eeare seers. i asa ayia ob ate Har ie junit oe | sitet Ky ba! 414% stn had eat Ne VAY hs sah vier iptida so ps 19 byt P-8htetendet N26 obey ai he 4 3} ‘ » of; “feh militant ts seek is vires atavatinaisectastod oh 44 et ad Wh Pia nh re et sia i s +) ea is cfs y An ta +4 a ph ody ur) 4 ay it ae =i oF ol ih eth ol oe i papal Peel Wdaeaye v6 it . rye " se hte js ot) Mhaaan a rhanet 9 iiattt Ho th 4D oN thi oy i rieteopsieth Prada) eens ny wiriae ata ‘Fecha d eee yl ia yh dy) E>. ee i rset ie ee has aby sSisininid AW sin chee ee " ashe 4 4 Wy Mab o4 i. ab sh seth | ne hifi "ie nt ve deded sich) ~ mes Se postal wie ‘ Saab AaAGE ul anseieeriegertis ‘ 0 ath abe: 1649 Mipe PSA gd Dd) ei eind of Hen ed Pe yinntne ts ‘ hi tr a 3 pit ? pre fe a Husdeste y eertee als NE Shih batihes miss Wathen itaae ishila isin ayise pitas ‘ iin bl bask ty 4 1 char tt v ase { Mdedvhsbablacbibunree wie abba ttt Hbh ' HW bad bv en ¢ vb pet) had Wig sete sjotet + y +hepy) reat rye tr rd or ay ui (ae Henenly ones PF NGA e edie soehed vaehtoearstiogy habehitabie | aks ote ” ites yet st ene) Pad she pees by ie situinertatpnaeon y tyes oh 0 Oh np pba sae PURAR! ptronihe sates i iyitsietrtabet rhee bevy Aa ca ry iia dy vbspeeeaatert net . oral ha ry ‘rial sete ai adh abe Mesa a ak hs sevtohaing + ¢v eicerpenes raiht8 06 96 be pain. otlese 89) alias wage bobs Fal SHE a bed 4 ated rite hth ie ae + hy At . ate 1 rita aegiteas fe ; as + freghige tithe bshwasenitete APD dr oh y bunt » onal ee thetvilos Bitatacaiaeretit +>) OP ATE 4 78 oi wre 4 yin ve mapbee. Rave weirel sie a4 ore bre 0 bh DOM es vO ri at or eh es shee reberergeyisetedin er phtiy Daeererr ety HEADS Eh bets y\ntr) iste ss otyahin veh « a Hyenay ne sh pete ran bests PERSE Pa ” A Pedra, > sid He ie btvie bye eestor Ay! pat pat Pit Hisar daliy 4 iaioseg. coma sl) hod af + PET bi ed yh. Aisa pnt i iit a rat ingest +09 28 an bo A Ji) a) + ste Pace bbs Ne iitihe iste Hh eM ronan i ae ee etal ng ak vi italia bea eit " “vl hob 8 at Taaesg Athy Hid} a ae he his eats » hts pte cint ie oint ad bubeh hie sass ahutaceerist frfubifvde ae: ona ae 4) eae yess Steet areidel wt» rein toiatate sicaant ath? ett sein rater Fatgsanes ritrteryae Saat ices os vt et) sits 4 Soi pedi bal) ie gies d Uy Pa wh Sib eth aA vay sh Ett) Pestana nh tire 29h pe eat ips edit Boat oly ea f olf vi vii pe + st sth + BY ee ania staf rah nie a piety mri bvby hi * Departed dre vinhenea be) % ca ie) ee 5s oC abe iad Dai) ona bt Ht tite sity dad e irintere Sebel chins ries ert - ut ich bobeteity¢) vane Atay tc, ahrirse 0s ah bea bak vio V4 04 P5289 9) = i a Sint ft be wdelyah meet? wer! 1 bet ubalipbyp ts, Maas SH yy ayo wekas srr pH o Sinan ity pyrite pao ke Wi a) 9 reesbamapeyty i piathh vatinnt My ah ps. nt yay seiaa teat rab eps 6 Say eaecamaet. (inet a) oat pasawientteass sth sere Haletshadeii dace babid eat 4 = Aehisetsrerraie psa ha eater rheasiataas baa taoaess Rectordeorss he Wibod ck ny Ne Hea ney ratehs Se daaihs su badaba den weal en ny rt satire teeta ah eaepered Ie MY vt papa betas ‘ WeboWee en yy sie tt Hd * isis) - Beatty mI r is oe coe pearieye pads iee arrears Sy pine 04 08 # ed > Cerrar Petes es tr ge! reo ba, pal dite Gora pte he Line pans apis Wasaepeies, ae Bie , neh Ltbeate sebhagose oy . tres FOR THE PEOPLE FOR EDVCATION FOR SCIENCE LIBRARY OF THE AMERICAN MUSEUM OF NATURAL HISTORY ! | 1 i ny N whi \ iM J 7 > or - te vey —— ee ee ee oe hie ea: WG QUARTERLY JOURNAL oF DA Ob \4 L\s MICROSCOPICAL SCIENCE. HONORARY EDITOR: Sir RAY LANKESTER, K.C.B., M.A., D.Sc., LL.D., F.BS. EDITOR: EDWIN 8S. GOODRICH, M.A., F.BS., PROFESSOR OF COMPARATIVE EMBRYOLOGY IN THE UNIVERSITY OF OXFORD; WITH THE CO-OPERATION OF