , Mett., nor- mal form. var cristatum. Gametophytic budding. Development of archegonial projections. Tracheides in prothalloid growths from archegonial projections. Similar to the normal form, but in addition a few apogamously-produced vegetative buds. Apogamy. Athyrium Filixfoemina , Bernh. var. percristatum. var. cruciatum cristatum. var. coronatum. Gametophytic budding. Development of archegonial projections. Development of cylindrical process from apex or from under-surface of the prothallus. ! Tracheides in process. Continuation of process as a leaf. Vegetative buds. Polypodium vulgar ey L., var. grandiceps. Gametophytic budding. I Isolated leaf-like growths, pogamy. j yegetatjve buc]s (numerous). Aspidiumfrondosum, Lowe(from Apogamy. Vegetative buds produced on short the Pits, Royal Gardens, cylindrical processes before the culture had Kew). been watered. After the culture was watered, normal em- bryos. In addition to the species mentioned in the Table above, cultures were made of crested and uncrested forms of Nephrodium Filix-masy Rich., representing the three sub-species, which are sometimes distinguished in this country. Some of these (both crested and normal) behaved in a similar manner to the species referred to in the Table, though only one instance of apogamy induced by long cultivation has as yet been found. Others (crested and normal forms) produced a single bud on the under-side of the prothallus which did not bear archegonia. Connecting this latter type of apogamy, which agrees with the description of De Bary and Kny, with the more normal prothalli, was one variety, the archegonia of which developed into typical arche- 254 Notes. gonial projections. In the place of the projection nearest to the apex a vegetative bud arose. It is possible to draw some general conclusions from this series of cultures. It is a striking fact that in every one of the species, prothalli, which under normal conditions would have produced normal embryos, became, after a longer or shorter period, apo- gamous. Further, there was a general similarity in the changes of form and structure of the prothallus, which preceded this result. This form of apogamy, occurring after prolonged cultivation of normal prothalli under special conditions, may be distinguished as induced apogamy , in contradistinction to direct apogamy , by which is meant the immediate production of vegetative buds by prothalli, which are usually incapable of being fertilized. Both forms occur in Nephrodium Filix-mas. The causes which appeared to induce apogamy in these prothalli were the prevention of contact with fluid water which rendered fertilization impossible, and the exposure to direct sunlight. Possibly the temperature also had some effect. The case of Nephrodium Filix-mas shows that the variable condition of the sporophyte, as indicated by cresting, &c., though possibly predisposing to the changes which lead to apogamy, does not stand in any necessary connexion with the phenomenon. That different degrees of apogamy are distinguishable was also shown by these cultures. The cylindrical process, arising from the apex of the prothallus, or from its under-surface, is to be regarded simply as a modification in form and structure of the gametophyte dependent on the altered conditions, and possibly a direct adapta- tion to these. The next stage is seen in cylindrical processes, which, while bearing sexual organs, also produce isolated members of a sporophyte (roots, ramenta, sporangia). It is to be borne in mind, however, that tissue differing from the rest of the process always occurred beneath the last-named structures. The final stage is the production of a vegetative bud capable of further growth as a typical sporophyte. In this a series leading from the bud arising by transformation of the tip of a cylindrical process, to buds pro- duced on or in the place of archegonial projections, and from this to buds situated on the under-surface of the prothallus itself, can be recognized. The readiness with which the intermediate form between gameto- Notes, 255 phyte and sporophyte and the early stages of vegetative buds re-assume the prothalloid form is worthy of note, as bearing on some cases of apospory. These departures from the normal development of the prothallus are not regarded as reversions in the ordinary sense, but as indica- tions of the capability of direct response to altered conditions possessed by the gametophyte. Their possible importance in relation to the theory of homologous alternation appears to the writer to be of this nature. If that theory be true, the sporophyte and gameto- phyteare modifications of a similar form. The gametophyte, especially the simple free living prothallus of the Ferns, has departed less widely from that form. Such an organism as a fern-prothallus would therefore appear to be suitable for experimental work, in the hope that its behaviour under altered conditions would afford hints as to the sort of changes which, in the original algal form, led to the evolution of the sporophyte. The altered conditions in this series of experiments are of a similar kind to those which are assumed by Professor Bower to have occurred on the spread of algal forms to the land, and to have conduced to antithetic alternation. The results may now be used in picturing the manner in which alternation of generations might have come about by the modifica- tion of originally similar individuals into gametophyte and sporophyte. It is assumed for this purpose that the sporophyte of the Vascular Cryptogams did not arise by the elaboration of a structure resembling a bryophytic sporogonium. It is recognized that the theory of antithetic alternation, as elaborated by Professor Bower, affords a consistent and satisfactory explanation, if the assumptions necessitated by the theory are granted. The present theory, which is put forward merely as a provisional hypothesis, is founded on another class of facts. With the spread of algal organisms to the land, where in the absence of any vegetation affording shade some at least would be exposed to more intense illumination, the flattened form would probably be assumed. Prolonged drought and the influence of direct sunlight, inducing directly a change of form into a cylindrical body, might be accompanied by the substitution of a reproductive organ forming dry reproductive cells (spores) for those adapted to an aquatic existence. The acquisition of more highly developed absorbent organs (primitive roots) would further the existence and 256 Notes. growth of this modified gametophyte. This spore-producing stage would at first follow the sexual stage in any individual exposed to dry conditions. It is possible to imagine, however, how the asso- ciation of the asexual with the sexual individual might come about. Absence of fluid water would prevent the liberation of motile spores from the zygote. The latter would be obliged to germinate in situ , and the fact that it did so under dry conditions would tend to the shortening of the sexual stage, and the speedy assumption of the sporophytic form and mode of reproduction. From the spore, which would always separate from the parent, a sexual individual would arise, since germination could only take place in a damp spot. As soon as, with the increase in size and complexity of the spore-bearing plant, a vegetation capable of affording shade came into existence, the conditions suitable for the persistence of the more primitive, alga-like, sexual stage in the life-history would be present. The latter has, of course, also been modified in various ways. In the concluding portion of this paper, the theories of antithetic and homologous alternation are compared by considering the ex- planations they afford of the facts. The general conclusion reached is that, while both afford a possible explanation of the facts of alterna- tion in archegoniate plants, any evidence which would render one or the other untenable is wanting. The reasons on which either is considered more probable depend on the views held as to the lines of descent which have been followed, and the degree to which the different groups of archegoniate plants have had a common origin, or represent actual steps in the process of evolution of the sporo- phyte. Under these circumstances the question must be regarded as an open one until the available lines of evidence have been more fully investigated. I am especially indebted to Dr. Scott and Professor Bower for their assistance and advice ; the work was commenced in the Jodrell Laboratory of the Royal Gardens, Kew, and subsequently carried on in the Botanical Laboratory of the University of Glasgow. THE LIGULE IN LEPIDOSTROBUS.— The presence of a ligule on the vegetative leaves of Lepidodendron is now a well- ascertained fact, and it can be readily seen in several of the slides preserved in the Williamson Collection at the Natural History Museum. The best description and figures of the ligule in the Notes . 257 vegetative region are those given by M. Hovelacque in his admirable memoir on Lepidodendron selaginoides 1, in which species the ligule is shown on the upper surface of the leaf-cushion, enclosed in a deep ligular pit penetrating far into the tissue of the leaf-base. This ligular chamber opens to the exterior only by a small pore situated just above the upper angle of the scar from which the leaf-blade separated ; the small pore probably corresponds with the mark first described by Stur2from the casts, and considered by him to be the ligular pit. The ligule itself is described by Hovelacque 3 as having the form of a triangular pyramid with blunt angles and point ; its insertion at the bottom of the chamber is very oblique, while it is so short as to be generally invisible from the exterior. Hitherto, however, the ligule in the strobilus ( Lepidostrobus ) has not been described, the only reference which I have been able to find in the literature being a not very clear statement by Solms-Laubach, who writes, referring to some impressions found in the Gegenort mine at Dutweiler : ‘ I observe on several of the leaves, close to where the line of fracture passes through their bases, a small obtusely triangular scar with a trace-point in its centre, which from its median position is probably the object discovered by Stur on the barren cushion, and called by him the ligular pit. Its occurrence in Lepidostrobi was till now unknown V Recently, in an examination of the Williamson Lepidostrobi under- taken under the direction of Dr. Scott, F.R.S., I have found the ligule sufficiently well preserved to enable one to make out not only its position but also some details of its structure. It can readily be seen in several of the Williamson slides (C. N. 568, 574, 1776 A, 1776 B, 1776 C), and I have also seen it in a slide (S. 615) kindly lent by Dr. Scott. In the accompanying figure (from C. N. 1776 C), the expanded, peltate, distal extremity of the horizontal portion of the sporophyll is shown at A , with part of the vascular bundle at B, and many of the short tracheides (barred cells) so characteristic of this region of 1 Recherches sur le Lepidodendron selaginoides , Stern : Memoires de la Societe Linneenne de Normandie, Caen, 1892. 2 Die Culmflora der Ostrauer und Waldenburger Schichten : Abh. d. k. k. Geol. Reichsanstalt zu Wien, Vol. viii, Heft ii (1877), Taf. xix, Fig. i. 1. 3 Loc. cit., p. 155. * Fossil Botany, English edition, pp. 234, 235. 258 Notes. the sporophyll at C. The ligule D is shown arising freely from the upper surface of the sporophyll just where the latter bends upwards to form one of the covering scales of the cone. At E the distal ends of two sporangia are shown, each filled with spores ; while at F is seen a ‘pad’ of tissue forming a portion of the attachment of the upper sporangium to the sporophyll, the dark bodies having probably been ‘ mucilage-cells/ Woodcut i. — Nearly radial section of the periphery of the cone of Lepido- strobus ( x 32 : semi-diagrammatic). A peltate expanded portion of sporophyll, with B its vascular bundle; C short tracheides ; D the ligule; E sporangia and spores; F pad supporting sporangium. In another slide (C. N. 568) the ligule is seen in transverse section, showing a triangular form with one angle directed inwards, and very similar to sections figured in Hovelacque \ The length of the ligule shown in the figure is about -5 mm., and the tissue consists mainly of a very small-celled parenchyma, the cells of which exhibit dense contents, making it probable that they were of the nature of mucilage-cells. Near the base is to be seen a transverse row of four or five larger clear cells possibly representing the glosso- 1 Loc. cit. ; cp. Fig. on p. 95, and T. xvii, PI. vii, Fig. 2. Notes . 259 podium described by Harvey Gibson in Selaginella \ These cells are well differentiated from the other cells of the ligule by their relatively large size and absence of contents, and between these clear cells and the ordinary tissue of the sporophyll are some large dense-looking cells, showing what appear to be conspicuous nuclei, which may be compared with the similarly situated dense cells shown by Harvey Gibson in Selaginella spinosa1 2. The base of the ligule is at a considerable distance from the leaf-trace bundle, and I can see no trace of tracheides or barred cells either in the ligule itself or in the tissue intervening between it and the leaf-trace bundle. There is no evidence of any special arrangement of the vascular tissues in relation to the ligule such as is described by Harvey Gibson in Selaginella helvetica and other forms, but there is seen a distinct convergence of the cell-rows towards the point of insertion as shown in the drawing. To sum up then, the position of the ligule in Lepidostrobus , with the sporangium between it and the axis, is identical with that in Sela- ginella ; but, whereas in the latter genus it is quite close to the axis of the cone, in the former the great elongation of the sporangium which had taken place in the radial direction had of course carried the ligule with it, and so the latter comes to be situated near the periphery of the cone and at a considerable distance (1.5 centim.) from the axis. The whole of the horizontal (sporangium-bearing) portion of the sporophyll thus appears to be homologous with the short leaf-base or cushion on the vegetative stem. The ligule differs markedly from those exhibited on the vegetative leaves of L. selaginoides , &c., in the complete absence of a ligular chamber ; although, as can be seen in the drawing, it is overarched by the protuberant distal end of the sporangium. In conclusion, I have to express my indebtedness to Dr. Scott, F.R.S., for kind help. ARTHUR J. MASLEN. Royal College of Science, S.W. 1 Contributions towards a Knowledge of the Anatomy of the Genus Selaginella , Spr. ; Part II, The Ligule : Annals of Botany, Vol. x. 2 Loc. cit., Figs. 3 and 6. The Fertilization of Onoclea1 2. BY WALTER R. SHAW. With Plate XIX. THE details of the process by which the sexual cells and their nuclei unite have not been exhaustively studied in very many plants. Most of the contributions to the subject are the results of investigations into the development during extended periods of the plants under observation, so that necessarily the particular subject of cell-conjugation and nuclear fusion has received proportionally less time and attention. This is more especially true with regard to the Characeae, the Bryophytes, and the Pteridophytes. Literature. The accounts of fertilization in plants all indicate that the process consists in the fusion of two nuclei in the resting condition. This has been described by Wager 2 for Cystopus , by Harper 3 for Sphaerotheca and ErysipJie , by Oltmanns 4 for Vaucheria , by Klebahn 5 for Oedogonium , and by Farmer and 1 Prepared under the direction of Professor Douglas H. Campbell in the Botanical Laboratory of Leland Stanford Junior University. 2 Wager ’96, p. 331. 3 Harper ’96, pp. 656 and 659. 4 Oltmanns ’95, p. 401. 5 Klebann ’92, p. 252. [Annals of Botany, Vol. XII. No. XLVII. September, 1898.] T 262 Shaw. — The Fertilization of Onoclea. Williams1 and by Strasburger2 for Fucus. The approaching sexual nuclei and the resulting spore-nucleus have been described by Klebahn 3 * for Closterium and Cosmarium and for Rhopalodia^ , and by Fairchild 5 for Basidiobolus. In Voucher ia the sperm-nucleus becomes nearly like the egg-nucleus before the two come in contact. In Oedogonium the chromatin granules of the sperm-nucleus are larger than those of the egg-nucleus at the time when fusion begins. In Fucus the sperm-nucleus becomes closely appressed to the egg-nucleus before the granular structure of the sperm-chromatin appears. From Strasburger ’s 6 work on the Gymnosperms it is evident that the sexual nuclei in these plants are alike and nearly equal at the time of fusion, and in this respect the sexual nuclei of Lilium Martagon are similar, according to the account of Guignard7. Mottier8 has recently found that, contrary to the description by Guignard of L. Martagon , the constituents of the male and female nuclei of Lilium candidum cannot be distinguished in the nucleus formed by the fusion. The general character of the sexual organs and the sexual cells of the Ferns is too well known to need reviewing here. The way in which the sexual organs open and the multiciliate spermatozoids make their way from the antheridium to the archegonium is described in many text-books9. It is with the development and structure of the spermatozoid with reference to its constituents and their distribution and activities that investigation has lately been mainly concerned. Guignard 10 had in 1889 concluded that the body of the spermatozoid is formed from cytoplasmic as well as nuclear substance: and, according to Strasburger11, Belajeff had in the same year advanced so far as to describe the spermatozoid 1 Farmer and Williams ’96, p. 482. 2 Strasburger ’97, p. 351. 3 Klebahn ’91, p. 440. i Klebahn ’96, p. 639. 5 Fairchild ’97, p. 292. 6 Strasburger ’78, pp. 50-51 ; see also Vines ’95, p. 472. 7 Guignard ’91, p. 198 ; see also Wilson ’96, p. 161. 8 Mottier ’97, p. 149. 9 See Strasburger and Hillhouse ’89, pp. 292-295 ; also Atkinson ’94, p. *0. 10 Guignard ’89, p. 379. 11 Strasburger ’92, p. 105. Shaw.— The Fertilization of Onoclea . 263 as consisting of a chromatic body formed from the nucleus of the mother-cell, and an achromatic band formed from the cytoplasm of the mother-cell. Strasburger himself described in detail the development of the spermatozoid of Osmunda 1} and the mature spermatozoid of Phegopteris Giesbrechtii 2 as it appeared when stained with fuchsin-iodine green. He found the two anterior coils to stain red, and these he considered to be of cytoplasmic origin ; while the posterior coil took the blue colour characteristic of the nucleus. The cilia he found distributed along the forward coil some distance from the end, and the middle coil, he suggested, might have contained the centrosome, if there had been one. At present 3 he is inclined to think that there is no individualized centro- some in the spermatozoid of Pteridophytes. He found by treating the spermatozoid of Marsilia 4 with the same stain and allowing it to fade, that only the hinder and larger of the ten or twelve coils contain the nucleus, which extends forward to the region where the cilia are attached. In following the development of the spermatozoid of Onoclea , Campbell 5 was unable to find that so large a part of the forward end is of cytoplasmic origin as Strasburger0 described for Osmunda and Phegopteris. ‘ The body of the free spermatozoid,’ the former writes, ‘ has the form of a flattened band with thickened edges, which tapers to a fine point at the anterior end, but is broader and blunter behind V This description fits the living spermatozoid as it is to be seen when caught in the slime about the open archegonia. The principal accounts of the spermatozoid after its entrance into the egg are those of Campbell. In studying the develop- ment of Pilularia 8 he found the spermatozoid within the egg close to the egg-nucleus. It had become a spherical granular nucleus darker than the egg-nucleus and with little more than 1 Strasburger ’92, p. 114. 2 1. c., p. 116. 3 Id. ’97, p. 420. 4 Id. ’92, p. 122. 5 Campbell ’95, p. 312. 6 Strasburger ’92, pp. 115-116. 7 Campbell ’95, p. 313. 8 Id. ’88, p. 249 ; ’95, p. 405 ; for fig. cf. also Wilson ’96, p. 106. T 2 264 Shaw.— The Fertilization of Onoclea. half the diameter of the latter. In following the development of Osmunda 1 he found that the spermatozoid loses its pyramidally coiled form after it comes in contact with the egg-nucleus and before the two fuse. In his work with Marattia 2 he found the spirally coiled spermatozoid in contact with the egg-nucleus. In a later stage he found two nuclei present which seemed to be the male and female nuclei. They were not very different in size or appearance. He concluded that the female nucleus had become smaller, and that the spermatozoid had changed into a similar nucleus. ‘ The two nuclei,’ he writes, ‘ then gradually fuse, but all the different stages could not be traced. Before the first division takes place, however, but one nucleus can be seen, and this nucleus resembles the nucleus of the unfertilized egg.’ From these accounts it has been concluded that the spermatozoid of the Ferns enters the egg in the condition in which it swims free, and then, on coming near to or in contact with the egg-nucleus, it undergoes a change in structure which makes it more like the latter and also like the resting nucleus in the sperm-cell of the antheridium. In this respect the details of fertilization are much alike in many plants, and the behaviour of the animal spermatozoon is strikingly similar. This is shown by Van Beneden’s account of fertilization of Ascaris 3, in which the spermatozoon, after entering the egg, rapidly changes and forms a typical nucleus exactly similar to the egg-nucleus. Preliminary Investigation. The differences in size, structure, and habits between the male and the female gametes are greatest in those plants in which the condition of the gametes during fertilization has been least studied : viz. the Archegoniatae and the Characeae. In the fall of 1895 the writer made some microtome-sections 1 Campbell ’92, p. 70 ; ’95, p. 348. 2 Id. ’94, p. 9 ; ’95, p. 261. 3 Wilson ’96, p. 133 (refers to Van Beneden 1883-87-88). 265 Shaw. — The Fertilization of Onoclea. of the oogonium of Chara with a view to finding the sperm a- tozoid on its way to the egg-nucleus. In the species which were studied, no sign could be found to indicate at what stage the spermatozoid entered the oogonium, and no spermatozoids were found in the sections. At that time Dr. D. H. Campbell suggested that some of the Ferns would be better subjects for following the stages which were sought, and he kindly supplied spores of Marsilia vestita and also another species of Marsilia ( M. . Drummondii ?) from Australia and Onoclea sensibilis. When the spores of Marsilia are sown in water, the prothallia become fully developed in about fifteen hours 1 at ordinary temperatures ; at the end of that time the spermatozoids are set free, and make their way to the female prothallium and through the large mucilaginous funnel to the archegonium, where they sometimes arrive before that organ is open. The first division of the egg takes place within about an hour 2 after the entrance of the spermatozoid into the archegonium. It was found difficult to mark the exact time at. which the spermatozoid enters. While en- deavouring to become sufficiently familiar with the plant to overcome this difficulty the writer found that in some cases the oospores of the Australian Marsilia developed into embryos without any evidence of the presence of sperma- tozoids about the archegonia. The spores had been sown about ten o’clock at night, and were first examined about half- past eight the next morning. No spermatozoids were seen during that morning, and there were no remains of sperma- tozoids about the mucilaginous funnel in front of the arche- gonium. The prothallia were kept, however, and after a few days they were found to bear gmbryos of considerable size. This suggested the idea that the embryos might have been developed without fertilization, and a short series of ex- periments 3 was made with isolated macrospores for the purpose of testing the matter, and indicated that such was 1 Campbell ’95, p. 405. 2 1. c., p. 407. 3 Shaw ’97, p. 1 14. 266 Shaw — The Fertilization of Onoclea . the case. In the light of this result the writer’s experience with Chara affords ground for questioning whether partheno- genesis in that genus is confined to the dioecious species, C. crinita. The spores of Onoclea sensibilis had been brought by Dr. Campbell from Massachusetts. They were sown in the laboratory, some in September, 1895, and others in January, 1896. When the earlier culture was examined in January, some of the female prothallia were found to have formed mature archegonia, but when they were placed in water most of the archegonia were slow to open. Spermatozoids were set free in a few minutes after male prothallia were placed in water. On February 26 spermatozoids crowded into and about the open archegonia within eight minutes after the prothallia were placed in distilled water. In Marsilia the first division of the embryo takes place within an hour after fertilization, and in Pilularia after about three hours1. Dr. Campbell, by his extensive studies of Fern embryology, was led to believe that the first division of the egg of Onoclea occurred about twenty-four hours after fertilization. It was desired to obtain a series of prothallia which would represent as completely as possible the period intervening between the entrance of the spermatozoid into the archegonium and the first division of the egg, for it was the intention at the outset to describe the nuclear and cytoplasmic structures and changes within the egg during that period. On February 2 6 and the two following days, prothallia were left on the surface of distilled water for a variety of periods of from about ten minutes to fourteen hours and then fixed. On March 4 a similar series was made, which extended the period to twenty-four hours. From the cultures sown in the middle of January, prothallia were used on March 23 to extend the period to thirty hours, and on April 23 to seventy-two hours. Prothallia from each series were sec- tioned, but in none of them were any embryos or signs of 1 Campbell ’95, p. 407. Shaw. — The Fertilization of Onoclea . 267 division of the egg found. Finally, a double series was begun on May 1 1 by flooding some of the prothallia in the saucers in which they had grown, and by transferring others from the soil to the surface of distilled water, as had been done with the specimens of all the previous series. Those left on the soil were flooded for a time sufficient for the impreg- nation of all the mature eggs and then drained. Specimens from the soil were fixed at the end of every twenty-four hours for fifteen days, and from the water at the same times for the first seven days. It was thought, naturally enough, that seven days would be more than enough for the develop- ment of embryos, and it was intended that this double series should show whether the growth of the embryos was pre- vented or retarded by removing the prothallia from the surface of the soil and leaving them on the surface of distilled water. The material was preserved in alcohol, and when sections were made, in the following September, no embryos were found in the prothallia fixed at the end of seven days, and the writer made the mistake of not sectioning a large number of those from the fifteen-day period. Dr. Campbell then obtained from Michigan spores of O. Struthiopteris , and several cultures were started, some in the saucers used in the preceding season, some in a box on soil already in the laboratory, and some in a similar box on a fresh supply of black soil, a mixture of leaf mould and adobe. The plants in the saucers were not well lighted, and for this and other reasons they did not grow so rapidly or uniformly as those in the boxes. Each box was covered with a pane of glass. The spores on the old soil were sown on November 3, and the box was inclined at an angle of 30° before a deep south window about 60 cm. distant. Those on the fresh soil were sown on November 17, and placed in the same way before an east window. The tilted position of the boxes secured a better illumination, and kept the surface of the soil well drained. It also prevented the water which condensed on the inside of the glass cover at night from dripping on the mature prothallia and thus fertilizing 268 Shaw . — The Fertilization of Onoclea. the ripe archegonia. Many of the plants of the culture of November 3 were ready for fertilization in less than two months. The soil of this culture was poorly mixed, so that the plants on different parts of its surface did not develop uniformly, but they were quite capable of producing embryos. The first series (No. I) of O. Struthiopteris was begun December 29, 1896. Three lots were fertilized in as many ways : lot 1 included prothallia which were removed from the soil, fertilized under the microscope, and replaced on moist soil in watch-glasses ; in lot 2 were plants which were removed from the box with soil attached to their root-hairs, and fertilized by flooding them in watch-glasses, where they were drained and left ; lot 3 was fertilized by flooding the prothallia on the soil. Others were treated like the first lot, except that they were not passed under the microscope. Ten specimens from lot 2 were placed in water under the microscope eight days afterwards, and archegonia opened on only two of them. This indicated that the formation of mature archegonia had ceased on most of the prothallia of this lot, probably because the fertilization was effective. Plants from different lots were fixed on each successive day for ten days. Specimens representing the whole series were sectioned, but no embryos were found. On February 15, 1897, the culture of November 17, 1896, wras well advanced. The prothallia were the greenest and most regularly formed of all that had been raised. The box was placed in a horizontal position, and filled for about fifteen minutes with water enough to flood half of the culture. The box was then drained and returned to its original position so as to minimize the geotropic disturbance. A series (No. Ill) was then taken from the box after varying intervals up to nine days and twenty-three and a half hours. When the oldest of these were sectioned, one egg with the nucleus in the anaphase of the first division, and another with the first division complete, were found. In the next younger lot of this series, which was six days and nineteen hours old, no dividing egg or embryo was found. With this clue to Shaw. — The Fertilization of O nocle a. 269 the fact that the first division of the egg takes place nine or ten days after fertilization, more of the older stages from Series I were sectioned and the following young embryos were found : — Embryos of Onoclea Struthiopteris- from Series I. Fertilized 9 days, 3 embryos of 2 cells. ( 2 „ 2 „ r> JO » \ 2 jj 4 v ' 1 „ « » On March 18 both box cultures were flooded (one had been half-flooded before), and from each specimens were fixed daily to form a series, one covering fifteen and the other twenty days. With thermometers, which hung near the cultures, the temperature of the room was noted each day when plants were fixed, and recorded with the maximum and minimum temperatures for the preceding twenty-four hours. It ranged from io° C. to 250 C. The material yielded the following : — Embryos of Onoclea Struthiopteris from Series IV and V. embryos of 4 cells. 1 » 6 33 Fertilized 10 days 1 53 15 33 2 >3 1 6 33 ^1 33 17 33 „ 12 ,, 1 >3 16 33 The results were enough to satisfy the writer that, under favourable conditions in the laboratory, the egg of Onoclea Struthiopteris is not likely to divide sooner than a week after fertilization ; and they indicated that uprooting, handling, and replanting the prothallia did not retard the formation of the embryo. After the embryos of O. Struthiopteris had been found, attention was again turned to the longest series of prothallia which had been killed in 1896, and by sectioning a large number of prothallia the following were obtained : — 270 Shaw . — The Fertilization of Onoclea . Embryos of Onoclea sensibilis from Series G. Fertilized 8 days, 1 embryo of 2 cells. ») 3 { 3 1 This showed clearly that we should not expect to find the first division of the egg in these prothallia less than a week after fertilization, and, conversely, that the absence of embryos in the shorter series which had been prepared does not by itself indicate that they were abnormal. It was unfortunate that the specimens kept on the surface of the water did not cover a period longer than seven days. This failure was repaired as far as possible by fixing such a series from the cultures of the second species studied, but these plants were killed too late to be available for present purposes. It is also regretted that the long series of O. sensibilis was obtained after the prothallia had grown to a large size, and bore such large numbers of archegonia as to make the study of fertilization in these plants a more complex problem ; for at some certain time after reaching maturity the egg, if unfertilized, becomes atrophied to such an extent as to render normal fertilization impossible. Yet spermatozoids enter such eggs, and it has been suggested that their presence may then have some retarding influence on the development of the normally fertilized eggs on the same prothallium. On account of the long period occupied by the stages which it was proposed to follow, these Ferns are less suitable for the research than was expected. Methods. The cultures of Onoclea sensibilis were obtained by sowing spores on sterilized soil in earthenware saucers. Cultures Shaw. — The Fertilization of Onoclea. 