SYDNEY J. HICKSON, M.A., F.R.S., BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER ; GILBERT C. BOURNE, M.A., D.Sc., F.B.S., LINACRE PROFESSOR OF COMPARATIVE ANATOMY IN THE UNIVERSITY OF OXFORD; J. GRAHAM KERR, M.A., FRS., REGIUS PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF GLASGOW ; E. W. MACBRIDE, M.A., D.Sc., LL.D., F.B.S., PROFESSOR OF ZOOLOGY AT THE IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY; G. Po BIDDER, M.A., Sc.D: VOLUME 66. New Series. WITH LITHOGRAPHIC PLATES AND TEXT-FIGURES. OXFORD UNIVERSITY PRESS, HUMPHREY MILFORD, AMEN CORNER, LONDON, E.C. 4. 1922 bs raat ‘J OM StS Page 0 ee aay ee AGODA | TYROVEIK PBN k ~ rid AAED: LON: TRAN TE ey \ an WUE At) ke Td 1 TObeud - ' we ae . ! ht | Bon 3 . fs ‘ i . S or | _ 2>-Q1On| Wrowely 23 < PRINTED IN ENGLAND _ AT THE OXFORD UNIVERSITY PRESS — CONTENTS CONTENTS OF No. 261, N.S., March, 1922. MEMOIRS : PAGE The Cytoplasmic Inclusions of the Germ-Cells: Part X. The Game- togenesis of Saccocirrus. By J. Bronrii Gatensy, B.A., D.Phil., D.Sc., Professor of Zoology and Comparative Anatomy, Trinity College, Dublin. (With Plates 1-4 and | Text-figure) . : ' 1 On the Development of the ‘ Enteronephric’ type of Nephridial system found in Indian Earthworms of the genus Pheretima. By Karm Narayan Baut, D.Sc., D.Phil. (Merton College, Oxford). (Plates 5-7 and 8 Text-figures) . : 3 : : ‘ : The Occurrence of Situs inversus among artificially-reared Echinoid Larvae. By Hrrosur Onsuima, Assistant Professor in the Department of Agriculture, Kyushiu Imperial University, Fukuoka, Japan. (With 3 Text-figures) . : : Os Note by Professor E, W. MacBride on Mr. H. Ohshima’s paper . =a 149 The Behaviour of the Golgi bodies during nuclear division, with special reference to Amitosis in Dytiscus marginalis. By RecinaLp James Luprorp, B.Sc., Department of Zoology and Comparative Anatomy, University College, London. (With 4 Text-figures) 2 451 On the anatomy and affinities of Paludestrina ventrosa, Montague. By Guy C. Ropson, B.A. (With 12 Text-figures) . 159 49 CONTENTS OF No. 262, N.S., June, 1922. MEMOIRS : The Gastric Mucosa. By Ropert K. 8. Lim. (From the Department of Physiology, Edinburgh University.) (With Plate 8 and 1 Text- figure) “ ‘ : : : - : . ‘ : On the Labral Glands of a Cladoceran (Simocephalus vet ulus), with a description of its mode of feeding. By H. Granam Cannon, B.A., Demonstrator in Zoology, Imperial College of Science, South Kensington. (With Plates 9 and 10 and 2 Text-figures) . , ea Surface Tension and Cell-Division. By J. Gray, M.A. (With 9 Text- figures) : : ; : : : : ; On the Classification of Actiniaria: Part III. Definitions connected with the forms dealt with in Part II. By T. A. SrepHENson, M.Sc.. University College of Wales, Aberystwyth . - 4 é 947 On the Post-Embryonic Development of certain Chalcids, Hyper- parasites of Aphides, with Remarks on the Bionomics of Hymen- opterous Parasites in General. By Maup D. Havinanp, Research Fellow of Newnham College. (With 7 Text-figures) . 3 Siege Animal Chlorophyll: its Relation to Haemoglobin and to other Animal Pigments, By Joun F. Fuuron, Jr., Magdalen College, Oxford : ; , 3 : : ; ; , 187 339 CONTENTS OF No. 263, N.S., September, 1922. MEMOIRS : Further observations on Chromosomes and Sex-determination in Abraxas grossulariata. By the late Professor L. DONCASTER, F.R.S., University of Liverpool . : : ; : , ; The Infra- cerebral Organs of Peripatus. By Wuouiam J. Daxry, D.Sc., F.Z.8., F.L.S8., Derby Professor of Zoology, University of Liverpool. (With 4 Text- figures) A Critical Study of the Facts of Artificial Fertilization Bee Normal Fertilization. By J. Gray, M.A., Fellow of Fees