271 started in September were ready for fertilization-experiments in the following February, and those started in January were sufficiently mature for the same purpose two months after the spores were sown. A series of trials showed that when female prothallia bearing mature archegonia were placed in water, the archegonia opened within five minutes, and when male prothallia were present the open archegonia were all entered by spermatozoids within three minutes. It was found convenient to carefully transfer large numbers of prothallia from the soil to the surface of water in watch-glasses. A few minutes or an hour later, the prothallia were examined, and those with no open archegonia were rejected. Then, after the lapse of the desired periods, specimens were taken from the watch-glasses and placed in the fixing agent. All of the series used, except Series G, were fertilized in the manner described. The plants of the latter series, as already explained (p. 2 67), were fertilized on the soil and not removed until the time for fixing. Unless otherwise specified the observations refer to Onoclea sensibilis. In the following table of the series of prothallia killed, the italics show the material used : — isj .9 8 Sowing of culture. Series Beginning of Series. Periods between fertilization and fixation. ^ 3 .y 0 6 w J.S w bfl u d 0 w> ffi.S ^ X M3 Sept. ’95 B Feb. 26, ’96 10, 20, 30 minutes, 1, 2, 3, 4, 5,0, 8, 10, 12, 14 hours. 1% 18-24 Sept. ’95 C March 4, ’96 6, 8, 10, 12, 14, 24 hours. 1% 24 Sept. ’95 D March 9, ’96 1, 2, 3, 4, 5,6, 7, 8, 9, 10 hours. 1% 6-8 Jan. 10-17, ’96 E March 23, ’96 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30 hours. 1% 8-24 Jan. 10-17, ’96 F April 23, ’96 1, 2, 3, 4, 5, 6, 8, 10, 12, 14,, 16, 18, 20, 22, 24, 28, 32, 36, 40, 44, 48, 60, 72 hours. 24 Jan. 10-17, ’96 G May 11, ’96 h 2, 3,4> 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 days. 1% 24 2J2 Shaw.— The Fertilization of Onoc/ea. For fixing1 the prothallia they were placed whole in \ °/o or i °/ chromic acid. The best results were obtained by leaving the specimens in \ °/o chromic acid for eight to twenty- four hours. In most cases dilute sulphurous acid was used for removing the last traces of chromic acid. Most of the material was stained with Czokor’s alum- cochineal, in which the prothallia were allowed to remain for eighteen to twenty-four hours. Some of the plants were stained in alum-carmine, and others in diluted Delafield’s haematoxylin. Some of the sections of plants which were stained in toto in alum-cochineal were counter-stained on the slide with Bismark brown. A few sections were restained on the slide with Flemming’s triple stain 2, safranin-gentian violet-orange G, and a few others with Heidenhain’s iron- haematoxylin. The prothallia were sectioned in paraffin. At first turpen- tine was used as a medium between alcohol and paraffin, but it was soon discarded, and xylene used according to the method recommended by Zimmermann 3 in his handbook. The wings were cut off with a razor before the prothallia were imbedded. The sections were cut io/x thick on a Minot microtome of the older style. They were cleared with clove oil and mounted in Canada balsam. Investigation. The spermatozoids, which were usually held in large num- bers in the slime before the mouth of the archegonium, remained unchanged for a long time, and were favourable objects for study. They were in a position to be acted upon very quickly by the fixing agent, and also by the stain. Those fixed after fourteen, eighteen, and twenty-four hours had the same appearance as those fixed after a few minutes. They had all lost the nutritive vesicle, and become a little 1 For fixing the series of 0. Strut hi o^teris most of the agents given in Zimmer- mann’s Botanical Microtechnique were used. 2 Zimmermann ’93, p. 186. 3 1. c., pp. 32-33- 273 Skawr — The Fertilization of Onoclea. drawn out on entering the slime. With all the stains used the darkest part of the body is a corkscrew-shaped homo- geneous rod, oval in cross-section, which tapers gradually forward and abruptly backward. It does not terminate in a sharp point at either end. In all the specimens which were noted closely the spiral turns of the body are in the direction of a left-hand screw. Webber 1 finds that the helicoid ciliated band in the spermatozoid of Zamia is coiled in the same direction. The number of turns varies with the length to which the corkscrew is extended : when shortened, like a watch-spring, it makes two turns (Fig. io b) ; just outside the archegonium it usually makes about three and a half turns (Fig. 5) ; in the canal, when it becomes most extended, it may make five turns (Fig. 1). In all the forms an almost stainless band can be seen attached to the forward edge of the corkscrew, and extending beyond its forward point. This band is broadest toward the forward end, and it tapers backward to the thickest part of the corkscrew rod. The outer edge of the band appears to be thicker than the rest, and the thickening is greatest near the forward end. The dark rod is assumed to be the nucleus, and the wing-like band the cytoplasmic portion of the spermatozoid. The cilia did not usually stain enough to be visible ; but when Heiden- hain’s iron-haematoxylin was left rather dark they could be seen distinctly enough to show that some extended forward and others backward. They could not be traced to their respective points of attachment, and so they were omitted from the drawing (Fig. 8) which was made from such a specimen. No centrosome was seen. The nucleus of the spermatozoid has the same form as that figured by Campbell2 for the developing spermatozoid of O. Struthiopteris . It extends through the greater part of the length of the body. In this respect it differs from the nucleus which Strasburger 3 found in the spermatozoid of Phegopteris . In the latter it occupied only the larger pos- 1 Webber ’97, p. 17. 2 Campbell ’95, p. 131. 3 Strasburger ’92, p. 116. 274 Shaw . — The Fertilization of Onoclea. terior turn of the corkscrew. The spermatozoid of O. sensibilis is like that of Chara , as described by Belajeff1, in having the cytoplasm extending along the nucleus. The thickened outer edge of the cytoplasmic band may correspond to the ‘ Riickfaden ’ which Belajeff was able to distinguish in the forward cytoplasm of the spermatozoid of Chara until a late stage in its development. Long before the archegonium opens the egg-nucleus comes to the resting condition, and contains one or more nucleoli. The ventral canal-cell is the smallest cell in the archegonium. After it is formed, the pressure of the egg makes it still smaller and concave, and its nucleus becomes flattened. In later stages the walls of the ventral canal-cell swell up, and by pressure cause the egg to become concave on the outer side, which later forms the receptive spot. In these stages the egg-nucleus also is flattened and concave. The egg is in this condition when the archegonium opens. In the living sections, under the microscope, the writer observed the egg to swell as soon as the canal was cleared of its dissolving contents, and fill up the venter. If many spermatozoids were near, they swarmed into the canal, and a large number made their way into the venter, where they swarmed about freely, quite differently from those in the close quarters of the neck, which were motionless or moved slowly. Often spermatozoids which had entered the venter found the canal again, and made their way out, slowly through the narrower portion of the canal, but rapidly in the wider part. On first entering the slime discharged by the archegonium, the spermatozoids left their trophoplasmic vesicles behind, and their motion was retarded. It was the resistance of the slime which pulled off the vesicles. In the canal the slimy mucilage seemed to be denser, and when a spermatozoid entered it, the corkscrew spiral became drawn out, and the number of turns increased, and the forward motion of the spermatozoid was accompanied by a rotation which corresponded to the pitch of the screw. 1 Belajeff ’94, p. 43. Shaw. — The Fertilization of Onoclea. 275 All the sections of prothallia that were killed within an hour after the entrance of the spermatozoids into the arche- gonia, show the eggs in a collapsed condition (Figs. 1 and 2), concave on the outer side, and the nucleus in each conforms to the shape of the cytoplasm. In this state the shape of the egg is about the same as it is in the unopened archegonium after the canal-cells have swelled. The return to this form may have been due to the action of the fixing or imbedding agents, the egg in this stage being more susceptible to their shrinking influence, either because it is not at this time in the state of tension which it acquires later, or because the open canal permits the rapid access of the plasmolizing agents. There are reasons to believe, however, that the collapse is not an artificial plasmolysis, but that it takes place as soon as the spermatozoid enters the egg. The mature egg has been described (for the other species) as having a large hyaline receptive spot 1. The concavity of the collapsed egg occupies the position of that spot. That it was formed before the plants were killed seems evident from the movement of a number of spermatozoids in the venter. This can be seen in the living plants. That the number of these spermatozoids is large is shown by the specimens stained and sectioned. They can hardly have been carried into the venter by the fixing agent, for those in the canal were fixed first, in the extended condition, and those in the venter afterward, in the contracted form. From the evidence at hand it appears that as soon as the egg is entered by a spermatozoid it loses its turgidity, and the spermatozoids which come into the venter afterward meet with little or no resistance from the egg. It may be that the turgid condition of the egg, in the first place, offers mechanical facility for the screw-like sperma- tozoid coming through the narrow base of the neck to force itself into the cytoplasm of the receptive spot ; and that the plasmolytic condition of the egg afterward deprives the following spermatozoids of this advantage, and protects Campbell ’95, Fig. 159. 276 Shaw. — The Fertilization of Onoclea . the egg from injury or from multiple fertilization by them. A few careful experiments in fixing archegonia before and after the first spermatozoid enters the egg ought to reveal the truth of the matter. Within half an hour after the entrance of the sperma- tozoids into the archegonium the canal is practically closed by the expansion of the four proximal neck-cells and the four just beyond them. The egg gradually recovers its turgidity and forces the free spermatozoids against the outer wall of the venter (Fig. 3). The membrane which, from analogy with the development of the eggs in other plants, we would expect to form around the egg immediately after the entrance of the spermatozoids, was not seen in any of the earlier stages. If the membrane is of the nature of cellulose it ought to be brought out distinctly by the Bismark-brown with which the subject of Fig. 1 was stained on the slide. This stain colours even the cytoplasm in this case. In all the preparations which were examined there was no evidence that a membrane of appreciable thickness is formed im- mediately after the entrance of the spermatozoid or for some time afterward. A conclusion so contrary to analogy must remain in doubt. It may be that the chromic acid used for too long a time destroyed the membrane. The early history of the spermatozoid inside the egg was not satisfactorily followed. After the ten- and twenty-minute periods the collapsed state of the egg interfered with the study of the enclosed spermatozoid, and the stains used for these stages were not the best. The difficulty was increased by the free spermatozoids crowding into the concavity of the egg. So the mode of entrance of the spermatozoid into the egg cannot now be described. In one case the sperma- tozoid appeared to be still outside the egg-nucleus, against which it lay in an open coil after an hour, but in other cases it was found inside the egg-nucleus within thirty minutes. The nucleus of the spermatozoid undergoes no visible change in structure while in the egg-cytoplasm. Whether the cytoplasmic wing and the cilia of the spermatozoid are taken Shaw . — The Fertilization of Onoclea. 277 into the egg-nucleus could not be shown with the faint cyto- plasmic stains used at these stages. We may reasonably expect to find that they are left outside in the egg-cytoplasm, but it is a question which will require to be settled by actual observation. In Figs. 3 and 11, which were drawn before the structure of the free spermatozoid was understood, there is something in the egg-cytoplasm on the outer side of the egg-nucleus which strongly suggests by its general appearance that it is the remains of the sperm-cytoplasm, accompanied in the case of Fig. 3 by the cilia. The writer has lately observed, in a specimen stained with Heidenhain’s haematoxylin, something that looks like a loose bunch of cilia in the same position ; and in some horizontal sections the outer part of the egg-cytoplasm, when seen from the outer side, shows radiations which have a spiral twist, such as the cilia of the spermatozoid sometimes show when that body is viewed from in front 1. The egg-nucleus at the beginnifig of fertilization, and all through the process, is in the typical resting condition. The nucleoli, for there are generally several, are the most conspicuous structures in the stained sections. They vary in size and present a peculiar porous structure. The linin appears as a delicate network which bears the very small chromatin bodies. This was not demonstrated for all stages, but the general appearance of the female nucleus is the same throughout the process, and in later stages some well-stained examples made it possible to observe the linin network and chromatin bodies minutely (Fig. 12). Nothing was seen to indicate that the whole nuclear c membrane/ if it may be called such, was dissolved. There were indications that it dissolves or is ruptured at the place where the sperm- nucleus enters. The most remarkable fact observed, and one about which there is no doubt, is that the sperm-nucleus enters the egg-nucleus before it changes in form or visible structure. This is clearly shown in the section represented 1 Webber has shown (’97, 2, p. 227) that in the fertilization of Zamia the cilia are left behind soon after the entrance of the spermatozoid into the egg-cell. U 278 Shaw . — The Fertilization of Onoclea . in Fig. 1 and sections of the other eggs on the same slide. In Fig. 1 the small end of the spermatozoid is directed toward the base of the archegonium. The lower portions of two coils were cut off by the razor, and are to be seen in the next section on the slide. The sperm-nucleus within the egg- nucleus becomes granular and the granules slowly separate. Thirty minutes after the entrance of the first spermatozoid into the archegonium the sperm-nucleus may show traces of the granular structure, as in Fig. 2, but usually it is not evident until two hours have elapsed. It shows plainly in Fig. 3, from a specimen fixed after three hours. Fig. 4, after twenty-four hours, shows little, if any, advance in this respect. Fig. 6, after twenty- four hours, shows the sperm-chromatin distributed in one quarter of the egg-nucleus very much as it was found in Fig. 9 after sixty hours. In most of the preparations the chromatin of the two nuclei can be seen distinctly, but in only a few cases was the linin also clearly distinguished. Among the best of these is one represented in Fig. 12, which was fixed in J °/o chromic acid for twelve hours, washed two days in water, one and a half days in dilute sulphurous acid, and three hours in water, and then stained with alum-carmine. The linin threads connecting and supporting the chromatin granules of the female nucleus are especially distinct, and had one the time one might almost construct a complete map of the network system. In the sperm-nucleus the granules and the threads are so closely packed that the courses of the threads cannot be followed in detail. The sperm-nucleus often retains the spiral arrangement of its substance for a long time, as in Fig. 13, after twelve hours. It may lose this arrangement early, as in Fig. 7, after fourteen hours. Here the larger and smaller end can still be identified. In the latest stages in which the sperm-substance could be recognized it had become distributed through a larger part of the female nucleus (Fig. 9). The different eggs fertilized on one prothallium at the same time do not have the sperm-nucleus in the same Shaw. — The Fertilization of O hoc lea. 279 condition. This makes it impossible to say without further study what is the rate of normal 4iuclear fusion. Among the few early stages of O. Struthiopteris that were sec- tioned and successfully stained, one after two days showed the male nucleus but little further advanced than that of Fig. 14, which was killed after thirty-six hours. Another egg of that species contained, after three days, a nucleus which had slightly enlarged but contained nothing that could be identified as male chromatin, although there were outside of this egg the crowded remains of free spermatozoids which must have entered the venter when the canal was open and the egg in a receptive condition. So it appears that after three days the nuclear fusion may be complete. The cytoplasm of the egg becomes vacuolated as the cell becomes turgid, and may after a time be pretty evenly distributed around the lumen (Fig. 11) ; or it may be denser on two sides of the nucleus (Fig. 4). In the material stained with alum-cochineal and Delafield’s haematoxylin, the cyto- plasm appeared to be composed mainly of spherical bodies. But in the beautiful alum-carmine preparations the cytoplasm, although very slightly stained, showed, with the most favour- able illumination, a reticular or alveolar structure resembling that of the nuclear network, but with larger meshes. This was not exhibited distinctly enough to be represented in the drawing. No centrosome or radiations, such as are formed about one, were seen in the egg-cytoplasm. In most of the eggs that are fertilized the protoplasm decreases in quantity after two or three days, and retreats to the inner side of the venter with the nucleus, which becomes smaller. Whether any such as these afterward divide is not known. A large proportion of the fertilized eggs never divide. It might be supposed that the rather peculiar treatment of the prothallia in these experiments was responsible not only for an increase in the proportion of sterile eggs, but also for a wider variation in the rate of nuclear fusion. Opposed to such a supposition is the fact that many of the prothallia left on water showed more U 2 280 Shaw. — The Fertilization of Onoclea. nuclei in various stages of mitosis than those which were not disturbed until they were fixed. In an early stage of the segmentation of the egg, an exception was found to the order of the divisions usually described x. This was in an embryo in which the epibasal octants had already formed, but the hypobasal quadrants wrere separated by a median instead of a transverse wall. On one side of this two octants had formed, and on the other the nucleus was in the metaphase of division. In other cases the quadrant walls were more or less oblique. The absence of radiations in the cytoplasm of the egg during fertilization is a character in which the Fern resembles the lower plants in which the process has been described. Strasburger’s suggestion 2 that their absence in Fucus , where they occur during mitosis, may be due to the fact that the cell division does not immediately follow fertilization, will be equally applicable to this Fern if the radiations can be found during mitosis. We were led to expect from the accounts of fertilization in other plants that the sperm-nucleus would become more or less like the egg-nucleus before the two united. The entrance of the unchanged sperm- nucleus into the egg-nucleus in Onoclea is so notably different from what has been said to occur in the eggs of other Ferns, that a further study of these is very desirable. Marsilia and Pilularia , while presenting some difficulty with regard to regulating and marking the time at which the spermatozoids enter the archegonia, have the great advantage that the nutrition of the eggs is well provided for, and there are not several eggs to contest for the food from one prothallium. It was thought that the disadvantages of working with the Onoclea prothallia could be avoided by selecting young ones on which only one archegonium was ripe, but this was not found practicable when a large number were required. The unsuspected fact that the egg did not divide for more than a week after fertilization was the greatest hindrance to the 1 Campbell ’95, p. 316. 2 Strasburger ’97, p. 419. Shaw, — The Fertilization of Onoclea, 281 completion of this account of fertilization. The quiescent period may be partly or entirely due to imperfect nutrition, and if so it is likely to vary in different closely related species, and in the same species under different conditions. We have found the period to be long in three cultures and two species. Many of the prothallia used for the present study probably bore, in addition to the eggs fertilized at the recorded times, others which matured later and were then fertilized. This might have been prevented by draining the prothallia, as was done with those which were fixed during the next season. This may be accomplished when the prothallia are removed from the soil to be fertilized, as must be done if they are to be carefully examined, by placing them on moist filter paper in a moist chamber. It will be desirable not only to prevent the fertilization of archegonia which mature later, but also any which have matured earlier than the beginning of the experiment. The most trustworthy account of fertilization will be based on eggs on prothallia on which only one egg is fertilized and that at maturity. Such specimens can be obtained by fertilizing large numbers of prothallia and selecting after ten minutes those on which only one archegonium opens and attracts spermatozoids. Summary of Results. j. The body of the free spermatozoid consists of a long corkscrew-shaped nucleus which stains homogeneously, and a lateral band of cytoplasm which extends a short distance in front of the nucleus. 2. The sperm-nucleus enters the egg-nucleus before it changes in form or visible structure. 3. Within the egg-nucleus the chromatin-granules of the' sperm-nucleus slowly separate as the meshes of the linin- network slowly enlarge. 4. Throughout the process of fertilization the female nucleus is in the resting condition. 282 Shazv. — The Fertilization of Onoclea. 5. The first division of the egg was in no case found until more than a week after fertilization. 6. It is suggested that the egg becomes plasmolysed as soon as the first spermatozoid enters it, and that this serves as a provision against injury by following spermatozoids. The work which forms the basis of this paper was done in the Botanical Laboratory of Leland Stanford Junior Uni- versity, with the kind advice and encouragement of Professor Douglas H. Campbell, for which the writer takes pleasure in here expressing his thanks. The paper was written and most of the figures were drawn during the summer of 1897, in the Hopkins Seaside Laboratory, a branch of the University located at ‘Pacific Grove, California. A detailed account of the development and structure of the spermatozoid of the Ferns was given by Belajeff1 before the above paper was written, but it had not then been seen by the writer. In that account Belajeff brought out the fact that there is a specially differentiated body in the cyto- plasmic band of the spermatozoid which gives rise to the cilia. He described the development of this body (the ‘ Nebenkern ’), from a small body near the nucleus of the ‘ spermatozoid mother-cell,’ into a thread-shaped body in the mature sper- matozoid. The writer2 of the present paper subsequently found these ‘ Nebenkerne ’ in the antheridia of Onoclea before and during the last cell-division by which the so-called c sper- matozoid mother-cells ’ are formed. But the cilia-bearing portion of the spermatozoid of the Fern, like that of the Cycad as described by Webber3, takes no active part in what may be regarded as the essential process of fertilization, and therefore no extended reference to these works need be appended to the foregoing account of fertilization. 1 Wl. Belajeff, Three preliminary papers in the Berichte d. deut. Bot. Gesell., *$97, P-337 ff- 2 W. R. Shaw, liber die Blepharoplasten bei Onoclea und Marsilia, 1898. Soon to appear. 3 Webber, ’97, 2, 227. Shaw. — The Fertilization of Onoctea . 283 Literature cited. Atkinson, G. F., ’94 : The Biology of Ferns. Belajeff, Wl., ’94 : Ueber Bau und Entwicklung der -Spermatozoiden der Pflanzen : Flora, Erganzungsband, p. 1 . Campbell, D. H., ’88 : On the Development of Pilularia globufera ; Annals of Botany, ii, p. 233. ’94 : Observations on the Development of Marattia Douglasii , Baker; Annals of Botany, viii, p. 1. ’95 : The Structure and Development of the Mosses and Ferns. Fairchild, D. G., ’97 : Ueber Kerntheilung und Befruchtung bei Basidiobolus ranarum, Eidam ; Jahrb. fur wiss. Bot., xxx, p. 285. Farmer and Williams, ’96 : On fertilization and segmentation of the spore in Fucus ; Annals of Botany, x, p. 479. Guignard, L., ’89 : Sur les antherozoides des Marsiliacees et des Equisetacees ; Bull, de la Societe Bot. de France, xxxvi, p. 378. ■ ’91: Nouvelles etudes sur la fecondation ; Annales des Sciences Naturelles, Botanique, Ser. 7, xiv, p. 163. Harper, R. A., ’96: Ueber das Verhalten der Kerne bei der Fruchtentwickelung einiger Ascomyceten ; Jahrb. fur wiss. Bot., xxix, p. 655. Klebahn, H., ’91 : Studien iiber Zygoten : I. Die Keimung von Closterium und Cosmarium ; Jahrb. fiir wiss. Bot., xxii, p. 415. — ■" ’92 : Studien iiber Zygoten : II. Die Befruchtung von Oedogonium Boscii ; Jahrb. fiir wiss. Bot., xxiv, p. 235. ’96: Beitrag zur Kenntniss der Auxosporenbildung : I. Rhopalodia gibbia (Ehrenb.) O. Muller; Jahrb. fiir wiss. Bot., xxix, p. 595. Mottier, D. M., ’97 : Ueber das Verhalten der Kerne bei der Entwickelung des Embryosacks und die Vorgange bei der Befruchtung; Jahrb. fiir wiss. Bot., xxxi, p. 125. Oltmanns, F.,’95: Ueber die Entwickelung der Sexualorgane bei Vaucheria\ Flora, 1895, p. 388. Shaw, W. R., ’97 : Parthenogenesis in Mar si Ha ; Botanical Gazette, xxiv, p. 1 14. Strasburger, Ed., ’78 : Ueber Befruchtung und Zelltheilung, 1878. » ■ ’92 : Schwartnsporen, Gameten, pflanzliche Spermatozoiden und das Wesen der Befruchtung : Histologische Beitrage, Heft 4, Jena. ’97 : Kerntheilung und Befruchtung : Jahrb. fiir wiss. Bot., xxx, pp. 351 and 406. Vines, S. H., ’95 : A Student’s Text-Book of Botany. Wager, H., ’96 : On the Structure and Reproduction of Cystopus candidus : Annals of Botany, x, p. 295. Webber, H. J„ ’97, 1 : The Development of the Antherozoids of Zamia ; Botanical Gazette, xxiv, p. 16. ’97? 2 : Notes on the fecundation of Zamia and the pollen-tube apparatus of Gingko ; Bot. Gazette, xxiv, p. 225. Wilson, E. B., ’96 : The Cell in Development and Inheritance; London and New York. Zimmermann, A., ’93: Botanical Microtechnique; English translation by Humphrey. 284 Shaw. — The Fertilization of Qnoclea. EXPLANATION OF FIGURES IN PLATE XIX. Illustrating Mr. Shaw’s paper on Onoclea. Fig. 1 was drawn with a Leitz objective No. 7, and a Leitz drawing-ocular ; all the others were drawn with a Zeiss apochromatic 2-0 mm objective, apert. 1*40; Figs. 5 and 12 with a Zeiss compensation ocular No. 12 ; the others with a Zeiss compensation ocular No. 8. For sketching the figures (except Fig. 1) the prism section of a Leitz 450 drawing-ocular was screwed to the Zeiss compensation-ocular in place of the upper diaphragm holder. Magnification determined with a stage micrometer : — Fig. 1 x 500 ; Fig. 5 X 2,000; all others x 1,200. All figures are from sections 10 ju thick. The arrow indicates the direction of the growing point of the prothallium. The time given is that which elapsed between the entrance of the first sperma- tozoids and the fixing of the prothallium. Onoclea sensibilis. Fig. 1. One per cent, chromic acid : alum-cochineal and Bismark brown. Vertical section through an open archegonium probably within ten minutes after the entrance of the first spermatozoid. One unchanged spermatozoid is inside the egg-nucleus. Parts of two coils of this spermatozoid were cut off, and appear in the next section of the series. Fig. 2. One per cent, chromic acid : alum-cochineal. Thirty minutes. Vertical section of the venter of an archegonium containing spermatozoids, and the collapsed egg with a spermatozoid within the nucleus. The canal of this archegonium is almost closed. Two spermatozoids in the section were not drawn. Fig. 3. One per cent, chromic acid: alum-cochineal. Three hours. Nearly vertical section of the egg. The outside spermatozoids are forced against the venter wall by the expanding egg. Portions of the egg-nucleus, one containing a nucleolus, are in each of the adjoining sections of the series. Fig. 4. One per cent, chromic acid : safranin and orange G. Twenty-four hours. Vertical section of an egg in which fertilization is not much in advance of the preceding case. Fig. 5. Half per cent, chromic acid : Delafield’s haematoxylin. Eighteen hours. Spermatozoid caught in the slime outside an archegonium showing the kino- plasmic band extending as a wing along the forward side of the two anterior coils of the corkscrew-like nucleus. Fig. 6. One per cent, chromic acid : safranin, gentian-violet, and orange G. Twenty-four hours. Two vertical sections of an egg. The male chromatin partly distributed in the egg-nucleus. The lightly shaded spermatozoid is in a lower focus. Shaw . — The Fertilization of Onoclea. 285 Fig. 7. Half per cent, chromic acid : Delafield’s haematoxylin. Fourteen hours. Horizontal sections of an egg. The large nucleolus is shaded lightly because it occurs in a lower focus than the adjacent male chromatin. Fig. 8. One per cent, chromic acid : Heidenhain’s haematoxylin. Fourteen hours. Two spermatozoids which were held in the slime outside an open archegonium. The cilia could be* seen, but were not drawn because they could not be traced to their points of attachment with certainty : some were directed back- ward, but most of them forward. One nucleus was broken into two and the other into three pieces by the razor. Fig. 9. One per cent, chromic acid: Heidenhain’s haematoxylin. Sixty hours. Egg-nucleus from an oblique horizontal section. The large nucleolus is shaded lightly because it is behind the male nucleus. Fig. 10. Half per cent, chromic acid : alum-carmine. Twelve hours. Sperma- tozoids held in the slime outside an archegonium, showing the nucleus and kinoplasm. Fig. 11. Half per cent, chromic acid : Delafield’s haematoxylin. Twelve hours. Vertical section of an egg with the enclosed spermatozoid still showing the coiled spiral form. Fig. 12. Half per cent, chromic acid: alum-carmine. Twelve hours. Hori- zontal section of an egg. Both the male and the female nucleus have a linin- network bearing chromatin bodies. The cytoplasm under most favourable illumination shows an alveolar or reticulate structure. Fig. 13. Half per cent, chromic acid: alum-carmine. Twelve hours. Vertical section through an egg-nucleus. x marks the direction of the archegonium canal. Fig. 14. Half per cent, chromic acid : Heidenhain’s haematoxylin. Thirty-six hours. Nucleus in a horizontal section. The distribution of the male nucleus is not much further advanced than in Fig. 12 after twelve hours. The lighter large nucleolus is not in the same plane as the end of the spermatozoid. / Annals of Botany. W.R.Sliaw, ad nat. del. SHAW. — JnnaLs of Botany Vot.XII,PL.XIX. W R. Shaw, ad nat. del. University Press, Oxford. SHAW. — FERTILISATION OF ONOCLEA. Some Thames Bacteria. BY H. MARSHALL WARD, Sc.D„ F.R.S., Professor of Botany in the University of Cambridge. With Plates XX and XXI. I. A short colourless Bacterium, forming stearine-like colonies 1 : type of Bacterium ureae (Jaksch). (PI. XX.) HIS form is apparently not uncommon in the river; JL I have isolated it several times, but have only culti- vated it twice through all media. It occurs on the plates as cocci, about i /x diam., not motile, and grouped in pairs, or rows of four, or isolated or in heaps, and evidently developed from the breaking up of short rods 2 x i j a found with it (Fig. i). On old agar-cultures the cocci alone are found ; but in actively-growing gelatine-cultures the rodlets prevail and can be seen to be breaking up to cocci in all stages. No spores have been found in any medium. The rodlets stain easily by ordinary methods — e.g. Loeffler’s methylene blue, carbol-fuchsin, &c. — but they are easily decolourized by Gram’s method. 1 This is the type of Group I, referred to in Proc. R. S., Vol. 6 1, 1879, P* 4 x7> [Annals of Botany, Vol. XII, No. XLVII. September, 1898.] 288 Ward.— Some Thames Bacteria . On plates at 12-15° C. it grows slowly as white, somewhat typhoid-like irregularly circular fronds, beautifully zoned and marked with radial lines. Under the J obj. these are contoured, hyaline at the edges, and yellower in the thicker, central portion. There are no pronounced blue or green sheens or iridescence, but after a week or so the fronds appear dull (matt), and like thin drops of stearine with irregularly contoured edges (Fig. 2). No liquefaction even after two to three months. Plates at 20° showed in forty-eight hours as white discs, like flattened milk-drops, 1-2 mm. diam. Under the § these are yellowish, contoured and opaque, the submerged colonies being very granular. On the third day the diam. = 2-5 mm., and with the typical opaque, stearine-like appearance. Fourth day = 9 mm. diam., and similar appearance. On holding up to the light a beautiful shagreen-like appearance, bossed in the centre, edges like ground glass. Sixth day =15 mm., opaque. Indented edges to the other- wise circular frond. Traces of zones and radial structure in some. Fourteenth day = 20 mm., zoned and radiate and with elegantly indented edges. Yellowish. Microscopic cultures in hanging-drops of gelatine were made, but it was found to be impossible to measure the growth. The short rodlets break up as soon as division is completed, and fall apart to make the colony of cocci. Fig. 3 shows a case where the isolated rodlet (a) in gelatine at 20° C., at 3 p. m. had divided into two, and one of these was dividing again at 10 p.m. (b). At 11 a.m. next day a colony, oval, pale, and with the normal characters, had formed, measuring 99 x 85 fi in length and breadth, and several rodlets thick (c). If it was ten rodlets thick, such a colony reckoned as a rectangular one would contain about 99 x 85 x jo = 42P75 rodlets, which indicates fairly rapid rate of growth. At the same hour (11 a.m.) the following day the colony Ward. — Some Thames Bacteria. 289 measured 300x180/;, which, calculated as before, gives us 540,000 cocci or 270,000 rodlets. Numerous other cultures only confirmed these results, and neither filaments nor spores could be obtained in any gelatine medium ; while the minuteness of the organism rendered futile all my attempts to directly observe its growth in liquid media. During active growth the rodlets 2x1/; prevail, but as the colony ages these all break into cocci. Stab-cultures at 12-150 show up in three days, but the development is slow. Yellowish-white dots appear in the tunnel, and a thin, dull, ground-glass-like frond above. This is hard and tough, like stearine, and difficult to scrape off. In fourteen days the surface is nearly covered, the frond with beautifully indented margins and radiate structure. No further growth in the tunnel. The matt appearance is due to a rough shagreen-like surface (Fig. 4). At 220 the development is equally good on acid or alkaline gelatine, and on the third day the characteristic matt frond appears above. The dots in the tunnel coalesce a little, indicating feeble growth. No trace of liquefaction. In old cultures the gelatine is slightly brownish-tawny above, and the colony has a faint greenish yellow tinge. The stab is sword-shaped. No trace of liquefaction even after three months’ culture : the yellowish white, thin, waxy, shagreened growth just covers the top of the gelatine, and has delicate fimbriate margins. Colonies submerged in the gelatine exhibit no growth. Streak-culture at 20° spreads fairly quickly, as a white, thin, matt film, like stearine or ground-glass, in forty-eight hours. In seven days nearly the whole surface is covered with a beautifully frondescent very thin film, spreading from the thicker streak, flush with the surface, greyish-white like ground-glass (Fig. 5). No sign of liquefaction even after two months. Edges very thin and fern-like. On agar at 30-35° a faint streak and one or two thin spots appeared in two days : these hardly increased in eight days. 290 War'd.' — Some Thames Bacteria . On keeping at lower temperatures a tough paste-like dirty white patch slowly spread in a month. At 22° white granular discs appeared in twenty-four hours, and coalesced to a thin spreading yellowish white film in two to three days. Slight white deposit in drainage. In a week there was little progress : the individual colonies are thin fronds like those on gelatine plates, but remain small. They are often polygonal where their edges touch, and give a curious mosaic-like or scaly look to the growth (Fig. 6). Under the lens these resemble the scales of a Turbot. In other cases the film is continuous. Potato at 340 gave no results in a week, but at 22° a dull, white yellowish streak appeared in twenty-four hours, with the peculiar stearine-like look of the gelatine colonies. In forty-eight hours this was thicker, dry, yellowish grey, and in four days had crenate and somewhat mesenteric edges. The colour darkened with age — buff, and like dull wax. In a week this consisted of rodlets 2 x i breaking to cocci i \x in diameter. Old cultures at 34° showed a slight growth after one month at lower temperatures. After two to three months the growth turns umber-brown. Broth at 250 formed dense flecks above in twenty-four hours, which easily shook down. In forty-eight hours a dense greasy-looking flocculent veil above, falling at the slightest shake : abundant deposit. These flocculent veils are renewed, the intermediate liquid remaining clear. The very abundant flocculent deposit acquires a slightly buff-white tinge with age. Milk at 25° showed no change in a month, beyond a distinctly acid reaction. On keeping three months still no change observable. Glucose at 30° formed white flocks below in twenty-four hours, and this deposit increased in two to three days. But no bubbles or turbidity resulted. At 250 the flocks and deposit were slightly more abundant, and resembled the broth cultures. No turbidity. The deposit slightly yellowish, but not very abundant even in three weeks. Urine at 25-30° gave very slight traces of turbidity in Ward. — Some Thames Bacteria. 291 twenty-four hours, and subsequently a few bubbles formed, and granular flocks were deposited. No definite turbidity, and no increase in three weeks. The most likely form I have been able to trace resemblances to is Jaksch’s Bacterium ureae , which may be distinct, as he believes it is, from Pasteur and Van Tieghem’s Micrococcus ureae. It agrees with Jaksch’s form in the size of the rodlets, the general characters of the plate cultures, and particularly in the dull (matt), ground-glass appearance. The stab-cultures also agree fairly well, though I have never detected the smell of ‘ Haringslake ’ noted by Jaksch. The growth is described as ‘ ungemein langsam V Of course this is very indefinite : in my form the growth is slow but not uncommonly so. The general behaviour as to temperature agrees, so far as can be gathered from the meagre information to hand. I prepared tubes of Jaksch’s fluid as follows — per 1 litre water : — iV gram. Mg S04. 1 8 » . . . K H2P04. 5 * . . . Rochelle Salt. 5 „ . . . Urea. In this perfectly clear liquid the organism grew very slowly, forming grease-like flecks and films on the surface and a very white deposit. Better at 250 than at 350. No odour could be detected, and it is pretty evident that if this is Jaksch’s form, it grows but feebly in the liquid given. M. ureae seems to differ from the rodlets chiefly in growing more quickly and at higher temperatures, and in the cocci — which may also be in pairs or tetrads or chains — and in the occasional formation of zoogloea. The stearine-like plate-cultures are very suggestive, and the stab-cultures agree well, except that I have not noticed the odour described. 1 Zeitschr. f. Phys. Chemie, Bd. 5, p. 395. 292 Ward.— Some Thames Bacteria . In both cases we are devoid of information as to the behaviour in other media than gelatine and urine or Jaksch’s fluid, so that it is impossible to be sure of the identity of these forms. There are some distinct resemblances also to Zimmermann’s M. concentricus 1, but he does not note the dull stearine-like appearance which is so striking in all my cultures. As Tatarofif 2 himself remarks, his ‘ Perlmutter-glanzende Diplococcus’ may be M. ureae , and the resemblances are noted. The following tabular summary gives the salient characters of this Bacterium. Habitat. Morphological characters. Plates. Stab. Streak. Agar. Potato. Broth. Milk. Glucose. Urine. Not uncommon in the Thames. Cocci about 1 n, not motile, single or grouped in pairs, rows of four, or in heaps, and formed by breaking up of rodlets 2 x 1 At. At 1 2-1 50 C. forms slowly growing white, typhoid-like, irregularly- circular, contoured, zoned, and radially-marked fronds : edges hyaline, centre yellowish. Become dull, matt, and like thin stearine drops. 1 2-1 50 yellowish white dots in the tunnel, and a thin frond-like ground-glass above. This is hard and tough, and matt, as in plate colonies. Quicker at 20-2 20 C. At 200 grow fairly rapidly as a white, thin, matt film, nearly covering the whole surface in a week. The streak thicker. At 30-35° grows slowly as a dull white and dryish layer. At 22° the characteristic white waxy matt film forms more rapidly. In some cultures isolated colonies form on the surface and coalesce to form the film, as polygonal Turbot- scale-like mosaic. No growth at 34° At 22° a dull, yellowish-white, waxy, dry streak-like stearine, darkening to buff, and after some weeks to umber. At 25° forms dense floating flecks and greasy white films, which fall as an abundant deposit, eventually buff- white. The films are renewed and liquid remains nearly clear. The greasy films adhere to sides of tube above. No change at 25° beyond acid reaction. No turbidity, but white veils and flocks form and fall as in broth, but less abundant. Better at 25° than 30°. At 25-30° a few bubbles and granular flocks only, and only slight traces of turbidity. 1 Die Bakterien unserer Trink- und Nutzwasser, Chemnitz, 1890, p. 86. 2 Die Dorpater Wasserbacterien, 1891, p. 71. Ward . — Some Thames Bacteria . 293 Jacksch’s liquid. Slow growth and greasy flecks, falling as a very white deposit. Better at 250 than at 350. No odour. Pathogenicity. Not pathogenic to guinea pigs, according to Professor Kanthack’s report l. After remaining from May 28 to June 8 of the following year, i. e. over twelve months, it was found quite easy to revive this form from an agar tube. Good plate-cultures resulted in four days at 20-22° C., and the colonies were quite characteristic. Further cultures in gelatine, agar, potato, broth and milk tubes confirmed this, and the results at 25° and 35° respectively were as above. This form must therefore be regarded as a very constant and persistent one, in marked contrast to many of the others I have had to deal with. II. A colourless capsuled Coccus or Bacterium2. (PI. XX.) An interesting form, isolated and cultivated through all stages at least twice from the river in the autumn, was one which occurred on the plates as small, short, oval, non-motile rodlets over 1 n long by 0-75 to 1 /x broad, and invested by a tough dense zoogloea or capsule, which occurred round the groups of dividing rodlets — then biscuit-shaped — as well as round individual cocci or rods. If rapidly stained by Gram’s method the capsules are decolourized, and the rodlets coloured : but they are easily decolourized. The stained rods = 1 x o«6 n to cocci about o*6 fji. The capsule = about 6 to 10 ju. On plates at 12-15° C. the colonies are white, porcellanous, shining discs or fronds, with a central spot and faintly zoned. To the unaided eye the colonies look bluish-white and translucent if held up, the zone or zones appearing yellower : the zones sinuate in agreement with the indented 1 I have to thank my colleague, Professor Kanthack, for kindly examining a number of these bacteria for me in respect of their pathogenic properties. 2 Referred to as the type of Group XI in Proc. R. S., Vol. xli, p. 420. X 294 Ward. — Some Thames Bacteria. margins. Under the J obj. the whole colony looks yellowish, granular, and gradually becomes more opaque in the centre as the frond thickens, the margins thinning out and paler. The microscopic examination also shows the colonies marked by irregularly and curiously contorted lines and streaks (Fig. i a), and scattered sets of brighter, rounded, sausage-shaped and vermiform areas. In the older colonies these are less visible in the centre, owing to the opacity as the colony thickens, but the zoning is found to be principally due to these brighter areas nearer the margins. The submerged colonies are yellow, granular, opaque, and irregularly lobed like a complex glandular acinus or salivary gland (Fig. i b , c). As the colony ages — three to four weeks — the white becomes tinged with a tawny hue, and a tendency to soften and sink into the gelatine is evident. Closer examination with a higher power shows that the bright vermiform and rounded areas are dense zoogloea masses embedded in the granular matrix of the colony, and that the glandular submerged colonies and the dark central part of the emerged ones are simply dense and irregularly- lobed zoogloeas containing the cocci and short rodlets, the rest of the colony consisting of irregularly and closely-crowded escaped cocci without any evident capsules (Fig. i d). These imbedded zoogloeas are so obviously similar to Cohn’s Ascococcus Bilrothii that I referred to them throughout my notes as Ascococci , but — though unfortunately we do not know the size of Cohn’s form — the cells seem to be larger, and they are certainly not permanently cocci, as will be seen later. At 20° the colonies were visible in twenty-four hours as minute grey points, yellowish and granular under the J-. A higher power (Zeiss. D) showed them already lobed and capsuled. On the second day they form white opaque irregular circles 2 mm. diam., and like milk : under the ^ the submerged colonies are lobed and glandular, the emerged Ward. — Some Thames Bacteria. 295 ones form discs, with the Ascococcus-like groups im- bedded. On the third day they are like irregular milky drops, too thick to show structure. Stab-cultures at 12-150 form a wet, glistening, thin white frond above and yellowish-white, dense, dot-colonies in the tunnel. In a week the frond has nearly covered the surface of the gelatine, and is depressed in the middle, slightly sinking into the gelatine ; while the colonies along the tunnel enlarge and tend to radiate into the surrounding gelatine. The sinking goes on until the frond lines the sides of a distinct funnel, devoid of liquid however ; and the sub- merged colonies form cloudy outgrowths and widen the tunnel. The sinking and softening of the gelatine continue, and are very decided in a fortnight to three weeks (Fig. 8). At 20° C. the phenomena are similar but quicker. In five days the softening of the gelatine is pronounced, the sub- merged colonies confluent ; and a good funnel with signs of liquefaction and running are evident in ten days. The growth is equally good — or even a little better — in slightly acid gelatine, as compared with slightly alkaline. The growth is easily removed by the needle, but does not lift as a whole membrane, and is firm and waxy or slightly slimy. Even after ten weeks there is no real liquefaction of the gelatine, but the cloudy white growth was penetrated far in. At 20° in sugar-gelatine, a milk-like spreading drop formed above, and a considerable confluence and growth in the tunnel in three days. Streak-cultures at 20° show a dull, translucent, white growth, yellowish if held down, bluish by transmitted light, thin at the margins, spreading slowly, and softening the gelatine in eight or nine days, and beginning to sink along the axis (F>g- 9)- In a month a deep spoon-shaped scooping has occurred, in which the cloudy white growth floats in viscid softened gelatine. X 2 296 Ward. — Some Thames Bacteria. Agar. At 20° C. a copious, thick, spreading, glistening, pure white streak with iridescent edges, extending to a fron- descent film all over in twenty-four hours. In three days a thick, shining, translucent, waxy, yellowish-white layer. On the fifth day this is a wrinkled membrane, and a white wrinkled veil and precipitate are seen on and in the liquid of drainage. This glassy-looking membrane is tough and lifts as a whole, and the microscope shows it as a dense zoogloea, with rodlets breaking up to cocci in Ascococcus-like masses. Staining with acetic acid and dahlia-violet shows that the capsules enclose both single rodlets and colonies (Figs. 3 and 10). At 340 C. the growth is similar, but less rapid. In four days the gum-like, translucent membrane is formed, but even in eight days it had not covered the surface. In strong growths the Agar is evidently diminished in a few days, serving as food-material. After a week or two at low temperatures the corrugated membrane is renewed on the stripped Agar. Potato. In forty-eight hours, at 220 C. a thin, wet, spreading, glistening film is formed, white at the thin fimbricated spread- ing margins, very pale yellow inwards, and with a greyish cast where thickest in the centre. About the fourth day the thin white margin disappears, and the whole patch is wet and slimy (Fig. 11). The microscope shows that the wet sulphur-white to yellowish-grey slime consists of rods 1-5 to 4/ux 1 y, motion- less, embedded in a tough slime which draws into long strings on the needle. At 22° alcaline potato is an equally good medium with normal, the colour of the copious whitish slimy growth being perhaps less grey and more sulphur-yellow in hue. Artichoke at 250 C. A white irregular patch was formed in two days, and spread all over as a white film on the fourth day, after which no further growth was noted, even in fourteen days. Ward . — Some Thafyes Bacteria. 297 Carrot at 250 gave good results. In twenty-four hours a rapidly-spreading gum-like layer formed, and extended all over as a wet, thin, watery layer in forty-eight hours. No change on the fourth day, and matters were the same at the end of a week. In fifteen days the tubes were discarded — no further growth. Turnip at 250 gave no certain results in four days, and even after a week no growth was observable. Kept for three weeks — no further results. Broth. No growth at 350 in three days. The liquid remains perfectly clear. At 250 C., however, the broth is turbid in twenty-four hours and with a dense precipitate, which is white in three days. Even after a fortnight the liquid is still densely turbid, and a copious flocculent pre- cipitate has fallen. Glucose at 250 showed no trace of activity in three days. A tube put in at 30°, when the temperature was falling to 25°, gave a slight turbidity in three days, and traces of white precipitate, but no fermentation visible ; this remained the same on the ninth day. In other cases no results were obtained in two or three weeks at 25° C. Milk at 25°C. No change to third day, but in fifteen days a thick custard is formed, and the tube can be upturned. In eighteen days the casein falls. The reaction is acid. No signs of solution in five weeks at 250 C. Urine at 220 gave a slight turbidity in five days, with traces of a ring, but no signs of further growth. P athogenicity . This form was kindly examined for me by Dr. Lazarus Barlow and gave pathogenic results. A guinea- pig inoculated in the peritoneal cavity with about 10 c.c. of a four days’ old beef-broth-culture died in twenty-three hours. On examining Dr. Lazarus Barlow’s preparations, I found them exactly to type. That from the peritoneal fluid showed the capsules, faintly but distinctly, but in the others they were almost invisible. It should be noted that no information had been given when the tubes were handed on, and so no 298 Ward. — Some Thames Bacteria . attempt to bring out the capsule had been made (Figs. 4 and 5). Professor Kanthack also found that this form is patho- genic to guinea-pigs, though as he was working with smaller doses, the results were not so fatal. Inoculation into the thigh produced a large swelling, intra-peritoneal injection made the animal very ill for a day. In both cases, however, the guinea-pigs recovered. The experiments gave the same results on repetition. It is worthy of note that the form sent to Professor Kanthack had been much longer in culture than that sent to Dr. Lazarus Barlow. On being revived on July 13 from an Agar-culture which had stood since May 13 of the previous year, i. e. fourteen months, the plate-cultures gave normal colonies, showing the characteristic zoogloeas embedded in the mass. On potato also the cultures showed the characteristic yellowish pasty growth with a broad white marginal area on the brownish- grey potato, and the other cultures were normal, and no question could arise as to identity. Even the custard in milk was developed in fourteen days at 30° to 350 C. All attempts to revive another culture failed. De Toni and Trevisan1 have attempted a classification of capsuled micrococci along the following lines : — They group all the forms under the head of Ascococceae. Then they cut out Winogradsky’s Amoebobacter , chiefly on account of the amoeboid movements and arrangement in series. The remainder are divided up, first according as the ‘ capsule 5 is general and around whole colonies, or special , i. e. around each individual coccus. Further subdivisions depend on whether the c cysts ’ or ‘ capsules ’ are lamellated or not, whether the colonies or families consist of few or many individuals, whether the divisions are in one or more planes, and so on. It seems difficult to accept the details, but no more con- sistent attempt is to hand, so far as I am aware. Sylloge Schizomycetum, p. 1035. Ward. — Some Thames Bacteria. 299 There are nine genera, as follows - Lamprocystis , which includes only Lankester’s Bacterium rubescens , with its numerous (real or assumed) synonyms. Ascococcus, again confined to one form — Cohn's A. Billrothii . Bollinger a, comprising two species of B. equi (. Micrococcus ascoformans of Johne, M. botryogenus of Rabe) and B. Vacchetae (Trev.). Cenomesia , also with two species — C. albida and C. lilacina , both from sulphur waters. Thiocystis , again comprising two forms — Winogradsky’s T. violacea and T. rufa , both from sulphur waters. Thiothece , including Winogradsky’s T.gelatinosa , from sulphur springs. Leucocystis , with only Schroeter’s L. cellaris , found in caverns, &c. All the above are regarded as having a general capsule, common to whole colonies or families : the following are devoid of this, but each coccus has its own special invest- ment : — Chlamydatomus includes the two species : C. Beigellii. , first described by Beigel as a Gregarina found on hair, and C. cellaris, found by Hansgirg in cellars. Gaffkya includes four species : G. grandis , the Microcoque des reins et des niches syphilitiques de la peau of Babes and Cornil. G. tetragena (Gaffky), Micrococcus tetragenus , found in phthisical sputum. G. Mendozae (Trev.), M. tetragenus mobilis ventriculi, a motile form which gives an odour of skatol in cultures. G. Archeri (Trev.), Archer’s Black Micrococcus , a deeply pigmented form found on potatoes. 300 Ward . — Some Thames Bacteria. But these are not the only micrococci described as having these ‘ capsular * investments, as the following list shows Micrococcus of Bovine pneumonia, Poels and Nolen, from the lungs of cattle infected with pleuro-pneumonia, and resembling Friedlander’s bacillus in many respects. Diplococcus of Horse pneumonia (Schiitz), a similar but im- perfectly described form. Haematococcus Boris (Babes) ; Pseudodiplococcus pneumoniae (Bonome), indistinguishable from M. pneumoniae crouposae except in its growth at lower temperatures ; M. ureae (Pasteur) ; M. luteus (Cohn.) ; M. viticulosus (Katz), are other species described as capsuled or forming investing zoogloea masses. I am unable to refer my Thames form to any of the fore- going with certainty, and am inclined to suggest that it should receive a name as a new ‘ species.’ From a Petri-dish, in which a plate-culture had been made from a drop of water impregnated by shaking up a zoogloea-mass grown on Agar, I removed a little of the gelatine-film with a loop, and transferred it to a culture- cell, suspending it from the cover- slip as for a hanging-drop culture. The plate-culture had been going twenty-four hours at 20°, and the colonies were just visible — hardly so without a lens— and my idea was to watch the behaviour of a rodlet at the thin margin of a colony. To do this, however, it was necessary to raise the tem- perature of the culture chamber just sufficiently to soften the gelatine and make it spread a little, for no matter how carefully one prepares such a culture as the above, the play of lights reflected and refracted at the conchoidal fractures of the solid splinter of gelatine interferes seriously with observations under high powers. Consequently it was necessary to warm the whole to nearly 250 C., and then let the minute-drop solidify again. Ward . — Some Thames Bacteria . 301 This was done, and several well-isolated rodlets were now found near the margin of the colony and clear of it. I now focussed a pair, lying close together but sufficiently apart for distinct observation : their position was fixed by means of the micrometer, and they were drawn at 10 a.m. ; the temperature being 21*5° C. Their behaviour at subsequent periods of observation is given in Figs. 16 a-f. At 30.20 each had divided, though the two halves were still joined: at 10.35 they were free, and now there were four rodlets in place of two (see Fig. 16 e). At 1 1. 10 a left-hand rod was dividing, as shown by its biscuit-shape, and at 11.40 there were six rodlets; at 12.20 a rod below, to the right, was dividing, and by 12.45 there were eight rodlets. Now it was evident that in the successive divisions the sister-halves were not equally capable of dividing. The question arises whether this is due to position, or some other cause. I am strongly inclined to regard it as due to position ; in each case the new divisions occurred first in cells nearest new territory , i.e. advancing away from the colony into unexplored gelatine. The above observations had now to be interrupted, and on resuming them at 3.20 p.m. a startling discovery was made — all the free bacilli were in active swarming move - ments . The temperature had slowly risen to 23*5 and remained there, and the gelatine-drop had absorbed a great deal of water : these factors, taken with the liquefying power of the colony, explain why the drop was now liquid. But the swarming was an unexpected phenomenon. I had got over my surprise at the isolated rodlets, above described, showing no capsules, because earlier examina- tion of the gelatine colonies showed that not all the cocci or rodlets are capsuled. Hitherto, however, they had shown no signs of movements. The obvious suspicion arose that an intruding swarmer had got into my hanging- drop. 302 Ward. — Some Thames Bacteria. That was not the case, however, as the following observa- tions show. As we have seen, the temperature had been slowly rising all the morning, as follows io*o a.m. temperature = 21-5 10-20 33 = 22 io*35 35 = 22-2,5 11*10 33 = 22-5 11-40 35 = 32-5 12-20 55 = 22-75 12-45 53 = 23-25 3-20 33 = 23-5 And I allowed this rise to go on. The numbers of swarmers increased enormously, and I suspected this was not due merely to the rapid division of those already in motion, but that the increase was partly due to reinforce- ments from the colony of resting forms. After some search — principally due to the difficulty of focussing now the drop was enlarging- — I got a very typical capsule enclosing six rodlets^ under observation at 3.40. The temperature was 24*5°, and remained there. But the rodlets inside this cap were no longer quiescent: they were slowly moving, tumbling over one another within the hyaline prison of the capsule. Numerous free swarming rodlets were now in the neigh- bourhood, and one saw here and there groups of about six to ten of apparently free ones moving about each other, gliding and tumbling one over the other in the same way as those imprisoned in the capsule referred to. This capsule was kept under observation from 3.40 to 4.40, and notes made at 3-55, 4.15, and 4.30. The slow swarming at 3.40 became more and more active as time went on, and at 4.15 was as active as in the apparently free swarming groups around, but the enveloping capsule was now swollen, and so transparent that it could only be known to be there by the limits its presence placed to the swarming movements of the imprisoned bacilli. At Ward . — Some Thames Bacteria . 303 4.30 the diffluent walls had softened and dissolved, and the imprisoned swarmers escaped, and were swimming about as actively as any of the other free rods. These observations were confirmed several times — as soon as the temperature rises to about 23-5 to 250 active swarming begins. In some cases a pair of recently divided rodlets behave in a peculiar manner, and this I have seen not once, but several times. Each is capable of movement on its own account, but in some cultures (gelatine) the newly-separated rodlets separate by backing away in a straight line, and then come together again, end on, and remain a few seconds as if they had never separated at all Thus, in Fig. 13 I have sketched the relative positions of one of these pairs at four stages of their oscillations. At first they Were closely applied pole to pole in a straight line ( a ), then they suddenly darted asunder ( b ), till separated by about three times the length of either : after a few seconds they flew together again (< c ), and then again flew apart ( d ) ; and this went on for at least half an hour. It is quite a common event to find the rodlets swarming in this way, though the pairs do not invariably approach and recede in the same straight line. It was noticed that in one and the same case the movements in either direction — separation or flying together — might concern both, or only one, or sometimes one and sometimes the other, and by no means equally. If we call the rodlets A and B, sometimes they darted apart five or six times the length, equally distant from the point where their poles joined, and next time A would dart off and leave B quiescent, or A would move twice or three times as far as B : similarly on darting together. I could find no rhythm about this phenomenon, and do not understand its meaning. At one time I thought the darting together might be due to an elastic cilium which they were tugging at, but it seems improbable. In most cases, however, the swarming is in and out irregularly when several are concerned, and it seems to 3o4 Ward . — Some Thames Bacteria. depend on the temperature. Fig. 14 gives a case where the single rodlet at 10.25 p.m. had divided at 11 p.m., and the two halves separated at 11.5 p.m. : these two oscillated away from and towards one another as above described, and went on doing this and dividing through the night, and at 7 next morning had developed into a colony about 25 \x in diameter, and containing probably 10,000 to 15.000 of the short rodlets, all swarming actively, in the circumscribed space of their own gelatinous investment. If the temperature does not rise beyond about 20° the colonies are developed without any swarming. Thus Fig. 15 shows a case where a rodlet was fixed at 4 p.m. in io°/o gelatine and remained at 150 C. through the night. Next morning at 8-i 5 it had formed a small colony about 3 m in diameter, and consisting of eight to ten rodlets, so far as I could make out — possibly twelve. The temperature was then allowed slowly to rise to 20-25° C., and at 3 p.m. the lobed colony of quiescent short rodlets and 1 5-16// in diameter, shown in Fig. 15 (c), had formed. At 8.30 p.m. the whole colony was in active swarming, but next morning was quiescent again. Numerous attempts to cultivate this form further were made without success. III. Rose-pink Micrococcus : Type of M. carneus . (Zimm.) (PL XXI.) A very pretty rose-pink form was isolated several times and studied during the winter of 1894-95, when it seemed fairly common. It is by no means one of the more frequent forms in the Thames, however. I was for some time puzzled by it, for at one period its alliances seemed doubtful. It occurs as spherical cocci of variable sizes, from 0-5 to i*o ju, or even occasionally up to 1 • 5 /x or nearly so, in diameter, in irregular botryoidal groups, and perfectly quiescent. It stains easily, and well-stained specimens may show a darker more or less central spot, and a paler halo round Ward ’ — Some Thames Bacteria. 305 the cells. Specimens in water sometimes seem to be dis- tinctly vacuolated, or even to have granules in them, and some of these characters at first led me to suspect its being an extremely minute yeast-form — for instance, the vacuola- tions, the paler halo, and the grouping — but I have been able by cultures to determine that this is not so : it is a true Schizomycete. The fact that it does not ferment glucose solutions is, so far as it goes, evidence against the yeast view ; but of course it is far from conclusive, since plenty of yeasts do not ferment sugars. In the absence of any proof of budding I considered this form as probably a Micros coccus, and the occurrence of diplococci and rows of nearly or quite equal-sized cocci point to the same conclusion. On cultivating it at 19-20° in broth-drops under the j^th immersion it proved to be a Sarcina- like Micrococcus 1, which divides in all three directions, but the progeny frequently partially separate later on, and only remain united in zoogloea-masses, and so form irregular botryoidal groups of cocci each 1-2 n in diameter. The high refrangibility of the gelatinous zoogloea investment makes it impossible directly to see the actual act of division, but enough evidence was obtained (see Figs. 8 and jo) to determine the nature of the organism. After being sown about twenty-four hours, the cocci are found dividing very regularly in the Sarcina- form (Fig. 8), but in the course of another twenty-four hours the cocci partially separate as they rapidly divide, and, rounding off, remain agglomerated in the characteristic grape-like manner shown in Fig. 8 d and Fig. 9. As time goes on, the separa- tion is more and more complete, and isolated cocci and diplococci are common in the drop. The series figured in Fig. 8 ( a to d) will show this. At 11.50 a.m. a group of three Sarcina- masses was isolated (< a ) and watched : at 2.30 p.m. the Sarcina- divisions had increased as seen in ( b ), though it was impossible to accurately 1 The type of Group XVII in Proc. R. S., Vol. xli, p. 421. 