College, Cam- bridge. (With 1 Text-figure) Calma glaucoides: A study in sda piation! By T. J. Evans, M. ‘. (Oxon.), Lecturer in the Medical School, Guy’s Hospital, London University. (With Plate 11 and 3 Text-figures) The Segmentation of the Head in Squalus ocarthwam By G. RYLANDs DE Beer, B.A., B.Sc., Demonstrator in the Department of Zoology and Comparative Anatomy, University Museum, Oxford. (With 13 Text-figures) Some Notes on the Gametogenesis of Ornithorhynchus. paradoxus. By J. Bronré Garensy, M.A., D.Phil. (Oxon.), D.Sc. (Lond.), Professor of Zoology and Comparative Anatomy, Dublin University. (With Plates 12, 13, 14, and 1 Text-figure) . Note on the Comparative Effects on Tissues of Tp beutie Saline endl Distilled Water when used as Solvents for Mercurie Chloride and Formol in Histological Fixation. By H. M. Car.etron, Lecturer in Histology, University of Oxford On the Structure of the Alimentary Canal and its Foonesias in the Bee (Apis mellifera, L.). By E. N. Paviovsxy, M.D., D.8c., Professor at the Military Academy of Medicine, Petrograd, and E. J. Zarty, M.Phar., Professor at the University of Latvia, (With 3 Curves and Plates 1H soca Way 8 () Se : CONTENTS OF No. 264, N.S., December, 1922. MEMOIRS : The Nature of certain Ovum-like Bodies found in the Seminiferous Tubules. By F. A. E. Crew and Honor B. Fern. From the Animal Breeding Research ec ci aicame tia = of Edinburgh. (With Plates 18-23) . : Glossobalanus inate a new See = Entersipaesiaiat from the North Sea. By ALEXANDER MEEK. (With 14 Text- figures) : : A : : : Studies on Insect Speerontapenesis 'V. On the Formation of the Sperm in Lepidoptera. By Ropert H. Bowen. From the Department of Zoology, Columbia University. (With Plates 24-26) . On the Biology and Structure of the Larvae of Hy droplates caraboides, L. By E. N. Paviovsxy, M.D., D. Sc., Professor of Zoology at the Military Academy of Medicine, Petrograd. (With Plate 27 and 16 Text-figures) A further Account of the a eee of Tie By H. Gaal Cannon. (With 1 Text-figure) . Cannibalism in Amoeba vespertilio (Penara): By Gace Lapace, M.Sc., M.B., Lecturer in Zoology, The V. ictoria University of Manchester. (With Plates 28, 29, and 3 Text-figures) PAGE 397 409 419 439 457 501 509 627 657 669 The Cytoplasmic Inclusions of the Germ-Cells: Part X. The Gametogenesis of Saccocirrus.! By J. Bronté Gatenby, B.A., D.Phil. (Oxon.), D.Se. (Lond.), Professor of Zoology and Comparative Anatomy, University of Dublin. | el H co Or Ms 8. With Plates 1-4 and 1 Text-figure. CoNTENTS. . INTRODUCTION . Previous WORK . : . TECHNIQUE AND MATERIAL . . SPERMATOGENESIS (a) The Spermatogonium of Saccocirrus (b) The Spermatocyte (c) The Spermatocyte Divisions . (d) The Newly-formed Spermatid (e) Spermateleosis (f) The Fate of the Golgi Apatite ani ing f ‘Sper fiBbcledais . THe Fate OF THE TAIL OF THE SPERMATOZOON DURING ENTRY . THE OOGENESIS OF SACCOCIRRUS . (a) The Nucleolus during Oogenesis (b) The Mitochondria (c) The Golgi Apparatus (d) The Formation of Fatty Yolk (e) Chromophility during Oogenesis (f) On Peri-nuclear Activity : SomE CHROMOPHILITY REACTIONS OF THE ous Guisconme OF CYTOPLASMIC GRANULES CENTRIFUGE EXPERIMENT ON THE Geienin Goon PAGE Oo ho “T-1 DG GS (eo) 11 14 14 15 16 17 17 18 1 The materials used for this research were purchased by a Government Grant of the Royal Society, for which thanks are expressed. No. 261 B 2 J. BRONTE GATENBY PAGE 9. CELLS FOUND IN THE MALE, INTERMEDIATE BETWEEN SPERMATO- CYTE AND OOCYTE . : : : - ; : owe 10. Discusston : ; : : 2 : ; : ee: | (a) General : : 21 (b) The Preciseness of Modern Peon ‘a the Oytontaciaae Inclusions and for Chromatin : 22 (c) On the Supposed Chromatinie Nature of Kaseniod Nucla lar Material ; : 24 (d) On the Special Part play a ty dig Nuslenk duels Oogenesis . ; ; ; - : Pere (e) Schaxel’s C Groiatin iaaaion ae 30 (/) Centrifuge Experiments in Annelid Dev slpatoae aad what they demonstrate - 32 (g) The Probable Part played by the Mitectinndiein oad Golgi Apparatus in Heredity : : : : : ae 11, SUMMARY . : ; : : ‘ : , : o! som 12. BIBLIOGRAPHY . : F : : ‘ : ; . 