306 Ward . — Some Thames Bacteria . count the cells. The group had rotated through about i8o° in the interval. At 4 p.m. the further development seen in (c) had taken place, and signs of loosening of the individual cells were evident, and at 9.50 p.m. the group was a rapidly increasing botryoidal mass as shown at ( d ). Next morning it was a loose mass of groups like Fig. 9. The best series, however, is the one in Fig. to, where I traced the whole course of development under the ^th immersion. The gelatine-drop was prepared at 5 p.m., and after allowing time to solidify, &c., the single coccus drawn in Fig. 10 [a) was fixed at 5.55, t = 2o°C. At 8.5 this had grown to the biscuit-shaped figure shown at (b), and at 11.40 p.m., the temperature having fallen to 3 9°C., there were four cocci in focus (c). Whether growth had occurred in the plane at right-angles to the paper I could not with certainty determine, but was of opinion that it had. During the night the temperature fell to 160 C., but was at 18° by 9.10 a.m., when nine cocci were clearly visible (d), and certainly some existed in the depth, but I could not focus down to them. By noon, growth was rapidly advancing, and two groups of four, one of two, and some behind were visible (e) : the temp. = 19° C. At 2.10 (t.= i9°) the group was loosening (/), and this went on as the growth and division rapidly advanced (g= 4 p.m., t. = 2i°), till at 9 p.m. (t. 22°) there was a mass like a bunch of grapes (h). Plate-cultures at I2-I5°C. show slowly-developing, raised, dry, rose-pink points, which even after three weeks are not more than 1-2 mm. in diameter, and do not as a rule liquefy. In a week the submerged colonies, under the Jrd objective, are irregular, roundish, dull-pink and granular ; while the emerged ones are prominent, rose-pink, opaque drops, showing a deeper centre, and a paler granular zone around. Even after two months the rose-pink, slightly sunk, projecting points are not bigger than in three- weeks’ plates (Figs. 1-3). Under the frd objective the older emerged colonies show Ward. — Some Thames Bacteria. 307 as granular discs with a colourless margin and deeper centre, others are distinctly zoned, pink, and considerable variation in the depth of colour occurs, from pale-brick red to lavender- tinted rose-pink. Gelatine streak at 20°. In twenty-four hours the growth begins as a dry, pale lavender, tinted rosy streak, much the colour of almond petals. In a fortnight it has a curious appearance of striping, like fresh-cut muscle under a lens. The colour gets more like sealing-wax at the thickened base. The transversely striped appearance — due to ridges — seems a constant character. In the course of a month or more it slowly liquefies, and in six weeks seven-eighths of the gelatine is quite liquid. Stab-cultures at 12-15° show small dots in the puncture- line in five days, and a protuberant dry pink button above. The colour deepens to plum-pink as the button widens, and in eighteen days no trace of liquefaction occurs. At 200 the growth is similar (Fig. 4), but traces of sinking are found in five weeks ; in six to seven weeks the gelatine is liquefied half-way down, and even more. It requires three months or more to complete the lique- faction to the bottom. Agar. In forty-eight hours at 250 a dryish rosy streak of isolated and conjoining raised dots. In three weeks confluent to a shining, pasty, rose-pink, broad streak, with thicker axis, and flattened, radiately striated mesenteric and indented margins (Fig. 5). Consistence pasty. The hue is a lavender-tinted rosy pink, much like almond petals. After being in culture some time on Agar at 25°, numerous minute dot colonies are formed, hardly showing trace of pink in six days : faint pinkish deposit. The growth at 35° is still more faint, minute pink dots appearing in ten days. Potato. At 20-22° forms a pink, rather moist, spreading layer in three days, which in five days becomes almost vermilion, thin, and spread all over. The colour is very peculiar; perhaps carmine is the nearest hue (Fig. 6). The growth on alkaline potato is extremely slow, or even nil. 308 Ward. — Some Thames Bacteria. Normal potato at 350 shows very slight growth in forty- eight hours. Little progress in four days to a week : merely a few extremely minute red spots in a watery film. Carrot at 250 gives a thin and very poor pinkish-white growth in fourteen days. Artichoke. No growth in fourteen days at 250. Turnip. No growth in fourteen days at 250. Broth. No growth at 350 in a week, nor after a month’s subsequent keeping at ordinary temperatures. At 250 a faint pinkish deposit in three days, but no turbidity. In a fort- night the deposit is increased — granular and flesh-pink : no turbidity or other change. The pink slowly deepens in hue. Milk. At 250 showed no change in fifteen days beyond traces of pink in a small white deposit. This had not in- creased by the third week, when the liquor was faintly but distinctly alkaline in reaction : no other change, but in the course of two or three months there are traces of peptonization without coagulation. Glucose at 250. Showed no growth in fourteen days. Not proved to be pathogenic for guinea-pigs according to Professor Kanthack. The results are doubtful. The following pink, non-liquefying micrococci and yeasts are on record. Micrococcus cerasinus siccus (List :) is a very minute form, 0*25 to 0-32 fj, in diameter, found in water, but growing best at high temperatures — e.g. 370 C. — and not doing well on gelatine. It is interesting to observe that this form is also noted as resembling a Torula in some cases, but it is in- capable, according to Adametz, of inducing fermentation. The description to hand is very meagre, but the size, temperature, and other characters seem different from those of the Thames organism. M. carneus (Zimmermann)2. This form, found in the Chemnitz water-supply, presents some striking resemblances to the Thames one. The cocci average about o-8 /x in 1 Eisenberg, Bakteriologische Diagnostik, p. 34. 2 Zimmermann, 1. c., p. 78. Ward. — Some Thames Bacteria. 309 diameter, and are arranged in irregular botryoidal clumps. It grows best at ordinary temperatures, and poorly at 30- 330 C. The growths on Agar and Potato are strikingly similar to my results, but there are minute differences in the description of the plate-colonies, possibly due to differences in the temperature of our cultivations. Lustig 1 describes a red form ( Coccus ruber) which Maschek found in water, and which he regards as probably identical with Zimmer- mann’s species. The differences in the two descriptions are nearly, if not quite, as great as those between Zimmermann’s account and mine, only Lustig gives too few particulars (e.g. as to temperatures, &c.) for a decisive judgement. Another red Micrococcus is Flugge’s M. cinnabareus 2, also found in water and air. Excepting that the cocci are described as ‘ large/ and frequently in pairs or in tetrads, and that the plate-colonies are red-brown under the low power, there is nothing in the short diagnosis to separate this form from the above, and we may well suspect that they are one and the same form, for the naked-eye colours of Flugge’s species agree very well. Of course much depends on what 4 large ’ means in his diagnosis 3. Mace4 describes under the name M. roseus (Flugge) a common air-form, in twos, threes or tetrads, with flat faces, about i*4ju along the greatest diameter. It does not liquefy, but the description is too meagre to make much of. Mace also points out how similar these forms appear to be, and remarks that the form termed M. cinnabarinus of Zimmermann cannot be distinguished from Flugge’s M. cinnabareus. This M . roseus of Flugge must however be distinguished from the M. roseus of Eisenberg referred to below, as well as from the M. roseus described by Maggiora 5, a non- 1 Diagnostik der Bakterien des Wassers, p. 40. 2 Flugge, Die Mikroorganismen, 1886, p. 174. 3 Mace, Traite pratique de Bacteriologie, p. 335, gives 0.9 /x, which would strengthen the force of the above. 4 Mace, p. 334. 5 Giorn. d. Soc. ital. d’igiene, Anno XI, 1889, p. 356, No. XXII. Y 3io Ward. — Some Thames Bacteria. liquefying form, o *6 ju in diameter, associated in irregular glomeruli, and forming a pale rose pigment. Mention may also be made of M. agilis (Ali-Cohen 1), a motile form, i \x in diameter, and which sometimes liquefies slightly after a long time : a pink layer is formed on Agar and potato. In addition to the foregoing non-liquefying forms, may be mentioned a series which liquefy the gelatine : — Bumm’s Diplococcus roseus 2, a liquefying air form ; Sarcina rosea (Schroeter 3), also a liquefying aerial form ; M. roseus (Eisen- berg 4 5 6), a slowly liquefying form found in sputum. Sarcina mobilis 5 (Maurea), said to be motile (?), liquefies, and will not develop on potato. Finally, reference may be made to Bacillus prodigiosus , which is often termed Micrococcus prodigiosus , owing to the shortness of its rodlets : this seems identical with M. haema- todes described by ZopfG as the form concerned in bloody sweat. The resemblances to Zimmermann’s M. carneus , which he regards as probably identical with Maschek’s form 7, is so marked that only one point of importance indicates lack of identity. This is as regards the mode of division. Zimmer- mann says (1. c. p. 78) the divisions occur in one direction only, but I find the divisions occur in all directions, and that in certain stages the groups resemble a Sarcina. It is an interesting point that Maschek’s form (I quote from Lustig, 1. c., p. 40) presents the same similarity to a Sarcina that mine does, and we have seen that Zimmermann regards Maschek’s 1 Central -bl. f. Bakt., 1889, VI, p. 36. 2 D. Mikroorg. d. gonorrhoischen Schleimhauterkrankung, 2. Ausg., Wiesbaden, 1887 (Eisenberg, 1. c., p. 12). 3 Eisenb., p. 16. 4 Eisenb., p. 408 (quite distinct from Flugge’s form : see above). 5 Sternberg, Manual of Bacteriology, p. 72°* 6 Spaltpilze, p. 60; see also Cornil and Babes, Bacteriologie, p. 142, and reference to a form mentioned by Pasteur. 7 Zimmermann, 1. c., p. 79 ; Maschek, Bakt. Unters. d. Leitmeritzer Trinkw., p. 60; Adametz, Die Bakterien d. Trink- u. Nutzwasser, No. 17. Ward.— Some Thames Bacteria . 31 1 form and his own as probably identical, and Lustig takes the same view. It may be worth while to raise the question whether the Sarcina- form and the Staphylococcus-ioxm of Micrococci are more than growth-forms of one and the same organism. If this turned out to be true, Schroeter’s Sarcina rosea — and possibly Menge’s Sarcina of red milk 1 is the same organism — would have to be examined in this connexion. Several of my micro-cultures in broth-drops showed, as we have seen, that this Micrococcus forms evident Sarcina- like groups when young and growing slowly, but that the botryoidal ( Stapkylococcus-Yike ) growth prevails later on when development is rapid. It is perhaps not incorrect to say that the few known forms of Sarcina all come from sources (acid media, air, water, &c.) which may be regarded as poor pabula for such organisms. In any case there is nothing absurd in the suggestion, because it is known 2 that Sarcina- forms may so alter their habits on certain food-media that the cells become isolated by dissolu- tion of the membranes and only single Micrococci , or (when dividing) Diplococci , are found, though the ‘ packet-form 5 can be obtained by another alteration of the food-medium. I regard the case as not only interesting, but of some importance, for no one would have been able to infer the existence of the two conditions without actual culture in hanging-drops. This form was easily revived on July 13 from an Agar culture of the preceding Aug. 14— i. e. eleven months — and soon came up normal. Its peculiar cherry red (cerise) colour and other characters were as before, and it was interesting to see how the differences between it and certain other red species — e.g. B.prodigiosus — were maintained. 1 Central-bl. f. Bakt., VI, p. 596. 2 E. g. Mace, 1. c., p. 364. 3 1 2 Ward. — Some Thames Bacteria . IV. A Pseudo-bacillus 1 . (Plate XXI.) This occurs as irregular and often curved rods 4 x 1 in water, motionless, often with spore-like darker spots in them, and breaking up into cocci. In old gelatine-cultures only the cocci are found, in chains or groups, or as diplococci and single cells, about 1 p or a little less when stained. They stain by Gram’s method. No true endogenous spores have been found, though easily stained oval bodies occur in the rods as described. In broth the motionless rods are often slightly curved, and measure 2-3 x 1-1*2 ju, and grow out to short filaments 10-12 ju, and segmented. In some cases inflated involution forms occur, nearly 2 n thick. Plate-cultures at i2-i5°C. show in four days as raised yellowish-white colonies, fairly quickly growing, and already coalescing. The submerged ones are very opaque, yellowish white, not zoned. Liquefaction begins in a week, as a slight sinking, but does not progress (Fig. 1). After three months in culture, plates at 18-20° showed nothing to the unaided eye until the third day, but in forty- eight hours the J detected minute pale discs. On the third day just visible as white points, which under the J are greyish, hyaline, coarsely granular. On the fifth day they look like raised drops of milk, 1 mm. diameter, domed, opaque, glistening yellowish white. Under ^ course, grey-yellowish, and opaque. On the sixth day they are 1*5— % mm., on the seventh 2-3 mm., opaque, cream-coloured, flattened domes. On the ninth day 3-4 mm., shining and like drops of cream. No trace of sinking, though some run together when in contact. The peculiar glistening appearance of the colonies is due to their wetness — as if sweating water on the surface. Stab-cultures at 12-15°. In two days a raised dome-like button, porcellanous white, and slight yellowish dots in tunnel. 1 Referred to as type of Group XVIII in Proc. R. S., Vol. lxi, p. 421. Ward —Some Thames Bacteria. 313 In a week the colony above is a pure white, much raised, and shining like wet glazed porcelain. In a month it becomes cream-like and soft. At 200 it grows equally well on acid and alkaline gelatine. In three days it is a very white raised button, 2-3 mm., with slightly confluent dots in tunnel. On the fourth day it is like porcelain, thick, glistening, raised. After about the sixth day it acquires concentric zones and a cream-colour, and looks as if turned (terraced) out of cream-coloured porcelain. No liquefaction, even in ten weeks. Streak-culture at 20°. Cream-coloured, raised, glistening streak in forty-eight hours, and this grows fairly rapidly (Fig. 9). In a week it is a thick, glistening, creamy porcelain-like patch, broader below. No liquefaction in two months. Agar at 30°. Forms a feeble streak, very thin, which makes no progress after forty-eight hours, but fades out as a trans- parent film. Invisible in eight days. At 350 also no growth in five days, whereas cultures at the same time at 23-250 formed a milk-like, broad, thin, shining, gummy or waxy streak with dense yellowish-white deposit all through the drainage (Fig. 10). Potato at 220. In twenty-four hours a wet spot, like dew. In three days this is a diffuse thin streak like milk and water. It thickens on the fourth day to a grey paste, and in a week is a not very extensive patch of cream-like, flesh-coloured paste (Fig. 8). On normal potato the growth is much more raised and distinctly flesh-coloured than on alkaline potato, perhaps because the potato acquires a pale violet hue showing it up. In ten days or so, both cultures are like rich buff or flesh-coloured cream. At 340 the growth fails. A dew-like patch forms at first, but shows no advance in six days. But on keeping the tube at lower temperatures, the characteristic flowing cream-like patch forms after some time. Broth at 250 shows traces of turbidity in forty-eight hours, and a slight deposit in three days. On the fifth day a copious yellowish-white deposit. In a week, still turbid and a white 3 H Ward. — Some Thames Bacteria . ring. In three weeks still turbid, white ring, and copious yellowish or buff deposit. Milk at 250 shows no change in fourteen days. In three weeks it is just acid, but no apparent alteration. Glucose at 250. A slight white deposit in three days. In three weeks greasy flecks above and a fairly abundant yellowish-white deposit. This form is not pathogenic to guinea-pigs, according to Professor Kanthack. It was easily revived from an Agar-tube which had laid quiescent from May to June in the following year — i. e. thirteen months. It came up very pale and weak at first, but soon recovered all its normal characters as described. From the sum of the characters, including the results of microscopic cultures below, this form presents resemblances to B . diphtheriae which cannot be neglected, but it is not a Bacillus. When I came to make micro-cultures of this organism in hanging-drops of gelatine and of broth, some unexpected results were obtained of considerable interest and importance. The following examples will illustrate this : — A gelatine drop-culture twelve hours old had a rodlet 4x1 /x at 8 a.m. (t. = 2i°C.), which was fixed and observed under the Zeiss E as shown in Fig. 6 ( a-k ). At 9.30 the much longer and sharply bent rod was behaving very curiously for a Schizomycete, for it appeared to be putting out a branch at right angles from its lower segment (Fig. 6 c). At 10.30 the much diluted gelatine was nearly fluid and an end-segment had broken off to the right and floated somewhat to the middle of the parent rod and there divided. The further course of the formation of the colony is visible in the drawings ( d-h ). At 4.40 p.m. a circular colony 24 /x in diameter had been formed (i) : at 8 p.m. this was 32 ^ in diameter (/). Next morning at 9 o’clock it measured 75 /x across (fc) 1 , and by noon it was 90 /x in diameter and quite typical. 1 Sketched under a lower power. Ward. — Some Thames Bacteria . 315 Now it is pretty clear that apart from that curious lateral branch, there is very little to denote that this is not a typical Schizomycete, the segmentation of which is at first into rather long rods (10-12 ju) and then into shorter ones (about 3-4 f). But there was no doubt that the branch was a true branch, and further examination in hanging-drops under the y^ and 2V immersion led to the proof that this organism is not a true Schizomycete at all , but an oidium-stage of an extremely minute fungus. The following series (Fig. 5), traced under the y1^ in a broth- drop, will suffice to demonstrate this. At 6. 1 5 a.m. a rodlet (a) 3 x 1 ju. was fixed, and at 8.40 a.m. it had grown out to a short curved filament (b) about 12 ju long : this was longer and distinctly segmented (c) at 9 a.m., and just before 10 o’clock ( d ) the longer segment was forming two branches, which had grown considerably by 11.10, and at 12.30 p.m. had crossed one over the other (e and /). The long segments now showed several septa, not easy to see but certainly visible with careful focussing. At 2 p.m. (g) the segments were breaking apart, after further growth of the terminal ones , i. e. the growth was not intercalary . At 7 p.m. (h) quite a large colony of separated segments, like rods, had formed, only part of which is figured. And next morning the still more broken up rod-like segments — some curved — ■ had spore-like, brilliant oval bodies in them (/). These are of the nature of oidia or chlamy do spores, in fact. These stain easily, with the ordinary alkaline methylene-blue, for instance. As the figures (Figs. 4, 5 and 7) show, these stainable points appear before the final segmentation of the rods into coccus-like short joints— oidia — and then the membrane appears to thicken round them, converting them into spore- like chlamy dospores. Now, the point of special interest is that here we have an organism which, according to all its properties as tested by ordinary bacteriological methods, is a Bacterium. Its micro- scopic appearances, as shown in stained preparations, its behaviour in plate-cultures, and on and in all the usual media 3 1 6 Ward . — Some Thames Bacteria . employed by bacteriologists, all suggest its being a Schizo- mycete. Nothing but cultivation in hanging-drops could have demonstrated the fact that it is not a true Schizomycete at all, but an extremely minute Fungus — at least, I presume no one will dispute that its apical growth, acropetal mode of branching, and other morphological characters constitute more important tests than the bacteriological ones. Such forms are quite common among the Basidiomycetes 1. In case any one should dispute this, however, it rests with him to construct a new definition of the Schizomycetes. Meanwhile, I emphasize the point — a point which I have insisted on elsewhere — that minuteness, staining reactions, rapid growth and the characters obtained in plate-cultures do not prove that an organism is a Schizomycete, and nothing but micro-cultures, difficult as they may be and are, can ever decide the point. This point raises another matter of considerable interest, however, viz. that of the multiple origin of the group commonly known as Bacteria, by which I mean not only the Schizomy- cetes proper, but the totality of micro-organisms usually grouped with them. Excellent evidence exists for the view that the true fila- mentous Bacilli (the Eu-bacilli or Endosporous Bacteria of De Bary) and the segmenting Bacteria which form no Endo- spores (De Bary’s Arthrobacteria) must be regarded as having their origin from among the lower Algae, and it is customary to refer the former to groups like Oscillatoriae and the latter to forms like Nostoc. Whether the group to which Cladothrix , Leptothrix , and Beggiatoa are commonly referred can be joined to these is a debatable point. For the various forms of Sarcina and Micrococcus , again, it is not difficult to find analogous forms among the lower Algae, e g. Chroococcaceae and Palmellaceae, though it must not be forgotten that Micrococci are often merely the ultimate segments of anthrosporous filaments. Without entering into the discussion as to alliances between 1 See Brefeld, Unters. aus d. Gesammtgebiete der Mykologie, Heft 8. Ward . — Some Thames Bacteria . 317 various forms of Bacteria (in the wide sense) with Protozoa and Myxomycetes, and merely admitting that such alliance may well exist among the group, I would simply refer to a possible source of confusion which has become more and more probable since Brefeld has made us acquainted with the frequency of oidium-forms and chlamydospores among the Fungi1, namely that these forms when very minute may easily be confounded with Schizomycetes. The only test is the acropetal mode of growth. That minute yeast-forms are also liable to be mistaken for Micrococci is evident. I had recently in my laboratory a minute organism which grows in Canada-balsam, and am as yet unable to say with certainty whether it is a yeast-form or a Schizomycete. Here then we have a good deal of matter for further re- search, for it is almost certain that minute organisms which will grow in gelatine and other media, and which stain by ordinary methods, are continually being described as Schizo- mycetes without the application of the only test which really decides the question. I am strongly inclined to the opinion that we shall have to revise our views as to the divisions of the accepted Schizomy- cetes very much before long. P'or instance, Fischer’s recent work on the cilia of Bacteria2 seems to raise the question whether we must not assume a different origin for the ciliated forms of ‘ bacilli ’ and for the non-ciliated ones ; and, in view of Ali-Cohen’s discovery of a ciliated ‘ Micrococcus 3 ’ (M. agilis ), the same applies to the Micrococci. In any case it is difficult to avoid the conclusion that the organisms grouped under the common denomination ot Bacteria (in the wide sense, but including obvious Fungi) are a heterogeneous collection of organisms with very different alliances, some of which have been indicated 4. 1 Brefeld., 1. c. 2 Unters. lib. d. Bau d. Cyanophyceen und Bakterien. 3 Centralbl. f. Bakt., Band vi, p. 33. 4 Migula, System d. Bakterien, 1897, and Fischer, Vorlesungen liber Bakterien, 1897, have recently proposed extensive revisions of the classification, and have raised similar questions, but not quite the same points as 1 have here suggested. 3 1 8 Ward. — Some Thames Bacteria . Another point of importance, however, concerns those endosporous bacilli which are never motile, e.g. B. Anthracis , and those which have cilia, e. g. B. subtilis. I believe no one has suggested that the former may have had a totally different origin from the latter and that both may have been derived from ancestors other than Cyanophyceae ; but it seems not impossible that minute reduced forms of Zygnemaceae and allied Conjugatae may have given rise to the non-motile bacilli. In such an event the endospores are probably homo- logues of azygospores 1, the intercalary growth, division, shapes of cells, and even the tendency to gelatinization of the cell-walls remaining the same. Indeed we may go further. Many Ulothricaceae would serve as prototypes of ciliated bacilli if they lost their chlorophyll and became reduced. It is not impossible that we may have to abandon the Cyanophyceae as probable ancestors of endosporous forms altogether, for none of the Oscillarieae develop ciliated cells, while many Chlorophyceae have intercalary growth and gelatinous walls. Even the curious pedicellate bacilli, which form one-sided growths or stalks of gelatinous consistence, such as my B. verniiforme 2 and the B. pediculatum of Koch and Hosaeus3, are not without possible parallels among Chlorophyceae, e. g. Naegeli’s Oocardium 4 and other Tetrasporeae. Moreover, it would seem probable that some of the Chlamy- dobacteriaceae have had a totally different origin from any of the other Schizomycetes, as is especially evident when forms like Phragmidothrix are compared with Bangia and its allies. The development of endospores has undoubted analogies with the formation of cysts in certain Flagellatae — e. g. Chro - mulina and Monas — as Migula has pointed out5, and there are several other cases. 1 Klebs, Die Bedingungen d. Fortpflanzung einiger Algen, &c., p. 255. 2 Phil. Trans., Vol. clxxxiii, 1892, p. 149. 3 Lafar, Technische Mykologie, p. 247. 4 Pflanzenfamilien, 1. Th., 2. Abth., p. 51, Fig. 33. 5 Pflanzenfamilien, 1. Th., 1. Abth. a., p. 11. Ward . — Some Thames Bacteria. 319 The Schizo-saccharomycetes, again, form a group which suggest obvious relationships to the yeasts, while Thaxter’s Myxobacteriaceae point to alliances with the Myxomycetes. All things considered, therefore, I think we should be prepared to accept that the morphological relationships of the minute organisms grouped together as Schizomycetes are neither few nor simple, and that their phylogeny is probably not even comparable with a complex tree-form, but is multiple in origin. 320 Ward . — Some Thames Bacteria . EXPLANATION OF FIGURES IN PLATES XX AND XXL Illustrating Professor Marshall Ward’s paper on Thames Bacteria. PLATE XX. I. Short Colourless Bacterium. Fig. i. Rodlets and cocci a actively growing on gelatine at ordinary tempera- ture ; b on gelatine at 20° C. ; c form from an old Agar-culture. Fig. 2. Gelatine plate-colonies at 20° C. a on the third day, b on the sixth day after making plates. Fig. 3. Culture from single rodlet. At 3 p.m. a rodlet (one of two) = 2 x 1 ft was fixed in gelatine at 20° C. a , at 10 p.m. this had divided b , and at n a.m. next morning it had formed the colony c: see p. 288. Fig. 4. Stab-gelatine, one month at 20° C. Fig. 5. Streak-gelatine, one week at 20° C. Fig. 6. Agar-culture, four days at 20° C. The layer in a consists of coalescent colonies shown slightly magnified in b. I II. Capsuled Coccus or Bacterium. Fig. 1. Plate-colonie^ at ordinary temperatures, a week old. a an emerged colony under ^ showing the characteristic streakings ; b a submerged colony natural size, and c the same under i showing the gland-like appearance; d a portion of a under E/4, showing the embedded zoogloea-masses. Fig. 2. Three submerged colonies, two days at 20° C. Fig. 3. A piece of Agar-culture, showing embedded zoogloeas : a under ^ ; b under XV imm. in water ; and c the same stained with methylene blue, showing the ‘capsule’ round the masses embedded in gelatinous matrix. The capsuled masses average 4-6 /x to larger and smaller : the organisms, 2xi/xtoixi^x. Fig. 4. Stained specimen after passage through animal ; preparation from plastic lymph ^ imm. Capsules hardly visible ; cocci 1 x 0-75 to 1 x 0.9 /x. Fig. 5. Similar from peritoneal fluid: the ‘capsules ’ visible. Cocci about 0-75 to I-OJU. Fig. 6. Similar preparation from Agar-culture stained by Gram’s method. ‘ Capsuled’ masses 6-10 ; rodlets 1 xo>6 to cocci about o*6 to 0-75 /x. Fig. 7. Colonies after one year’s rest, seven days’ plate at 18-20° C. Nat. size. Fig. 8. A ten days’ stab-culture showing commencement of the liquefaction. Fig. 9. A nine days’ streak-gelatine-culture. Fig. 10. An Agar-culture after revival. Five days at 35° C. Ward.- — Some Thames Bacteria . 321 Fig. ir. A potato-culture (revived), three days at 25° C. Fig. 12. a, a rodlet 2*5 x 1 /J. sown at 3 p.m. in gelatine-drop at 25° C ; b, the same, divided into four rodlets at 10 p.m. imm. Fig. 13. A similar culture showing the oscillating movements in the partially liquefied gelatine. In a the pair of rodlets together end to end ; b they have flown apart ; c come together again ; d again apart. This oscillating movement may concern both or only one rodlet at a time. Fig. 14. Similar culture, a rodlet fixed at 10.25 p.m.; b the same dividing at 11 p.m. ; and c oscillating apart at 11.5 p.m. At 7 next morning the colony df 25 jx in average diameter, had been formed. Fig. 15. a, a rodlet fixed at 4 p.m. This remained at 150 C. through the night : at 8.15 a.m. next day it had given rise to 8-10 bacilli b forming a minute colony 3 n diam. The temperature then slowly rose to 25° by 3 p.m., when the colony c had resulted, about 15 /x diam. At 8.30 p.m. the rodlets were actively swarming, but came to rest during the night. Fig. 16. A series showing division & c. of rodlets in gelatine, a at 10 a.m.; b at 10.20 ; c at 10.35 ; d at 11.40; e at 12.20 ; f at 12.45. t. = rising from 21*5° to 2 3. 50 (see p. 301). XV immersion. PLATE XXI. III. Rose-pink Micrococcus. Fig. 1. Plate-colonies at 15-20° C. for ten days : nat. size. „ Fig. 2. The same in a month : nat. size. Fig. 3. The same, six weeks, under £. Fig. 4. Stab-culture, five days at 20° C. Fig. 5. Agar-culture, three weeks at 250 C. Fig. 6. Potato-culture, fourteen days at 20° C. Fig. 7. Groups of cocci under Tax : a fresh ; b stained. Fig. 8. Broth-drop culture under XV, showing Sarcina-form : a at 11.50 a.m.; b at 2.30 p.m. ; c at 4 p.m. ; d at 9.50 p.m. Fig. 9. Characteristic groups from a strong culture at 20° C. on third day under showing the glandular botryoidal masses. Fig. 10. Gelatine-drop culture under imm. a at 5.55 p.m. ; b = 8.g p.m. ; £=11.40 p.m.; ^=9.10 a.m. next day; £ = 12 noon ; /= 2.10 p.m. ; ^=4 p.m. ; h~ 9 p.m. Temperatures fell from 20° to 160 C., then rose to 220 C. IV. A Pseudo-Bacillus. Fig. 1. Plate-colonies : a on fourth day at 15-18°, under £; b after fourteen days at 20° C. Nat. size. Fig. 2. Plate-colonies at 20° after seven days. Nat. size. Fig. 3. Rodlets from an old gelatine-culture, x*x imm. Fig. 4. Rods stained with methyl-blue, imm., showing spore-like bodies. Fig. 5. A rod a in broth-drop at 17-5-19° C., under imm., showing 322 Ward ’ — Some Thames Bacteria. growth, a at 6. 15 a.m. ; b at 8.40; c at 9.0 ; d at 10 ; e at 1 1.10 ;/at 12.30 noon ; g at 2 p.m. \ h at 7 p.m. ; j at 11 a.m. next day. Fig. 6. A series at 21-230 C. in gelatine-drop under E/4, a at 8 a.m. ; b at 8.50; c at 9.30 ; d at 10.30 ; e at 1 1.25 ; f at 12.50 p.m. ; g at 2.15 ; h at 3.50 ; i at 4.40 ; j at 8 p.m. ; k at 9 a.m. next day ) . Fig. 7. Preparations in broth a, b, and c, and stained with methylene blue d under -fa imm. Fig. 8. A potato-culture, three days at 20° C. Fig. 9. Gelatine-streak, three days at 20° C. Fig. 10. Agar-culture, a week at 250 C. tAsruzals of Botany I; Short colourless Bacterium. H.M.W. del WARD. - THAMES BACTERIA. Vol.Hl Fl.XX. II, Capsuled Coccus or Bacterium. University Press, Oxford. Annals of Botany H.M.W. del. Ill, Rose-pink Micrococcus. WARD. - THAMES BACTERIA. IV, A p s eudo - Bacillus. University Press, Oxford. On the Roots of Bignonia. BY T. G. HILL, National Scholar in Botany , Royal College of Science, London. With Plate XXII. PAVING had the opportunity of investigating the root- X 1 structure of a few species of Bignonia , it seemed desirable, as there exist some doubts as to whether the roots of these plants follow the anomalous structure of the stem, to put the results obtained on record. As far back as 1850 Criiger1 drew attention to the curious tuberous growths found on the roots of Bignonia Unguis , and he also stated that the same anomaly occurred in the roots as in the stem. This was subsequently denied by later investigators. Thus Bureau 2 says, ‘ M. Criiger ... II a suivi, dit-il, ce developpement sur le Bignonia Unguis. J’ai pu examiner une racine de cette meme espece, et je dois dire que mes observations ne s’accordent point avec celles de M. Criiger.’ Further on in his monograph he states: ‘On ne voit point le tissu ligneux s’arreter dans son developpement sur quatre points opposes en croix, et le liber remplir les vides 1 H. Criiger, Bot. Ztg., Vol. viii, pp. 109- iio. 2 Bureau, Monogr. des Bignoniacees, 1864, P* I49* [Annals of Botany, Vol. XII. No. XLVII. September 1898.] 324 HilL — On the Roots of Bignonia . ainsi formes : mais il y a neanmoins une penetration de 1’ecorce dans le bois.’ It may, however, be stated that Bureau remarks that he had some doubts as to whether his plant was really Bignonia Unguis and not Stigmaphyllon ; he decided it was the former, on account of the character of the sieve-tubes. Judging from further remarks which he makes on the root-structure of this plant, it is very probable that his material came from a wrongly named specimen. De Bary1 states that he was unable to find the characteristic stem- structures in the roots of Bignonia capreolata. Van Tieghem 2 states at the end of his description of the stems of these plants : ‘ Les racines de ces memes Bignoniacees ne paraissent pas presenter ces anomalies.’ Before passing on to the root-structure of these plants it will, perhaps, be well to briefly describe the structure which obtains in the stem for the sake of comparison, Secondary thickening begins quite normally, as in other dicotyledonous woody plants. Sooner or later the develop- ment of secondary wood slackens very much at four points arranged cross-wise ; and as the secondary growth of wood at the intermediate portions of the circumference continues at the same rate as before, it follows that four depressions are left dipping down into the secondary wood : but inasmuch as the formation of phloem at these four points is increased in inverse proportion to the decreased formation of wood, it is obvious that the depressions are quite filled up with phloem as far as the outer limits of the bast, so that the general external form of the stem remains similar to that of an ordinary plant. These four phloem-wedges are very charac- teristic of many Bignonias. The cambium at the bottom of each depression still slowly forms xylem-elements, and in the cortex of the stem numerous sclerenchymatous masses occur. An essentially similar structure obtains in the root. In the ordinary root four phloem-wedges are found, and their development is identical with those of the stem. 1 De Bary, Comp. Anat., Eng. ed., p. 573. 2 Van Tieghem, Traite de Botanique, p. 823. Hill. — On the Roots of Bignonia . 325 We will now turn to a more detailed description of the root-structure in various species. Bignonia Unguis. In this species the anomalous structure of the stem is undoubtedly present in the root. The structure of the normal root without any sunken phloem is illustrated by Fig. 1, the protoxylem and the medullary rays are well marked and quite typical. With this illustration might be compared Fig. 2, which is a somewhat diagrammatic repre- sentation of a transverse section of an older part of the root figured in Fig. 1. It will be observed that the formation of sunken phloem has gone on to some extent. The first indication of the sunken phloem makes its appearance in quite small roots from -8 to 1 mm. in diameter, with five to ten cells in the radial rows of secondary wood. Such a stage is indicated in Fig. 3, which illustrates a trans- verse section of a young root. It will be seen that that protoxylem ( pxy ) is well marked, and is typically that of a root both as regards structure and position ; secondary growth of wood has not gone on to any great extent ; the phloem exhibits very fine sieve-tubes and companion-cells; and finally the cortical sclerenchyma is made up of fibres lignified to a very great extent, so that the lumina are quite small. In these fibres very fine simple pits were to be seen in the preparation from which this figure was taken. On the other hand, in some roots with about 2 6 cells in the radial rows of secondary wood the process has not gone much farther; thus it seems that the development of the sunken phloem does not originate at the same time in equally developed roots. In roots of about 1*3 mm. diameter the development has proceeded until, with a diameter of 1-76 mm., the stage has reached that figured in Fig. 4. In some roots of a smaller diameter than the last, viz. 1-5 mm., the development has gone on to a greater extent, the sunken phloem being -28 to *3 mm. in depth (measured from the outer limit of the wood). Such a stage is indicated in Fig. 5. Criiger1 states that the anomaly is not so regular as in 1 Loc. cit. Z 326 Hill.— On the Roots of Bignonia. the stem. I have been unable to verify this. Most of my preparations show a very regular structure, e. g. those of the stage indicated in Fig. 5? and only in one case have indi- cations of six phloem -wedges been seen. The roots of Bignonia Unguis have a further interest in the fact that, at intervals, they swell out into tuberous growths resembling those occurring in the roots of certain species of Asparagus ; they may attain a diameter of about 1 cm. and a length of about ii cm. In a transverse section it is seen that there are many points of similarity between the tuberous and the ordinary roots. For instance, there is a well-developed periderm, and in the cortex there are numerous masses of sclerenchymatous elements. In the tuberous roots, however, they are arranged more regularly in concentric rings, and the masses also grow smaller towards the periphery ; the sieve-tubes and com- panion-cells are well marked ; and finally there are well- marked protoxylem-groups. The chief points of dissimilarity between the tuberous and other parts of the root lie in the great development in the former of parenchyma in the cortex and, to a lesser extent, in the pith ; and also in the breaking up of the xylem into separate masses, often by a certain amount of dilatation parenchyma. There can be no question as to the structures described above belonging to the roots. In the first instance the material was carefully examined for any evidence of stem- nature in the shape of buds, &c., but without any success ; then again the tuberous growths on the root are characteristic of the genus, the externally placed protoxylem is typically that of the root, and finally no phloem was found opposite the groups of protoxylem. Bignonia venusta closely corresponds with Bigonia Unguis in the possession of sunken phloem in the roots. The first indications of the anomaly were found in roots of about •97 mm. in diameter : in other roots of about 1 mm. diameter the phloem-wedges were about six cells in depth, while in Hill. — On the Roots of Bignonia. 32 7 roots of i*4 mm. in diameter the depth was about twice as great. Bignonia capreolata is the only other species which has been examined ; its roots do not show the anomalous structure of the stem ; and although my material reached a diameter of 3 mm. and the outline of the phloem was distinctly waved, still there were no indications of the forma- tion of sunken phloem. The general structure of the root resembles that of Bignonia Unguis, the chief differences being the great amount of cortical tissue, and the less abundant cortical sclerenchyma not grouped together in masses, but in many groups with a few fibres each, in this species. 328 Hill . — On the Roots of Bignonici . EXPLANATION OF FIGURES IN PLATE XXII. Illustrating Mr. Hill’s paper on Bignonia. All the Figures refer to Bignonia Unguis. Abbreviations : camb., cambium ; c. c., companion-cells ; co., cortex ; cr. ph ., crushed phloem; m. r., medullary ray; ph„ phloem; pxy protoxylem ; p. w., wedge of sunken phloem ; scl., sclerenchyma ; xy., xylem. Fig. i. Transverse section of a root. Fig. 2. Transverse section of an older part of the same root as the last (somewhat diagrammatic). Fig. 3. Similar section from another root. Fig. 4. Transverse section of an older root showing an early stage in the development of the sunken phloem. Fig. 5. Diagram of a transverse section of a root showing the sunken phloem well developed. Cound. University Press, Oxford . Cupressinoxylon vectense; a fossil Conifer from the Lower Greensand of Shanklin, in the Isle of Wight. BY C. A. BARBER, M.A., F.L.S., Lecturer in Botany in the Royal Indian Engineering College , Cooper's Hill . With Plates XXIII and XXIV. Introduction. IN the study of the anatomy of plants few subjects have received such minute attention as the structure of Coniferous wood. The simple and unique character of the elements composing such wood has rendered it peculiarly suitable for examining the growth of timber, the effect of seasonal variations and the formation of annual rings, the ascent of water, and other life-problems. But perhaps a greater inducement, at any rate among the older writers, was found in the useful and ornamental character of many of the fossil woods — coal, lignite, agate, opal, amber — which led to the formation of collections of these substances and their examination by the curious. It is hardly necessary to point out that the bulk of these woods belonged to the Coni- ferous type, this being due to their much wider distribution [Annals of Botany, Vol. XII. No. XL VII. September, 1898.] 330 Barber . — Cupressinoxylon vec tense. in former periods than at present, and the comparatively late appearance of the Dicotyledons in geological time. Unfortunately the conditions required for the preservation of the minute anatomy of pieces of stems and roots of fossil plants differ widely from those under which the leaf- and bark- impressions have been handed down to us, and the two are therefore hardly ever found together1. This is not entirely due to the phenomenon of leaf-fall. The imprinting of the delicate venation of decaying leaves is largely a mechanical process, and pictures to the mind still lagoons and slowly-flowing streams ; whereas the fragments of wood frequently bear the marks of much tossing to and fro before they were subjected to the chemical replacement of their cell- walls by silica, lime, or iron pyrites. It is comparatively rare that the bark or cortex is left in fossils, and it thus happens that the sole guide to their systematic position is to be found in the minute structure of the wood. The uniformity of the elements in Coniferous wood has already been noted, and the problem here presented is seen to be of no common difficulty. Some of the best workers on the anatomy of plants have devoted their energies to its elucidation. Unger, Schleiden, Schacht, Mohl, Goppert, T. Hartig, Mercklin, Cramer, Kraus, Stenzel, Schenk, Con- wentz, Felix, and many others, have produced a mass of monographs and treatises, almost all of them originating in an attempt to describe fossil woods, the necessity soon arising of extending their researches to an exhaustive study of recent Conifers. In spite of these laborious investigations, the results have frequently been anything but satisfactory. Much labour was wasted by the earlier writers in attempting to unite each fossil wood with leaf or fruit remains occurring in strata of the same age. It was however soon demonstrated that classifications founded upon the structure of the wood could not be made to agree with those recognizing the natural 1 Schenk, in Zittel, Palaeophytologie, p. 873, mentions Sequoia Couttsiae among Coniferae. See also Elate austriaca in Unger, Chloris Protogaea, 1847. Barber. — Cupressinoxylon vec tense. 331 affinity of plants. The independence of these systems was first insisted upon by Goppert, and to him therefore we owe our first real advance in the subject1. Goppert instituted certain classes or types of wood which he termed c genera,5 but which, so far from being synonymous with the genera of systematists, frequently united members of the most widely separated groups. Goppert’s genera, emended by Kraus 2, form the basis of all recent classifications of existing and fossil Coniferous woods. These may be arranged according to the following types : — 1. Araucaria Type. Araucarioxylon. Bordered pits small, touching, usually mutually compressed, several rows in each tracheide, the pits in adjacent rows placed alternately. 2. Cupressus and Abies Type. Cupressinoxylon and Cedro - xylon. Bordered pits separate, in one row, or, if in more than one, the pits of different rows opposite one another, resin-canals absent, but strands of wood-parenchyma contain- ing resin. In Cupressinoxylon the resin-parenchyma is abundant, in Cedroxylon scarce or absent. 3. Pinus Type. Pityoxylon. Bordered pits as in 2, resin- ducts with sheaths of wood-parenchyma among the tracheides and in the medullary rays, no separate wood-parenchyma strands present 3. 4. Taxus Type. Taxoxylon. Bordered pits and wood- parenchyma as in 2, no resin-ducts, tracheides with well- marked spiral thickenings. Of these four types the first two are far more frequent in the fossil condition, the Araucarian with its allies extend- ing from the Devonian to the present time, while Cupressino- xylon and Cedroxylon are more characteristic of the Secondary and Tertiary periods. 1 Goppert, De Coniferarum structura anatomica, Breslau, 1841. 2 Goppert, Monographic der fossilen Coniferen, Leiden, 1850; Kraus, Mikro- scopische Untersuchungen iiber den Bau lebender und vorweltlicher Nadelholzer, Wiirzburger Naturwissenschaftliche Zeitschrift, v, 1864. 3 Schroeter, Untersuchungen iiber fossile Holzer aus d. arct. Zone, FL fossil, arct., Oswald Iieer, vi, Zurich, 1880, 332 Barber,— Cupressinoxy Ion vec tense. It is doubtful how far it is possible to separate the last two forms of wood. Strasburger states that resin-parenchyma is to be found in all Coniferous wood except that of Taxus 1. Beust asserts that the frequency of resin-parenchyma cannot be used in classification, the only division possible being into those with and those without it 2. Certainly its absence from a few sections of a fossil would not demonstrate its typical absence from the species. Thus Felix, examining a specimen described by Schenk as Cedroxylon , because he found no parenchyma, was compelled to alter its name to Cupressinoxy Ion 3. It becomes very difficult to draw a sharp line between the two types because of the many transitional cases. In young branches of Cedrus Deodara and Cryptomeria japonica examined by myself, the number of wood-parenchyma-cells per sq. mm. of transverse section was greater in Cedrus than in Cryptomeria , being fairly abundant in both. Yet Cedrus is included in the Cedroxylon type and Cryptomeria is a typical Cupressinoxy Ion, Leaving out of consideration the collection of families in- cluded under Araucarioxylon , it is significant that Cupressino - xylon includes, among living forms, Cupressaceae, Taxo- diaceae, PodocarptLS , Dacrydium , Saxogothaea , Phyllocladus , Ginkgo , Abies Webbiana. The ‘genus’ therefore comprises the Cupressaceae in general as well as members of the Abietaceae, Taxodiaceae, and Taxoideae. Among Abietaceae are to be found genera or species belonging to Cupressino xylon and Pityoxylon , and among the Taxoideae are included Taxo - xylon and Cupressino xylon . These facts make it sufficiently clear that the character of the wood alone is of little taxonomic value. At the same time it cannot be altogether neglected. The section Pityoxylon offers characters distinct enough for determining 1 Strasburger, Ueber den Bau und die Verrichtungen der Leitungsbahnen in den Pflanzen, Histologische Beitrage, iii, 1891. 2 Beust, Untersucliungen liber fossile Holzer aus Gronland, Allgem. Schweiz. Gesellschaft, neue Denkschrift, xxix, 1884. 8 Felix, Studien liber fossile Holzer, p. 29, Inaug. Diss., Leipzig, 1882. Barber . — Cupressinoxylon vec tense. 333 the principal groups of Pinus and a certain number of species 1. Under Cupressinoxylon , the genera Sequoia , P hyllodadus) Ginkgo , Taxodium and Glyptostrobus are said by different authors to possess characters by which they may be separated from all others 2. A more careful study of the structure of gymnospermous wood might have placed this group between the Angiosperms and the Cryptogams instead of with the Dicotyledons as was done at first3, just as Mercklin separated Sequoia from the Taxaceae on account of its wood-structure at a time when it was regarded as a member of that group 4, and as Kraus detached Dammar a from Cunninghamia 5. In spite of the fact that the woods grouped under Cupres- sinoxylon are drawn from such widely different groups of Coniferae, there is a truly remarkable similarity in their structure. Thus Kraus mentions forty-six species of recent Conifers, belonging to various groups, with woods of this type which he states are indistinguishable. This is the more to be regretted because the bulk of fossil Coniferous woods from the Jurassic to the Tertiary periods, and especially the enormous numbers found in the Brown Coal, belong to Cupressinoxylon. The earlier workers do not appear to have appreciated these facts, and their diagnoses were lamentably meagre. They do not, moreover, appear to have had any adequate knowledge of the process of petrifaction, nor were they aware of the differences existing as regards wood -structure between different parts of the same plant. Branches, roots, stems, portions altered by chemical action or undergoing decay, were accordingly described as separate species, so that Kraus, in his analysis of work previous to 1864, ventured the extraordinary assertion that all Cupressinoxyla described 1 Schimper, Traite de Paleontologie vegetale. 2 Schmalhausen, Beitr. zur Tert. FI. Siidwest-Russlands, Pal. Abh., Dames und Kayser, i, 1884. See also Kraus, Schroeter, Schenk in Zittel, and Goppert, 1. c. 3 Knowlton, Fossil Woods and Lignite of the Potomac Formation, Bull. No. 56, U. S. Geological Survey, 1889. 4 Mercklin, Palaeodendrologikon Rossicum, St. Petersburg, 1855. 5 Kraus, 1. c. 334 Barber. — Cup res sin oxy Ion vec tense. up to that date might, from their diagnoses, very well belong to different parts of one and the same tree b Many of the characters formerly used in descriptions, such as the width of the annual rings, numbers and heights of medullary rays, sizes of tracheides, thickness of walls, &c., have now been demonstrated to differ as much in the same species as in different genera. To illustrate this fact, Kraus carefully examined a fossil stem with branch attached. He showed that these two parts differed so widely in the characters just enumerated that, following the lines of the old-time diagnoses, they would have to be put into well- separated ‘ species.’ The detailed study which has led to these results has indeed characterized the period inaugurated by Kraus (1864). Sanio, Schroeter, Schmalhausen, Russow, Kny, Schulze, Wille and others, have supplemented their general comparative study of many forms by a minute, intensive examination of individual species. In the process the earlier descriptions have been thrown into confusion, but the path has been marked out along which any work of value in this field must be followed. It is true that these writers have shaken, one by one, the pillars upon which the classification of Coniferous woods has been erected ; their work has, in this sense, been mainly destructive. The characters of absolute importance have, however, been rigidly defined, new characters have been raised from relative to absolute value ; and, what is perhaps of equal importance, the conditions have been determined under which the characters of relative value may be used in diagnoses. These conditions include a knowledge of the age and morphological character of the part as well as the mode of petrifaction. It is nevertheless only too apparent that the work of reconstruction has only commenced, and any description of fossil wood which is to be of use in the future must include an exhaustive analysis of all the characters, both absolute and relative, which have not been proved to be purely fanciful. 1 Kraus, 1. c. Barber . — Cupressinoxylon vectense . 335 In view of this formidable and discouraging literature — an abyss in which such a vast amount of useless labour has been sunk — the description of a new ‘ species ’ of Cupressinoxylon is not an enviable task. But, as will be seen, it is possible to determine fairly accurately the age and morphological character of the specimens selected for description in the present paper ; and it becomes in this way permissible to submit them to the same exhaustive analysis as would be made in the study of a recent wood. The pieces of fossil wood dealt with are taken from a series of silicified, rolled, and water-worn fragments, much perforated by teredo-like burrows, collected from the Lower Greensand of Shanklin in the Isle of Wight \ Some of these specimens are in a beautiful state of pre- servation, showing not only the bordered pits on the radial walls with great distinctness, but even their sections in tangential view with what appears to be the suspended torus. Since a pith is to be found in all of them, it is easy to determine their approximate age ; and, while it cannot be definitely stated which are stems and which are branches, certain of them show in a marked manner the arrangement of cells in the annual rings which are characteristic of roots. It has, further, been possible to obtain in each specimen several sections in the transverse, tangential, and radial directions (thirty-two sections in all), and by this means some of the minor peculiarities of individual parts have, it is hoped, been ruled out. 1 I am indebted to Mr. A. C. Seward, F.R.S., for the first and most beautifully preserved specimen, and to Dr. D. H. Scott, F.R.S., for the others. In both cases the specimens were accompanied by sections to which I have added a considerable number. The slides dealt with in the present paper have been numbered A. C. S. 4, 5, 6 ; D. H.S. 338, 344-349 5 C. A.B. 1, 2, 3, 7, 8, 9, 11, 12, 13, 17, 18, 23, 24, 29, 30, 35, 37_41 2* These sections have been cut by Mr. F. Chapman. As regards the expense of section-cutting, this has been defrayed out of a small Government grant from the Royal Society. By this means I have been enabled to obtain a far more complete series of sections than would have been the case otherwise. 336 Barber. — Cupressinoxylon vec tense. Description of Specimens. Specimens t-2 inches in diameter, with distinct, central, or excentric pith. Rings of growth well marked, averaging 1-2 mm. wide, composite, each with 2-6 bands of narrow dark summer elements. From the arrangement of the cells in the rings the specimens are considered to be young branches and roots. Pith well preserved, diameter in branches -9 mm., in roots *3 mm. to *4 mm., cells increasing in size towards the centre, 1 o /x to 50 /x in diameter, copiously pitted, with large triangular intercellular spaces 5-1 5 /x across. Medullary sheath : in the roots the rows of tracheides pass directly into the cells of the pith, in the branches they terminate in small groups of' cells irregularly arranged. Spring tracheides not differing much in branch and root, tangential width 12-25 M, radial 17-22 yuc ; summer tracheides, in the branches, radial diameter 10-ii/u, averaging 4-6 rows, in the roots radial diameter 12/x, averaging 2-4 rows. Bordered pits in a single row (rarely double in roots), free and rounded in branches, often touching and compressed in roots, outer diameter 7-14 /x, inner 3—5 /ot. Tangential pits frequent, occurring in 2-7 rows of summer cells, outer diameter 5-7 /x, inner 2-3 ft . Medullary rays simple, usually one, occasionally two cells broad, 1-16 cells high, the average being 2-3. Cells of ray 15-20 /x high, 12-16 /x broad, radial length various, covering 2-6 tracheides. Proportion of medullary-ray-tissue to the rest of the wood about 1 : 30. Resin-tissue, consisting of isolated rows of parenchymatous cells, abundant, equally distributed. Length of cells various, tangential width 19-28 /x, radial 11-20 /x. It has not been found possible to place this wood under any species of Cupressinoxylon already described. The peculiarity of the rings of growth is probably in itself sufficient to establish a new species. Besides this, however, the youth of the specimens and the detailed examination to which they have been subjected, leave few points of comparison with known fossils. It will be better to defer Barber. — Cup ressinoxy Ion vectense . 337 these comparisons until older specimens presenting the same peculiarities have been examined. From the frequent occurrence of this type of wood in the Lower Greensand of the Isle of Wight, I have decided to call it Cupressinoxylon vectense. Annual Rings. The rings of growth are so peculiar as to deserve an ex- tended description. They are compound in all the specimens (Fig. 1). They are readily visible to the naked eye, and each is seen to consist of a broad, inner, clear, and a narrow, outer, denser portion. The former, under a low magnifying power, is seen to consist of wide and more or less thin walled * spring ’ elements. The latter is, however, not homo- geneous, but is composed of a varying number of dark lines in a lighter matrix. The dark lines are narrow bands of thick-walled, flattened cells, and the lighter parts between them consist of wider cells with thinner walls resembling the cells of the inner portion of the ring but not so large (Fig. 2). Under the microscope the dark bands of each series strongly resemble the ‘autumn,’ ‘late,’ or better ‘summer’ wood of living trees ; but they are seen to vary much in extent and thickness, frequently anastomosing with one another or fusing together at short distances along the circumference of the section. What is one thin band at one place breaks up to from two to five in another or increases greatly in thickness. Besides these regularly grouped bands of flattened cells, other less definite ones may be traced at various points in the wide zone of spring tracheides. A close examination of this part of the ring shows that there are frequent changes in the radial width of the elements and the thickness of their walls, sometimes constant for some distance or indeed all round the section, at other times limited to a few cells only. In the majority of cases the narrow dark bands would be classed as ‘ false ’ or ‘ partial ’ rings. They show a gradual 338 Barber . — Cttpressi noxylon vec tense. transition from wide cells within to narrow thick-walled ones, the latter passing again gradually into wide thin-walled elements outside (Fig. 2). The outermost band in each series usually presents the character of a ‘ true,’ ‘ normal 5 or ‘ sudden ’ ring, that is a sudden change from flattened summer- cells to wide succeeding spring-tracheides. This sudden change in the character of the elements is usually regarded as correlated with a profound change in the vegetative activity of the plant ; and for this and other reasons the broad rings of growth described above have been treated in the present paper as rings in the ordinary sense of the term. It is not to be supposed that these rings of necessity represent annual periods of growth, for there are sometimes deviations from the above arrangement which would render such an interpretation full of difficulty. Thus it occasionally happens that the outermost dark band of a series is gradual, and more frequently one or more of the inner bands have the character of sudden rings. The same band at different parts of the section will change from sudden to gradual several times over. Lastly, in a few instances, a sudden ring appears in the midst of the wide spring-elements, can be traced for a short distance, and then completely dis- appears (Fig. 4). Taking into consideration this extra- ordinary variability, it will be seen that we have no proper data regarding the rate of growth of the plant to which the specimens belonged, especially as transverse sections taken at short distances from one another show great differences in the arrangement and composition of the rings of growth. The adoption of the term ‘ annual rings’ in the present paper is to be regarded as a matter of convenience in describing the other characters of the wood. This is not the place to review the voluminous literature dealing with the effect of climate on the structure of wood x. 1 For references to the principal papers dealing with this subject see especially Biisgen, Bau und Leben unserer Waldbaume, Kap. vii, Jena, 1897; Nordlinger, Deutsche Forstbotanik, i, 1874; and Seward, Fossil Plants as Tests of Climate, C. v, London, 1892. Barber. — Cupressinoxylon vec tense. 339 It is important rather to determine how frequently irregu- larities of the nature described occur in recent and fossil species. There appear to be no analogous cases among fossil woods already investigated. Conwentz mentions that in the amber-producing Finns succinifera the rings are distinctly visible to the naked eye, but under a lens they are found each to consist of a series of narrow rings, and he states further that he remembers having seen cases of a similar nature in other fossils 1. Fliche figures an anasto- mosis between two rings of Cupressinoxylon infracretaceum 2, and Seward notes the presence of partial rings in his recently described Pinites Ruffordi 3. But besides such isolated examples there appears to be nothing resembling our specimens among described fossils. It is well known that in Dicotyledons of warmer regions the ring-formation, so regular with us, is deficient or absent. Michelia and Avicennia may be selected among many as resembling our fossil in the sharp character of the stem- rings, which nevertheless anastomose with one another, and thereby render the counting exceedingly difficult. When however we turn to recent Conifers, we meet with much less irregularity. It has been considered worth while, in books devoted to forest-botany, to record cases of partial or indistinct ring-formation in this group. The anomaly is seen to be widely extended. Of more immediate interest to us are such plants or parts of plants in which this peculiarity occurs, if not habitually, at any rate more fre- quently than elsewhere. We are told that branches are more subject to irregularities in wood-formation than stems4. In roots again, according to Nordlinger, it is frequently impossible to count the rings 1 Conwentz, Monogr. d. baltischen Bernsteinbaume, 1890, p. 32. 2 Fliche, Note sur les Nodules et bois mineralises trouves a Saint-Parres-les-Vaudes (Aube) dans les gres verts infracretaces, Mem. de la Soc. Acad, de l’Aube, lx, 1896. 3 Seward, Pinites Ruffordi from the English Wealden Formation, Journal Linnean Society, Bot. xxxii, 417. 4 Felix, Beitr. zur Kenntniss foss. conif. Holzer, Engler, Botan. Jahrb. iii, 1880, p. 265. 340 Barber . — Cupressinoxylon vectense . of growth even in those plants which have rigidly-defined rings in their stems and branches h As will be seen directly, our specimens probably include both branches and roots. I have observed gradual rings in branches of Abies Pinsapo, Cryptomeria japonic a, juniper us Virginian a, . and especially Juniperus communis , although the rings are well separable and easy to follow. The sections of Cupressus sempervirens in Nordlinger’s famous series, on the other hand, show a con- fused arrangement of gradual and normal rings frequently anastomosing with one another, and rendering the counting a matter of considerable difficulty. The resemblance between these sections and our fossil is very striking. In the other cases just mentioned the gradual rings may be regarded as exceptionable disturbances in the wood-formation ; in Cu- pressus sempervirens the irregularity appears to be the usual state of affairs. Thus Hartig describes two branches of this Cypress, one fifteen to twenty years old, showing only three rings of growth, and one twenty-five years old with nine rings2. Through the kindness of Mr. Thomas Hanbury I have been able to examine the branches and roots of two trees of this species from his garden at La Mortola. A study of these specimens confirms the previous observa- tions. There are the gradual and the sudden rings, the anastomosis of the narrow dark bands, and even a tendency to the formation of bundles of small rings, usually two together, alternating with clear zones, which forms so marked a character of our fossil specimens. Such irregularity as this cannot be referred to external conditions, and, as we have seen, the rings of this Cupressus do not of necessity correspond to definite periods of time. I have accordingly felt justified in including the irregularity in the formation of the annual rings among the distinguishing characters of the fossil under discussion. We are sometimes able, from a study of the arrangement of the tracheides in a transverse section, to tell whether it 1 Nordlinger, 1. c. 2 Hartig, Vollst. Naturgesch. d. forstl. Kulturpflanzen Deutschlands, p. 86. 34i Barber Cupressi noxylon vec tense. is taken from root, stem or branch. In the rings of a Coni- ferous stem, according to Mohl \ there is a gradual decrease in the radial diameter of the cells from the first tracheides of the ring to the last, accompanied by an increase in the thickness of the walls. In a root-ring, on the other hand, a number of layers of wide spring-elements are succeeded suddenly by a band of flattened thick-walled summer-cells which terminate the ring. This difference is explained in the following way. In a wide ring, whether in stem, branch or root, there are three distinct zones which differ as regards the form and size of their cells — an inner layer of wide square cells, a middle zone of polygonal, usually hexagonal cells, and an outer zone of typical summer-cells which are again more or less rectangular. Stems generally have wider rings than branches or roots, and all three layers are met with in them. In branches the inner layer is variable, and in thin rings is frequently absent. In roots on the contrary it is the middle layer which varies, and in thin rings it is usually wanting. The thin rings which characterize smaller roots and branches are thus seen to be very different in appearance. In branches a set of hexagonal cells passes gradually into the thick, flat summer-cells ; in roots wide, square tracheides are followed without transition by the flatter layer, which, in this case, is often reduced to only two or three rows of cells. On applying these facts to our fossil sections we note that they can with little difficulty be divided into two sets (Figs. 2 and 3). In one of these, which I have called branch (1) and branch (2), the cells of the inner layer are polygonal and pass gradually outwards to from 2-12-20 summer cells. In the other specimens, while in places there are gradual transitions, many parts show a sudden passage from wide, frequently squared cells, to from 1 -2-4-6 flat dark tracheides. These I have called root (1) and root (2). It seems therefore probable 1 Mohl, Einige anat.und physiol. Bemerkungen iiber das Holz der Baumwurzel, 342 Barber . — Ciipressinoxylon veciense . that the specimens are smaller branches and roots, an assumption which is supported by a number of other characters, although a few difficulties are met with. It should be pointed out in passing that this distinction between root- and stem-structure must be used with a certain amount of caution. If, as Mohl holds, it has its explanation in fundamental differences in the function of the two organs, it will be capable of wide application h There are, however, instances of stems exhibiting the typical root-arrangement. Thus Strasburger, in describing the last- formed layers of an old moribund Larch, states that the stem-rings are small and consist of very wide spring-tracheides followed suddenly by narrow thick-walled summer-cells. The same arrange- ment was found in the wide rings of a perfectly healthy Larch forty-eight years old. It is indeed stated by Nord- linger to be a common feature of Larch-wood 1 2. Lastly, Mold expressly remarks on the small difference between the outer stem- and root-rings of the Larch. It cannot then be con- sidered altogether safe to use this character alone in deter- mining the root- or stem-nature of fossil woods. To give an example, Cupressinoxylon distichum , described by Mercklin 8, is now regarded as a root, and yet the rings (which are stated to be the outer rings of a very thick trunk) correspond exactly with those of Strasburger’s old dying Larch stem 4. In one of the specimens the rings are wavy, i. e. thrown into a series of irregular arches around the section (Fig. i). This has been noted as an occasional occurrence in many recent Coniferae ; and among fossils in Citpressinoxylon nodosum (root) by Goppert 5 and C. erraticum Terebinum and C. Fritscheanum by Mercklin 3. The arches render the measurement of the elements difficult, since the cells are much smaller where two arches meet, as if these regions were subject to pressure. Possibly this is due to the nearness 1 Mohl, 1. c. 2 Strasburger, 1. c., p. 24. 3 Mercklin, 1. c. 4 See also Kobbe, Ueber die fossilen Holzer der Mecklenburger Braunkohle, Inaug. Diss., Rostock, 1887. 5 Goppert, Monograph. Barber. — Cnpressinoxylon vec tense. 343 of small branches, as in the case of Goppert’s C. nodosum. See also his figure of a Larch-stem with many small lateral branches and wavy rings h The width of the rings of growth as defined above is somewhat variable. This character is not now regarded as of diagnostic value, since the rings will vary in width according to climate, age, soil, and the part of the plant. As the same ring differs considerably in different parts of its course, the measurements have been taken in all directions where the structure has been undisturbed. Branch (1). Average width of first ten rings 1*7 mm., varying from *6 mm. to 2-8 mm. Branch (2). Average width of first ten rings 1-45 mm., varying from -7 mm. to 2*3 mm. Root (x). Average width of first twelve rings 1-07 mm., varying from -4 mm. to 2-i mm. Root (2). The rings near the pith are quite indistinct. Then follow five rings with an average width of 1*7 mm., varying from 1 mm. to 3-4 mm. After this five narrow, dark bands occur at distances of from -2 mm. to -8 mm. Pith and Medullary Sheath. The pith has received careful examination whenever the sections have admitted. It is obvious that from its study great assistance might be expected in determining the root- or shoot-nature of organs. This expectation was only par- tially realized. In the first place the usual absence of pith from pieces of fossil wood has caused it to receive small attention in the literature of the subject, and secondly the state of preservation of the cell-wall did not allow the markings to be clearly seen at the critical points. The diameter of the pith in branch (1) and branch (2) was -9 mm. ; in root (1) it was -4 mm., and in root (2) -3 mm. This is in agreement with Nordlinger’s statement that, while in Goppert, 1. c., Tab. i? Fig. io, A a 2 344 Barber,— Cupressinoxy Ion vectense . branches and in primary roots the pith is distinct, that of lateral roots is point-like if present at all 3. An examination of the ‘ medullary sheath or crown ’ points in the same direction. While branch (i) and branch (2) have well-marked groups of primary xylem where the tracheides lose the radial arrangement (Figs. 5 and 6), roots (1) and (2) show rows of tracheides passing directly to the borders of the pith (Fig. 7). A search for the primary xylems in longitudinal and oblique sections of these specimens has not been very successful ; but in the tranverse section of root (i) there is a distinct appearance of tracheides with spiral thickenings in the pith opposite several of the primary medullary rays. In the radial sections the appearances are not so con- vincing. In branch (1) the tracheides with bordered pits are easily seen to be in contact with the spiral ones, these latter adjoining the pith-cells. This is what one would expect in a section of a shoot. But there is every appearance that the same is the case in root (1). The state of preservation in root (2) and branch (2) prevents their sections from throwing any light on the question. The number of primary medullary rays varies in the sections. It is not possible to determine them accurately, but they seem to be between 2 and 7. The cells of the pith in transverse sections are largest in the centre, from whence they decrease outwards to the ‘crown’ of tracheides (Fig. 7). The diameter of the former may be taken as about 45 /u, while the outer medullary cells measure 10-20 /x. In longitudinal sections the inner cells are seen to be flattened transversely, so as to extend across the pith, while the cells near the tracheides are elongated in the same direction as the latter. All the pith-cells are richly pitted. Where undisturbed they are circular in transverse section, and have well-marked triangular intercellular spaces 5-10-15 /x across. 1 Nordlinger, Forstbotanik. Barber. — Cupressinoxylon vectense . 345 The Tracheides. The general arrangement of tracheides, as seen in transverse section, is fairly regular and similar to that in most woods of this class. Radial rows of broad cells alternate with narrow ones wedged in between them. The former represent the tracheides cut across at their broadest place, the latter the chisel-shaped ends with richly pitted walls. While the cells of each radial row are nearly constant in tangential width, neighbouring cell-rows differ very much. In order to measure the tangential width of the tracheides it is obviously impossible to isolate the cells and measure them individually at their broadest part, as has been done by Schulze for recent woods 1. All that can be done is to obtain a general average of the tangential width by counting the number of cells in a measured distance2. This number will depend partly on the length of the tracheides ; the shorter these are, the more frequent will be the rows of narrow ends. It has been already pointed out that the cells vary in their size in wavy rings, being apparently much compressed where the arches join. It has therefore been found necessary to multiply observations in order to obtain a true average. The results appear to be in accordance with the general laws laid down by Mohl, Sanio, Schulze, Kraus, and Conwentz. The following figures give the measurements of from 100 to 250 tracheides in each ring in eight different sections, the number of observations depending upon the state of pre- servation. They have involved the counting of over 17,000 cells. 1 Schulze, Ueber die Grossenverhaltnisse der Holzzellen bei Laub- u. Nadelholzern, Inaug. Diss., Halle, 1882. 2 Sanio, Ueber die Grosse der Holzzellen bei d. gemein. Kiefer, Pringsh. Jahrb. viii, 1892, p. 403. 346 Barber. — Cupressinoxylon vectense. Tangential width of spring-tracheides in \x\ — Rings 1 2 3 4 5 6 7 8 9 10 11 Branch (i) 12.5 18-8 20.7 22.7 23-7 24-3 24-5 24 24.7 Branch (2) 14 i7-5 20 21 23*5 24 25*5 Root (1) 12.7 16 20 23 23-7 25 24 25.2 24.5 25 24-5 In root (2), the inner rings not being well defined, the tracheides were measured at about equal distances from the pith, and in successive rings where these were apparent. Root (2) | 10 | 16.5 | 21 | 24 [ 26 | 26 | 22 J 24 | 22 ft. We thus see that the tracheides in the first ring are very- small ; from the second to the sixth a rapid rise takes place, and from the sixth to the eleventh the tangential width is fairly constant. The radial width of the tracheides has been calculated by measuring the first ten tracheides in each ring in ten as widely separated regions of the section as possible. The resulting average of 100 spring-tracheides is assumed to. be a fair guide. The measurement is, however, rendered difficult by the indefinite limits of certain of the rings, and the consequent necessity of selecting suitable rows for observation. No figures could be obtained for the first few rings, and in root (2) it was found necessary to adopt another method. Radial width of spring-tracheides in y : — Rings 1 2 3 4 5 6 7 8 9 IO II 12 Branch (1) 17.2 J>- 00 19.1 20-6 21.5 22.3 Branch (2) 20.4 21 23-3 22-2 Root (1) 19.2 19-6 20-9 20-3 21-6 22-5 22-1 20-5 20-2 Barber . — Cupressinoxy Ion vec tense. 347 In root (2) well-defined rings are absent near the pith. Besides this, where the rings are visible, it is found that the largest cells are in the middle of the ring, those at the commencement being considerably smaller. I have, accordingly, measured the ten tracheides at regular intervals from the centre or in the middle of each ring. Root (2) | 10-5 [ 14 | 20 | 25 | 26 | 32 ] 29 | 26 | 33 | 26 | 26 | 26 ] 28//. The branches show the increase in radial diameter which was to be expected. The roots show first an increase in width, followed by a decrease in succeeding rings. Without laying much stress upon this latter peculiarity, because of the fewness of observations, it may be pointed out that the arrangement is in accordance with the results obtained by Sanio. He showed that while in branches the increase in size of the tracheides is maintained until a constant is reached, in roots these elements increase in radial diameter during the first few years, then decrease, later on again increasing until a constant is reached \ It is important to note that, according to Mohl and other observers, the elements of roots are larger than those of stems and branches. This does not hold good in the present case, a fact which throws some doubt upon the division of our specimens into roots and branches 2. An examination of the summer-tracheides, on the other hand, tends to support this division. The summer-elements of roots are stated by Mohl to be fewer in number and less flattened than those of branches and stems 3, and this appears to be the case in our section. I have collected for comparison the most flattened elements of each ring, and have in each case tried to get as many in a radial group as possible (cf. Figs. 2 and 3). 1 Sanio, 1. c. 2 Nordlinger, Forstbotanik, mentions cases where there is not much difference between the size of cells in stems and roots (Aspen, &c.). 3 Mohl, 1. c. 348 Barber . — Cupressinoxylon vectense . Measurements of summer -tracheides. No. of groups measured. Branch (i) 20 Branch (2) 13 Root (1) 31 Root (2) 25 Average number of tracheides in a group. Average radial diameter of tracheides. 4*3 5-4 3 3 hi M 9*8 12 /X 12*2 fJL The walls of the tracheides are frequently seen to be striated in longitudinal sections. This is especially the case where, from the appearance of the bordered pits and the presence of fungus hyphae, it is evident that decomposition had commenced before petrifaction. The striation in these cases is probably due to this cause. In tangential sections, however, the summer-wood is seen to be striated even in well- preserved parts where the spring-wood is not — a phenomenon not due to decay, but often met with in the wood of recent Conifers. The thickness of the walls of the tracheides, a character which was once considered to be absolute, varies a good deal with the state of preservation of the part. An average of 80 measurements gives 7*3^, ranging from 4ju to 14/1. Bordered Pits. The bordered pits have received the utmost attention from students of wood structure. By means of these, Coniferous wood is easily separated from that of Angiosperms, and of the former the Araucarian type is cut off from the rest. There are also subordinate differences in the sizes of the bordered pits and the numbers of rows per cell in stems, branches and roots. Since in Coniferae they occur ex- clusively upon the radial walls, the study of radial sections has acquired great importance in diagnosis. The bordered pits in our sections are arranged in one row — probably in great part due to the youth of the specimens. A comparison of the radial diameter of the tracheides and Barber . — Cupressinoxylon vectense. 349 that of the pits shows that there is barely room for two rows of pits in one cell. We should not therefore meet with a double row of pits unless two tracheides were placed exactly opposite one another. This arrangement of cells, as already mentioned, is more frequently met with in the root-specimens, and we accordingly find one or two instances in the root- sections of two bordered pits occurring side by side at the same level on the wall of one tracheide (Fig. 10). In branch (1) and branch (2) the pits are usually quite free from one another, very rarely touching or slightly compressed. They are however much more abundant at the ends of the tracheides. In root (1) the pits, while generally free, are often in close contact, especially at the cell endings, the result being that their shape is oval with flattened upper and lower sides. The pits are sometimes flattened in root (2), but the structure is not well enough preserved to institute comparisons. Bordered pits in Coniferae usually show two concentric circles in surface view, the outer and inner borders. It has been customary to determine the width of these circles most accurately in the hope of thus obtaining data for the separa- tion of different plants. From the extended researches of Kraus1, Wille2 and others, it does not seem that there has been much progress made in this direction. In the present specimens much difficulty has been experi- enced because of the varying state of preservation. At first this caused a good deal of unnecessary labour. It was found that in parts of some sections the pits showed more than two concentric circles (Fig. 15) ; in other cases the two circles could hardly represent the outer and inner borders of the pit, because of the large and varying size of the inner one. The appearances referred to evidently point to different stages of decomposition, some of them closely resembling for instance those figured by Hartig in his description of 1 Kraus, Beitrage zur Kenntniss. fossil. Holzer, II, Zur Diagnostik des Coni- ferenholzes, Abhandl. d. naturf. Ges. zu Halle, xvi, 1886. 2 Wille, Zur Diagn. d. Coniferenholzer, lialle, 1887. 350 Barber. — Cupressinoxylon vec tense. the decomposition of spruce-timber by Polyporus borealis 1 ; and the far from rare occurrence of hyphae in the specimens shows that some fungus had been at work before petrifaction. The case referred to is exactly analogous to the decaying bordered pits of a piece of rotten Sorbus Aucuparia described by Kraus, where the inner circle increased in diameter according to the stage of decay until it fused with the outer2. What at first sight was regarded as a remarkable variability in the bordered pits was thus found to be due to the state of preservation of the specimen, and all measure- ments made in such parts were subsequently disregarded. The inner pore, which is circular, is not usually well preserved in the specimens. The average of a number of measurements in the best places was 3-5 //, these limits being rarely passed. More detailed results were obtainable with the outer border. Many hundreds of measurements were taken with a view to determine whether, as in recent woods, any clearly marked increase could be noted in the size of the pits from the pith outwards. The small pits in the summer wood were left out of consideration. The results are as follows : — Greatest diameter of Bordered pits. Branch (1), 3rd ring, 7-0 n; 5th, 9.5 /*; 6th, 11.3 ju ; 7th, 14-0/4. Branch (2), 3rd ring, 8-4/4; 4th, 10-5/4; 5th, 12*8/1; 8th, 14-0/4. Ditto in another section, at intervals, 9-7, 10-3, n-i, 12-0, 12*2, 13*7/4- Root (2), successive rings, 10-5, 13, 13, 13.3, 13*2, 12-5, 12*7, 14, 13*5/4. Root (2), successive regions, 10-4, 12*2, 12*3, 13-6, 13 /4. These figures show how incompletely a single number would suffice to express the size of the pits. From our knowledge of their varying size in different organs, and in the same organ at different ages, we should expect to 1 Hartig, The Diseases of Trees, English Edition, Fig. 124. 2 Kraus, Mikrosk. Unters. 35i Barber — Ciipressinoxylon vec tense. find a series of measurements in every recently described fossil, where the size of the specimen and the state of preserva- tion allowed of this. Yet I have only met with one instance in which such a series of observations has been made h The average diameter of the bordered pits in young roots and branches cannot be definitely stated. In the oldest parts of our sections it may be given as 13-14 /x1 2. Bordered pits on the tangential walls are present in all the sections, both radial and tangential (excepting in root (2), where the structure is not well preserved). According to Strasburger, tangential pits are found in the summer-elements of all Coniferous woods which do not possess 6 tracheidal elements * in the medullary rays 3. The function of the latter is to supply a radial passage for the water, and the tan- gential pits accomplish this from one ring to another, there being no room in the last summer-elements for the radial pits. In support of this view Strasburger shows that tan- gential pits occur in other parts of the ring where the normal passage of water from one tracheide to another is interrupted, e.g. when a new medullary ray is formed, or opposite a resin- duct. A good example of this is seen among fossil woods in Pinites Ruff or di , where a solitary tangential pit is figured — opposite a resin-duct4. In Ciipressinoxylon , with its purely parenchymatous medul- lary rays, one would look for tangential pits in all the ‘ species/ But they have only been recorded in six of the sixty species summarized by Beust5. A certain amount of perseverance is necessary to find the pits in some of our sections, although they are very well seen in the others. They are much smaller than the radial pits. 1 Conwentz, 1. c., p. 40. 2 Kraus, Zur Diagn., gives 15 \x + as the usual size in Abietinae and Cupressinae, the elements of young parts being smaller. Schenk, in Zittel, states that the full size is generally reached during the first ten years. 3 Strasburger, 1. c., p. 9. 4 Seward, 1. c., Fig. 6. 5 Beust, 1. c. Barber . — Cupressinoxylon vectense. 352 Measurements of Tangential Pits . Branch (1). Average of 36, outer diameter 5 [x, inner 2-35 /x. Branch (2). Presence easily noted, but not sufficiently clearly for accurate measurements. Root (1). Average of 26, outer diam. 7.3 fx, inner 2-3 jx. Root (2). Structure of wall not well preserved : tangential pits not met with. The numbers of rows of summer tracheides in each ring bearing tangential pits have been carefully determined for recent Conifers, and the different groups are seen to vary considerably in this respect. Kraus recommends observa- tions in transverse sections, but it is only possible in our sections to use the radial ones. In the latter the tangential pits are sometimes very clear. The section of the pit has the form of a sharply cut lens-shaped space, and not in- frequently the torus is found stretched across it as a short black line. In other cases this space is without the torus. In a third set of examples the lumen of the pit is filled up by infiltration, but the torus is well seen as a short black line in the thickness of the wall (cf. Figs. 11-13). These three conditions are often found close together, all manner of transitions being seen. It becomes quite possible then, after careful study, to make out the tangential pits in comparatively thick sections with the help of a substage condenser. This has been of the greatest service in counting the number of rows of summer-tracheides with these pits, because this part of the wood is generally very dense. The observations have been most successful in branch (2). Here the tangential pits are sometimes found in the last seven rows of tracheides. This abundance of tangential pits is in agreement with the other characters of our fossil wood, showing us that in the Cretaceous period, where Cupres - sinoxyla are first met with1, the type of wood found in Cupressus , Sequoia and Abies was developed in its minutest details. In Abietinae with composite medullary rays these 1 Schenk in Zittel, 1. c. Barber,— Cupressinoxy Ion vectense, 353 pits are much less abundant, and are rare or absent in the Pinus group. The torus, so well seen in certain tangential pits, is still more clearly defined in the larger pits on the radial walls (Figs. 11 and 12). These, examined in tangential sections, frequently show the torus as a dark line suspended in the lens-shaped cavity of the pit or closely applied to one side1 — positions which might lead to interesting speculations as to the condition of the wood at the time of petrifaction 2. Medullary Rays. The medullary rays figure conspicuously in every published description of fossil Coniferous wood. While in Dicotyledons they are composed of similar elements but differ widely in form, in Conifers they are very uniform in shape but differ greatly in their component parts. This has led to a minute study of them in the latter class of woods. In the first place there is their difference of composition. Some of them have tracheidal as well as parenchymatous elements, and this difference is regarded as an excellent diagnostic character. Those woods further which possess vertical resin-ducts usually also have radial ones in certain of the medullary rays. The latter are then compound and are more than one cell broad. In living Conifers which have no resin-ducts, the medullary rays are simple and consist of a single vertical layer of cells. The two-rowed medullary rays, although occasionally met with, are rare 3 ; but they are much more frequently found in fossil woods without resin-ducts, and this constitutes one of the differences between recent and fossil species 4. 1 The torus is well seen also in Sequoia canadensis , Schrtr., and Pinus succinic fera, Con wentz. 2 Strasburger, 1. c., p. 32 ; also Russow, Zur Kenntniss des Holzes, insbesonderheit des Coniferenholzes, Bot. Centralblatt xiii, 1883, p. 37. 3 Beust, 1. c. 4 Kraus, Mikr. Unters., mentions Cupressinoxylon jissum. Araucarian woods with two-rowed medullary rays are called Pissadendron. Among woods with resin-ducts, Strasburger mentions Larix as not infrequently having double medullary rays without resin ducts. 354 Barber. — Cnpressinoxylon vec tense. In the fossils under examination the medullary rays are uniform in structure, consisting of parenchyma alone ; they are also generally one cell thick (excepting in root (2), which should perhaps be separated because of this and other differences in the medullary rays). The characters mentioned thus far are of great importance in descriptions ; as regards the constancy of those to follow we are a good deal in the dark. The earlier writers, in describing the medullary rays, were content with mentioning whether they were simple or compound, and indicating the height they might reach in cells vertically above one another, e.g. 1-18, 1-6, 2-36, &c. But the uselessness of these numbers will be at once apparent to any one who has examined a number of tangential sections of Coniferous wood, the maximum height of the medullary rays depending largely on the time taken in their examination. It has moreover been demonstrated that in this respect medullary rays vary according to their distance from the pith and height up the stem ; and stems, branches and roots of the same plant may have medullary rays of different heights1. Nevertheless we do meet with cases where the height of the medullary rays is more or less characteristic of the wood. They are usually low in Thuja , higher in Cupressus than in Juniperus , in Abietinae they have half the number of cells they have in Cupressinae (except Thuja :), &c. It is there- fore necessary to include the height of the medullary rays in all diagnoses; especially in those cases where we have a know- ledge of the age of the part and its morphological value. A similar series of remarks is applicable to almost every one of the characters of medullary rays which different writers have from time to time attempted to introduce. Their frequency in the wood, the height and width and radial length of the individual cells, the thickness and pitting of the walls and the general shape in cross-section — - all these have been demonstrated to be valueless for the 1 Essner, Ueber den diagn. Werth der Anzahl u. Hohe der Markstrahlen bei den Coniferen. Abh. d. natur. Gesell. zu Halle, xvi, 1886. Barber . — Ctipressinoxylon vectense . 355 general separation of species, and yet there are certain genera which show marked and apparently constant peculiarities in one or other of these respects. A careful analysis is therefore necessary of the medullary rays and their cells. Such has been attempted in our sections, following as far as possible the lines laid down by Kraus, Essner, and Beust in their investigation of living forms. At the instigation of Kraus, Essner has shown that the numbers of medullary rays per sq. mm. of tangential section in any species is greatest in the first year, and that it decreases gradually outwards, while on the other hand the medullary rays are lowest near the pith and increase in height towards the cortex. This has led him to the idea that the number of medullary cells per sq. mm. may be constant for the same species in different years. In his attempts to obtain characteristic numbers for different species, he found, however, that individuals of the same species varied more than did those of different species. This character cannot therefore be considered absolute. Beust, arguing from the intrinsic importance of the medullary-ray-tissue in the life of the wood, regards it as probable that some relation exists between the relative bulk of this tissue and that of the tracheides. To obtain this proportion he has combined the several factors deter- mined by Essner. It is however not certain at present how far the proportion will hold good for the different types of wood. I have attempted to apply these observations to the fossil sections. It is not easy to determine accurately the age of the part in a tangential section of a fossil. In my specimens they have been cut as a rule between the sixth and the tenth rings of growth. The following figures repre- sent a great deal of labour. Their worth will not be fully apparent until it is possible to compare them with those of other specimens. As their value greatly depends upon the number of medullary rays observed, I have where possible noted this. 356 Barber . — Cupressinoxylon vec tense. The first (p. 357) table gives the number of medullary rays per sq. mm., as well as their heights in cells. All the slides here examined were tangential. The method adopted was to count the number of medullary rays in the field of the microscope, the area of the field having first been accurately determined. In the case of medullary rays partly included those only were reckoned which were more than half in the field. The second table (p. 358) gives the number of medullary- ray-cells per sq. mm., as well as the average area of each cell. These combined give the proportion of medullary-ray-tissue to the whole wood. The observations on the radial sections are given as control measurements. The areas are calculated from the tangential measurements alone. Each medullary- ray-cell is practically an ellipse. In measuring its width it was impossible to decide how much of the cell wall belonged to the cell of the ray and how much to the adjoin- ing tracheides. The whole thickness of the wall on one side was therefore added to the lumen. Pits were observed only on the radial walls of the medullary- ray-cells. In both radial and tangential sections these were seen to be simple. In shape they were oval and obliquely placed, usually two or one per cell, occasionally three, or more rarely four. The higher numbers were, as usual, on the outer cells, the middle cells having frequently one pit each. This is in accordance with the Sequoia type of wood. The longest diameter of the medullary-ray-pits could only be accurately determined in two sections. In branch (1) it varied from 3 to 10 yu, the average of fourteen pits being 6*5 /x. In root (2) the diameter varied from 6 to 12/01, the average of thirteen pits being 8 /u. Intercellular spaces were very clearly seen in the medullary rays, their distribution being as in recent woods. Resinous Parenchyma. There are no resin-ducts. Wood-parenchyma is however to be found among the tracheides* and this, from analogy TABLE I. Barber, — Cupressinoxylon vec tense. 357 Total number of medullary rays ex- amined. cO VO 10 . O CO co 00 VO 00 Ui No. of double medullary rays. • * CO 6= 1 p Average number of cells per medullary ray. VO ON O 9s . CO CO 10 0 CO CO VO CO CO 10 VO 10 10 00 VO 10 - CO ON M -T CO M t}- JC^c CO cO e* CO c* CO VO e* X>. VO c* 00 CO CO tJ- CO CO CO VO CO Tt- i>. n- ’'d- CO CO - CO ON CO O CO CO 0 CO J>* j>. •urai *bs .iad sAua &xz\ -ppaui jo -o$[ esi VO 10 Th ON •d- OV CO 10 CO OO irj ^ Section (i) ' Section (2) j Section (1) * Cl 0 0 (D CO ! Section (1) 1 Section (2) Section (1) [ Section (2) *4 PQ pq 2 .S ’’d oj £ 9* S tuo "5a .2 4- 2 G o CD O S-, CD ts-° g JJ - S .2 *C G2 5 o ^ -G Jg'ii S « 3 -5 G s-< CL) 3 S £ l B, G CD 'G 1/3 3 rG S O rQ CU 2-3 (U s £ g o 23 Dh «S 3 .2 nb 0) -M G 3 O O • cu S UM > a m v ^ Cu 3 «> u in 5| r— i <4H 3 0 nd m a .3 * ■gig I g ^ 8^§ O Sa £ s ^3 * 3 g > 2 Z ^ *2 * cfl 0 0s ca'g 2 > 3 3,2 S “ § 3 tuO ° P , « £ ■“ * S -O £ .5 « .» tyO g TD « ^ O 5&). We are I think justified therefore in concluding that the increase in the stainable material of the nucleolus is due to the absorption of stainable substance from the surrounding protoplasm. Wager . — The Nucleus of the Yeast-Plant . 535 The nucleolus now begins to divide. Its outline becomes slightly irregular, and the deeply stained granular mass becomes more prominent. Then an elongation of the nucle- olus takes place (Fig. 60), and we have gradually formed a long row of granules surrounded by a lightly stained blue substance, derived from the nuclear body, stretching across the cell either in a longitudinal or a transverse direction (Figs. 61-63). These granules gradually become separated into two groups by constriction, but they remain connected together for some time by a less deeply stained substance drawn out between them (Figs. 64, 65, &c.). Finally complete separation is effected and two daughter-nucleoli are produced (Fig. 68). Each one then divides again in the same manner, but in such a way that the line of division of one is perpendicular to that of the other, so that the two dividing nuclei often present the aspect of a cross (Figs. 69, 70, 72). Further divisions may take place leading to the formation of as many as eight nuclei (Fig. 71), but in most cells only four are produced (Figs. 73- 75). Each of the four nucleoli thus formed becomes surrounded by protoplasm and a thin cell-membrane, and thus constitutes a spore lying free in the remainder of the protoplasm (Figs. 76-78). The spores are at first very small, but they gradually increase in size at the expense of the surrounding protoplasm, a thick cell-wall being produced around each, until finally they completely fill the mother- cell, the wall of which at this stage is in consequence not easily visible. They are then mature and enter upon a resting- stage. The process of nuclear division just described may perhaps be regarded as a case of direct division in which the chromatin- substance is previously taken up into the nucleolus and separated out in the form of granules, which ultimately divide into two equal or nearly equal groups. But it may possibly be regarded as a very simple case of karyokinesis, if we look upon the granules as chromosomes, and the lightly stained substance which surrounds them during division as of the nature of a spindle-figure. The difficulty of observing all 536 Wager. — The Nucleus of the Yeast-Plant . the details of the division is, however, so great that one must be very cautious in attempting an explanation of the facts observed. Summary. It may be useful to give here a short summary of the conclusions at which we have arrived as a result of this investigation. 1. All cells of Yeast contain a nuclear apparatus. 2. In the earlier stages of fermentation this consists of a nucleolus in close contact with a vacuole which contains a granular chromatin-network, and exhibits a structure in many cases like the chromatin-network of the nuclei of higher plants. 3. In the later stages of fermentation the chromatin- containing vacuole may disappear, its place being taken by a granular network or a number of chromatin-granules, which may be disseminated through the protoplasm or grouped around the nucleolus. 4. The nucleolus is present in all cells. It appears to be a perfectly homogeneous body, which may, however, at times appear granular owing to the granules around it. 5. In young cells numerous chromatin-vacuoles are often found. These appear to fuse together to form the single vacuole which occurs in cells during the early and sometimes later stages of fermentation. 6. In the process of budding, the division of the nuclear apparatus does not exhibit any definite stages of karyokinesis. It must, I think, be regarded as a direct division of the nucle- olus into two equal or nearly equal parts, accompanied by division of the chromatin-vacuole, network, or granules. 7. The nucleolus divides either in the neck joining the bud to the mother-cell, or more rarely in the mother- cell itself, one of the products of division passing subsequently into the bud. 8. In spore-formation, the chromatin disseminated through the protoplasm becomes absorbed more or less completely Wager . — The Nucleus of the Yeast-Plant . 537 into the nucleolus, which then divides by elongation and con- striction into two. • During the division deeply stained granules (chromosomes ?) appear surrounded by a less deeply stained substance, which remains for a time connecting the two daughter- nucleoli together. This may perhaps indicate a simple inter- mediate stage of karyokinesis. 9. Subsequent divisions take place resulting in the formation of four (sometimes more) nucleoli. Each nucleolus becomes surrounded by protoplasm and a delicate membrane, and thus the spores are formed standing free in the remainder of the protoplasm. 10. The spores are at first very small, but they soon increase in size ; the surrounding protoplasm becomes used up ; the spore- membranes increase in thickness until at last in the mature condition they completely fill the mother-cell. 11. In S. Ludwigii and vS\ pastorianus the structure of the nuclear apparatus is similar to that in vS. Cerevisiae , and its division during the process of budding appears to be also the same. O O 2 538 Wager. — The Nucleus of the Yeast-Plant. List of Papers Quoted. Nageli, 1844 : On the nuclei, formation and growth of Vegetable Cells. Trans- lated from Schleiden u. Nageli’s Zeitschrift f. Wissenschaftliche Botanik, 1844, by Henfrey, Ray Society, 1845, p. 224. Schleiden, 1849 : Grundziige der wiss. Botanik, 1849, p. 207. The principles of Scientific Botany ; or, Botany as an Inductive Science ; trans. by Lankester, 1849, p. 37. Brucke, 1861 : Die Elementarorganismen : Sitzungsb. d. k. Akad. d. Wiss. zu Wien, Math.-Nat. Classe, Vol. xliv. Schmitz, 1879 : Untersuchungen fiber die Zellkerne der Thallophyten ; Verhand- lungen des naturhistorischen Vereines d. Preuss. Rheinl. u. Westfalens. Band xxxvi. Errera, 1882 : L’epiplasme des Ascomycetes et le Glycogene des Vegetaux. These, Brussels, 1882. — 1885 : Sur l’existence du Glycogene dans la levure de Biere. Comptes- rendus, t. ci. p. 253, 1885. Krasser, 1885 : Ueber das angebliche Vorkommen eines Zellkerns in den Hefezellen : Oesterreich. Bot. Zeitschrift, 1885. Zalewski, 1885: Ueber Sporenbildung in Hefezellen: Verhandlungen der Krakauer Akad. d. Wissenschaften, Math.-naturwiss. Section, Bd. xiii, 1885. See Bot. Centralbl. xxv, 1886. Hansen, 1886 : Recherches sur la physiologie et la morphologie des ferments alcooliques (Meddelser fra Carlsberg Laboratoriet, Bd. ii, p. 152). Strasburger, 1887 : Botanisches Practicum, 1884 and I887, p. 339. Zacharias, 1887 : Beitrage zur Kenntniss des Zellkerns und der Sexualzellen. Bot. Zeitung, 1887 (p. 298). Zimmermann, 1887 : Die Morphologie und Physiologie der Pflanzenzelle, Breslau, 1887, p. 25 (figure on p. 23). Lindner, 1887 : Wochenschr. f. Brauerei, 1887, p. 773. See Jour. Roy. Mic. Soc., 1888, p. 156. Raum, 1891 : Zur Morphologie und Physiologie der Sprosspilze; Zeitschr. fiir Hygiene, 1891, Vol. x, p. 1. Moeller, 1892 : Ueber den Zellkern und die Sporen der Hefe, Centralbl. fiir Bakt. und Parasitenkunde, Vol. xii, 1892, p. 537. Mann, 1892 : The Embryo-Sac of Myostirus minimus, L. : A Cell-Study : Transactions and Proceedings of the Roy. Bot. Soc. of Edinburgh, 1892, P- 394- Moeller, 1893 : Weitere Mittheilungen fiber den Zellkern und die Sprosse der Hefe, ibid., 1893, Vol. xiv, p. 358. Moeller, 1893 a : Neue Untersuchungen fiber den Zellkern und die Sporen der Hefen, Ber. d. Deutsch. Bot. Gesellsch., Bd. xi, 1893, p. 402. Krasser, 1893 : Ueber den Zellkern der Hefe, Oesterreich. Bot. Zeitschr., 1893, p. 14. Hieronymus, 1893 : Ueber die Organisation der Hefezellen, Ber. d. Deutsch. Bot. Gesellsch., Bd. xi, 1893, p. 176. Janssens, 1893 : Beitrage zu der Frage fiber den Kern der Ilefezelle, Centralbl. fur Bakt. und Parasitenkunde, Vol. xiii, 1893, p. 639. Wager . — The Nucleus of the Yeast-Plant. 539 Zimmermann, 1893: Sammel-Referate aus dem Gesammtgebiete der Zellenlehre. Beihefte z. Bot. Centralbl., Bd. iii, 1893, p. 420. Dangeard, 1893 : Sur la structure histologique des Levures et leur developpement. Comptes-rendus Acad. Sci., 1893. — 1894 : La structure des Levures et leur developpement : Le Botaniste, 1894, p. 282. Macallum, 1895 : On the distribution of Assimilated, Iron Compounds, other than Haemoglobin and Haematins, in Animal and Vegetable Cells: Q. Jour. Micro. Sci., 1895, p. 243. Eisenchitz, 1895 : Centralbl. f. Bakt. u. Parasitenk., 1895, p. 67 4 ; Jour. R. Mic. Soc., 1896, p. 218. Crato, 1896 : Beitrage zur Anatomie und Physiologie des Elementarorganismus : Cohn’s Beitrage zur Biologie d. Pflanzen, Bd. vii, Heft 3, p. 495. Clautriau, 1895 : Etude chimique du Glycogene chez les Champignons et les Levures : Extrait du tome liii des Memoires couronnes et autres Memoires publies par l’Academie Royale de Belgique, 1895. Hartog, 1895 : On the Cytology of the Vegetative and Reproductive organs of Saprolegnieae : Trans. Roy. Irish Acad., xxx, 1895. Henneguy, 1896 : Lefons sur la cellule : Morphologie et Reproduction ; Paris, 1896, p. 137. Buscalioni, L., 1896 : II Saccharomyces guttulatus Rob. : Malpighia, 1896. Wager, 1897 : The Nucleus of the Yeast-Plant: Report of the British Association, Toronto, 1897, p. 860. Janssens and Leblanc, 1898 : Recherches Cytologiques sur la Cellule de Levure : La Cellule, t. xiv, 1898, p. 203. Bouin, 1898 : Contribution a l’etude du noyau des levures : Arch, a’anat. micro- scopique, t. i, 1898, p. 435. Macallum, 1898 : On the Detection and Localization of Phosphorus in Tissue- Elements ; Proc. Roy. Soc., Vol. 63, 1898. See ‘Nature/ Sept. 15, 1898, p. 483. Errera, 1898 : Structure of the Yeast-Cell. A paper read at the Bristol meeting of the British Association in 1898 : also p. 567 of this number of the Annals. Wager, 1898 : Structure of the Yeast-Plant. A paper read at the Bristol meeting of the British Association in 1898. 540 Wager. — The Nucleus of the Yeast-Plant. EXPLANATION OF FIGURES IN PLATES XXIX AND XXX. Illustrating Mr. Wager’s Paper on the Nucleus of the Yeast-Plant. All the figures have been drawn freehand, with the aid of Zeiss’s apochromatic 2 mm. aperture i- 4, and oculars 8 ( x iooo), 12 (x 1500), and 18 ( x 2250). In most cases the outlines of the cells, &c., have been drawn with the aid of the camera lucida. Figs. 1-3, 5-21, 28-32, and 41-53 have been drawn from preparations stained in methyl-green and eosin; Fig. 4, safranin ; Figs. 22-27, methyl-green and fuchsin ; Figs. 33-40 from living cells ; Figs. 54-57, 60, 62, and 63* gentian-violet ; Figs. 58, 59, 61, 64, 65, 68, and 72-78, carmine and nigrosin ; Figs. 66, 67, and 69-71, Heidenhain’s haematoxylin. Saccharomyces Cerevisiae. Figs. 1-5, 7-12, 14-18, and 27-32, after three hours in Pasteur’s solution. Fig. 1. Cell showing nuclear body and vacuole with network and one deeply stained granule. Fig. 2. Ditto showing three deeply stained granules in the vacuole. Fig. 3. Ditto showing a vacuole not much larger than the nuclear body. Fig. 4. Cell containing a vacuole which shows the nuclear-like network very clearly. Fig 5. Cell with vacuole full of deeply stained substance, partly enclosing the nucleolus. Fig. 6. Cell of compressed Yeast — the vacuole contains two deeply stained granules and delicate threads. Both nuclear body and vacuole are to some extent surrounded by unstained granules — Hieronymus’ granules. Two hours in sugar-solution. Fig. 7. Cell showing nucleolus as seen through a vacuole. The appearance is presented of a nucleus with nuclear membrane, &c. Fig. 8. Cell showing nucleolus, in part surrounded by a dense mass of granules. Fig. 9. Shows the nucleolus in contact with what appears to be a very definite chromatin-network. Whether this is contained in a vacuole or not, could not be made out clearly. Fig. 10. Shows the position of the nucleolus at the time when the cell begins to bud. The vacuole in this case contains very little, if any, stainable substance, but is surrounded by deeply stained granules. Fig, 11. The vacuole, with granular contents, is making its way into the young bud. The nucleolus still retains its position on the opposite side of the cell. Fig. 12. The nucleolus has made its way to the opening between the bud and the parent- cell. Fig. 13. Ditto. The vacuole in this case is surrounded by deeply stained granules. Seventy-two hours in Pasteur’s solution. Fig. 14. The nucleolus puts out a projection into the neck of the budding cell. Wager. — The Nucleus of the Yeast-Plant. 541 Fig. 15. Ditto, but a slightly later stage. The vacuolar contents are very abundant. Fig. 16. Slightly later stage than Fig. 15, just previous to the complete division of the nucleolus. The vacuole is small and irregular in shape. Fig. 17. The same stage as Fig. 16, but the vacuole has nearly disappeared, and in its place a deeply stained mass of granules nearly equally divided between the two cells. Fig. 18. Complete separation of the newly formed nucleoli to opposite ends of their respective cells. A granular thread is shown drawn out between the two from the granular network. Figs. 19-24. After seventy- two hours in Pasteur’s solution. Fig. 19. Shows the nucleolus beginning to divide by constriction in the parent- cell. Fig. 20. Later stage than Fig. 19, the division is completed even before the vacuole begins to divide. Fig. 21. Division of nucleolus in the parent-cell. A small vacuole only is present but there are a number of deeply stained granules, which are separated into two equal groups with a granular thread drawn out between them. Fig. 22. Shows the glycogen-vacuole beginning to form. The chromatin-vacuole is still present with one deeply stained granule, and near the nucleolus numerous deeply stained granules are to be seen. Fig. 23. Later stage — the large vacuole is the glycogen-vacuole. Fig. 24. The nucleolus, chromatin-vacuole, and granules restricted to the wall of the cell by the glycogen-vacuole. Fig. 25. S. Cerevisiac, Hansen I, shows nucleolus and granular network and large glycogen-vacuole. Fig. 26. S. Cerevisiae, Hansen I, shows a nucleolus lying on the wall of the cell surrounded by a lightly stained vacuolar protoplasm, containing a few deeply stained granules. The remainder of the thin layer of protoplasm lining the cell is granular. Fig. 27. Cell showing the gradual formation of the glycogen-vacuole, and the contraction of the chromatin-vacuole. Note the close attachment of the latter to the nucleolus. The protoplasm contains numerous deeply stained granules. Fig. 28. Young cell with numerous small vacuoles, each enclosing a deeply stained granule, and some granules with a vacuole apparently just forming around each. The whole of the cell contents stain deeply. Fig. 29. Young cell with three vacuoles, each containing a deeply stained granule. The cell contents stain deeply. The nucleolus is only visible after very careful staining. Fig. 30. Ditto, with four vacuoles. Fig. 31. Young cell with one vacuole, containing a deeply stained granule and delicate radiating threads. Near it a nucleolus only visible with difficulty. Fig. 32. Young cell with nucleolus in the midst of a peripheral ring of granules. The whole cell is pervaded by a deeply stainable substance. Figs. 33-40. Compressed Yeast examined in the living condition. The black granules represent the highly refractive granules described by Hieronymus. Fig* 33* Half an hour after being placed in 5 % sugar- solution. The nucleolus is visible, and is indicated by a slight depression in the vacuole. Few refractive granules present. 542 Wager. — The Nucleus of the Yeast-Plant. Fig. 34. The nucleolus is surrounded by the refractive granules, which are now more numerous. Two moving granules in the vacuole. Two hours in sugar- solution. Fig. 35. The bright granules more numerous. Two vacuoles present and a nucleolus. Two hours in sugar-solution. Fig. 36. Shows three pairs of granules and small groups on each side of the nucleolus. Two hours in sugar-solution. Fig. 37. Numerous vacuoles appear, as a preliminary to spore-formation. Three hours in sugar-solution. Fig. 38. Later stage — the vacuoles are more numerous, the bright granules surround the vacuoles. Three hours in sugar-solution. Fig. 39. The vacuoles disappear. The protoplasm, as shown by means of reagents, exhibits a foam-structure at this stage. Twenty-four hours in sugar- solution. Fig. 40. Two groups of granules on opposite sides of the nucleolus in a hyaline protoplasm. Twenty-four hours in sugar-solution. S. Ludwigii. Fig. 41. Cell showing chromatin-containing vacuole and nucleolus. Fig. 42. Ditto. Two lines of granules run from one end of the cell to the nuclear vacuole. The figure shows the nucleolus as seen above the vacuole, not inside it. Fig. 43. Young bud just forming. Nucleolus beginning to divide. One row of granules stretching from one end of the cell to the nuclear vacuole. Fig. 44. Cell showing two vacuoles and a nucleolus between them. S. pastorianus. Fig. 45. Cell showing two large normal vacuoles, and in the strand of proto- plasm across the middle a nuclear vacuole and a nucleolus on one side. Fig. 46. Shows the nucleolus as seen from above. On focussing down the vacuole could be seen. Fig. 47. Cell showing nucleolus as seen through the vacuole. Fig. 48. Cell with nucleolus and three small chromatin-vacuoles. Fig. 49. Cell showing nucleolus at one end surrounded by granules which are connected to a small vacuole at the other end by a deeply stained row of granules. Fig. 50. Nucleolus in the middle of the cell in contact with a curving line of granules running from a chromatin-vacuole at one end to the other end of the cell. Fig. 51. Cell with young bud. The nucleolus is in contact with a vacuole containing a deeply stained granule and network. Fig. 52. Later stage than Fig. 51 ; the nucleolus is beginning to divide. Fig. 53- Case of budding in which the nucleolus is about to divide in the neck joining the two cells. The daughter-cell contains a chromatin-vacuole in close contact with its share of the nucleolus. Spore-formation. Fig. 54. Cell just at commencement of spore-formation — protoplasm reddish blue, nucleolus light blue. Wager . — -The Nucleus of the Yeast- Plant. 543 Fig. 55. Cell in which the deeply stainable substance is becoming concentrated in a central mass of protoplasm. The nucleolus stains light blue as before, but a little more deeply. Fig. 56. Later stage. The nucleolus shows a deeply stained granular substance inside or in close contact with it. Fig. 57. The deeply stained granular mass inside or in contact with the nucleolus has increased in size. The deeply stained central portion of the protoplasm is surrounded by granules. Fig. 58. Later stage. The deeply stained mass of protoplasm is smaller. Fig. 59. Still later stage. The deeply stained nucleolus is now the most prominent structure. Fig. 60. Commencement of division. The nucleolus and its deeply stained mass begin to elongate. Fig. 61. The deeply stained granular mass at a later stage in the process of elongation. Fig. 62. The row of deeply stained granules stretching across the cell, surrounded by a faintly stained substance. Fig. 63. Shows the gradual accumulation of the granular substance at both ends to form the daughter-nucleoli. Fig. 64. Complete separation has now taken place. Fig. 65. Later stage in the division ; the two daughter-nucleoli are still connected together by a less deeply stained substance. Fig. 66. Division- stage, as shown in a preparation stained with haematoxylin, Heidenhain’s method. One portion of the protoplasm is still shown more deeply stained than the other, with a distinct line of separation between them. Fig. 67. Same method of preparation. The division is transverse. Fig. 68. The two groups of deeply stained granules completely separated, each surrounded by a less deeply stained substance with a faintly stained granular thread drawn out between them. Fig. 69. Division into four. Fig. 70. Slightly later stage than Fig. 68. Fig. 71. Shows a cell in which division is taking place into eight nucleoli instead of four. Fig. 72. Four groups of deeply stained granules, still connected together by less deeply stained substance. The division apparently has taken place at one end of the cell. Fig. 73. Stage similar to Fig. 71, but the groups of granules are more irregularly distributed at one end of the cell. Fig. 74. The four products of the previous two divisions are now lying in a slightly more deeply stained portion of the protoplasm at one end of the cell. Fig. 75. Later stage than Fig. 74. The spores beginning to separate out. Fig. 76. The spores now visible, each with a distinct outline due to the presence of a thin membrane. Fig. 77. Later stage. The membrane is more distinct. Fig. 78. Same stage as Fig. 77, but five spores shown, two smaller than the others. From observations made recently I am inclined to think that these two fuse together to form one. 'JfjuwZs of Bo lajiy VolM,PLl 'XIX. University Press, Oxford. lAjuwZs of Bolajyr H. Wager del. WAGER YEAST NUCLEUS Voi.xgPkm. : 63. 67. 75 . 64. 68. 76 . 62. 78. 72. 61. 60. University Press, Oxford. The Proteolytic Enzyme of Nepenthes (II). BY S. H. VINES. IN the Annals of Botany for December, 1897, I published a paper on this subject, in which I adduced a con- siderable amount of evidence to prove that, contrary to the opinions of Dubois and Tischutkin, the pitcher-liquid of Nepenthes contains a proteolytic enzyme. Since writing that paper I have continued my observations, of which I now give some account by way of supplement. Activity of Pitcher-Liquid. I have nothing to add that would in any way modify my assertion that I have never failed to obtain digestion of fibrin by the liquid in a relatively short time, provided that the liquid was duly acidified. My new observations refer to the effect of exposure to high temperatures, of treatment with alkalies, and of filtration of the liquid, upon its digestive activity. Heat . The following results will serve to illustrate the general effect of exposure to high temperatures. The method of experiment was to maintain the liquid for a given time at the required temperature, and then to institute a digestion- experiment, adding fibrin and the necessary acid ; in nearly [Annals of Botany, Vol. XII. No. XL VIII. December, 1898.] 546 Vines. — The Proteolytic every case there was a control digestion-experiment with unheated liquids: — (March i, '98.) Liquid heated to 70-80° C., for 15 minutes: digestion of fibrin (-05 grm.) not complete in 5 hours, though it eventually took place ; control-experiment, digestion complete in 2f hours. (March 3.) Liquid gradually heated from 6o°-8o° C., maintained at 80° C. for 5 minutes, then allowed to cool : total time of exposure to heat, 1 5 minutes ; digestion not complete in 5 hours, but within 20 hours; control, digestion complete within 2 hours. (March 8.) Liquid maintained at 8o° C. for 15-20 minutes: digestion not complete until morning of the fourth day (March 11); control, digestion complete within 3 hours. In a subsequent experi- ment (May 17) the effect of treating the liquid to 8o°C. for 20 minutes was less marked: in this case the time required by 10 cc. of the liquid to digest -05 grm. of glycerin-fibrin was just 24 hours. (March 15.) Liquid maintained at 78-83° C. for 30 minutes: digestion did not take place, although the experiment was continued for 4 days; in the control, digestion was complete in ij hour. In a subsequent experiment (March 19) I found that liquid which had been kept at 80° C. for 30 minutes gave no indication of digestive action on fibrin although the experiment was prolonged for a week ; it may be fairly concluded that the digestive power had been entirely destroyed. In the control, digestion was complete within 5 hours. With regard to the action of a boiling temperature (ioo°C.), I was surprised to find, on several occasions, that liquid boiled for some seconds did not lose its digestive power, though the rate of digestion was made very much slower. It seems, in fact, that to entirely destroy digestive power, the liquid must be kept at ico°C. for an appreciable time, say 3-5 minutes. Alkali . I have confined myself to the investigation of the action of sodium carbonate (Na2COs) upon the digestive activity of the liquid. The method of experiment was to add to a quantity (usually 5 or 10 cc.) of the pitcher-liquid an amount of the solid salt requisite to produce the desired degree of alkalinity : the alkaline liquid was then placed in the incubator, and maintained for any required time at any 547 Enzyme of Nepenthes (//). given temperature. After this treatment the liquid was neutralized, then acidified with HC1, and a digestion-experi- ment was made. At an early stage it became apparent that the results of such experiments are dependent upon the three following factors: — (i) the degree of alkalinity; (2) the duration of the period of alkalinity; (3) the temperature maintained during the period of alkalinity. I took as a starting-point the statement made in my paper of last year (p. 572), that the digestive activity of the liquid is destroyed by treatment with 5 °/o Na2C03 for 3 hours at a temperature of 35-40° C. I found, on repeating the experiment and prolonging it, that this conclusion is not accurate. In the original experiment the time allowed for digestion was 16 J hours, and in that period, it is true, no trace of digestive action could be detected : but in a sub- sequent experiment with a longer digestive period, evident signs of digestion were apparent in 20 hours, and digestion was completed in about 26 hours. The following is a summary of the principal experiments made : — 1. Temperature during period of alkalinity, 35-38° C. Treatment with 0-5 °/Q Na2C03 for 30 minutes ; retards digestion. 1% 3 3 33 2 2 hours ; about doubles time of digestion. » 1% 33 3 3 *7 ,, digestion much retarded. 2% 3 3 3 3 ^ 33 33 33 „ 4% 33 „ 2 ,, digestion complete with- in 20 hours. .. 5 °/0 3, „ I ,, digestion complete with- in 6 hours. „ 5% 3 3 3 3 3 ,, digestion complete with- in 26 hours. It will be observed that, in all the foregoing experiments, digestion, though often much retarded, eventually took place. It still remained to ascertain under what conditions total destruction of digestive power could be effected by treatment 548 Vines . — The Proteolytic with Na2C03. In view of the great retardation of digestion caused by treatment with 5 °/o Na2C03 indicated in the fore- going table, I determined to experiment especially with this degree of alkalinity, but at a higher temperature : the tem- perature upon which I fixed was 50° C. The following are illustrative experiments : — 2. Temperature during period of alkalinity , 50° C. (May 30.) Treatment with 5% Na2C03 for 45 mm.; digestion much retarded. (June 1.) „ ,, ,, ,, ijhour; no digestion in 48 hours. Hence it would appear that, in the second case, the digestive power had been totally destroyed. The following are some further comparative experiments to determine more closely the degree and duration of alkalinity, at 50° C., necessary to destroy digestive action. In each case there was a control-tube, the liquid in which was heated to 50° C. for the same length of time as the others, but was not treated with Na2C03, though diluted subsequently to the same extent as the others : the amount of fibrin used in the digestion was -oi grm. (a) Treatment with 5 °/0 Na2C03 for \\ hour; no digestion in 6 days. digestion complete in 4 days. control ; digestion complete in 3-I hours. (6) „ „ 5% „ „ 1 ; hour ; no digestion in 5 days. no digestion in 5 days. control ; digestion complete in a few hours. (c) „ „ 1% „ „ 1 hour; no digestion in 4 days. control ; digestion complete in a few hours. 549 Enzyme of Nepenthes (II). On comparing the results of a, h, and c, it would appear that treatment with I °/o Na2C03 for one hour at a tempera- ture of 5o°C. is an approximate index to the stability of the enzyme. Filtration. It occurred to me that some light might be thrown upon the bacterial explanation of the digestive activity of the pitcher-liquid by experiments with liquid which had passed through a Berkefeldt-filter. I found that liquid which has passed through such a filter has lost its acid reaction and its colouration. It still retains some digestive power, but is far less active than unfiltered liquid, the period of digestion being more than doubled. This result might be made use of by the supporters of the bacterial explanation, as affording some sort of evidence in favour of that view ; but even so, it would be far from con- clusive. However, in order to test the value of this evidence I instituted some experiments with regard to the effect of filtration through the Berkefeldt-filter upon liquids containing pepsin and ptyalin. With regard to pepsin, I found that an unfiltered solution of glycerin-extract of pig’s stomach digested *05 grm. of fibrin in a quarter of an hour, whilst an equal quantity of the same solution after filtration required nearly 5i hours to digest the same weight of fibrin. Similarly some diluted saliva caused the complete conversion of a small quantity of starch into sugar within a few minutes, whereas starch treated with filtered dilute of saliva continued to give more or less marked blue-reaction with iodine for four or five hours. It is clear that solutions of pepsin and of ptyalin are affected by the Berkefeldt-filter in much the same way as is the pitcher-liquid of Nepenthes. If it be argued that the diminished activity of filtered pitcher-liquid is due to the removal of Bacteria, the same argument must equally apply to the solutions of pepsin and of ptyalin ; but I do not think that any one would venture at present to attribute the action of gastric juice or of saliva to the presence of Bacteria. The obvious conclusion to be drawn from these experiments is 550 Vines . — The Proteolytic \ that enzymes are retained in a marked degree by the* Berkefeldt-filter. The Zymogen. In my previous paper (p. 578) I stated that I had not succeeded in successfully repeating the experiments of 1877 1, which demonstrated the presence of a zymogen in the glandular tissue of the pitcher. The method which I adopted in 1877 was as follows: some pitchers were treated with dilute acetic acid (1 °/o) for 24 hours previously to the preparation of the glycerin-extract ; comparative experiments between the glycerin-extracts prepared from pitchers so treated and the glycerin-extract prepared from pitchers gathered at the same time, but not treated with acid, showed that in every case the digestive power of the former was much greater than that of the latter, as indicated by the greater rapidity of digestion. In one experiment the acid- extract dissolved fibrin in 6 hours, whilst fibrin put to digest with the ordinary extract was but slightly attacked in that time. During the present year I have made some experi- ments with results confirmatory of those of 1877. They are as follows : — June 2, ’98. Took two unopened pitchers of N. Master siana\ opened and washed out the pitchers ; the pitcher-liquid strongly acid ; cut up the glandular portions of the pitchers into small pieces ; the whole weighed 8 grm. ; divided into two halves A and B ; A was rubbed up in a mortar with 20 cc. distilled water; B was rubbed up in a mortar with 20 cc. of ‘25°/0 solution of HC1; both A and B were then placed for 45 minutes in the incubator at 5o°C. ; the liquid was then poured off from each, the substance dried somewhat with blotting-paper, and then rubbed up each with 20 cc. glycerin, to prepare glycerin-extracts A and B , and left to stand. On June 10 the digestion-experiment was made: 5 cc. of each of the extracts were taken, and placed in a tube with 5 cc. of -4 °/Q HC1, together with -oi grm. of fibrin; the tubes were put into the incubator (temperature about 37° C.) at 10.30 a.m. By 6.30 p.m. the fibrin in tube B (acid-extract) was completely digested, whereas 1 Journ. Linn. Soc. xv. Enzyme of Nepenthes (//). 551 that in tube A (neutral extract) did not disappear until over 48 hours later. In another experiment of about the same date, in which, however, the pitcher-substance was treated with •&°/0 acetic acid for about 24 hours at ordinary temperature (about i5°C.), the acid-extract digested more rapidly than the neutral, but the difference was not so marked as in the preceding case. July 21, '98. Took two pitchers, one unopened, the other recently opened; cut up small; material divided into three parts of 2-7 grm. each: (1) was rubbed up at once with 20 cc. glycerin; (2) with 20 cc. of *25°/0 HC1; and (3) with 20 cc. distilled water: (2) and (3) were placed in an incubator at 36° C. and were kept there till next day (18-20 hours), the liquid was filtered off from them, and they were each rubbed up with 20 cc. glycerin. On Aug. 2 the digestion-experiment was made: 10 cc. of each extract were strained off through muslin; to each were added 10 cc. of *4°/0 HC1, together with -oi grm. fibrin, and all three were placed in the incubator (temperature 36° C.) to digest at 10.30 a.m. At 5 p.m. the fibrin in tube (2) (acid-extract) showed signs of digestion, and was almost entirely dissolved by 5 p.m. on the following day. The fibrin in tube (3) had all undergone solution by the evening of Aug. 5, whereas that in tube (1) still showed no sign of digestion. A second digestion-experiment with the same extracts was made on Aug. 30, with essentially similar results, though the period of dig^tion was longer. The experiment commenced at 10.15 and at 5 p.m. no indication of digestion could be seen in any one of the three tubes. Next morning (Aug. 31) at 9 a.m. the fibrin in tube (2) (acid-extract) was seen to be attacked ; the process of digestion continued slowly in this tube until it was complete (night of Sept. 2) ; the fibrin in the two tubes (1) and (3) underwent no perceptible change in this time. The foregoing results suffice to show that, under certain circumstances, previous treatment with acid causes the glands of the pitcher to yield a more active glycerin-extract, or to yield an active extract when otherwise the extract would be inactive ; and it can only be concluded that this must be due to the presence of a zymogen in the glands from which the enzyme is liberated on treatment with acid. However, I must P p 552 Vines . — T he Proteolytic admit that, as pointed out in my paper of last year, I have by no means always succeeded in obtaining a more active extract as the result of treatment with acid ; on the contrary, I have frequently found that previous treatment of the pitchers with acid diminished instead of increasing the activity of the glycerin-extract. I do not regard these apparent contra- dictions as wholly attributable to various conditions of the pitchers, for I have obtained sometimes quite opposite results with pitchers as nearly as possible of the same age, and sometimes quite similar results with pitchers of different ages ( e.g . opened and unopened). On the contrary, my results seem to show that the differences in the activity presented by the various acid-extracts are due rather to the mode of treatment of the pitcher-material. It would appear that the most effectual mode of decomposing the zymogen is to act upon the tissue with acid for a short time at a relatively high temperature (see example of June 2-10). More prolonged treatment at a lower temperature (say 350 C.) would seem to cause not only the liberation of the enzyme, but also its extraction from the glands in connexion with the digestion of the pitcher-tissue itself. I may just point out in conclusion that the marked acidity of the liquid in the unopened pitcher is no doubt to be connected with its high digestive activity ; whilst the acid is useless for digestive purposes until the opening of the pitcher, it is probably of importance in that it acts upon the zymogen, liberating the enzyme. The Products of Digestion. In my paper of last year I pointed out that the chief proteid product of digestion was a substance closely resembling deutero-albumose, and I stated further that I had failed to detect the presence of a true peptone, that is, of a proteid which is not precipitated on saturation with ammonium sulphate. I have, however, since detected the presence of peptone, though in relatively small quantity, among the products of the digestion of fibrin by the pitcher-liquid. 553 Enzyme of Nepenthes (//). The clue to the matter was found on this wise. The method which I had followed (see my paper of Dec. ’97, p. 579) in examining the products of digestion involved the precipitation of these substances by filtration into excess of alcohol. Whilst investigating the nature of the ultimate products, I evaporated a considerable quantity of the alcohol which had been used for precipitation, and this left a dark brown syrupy residue. Some of this residue, dissolved in a small quantity of distilled water, formed a brownish solution, giving no precipitate on boiling, but good xanthoproteic and biuret reactions. A portion of this solution was put to saturate with ammonium sulphate and gave a dense precipitate which brought down with it all the colouring-matter ; the clear, colourless liquid obtained on filtration still gave strong xanthoproteic reaction, and continued to do so after continued saturation for two days longer. Another portion of the brown solution was put to dialyze, and within twenty-four hours the dialysate gave strong xanthoproteic reaction ; on saturat- ing the dialysate with ammonium sulphate, there was a precipitate, the filtrate from which still gave the xanthoproteic reaction, and continued to do so on further saturation. These observations indicated the presence of peptone without, how- ever, absolutely establishing it ; for it might be the case that the precipitation of the deutero-albumose by means of ammonium sulphate had been incomplete, and that the proteid reactions were due to this substance rather than to peptone. It became necessary, therefore, to employ some method by which the separation of deutero-albumose and peptone could be certainly effected. Fortunately I applied for advice to Mr. Ramsden, Fellow of Pembroke College, who has an intimate knowledge of the chemistry of proteids, and he kindly directed me to a paper by Kuhne1 on this very point. Kiihne’s method consists in saturating the neutralized diges- tion-liquid with ammonium sulphate when boiling. After 1 Zeitschrift fur Biologic, 1892 (Erfahrungen iib. Albumose und Pepton). It may be asserted that, until the appearance of this paper, no true peptone had been obtained free from albumose. P p 2 554 Vines. — • The Proteolytic cooling and filtering, the liquid is again heated and made alkaline with ammonia, and is then again saturated with ammonium sulphate. It is once more cooled and filtered, then boiled to drive off the ammonia, again saturated with ammonium sulphate whilst boiling, and made acid with acetic acid. The excess of ammonium sulphate is then got rid of by adding alcohol to the liquid ; most of the ammonium sulphate is precipitated, and the supernatant liquid holds in solution whatever peptone is present, which may be ulti- mately obtained by repeated treatment with alcohol and decantation of the supernatant liquid. With this method at my disposal, I had no difficulty in demonstrating the presence of peptone in the digestion-liquids, though it was necessary to concentrate these liquids before proceeding to test them. Mr. Ramsden also found peptone in a quantity of digestion-products precipitated by alcohol, which he was good enough to examine for me. I have nothing to add with regard to the other products of digestion, beyond the fact that I have been able to confirm my statement that leucin is one of them . Conclusion. 1. The experiments relating to the action of high tem- peratures and of alkalies upon the enzyme confirm the statement made in my paper of last year with regard to its great stability; in fact, it would appear that it is the most stable of all known proteolytic enzymes. Whilst its activity can easily be much diminished by exposure to high tem- perature or treatment with an alkali, it still retains a sort of residual digestive power which asserts itself in very slow and prolonged digestion, and which can only be destroyed by relatively strong measures. 2. It may, I think, be fairly concluded from the facts given in this paper, in conjunction with those which I published in 1877, that the enzyme is derived from a zymogen present in the gland-cells. 555 Enzyme of Nepenthes (//). 3. The discovery of true peptone among the products of digestion facilitates the classification of the enzyme. Green 1 has found that the proteolytic ferment present in germinating seeds acts in an acid medium, producing a relatively large quantity of albumose together with peptone, leucin, and tyrosin ; that it is in fact a tryptic ferment, differing mainly from the trypsin of the pancreatic juice in requiring an acid medium for its digestive action. In all these respects (though I have not made out the production of tyrosin) the proteolytic enzyme of Nepenthes- pitcher closely resembles that of the germinating seed ; but it is much more rapid and energetic in its action, and apparently more stable in its nature. These two proteolytic enzymes are distinguished, by their activity in an acid medium, from those, such as papai'n2 and the enzyme in the fruit of Cucumis utilissimus 3, which are most active in a faintly alkaline medium. It is a remarkable fact that, whatever may be the reaction of the medium in which they can work, all these enzymes are essentially tryptic in their mode of action ; in fact it is not improbable that this may be a characteristic feature of all vegetable proteolytic enzymes whatsoever. 1 Phil. Trans., 1887. 2 See Martin, in Journ. of Physiol., V, 1884, and VI, 1885. 3 See Green, in Annals of Botany, VI, 1892. NOTES CHANGES IN THE SEX OP WILLOWS.— In the genus Salix flowers of both sexes are occasionally present in the same catkin, and one sometimes finds that the sexual organs are inter- mediate in structure between stamens and carpels. By using the published records, and by availing myself of the large accumulation of material for study in the Herbaria at Kew, Cambridge, in the British Museum, and at the Jardin des Plantes, Paris, I have gathered together a number of facts which may be of interest. Firstly, it is obvious that these abnormalities, though widely dis- tributed in the species of Salix, are much more common in some sections than in others. The section Capreae yields by far the greatest number ; and second to it comes the section Fragiles. In dwarf willows they seem to be very rare, and in the section Glaciales I have only found one abnormal catkin. Secondly, we notice that, though the two-staminal willows yield most freely these abnormalities, those in which the male flowers possess more than two stamens sometimes show them. I can instance S. peniandra and S. humboldtiana. That the male organs or the female organs are produced from the same rudiments is extremely probable ; and in the normal Salix we have an unisexual flower, which cannot, as in most Phanerogams, be shown to have had an origin from a hermaphrodite flower by abortion of one sex. In these abnormal willows, while we readily follow the change of the two stamens of one of the Capreae or Purpureae into the two carpels, it is not so easy to say what happens when five or more stamens have to be replaced by two carpels. 1 These Notes are abstracts of papers read before Section K of the British Association, at the Bristol Meeting, September, 1898. [Annals of Botany, Vol. XII. No. XLVIII. December, 1898.] 558 Notes. Lastly, of the several theories thus far proposed to account for the occurrence of the abnormalities, none is capable of wide application. Sometimes the abnormalities reappear year after year ; sometimes they prove inconstant. Had we a fuller knowledge, some explanation, partial or complete, might be forthcoming; for frequently, both in their distribution in the catkin and on the branch, the changes in sex show a tendency to arrangement. At times the male is above the female ; at times the reverse is the case. Rarely there are three or four belts of flowers on one catkin, male succeeding female, and female male, in definite order. Royal Gardens, Kew. I. H. BURKILL. THE AH ATOMY OF THE STEM OF SPECIES OF LYCO- PODIUM.— Ten species of Lycopodium have been examined; among these two types may be distinguished. 1. Type of Z. clavatum (L.). The oval stelic arrangement is marked by a considerable amount of xylem, broken up into patches by bands of phloem. Centrally these bands are strap-shaped, but at the ends of the long axes the areas of phloem are external, and occur as curved and flattened wedges. Large cells without contents, sieve- tubes, appear in the centre of the strap-shaped bands. Protophloems and protoxylems are external, forming a continuous ring, as figured by Hofmeister ; so that, using De Bary’s terminology, the arrangement of the bundles is radial. Pericyclic and the so-called endodermal cells occur in concentric zones, 1-3 cells broad. The former swell up, especially in glycerine or glycerine-jelly ; the latter are generally considerably lignified. The cells of the cortex lying just external to the endodermal cells are thickened and lignified, forming a third concentric zone several cells deep. To this type conform Z. alpinum (L.), Z. Phlegmaria (L.), Z. dendroides (?), and Z. cernuum (L.). 2. Type of Z. squarrosum. The type which contrasts most markedly with the former is found in Z. squarrosum (Forst.), Z. dicho- tomum (Jacq.), and Z. nummularifolium (Blume). The phloems occur as islands in the sea of xylem, or as inserted peninsulas. The phloems are centrally built up, with the apparent sieve-tubes in the centre. Protoxylems are well marked, and lie externally, but protophloems are not to be distinguished. Endodermal cells and pericycle are found as in the previous type. The sclerenchymatous sheath is wanting, or very slightly developed. Notes . 559 The two remaining species, Z. Dalhousieanum (Spring) and Z. Selago (L.), are, to some degree, intermediate types. The phloem in Z. Dalhousieanum shows both types, strap-shaped and centric. In the branches the structure becomes simpler. There are two narrow strips of xylem, with an intermediate strip of phloem, so that a prominent row of sieve-tubes occupies the very centre of the stelic cylinder. Z. Selago in its structure is modified on that of L . clavatum. An interesting feature of Z. Selago and Z. squarrosum is the occurrence of root-structures running through the stem. These consist of steles containing a crescent-shaped mass of xylem, with protoxylems to- wards each tip, while the concave portion is filled up with phloem. A characteristic sclerenchymatous sheath surrounds the stele. In Z. Selago these root-structures are found even above the point where the stem branches, but in Z. squarrosum they have fused with the central cylinder before branching occurs. C. E. JONES. University College, Liverpool. REPRODUCTION IN DICTYOTA DICHOTOMA. — i.Diclyola is an annual. In this country it germinates during the summer, remains small during the winter, grows very rapidly in June, and begins to form its reproductive cells in July. 2. The tetraspores are produced throughout the season, and all stages may be found together on the same plant. The sexual cells, however, show a remarkable periodicity. The formation, maturation, and liberation of each crop occupies a fortnight, the interval between two spring-tides. The sori are formed during neap-tides, and the cells are liberated during or immediately after the highest spring-tides. 3. When liberated the oospheres are not invested with walls. In this condition they strongly attract the antherozoids, become fertilized, and at once start germinating. The plantlets are similar to those figured by Thuret as resulting from the germination of the tetraspores. 4. If not fertilized the eggs lose the power of attracting anthero- zoids, they form walls, and, as already described by Thuret and Bornet, they germinate parthenogenetically. After one or a few divisions, sometimes accompanied by formation of a rhizoid-rudiment, the process stops and the plantlets die. 5. Towards the close of the season some sori fail to mature within Notes. 560 the usual period, and the crops become less regular ; the same effect is brought about during very cloudy and cold summers. 6. The same conditions bring about sterilization of certain of the sexual cells. Thus, patches of cells within the antheridial sori fail to divide. Cells at the margins of female sori remain barren, so that the usually borderless sori acquire partial or even complete borders. 7. There are strong reasons for concluding that the factor which determines the maturation and liberation of the sexual cells, and the fertilization of the oospheres, is the amount of the illumination to which the plants are subjected. 8. The cytology of the reproductive cells will be described as far as it has been made out. University College, Bangor. J- LLOYD WILLIAMS. CHANGES IN THE GLAND-CELLS OF DROSERA PRO- DUCED BY VARIOUS FOOD-MATERIALS.— The work is an extension of that previously undertaken by the authoress, an account of which has already appeared in ‘ Quart. Micro. Journ./ vol. xxxix. In the experiments now described, leaves were fed with egg-albumin, globulin, peptone, fibrin, milk, nuclein, nucleic acid, and calcium phosphate, the histological changes in the gland-cells being noted in each case. The results to be described were obtained with fixing-fluids, widely differing in their chemical constitution. Egg-albumin. The basophil cytoplasm becomes pink, and is reduced in twenty to thirty hours to a mere vestige. After two days it commences to recuperate, and ultimately becomes again basophil. The changes in the nucleus comprise — (1) those of the nuclear chromo- somes, (2) those of the nuclear plasm, and (3) those of the nucleoli. In the resting-cell the nuclear chromatin is scanty, but immediately after feeding it commences to increase, till in twenty to thirty hours large segments are formed as in mitosis. During recuperation the segments again diminish. The eosinophil nucleoli are large in the resting-cell; they diminish after feeding in direct proportion to the increase of the basophil chromatin, and finally enlarge when the chromatin-segments diminish. Peptone is absorbed much more rapidly than egg- albumin, and produces in one hour changes similar to those effected by egg-albumin in twenty to thirty hours. Notes . 561 Globulin also produces changes in twenty-four hours, but to a less marked degree than egg-albumin. Food passes into the tentacle between the lateral walls of the cells, and secretory products pass through the apical walls, thus producing an appearance of striae in the food which is in contact with the tentacles. Fibrin is digested slowly, and changes similar to but generally less pronounced than with egg-albumin are seen. Milk is absorbed rapidly and completely. The morphological changes are less marked than with any of the above-mentioned foods. The cell-plasm remains basophil throughout. Nuclein produces almost no effect ; the tentacles do not bend in, and do not secrete more copiously than before. No cytological changes are produced except very slight vacuolation of the cell-plasm. All the colour reactions are the same as those of controls. Nucleic acid produced rapid bending in of the tentacles, and ex- tremely copious secretion. The leaves reopen in one to three days, and although the quantity of nucleic acid given is not perceptibly diminished, there are great histological changes, consisting in an almost complete disappearance of the cytoplasm (which remains basophil throughout), and of the nucleoplasm. The basophil chromatin- segments remain unaltered. Calcium phosphate produces appearances very similar to those after feeding with egg-albumin, but the cytoplasm remains basophil. Control leaves, after the application of all the above substances, reopened in a perfectly healthy condition, as determined by their naked-eye appearances while living, and their microscopic structure after fixing by different methods. LILY H. HUIE. A POTATO-DISEASE. — I have for some time past had occasion to recognize here and there, in various parts ol England, a potato- disease which is not due to Phytophthora , and which has often been ascribed to bacteria. During the past two years my attention has been especially directed to testing its bacterial origin, and I am convinced that it is not due to bacteria, but to a true hyphomycetous fungus. Without going so far as to say there is no bacterial disease of the potato, I wish to express the conviction that the alleged cases of such 562 Notes. lately published are not convincing, and that a tendency exists to draw conclusions from imperfect evidence. I shall show that the way into the tuber is prepared for bacteria by fungus-hyphae, and the open passages of destroyed vascular bundles afford them ample space. The disease I have studied has appeared in a more or less epidemic form at least twice in my experience : it was very common two years ago, and this year has been abundant in various parts of England. In a subsequent publication I shall show that it is common and widespread, and even known in some countries, though not adequately recognized. Symptoms. — The shoots turn yellow and die prematurely during the summer, and before the tubers are anything like full. The disease starts from below and not from the leaves. The roots are few and poor, and soon rot away. The tubers are few, do not mature, and often rot in the ground. The leaves turn yellow and wither on the stems, with the symptoms of premature wilting , and often remain long hanging on the yellowing, glassy-looking, but still living stems. In very mild cases these symptoms are not obvious, and supervene slowly, and the case may be complicated by the coexistence of Phytophthora. In very severe cases, on the other hand, especially in wet situations, the stems and roots may be all rotten by the end of July, and casual observation may ascribe the damage to Phytophthora entirely. In ordinary cases, again, it is easy to suppose the damage due to some insect attack, or to drought. In advanced stages of the disease the stems either dry up to brown sticks, or putrefy on the wet ground ; very often bacteria have gained access to the tissues at a comparatively early stage. Microscopic Appearances. — Sections across the lower parts of the attacked stems show one, two, or more of the vascular bundles yellowish- brown — visible even without a lens — and the principal vessels of these contain branched, septate hyphae. In several cases I have traced these hyphae through every internode of the stem, into the petioles of the still hanging leaves, into the young lateral shoots, throughout the roots and subterranean rhizomes, and up to and even just into the tubers. In two cases I have done this in one and the same potato-plant, and so have no longer any hesitation in ascribing the disease to this fungus, the morphological features of which will be described in a subsequent paper. In advanced cases the brown vessels are stopped with a yellowish gum-like substance. Tyloses are common in the vessels of Notes. 563 the root. Those tubers which are not attacked while still very young, but which have already begun to fill with starch, may offer considerable resistance to the invasion of the fungus ; but eventually the vascular strands diverging from the point of attachment to the rhizome exhibit the tell-tale foxy-red or yellowish-brown colour, and in many cases the ripened tubers are to all appearance sound, except for microscopic reddish spots just at the points of entry of these bundles. During the winter the stored potatoes, with the fungus thus just lurking in them at the morphological base (the so-called heel) of the tuber, may undergo little change to all appearance if gathered and stored dry . But if wet, various kinds of rot may supervene, owing to the subsequent invasion of various micrococci, bacteria, fungi, &c., following the lines of weakness opened up by the fungus in question, and living as saprophytes on the stored reserves. In some cases even apparently dry tubers may undergo a curious rot — dry-rot — owing to the ravages of a particular bacterium or mould, perhaps more than one, which finds sufficient moisture for its purposes. The principal point is that the fungus I have especially studied leads the way for these purely saprophytic anaerobic and aerobic forms into the tuber: once in the mature tuber, its progress is necessarily slow until the reserves move in the spring. During the past winter I gave to Miss Dawson, who is working at such subjects in my laboratory, some of the tubers saved from plants attacked with this disease, to investigate the various fungal forms lurking in the diseased tubers. Her investigations are not yet com- pleted, but enough has been accomplished to convince us that after the fungus in question has opened up the way into the tuber, all sorts of bacteria and fungi can make their way down the destroyed vascular strands, and reappear in spring, when the tubers are replanted. But this is not all. The evidence shows that the fungus in question, once in the tuber, leads a dormant life during the early part of the winter, but gradually invades the new sprouts as they slowly appear in the early spring, and that the parasite is actually replanted by the farmer or gardener, when restocking the ground, in his new ‘ sets! If we reflect that the tuber is really a bud, there is nothing especially strange in this phenomenon ; the fungus enters the base of the bud in autumn, and takes some months to traverse its dormant tissues during 564 Notes. the winter and spring. A spotted tuber may give rise to some healthy and some diseased sprouts, according to the tracks of the fungus. A curious phenomenon was observed in some potato-plants very badly attacked by this disease this summer. In some of the badly diseased young shoots, quantities of beautifully developed cubical proteid crystals (crystalloids) were observed in the parenchyma of the pitch and cortex. It is due to Mr. W. G. P. Ellis to point out that he was the first to see these in some sections he was kindly cutting for me of this batch of specimens. On going further into the matter I find such crystalloids have been seen by Heinricher in the shoots of a diseased potato1, but he did not give any account of the disease itself. I find these crystals are not uncommon in the still green bases of the petioles of the withered leaves hanging on the diseased shoots, though they do not always occur. I ascribe their formation to the accumulation of proteids in the leaves, while still living and active, from which the passages of trans- ference at the nodes of the stem have been cut off by the fungus ; just as the eventual withering of the leaves is due to the blocking of their water-conduits when all the vessels are stopped up. At the same time, the attempts I have made to induce the formation of these crystalloids artificially have failed so far. Neither ringing, nor ringing combined with destruction 01 the pith with a hot skewer — to destroy the internal phloem — has given satis- factory results as yet, though the leaves of healthy plants withstand this drastic procedure much better than might be supposed. Here again I must reserve further particulars for the fuller paper. In conclusion, it is evident that the efforts of the potato-grower must be directed to the selection of sound sets, and to the careful preparation of his ground. I hope to show later that it is a fatal procedure, even with sound sprouts, to allow the young shoots to lie in contact with raw manures, as it is via wounds and small rotting spots at and near the collar that new infections occur. The same arguments apply to wet soils and situations, and the disease is particularly apt to increase when wet and cold weather supervenes on the early growths. H. MARSHALL WARD, Cambridge. 1 Ber. d. deutsch. bot. Ges., 1891. Notes . 565 PENICILLIUM AS A WOOD-DESTROYING EUNGUS.— - Spores from pure cultures of Penicillium were sown on sterilized blocks of spruce-wood, cut in March, and were found to grow freely and develop large crops of spores on normal conidiophores. Sections of the infected wood showed that the hyphae of the mould entered the starch-bearing cells of the medullary rays of the sap-wood and con- sumed the whole of the starch. The resin was untouched. In culture three months old the hyphae were to be seen deep in the substance of the wood passing from tracheide to tracheide vid the bordered pits. Control sections, not infected and kept side by side with the above, contained abundance of starch, and no trace of hyphae could be detected in them. The observation appears of interest in several connexions. Peni- cillium is one of our commonest moulds, and undoubtedly plays a part in the reduction of plant debris to soil- constituents; how far it can itself initiate the destruction of true wood, or how far it merely follows on the ravages of other fungi, bacteria, &c., is unknown. There are strong grounds for believing that it destroys the oak of casks, &c., but since these are impregnated with food-materials this is not very surprising. Trabut1 has shown that Penicillium will grow in solutions containing 2-9*5 Per cent* °f CuS04, and other evidence exists showing how remarkably resistant this mould is, and how little organic matter it needs for life. Dubois 2 showed that Penicillium , or a closely-allied form, not only lives in strong solutions of copper, neutralized with ammonia, but will erode metallic copper and bronze if transplanted thereon. Jonssen3 found Penicillium living in one-tenth normal sulphuric acid solution, and gives some interesting facts regarding the sulphur- containing oil-drops in its protoplasm, and other statements concerning oil in this fungus occur in the works of De Bary, Brefeld, Pfeffer, &c. Gerard4 gives proof that Penicillium can liberate butyric acid from mono-butyrine, and evidence that this is due to its power of forming a lipase or fat-splitting enzyme. Lesage5 gives striking instances of the resistance to externa 1 Bull, de la Soc. Bot. de Fr., xlii, 1895, 1. 2 Comp. Rend., 1890, cxi, p. 655. 3 Bot. Centr., xxxvii, 1889, p. 201. 4 Bull, de la Soc. Mycol. de Fr., xiii, 1897, p. 182. 5 Ann. des Sc. Nat., Ser. 8, T. i, 1895, p. 309. 566 Notes , influences shown by the spores on germination. Not only will they germinate and live for some time in water, and under almost anaerobic conditions, but he found them germinating in 26*5 per cent, solutions of common salt; 30 per cent, solutions were too much for them, however. He states also that the vapours of cedar-oil, iodoform, napthalin, camphor, and patchouli do not prevent germination; though those of clove-oil, ether, alcohol, chloroform, and acetic acid prevent it. The maximum for alcohol was somewhere between 4-2 and 6*2 per cent. In acetic acid they germinated in twenty-four days in solutions of 1 : 2 56, but failed to do so in solutions of 1 : 64, whereas in HC1 they germinated in two days in 1 : 4 solutions. As regards temperatures, it is well known how resistant the spores are. A striking instance of the hardships the mycelium can undergo is given by Woronin 1 : he found Penicillium vegetating on the melting snow, where the temperature at night fell below o° C. Bourquelot2 found invertase, maltase, trehalase, emulsin, inulase, diastase, and trypsin in the allied Aspergillus , and pointed out how suggestive this is in explaining the ubiquity of this mould. Probably Penicillium is equally rich in capacity for enzyme-production. Miyoshi 3 showed that Penicillium can bore through cellulose membranes, and no doubt similar chemotactic phenomena are con- cerned in the piercing of wood-elements by the hyphae. It certainly looks as if Penicillium may be a much more active organism in initiating and carrying on the destruction of wood than has hitherto been supposed, and that it is not merely a hanger-on or follower of more powerful wood-destroying fungi. It is also, doubtless, very independent of antiseptics. H. MARSHALL WARD, Cambridge. A METHOD 03? OBTAINING MATERIAL FOR ILLUS- TRATING SMUT IN BARLEY. — By sowing soaked, skinned barley that had been plentifully covered with Ustilago spores a supply of smutted barley may be ensured, and in such material it is easy to trace out the spore formation. Hand-sections of the ear when about § inch long showed the 1 Arb. d. St. Petersb. Naturf.-Ver., B. xx, p. 31. 2 Bull. Soc. Mycol., 1893, p. 231. 3 Bot. Zeit., 1894, H. 1. Notes 567 mycelium at the growing-points of the flower shoots, and in such sections the mycelium, at first intercellular, could readily be found becoming intracellular and of much greater diameter. Branches became very numerous, and in the hyphae and branches spores were formed. Towards the central parts spore-clusters were too dense for examination, but nearer the epidermis the branching and arrangement of the sporogenous hjphae could more easily be made out; and the teasing of the lateral flowers of each notch of the rachis was often more successful than if the central — and only flower of the ordinary ear— were taken. Sections were mounted in water, and some in 1 per cent. KOH, and it is but fair to say that such treatment has failed to show any septation of the hyphae as a preliminary to spore- formation. Material for microtome-sections was prepared as follows : — The leaves of a barley-shoot were stripped down so as to expose the apparently highest node, and the part an inch or two above this was cut off; then by a series of successively lower horizontal cuts the youngest leaves were removed until in the space they enclosed the tips of the awns or ear were seen ; then a cut was made through the node, and the removed ear was placed in Flemming’s or Rath’s solution for fixing, the ear thus being, for a very few seconds only, between plant and reagent. If a smutted ear be removed and kept floating on water, its spores continue to develop, and in several cases they matured first in the awn. It was by no means uncommon, on teasing out young fruits from such an ear, to find that the spores had germinated. I have not yet made similar observations for Tilletia as my bunted wheat was less forward than my smutted barley, but I am satisfied that by this method of working class material for illustrating Bunt and Smut may easily be obtained. W. G. P. ELLIS, Cambridge. STRUCTURE OP THE YEAST- CELL.— A study of the cells of Saccharomyces Cerevisiae has led me to the following conclusions, part of which merely confirm former researches : (1) A relatively large nuclear body exists in each adult cell. (2) Young cells contain no such body; a little later the old nuclear body divides, and one* of its two daughters wanders through the narrow connecting-channel into the young cell. (3) After the division is complete, the two cells are still kept together by a mucilaginous neck-shaped pedicel, which Q q 568 Notes. appears not to have been noticed hitherto. It may persist or not, thus explaining the occurrence of cell-chains or of isolated cells in different races of Yeast. (4) Carbohydrates are stored up in Yeast in the form of glycogen, which accumulates or disappears from the vacuoles very rapidly, according to conditions of nutrition and growth. The colour given by a known quantity of iodine-solution to a known amount of Yeast-culture shows these variations most sharply. The change of tint by heat after iodine-action, and the destruction of the intracellular glycogen by saliva, also give very clear results. L. ERRERA, Brussels. OSMOTIC OPTIMUM AID MEASUREMENTS. — Recent researches made by Dr. F. Van Rysselberghe in the Botanical Institute of Brussels have shown that vegetable cells generally answer an osmotic stimulus by an appropriate osmotic reaction , and that the relation between stimulus and reaction follows, within wide limits, the ‘law of Weber/ Hence results the possibility of predicting the existence and value of an osmotic optimum. Let n be the normal osmotic pressure in a given cell ; x the osmotic pressure of an external solution applied as stimulus ; R the reaction, i. e. the change in the osmotic pressure of the cell in response to this stimulus. Then one has, according to Webers law : X R = c log — ( c and ^ being constants). The total value of the osmotic pressure in the cell is of course R + n, and its excess over the pressure of the surrounding solu- tion is, y = R + n — x, X or y — c log — +n — x. It is easy to find by differentiation that this excess has a maximum value when x — c log e ( e being the basis of the Napierian logarithms = 2,7182818 . . .). Experiments made with Tradescantia , Symphoricarpus) Allium , Elodea , Spirogyra , agree most satisfactorily with these theoretical results. Additional interest arises from the fact that these values of x really Notes . 569 correspond to optimal solutions, in which the cells live longer than in any other. The investigations just alluded to have proved that de Vries’ constant isotonic coefficients , excellent as they are for a first approxima- tion, are not sufficiently exact for more minute experiments. Here it is advisable to use, instead of them, the coefficients of electric con- ductivity, which vary slightly with the concentration of the solution. Thus, osmotic pressures are not strictly proportional to the con- centration of the plasmolysing solutions, and these pressures ought no more to be expressed in molecule-grams of NOsK, as is now generally done. The use of an atmosphere as unit, though better, is also objectionable, as it varies from one place to another. I would therefore suggest to adopt the C.G.S. unit of pressure, viz. 1 dyne per sq. cm., or rather (to avoid useless decimals) 1 myriadyne per sq. cm ., i. e. the pressure of 10,000 dynes per sq. cm., the dyne being the force which gives the mass of 1 gram in 1 second an acceleration of 1 cm. per second. This unit is roughly equal to TJ