42 13. EXPLANATION OF PLATES : i ; ; : - 2 ot 1. INTRODUCTION. In this paper I have attempted to shed further light on that difficult problem, the oogenesis of an Annelid ; while Sacco- cirrus, as Goodrich has shown (26), is an Archi-annelid closely related to Protodrilus, it is probably much like other Annelida so far as its oogenesis is concerned. I was influenced to undertake a study of the oogenesis of an Annelid by three special reasons, namely, Schaxel has described ‘chromatin’ emission in Aricia foetida, a Polychaete ; Hempelmann and Buchner have both deseribed most peculiar peri-nuclear extruded bodies at one period of oogenesis of Saccocirrus; and, finally, Saccocirrus is an example of precocious entry of the sperm into the young oocyte: we know that the head of the sperm lies quiescent till the oocyte is ripe, and I believe that this would offer the opportunity of observing whether the tail of the sperm entered the oocyte, and (if so) whether it- took any special part in fertilization. GAMETOGENESIS OF SACCOCIRRUS 3) The behaviour of the cytoplasmic inclusions has been the object of special attention, and I have also gone into the question of the spermatogenesis, and showed in these stages the presence of peculiar yolk-spheres which are divided out between the daughter-cells at the maturation divisions. It is extraordinary the way such metaplastic bodies are shepherded into two groups during the divisions of the spermatocyte. This is the concluding part of this present series of papers. 9. Previous Work. The precocious entry of the spermatozoon into the young oocyte was discovered independently by three observers— F. Hempelmann, and by van Gaver and Stephan jointly. Hempelmann’s paper appeared a short time before that of the two other observers. Hempelmann has contributed two papers on this subject, one in 1906 (84), and one as recently as 1912 (85). His main conclusions are to be found in the latter paper. He shows that the sperms enter oocytes just after the end of the bouquet stage of the prophases of the heterotypic division. Only one spermatozoon enters one egg. Just after the last stage of the prophases of the heterotypic division, Hempelmann describes peculiar granules which originate from the nucleus and pass into the cytoplasm, eventually leading to the formation of yolk. These granules are at first quite solid, grow large, and gradually become vacuolated here and there, so as to form a large number of small sranules lying in a single vacuole. It is just after this has taken place that the egg cytoplasm becomes filled with yolk ; speaking of these curious peri-nuclear granules, Hempelmann says, “so scheint es mir kaum zweifelhaft, dass diese Zellbestandteile mit der Dotterbildung in Beziehung gebracht werden miissen ’. With regard to the origin of the peri-nuclear granules Hempel- mann writes, ‘Auch der Nucleolus beteiligt sich wohl mehr oder weniger an dieser Dotterbildung, denn gelegentlich zeigt er sich umgeben von kleinen, sich ebenso wie er selbst mit den B 2 4 J. BRONTE GATENBY gebriiuchlichen Kernfarbstoffen intensiv fairbenden Kigelchen, die wohl aus ihm hervorgegangen sind, und die an den Rand des Keimblischens riicken, um wahrscheinlich aus diesem in das umgebende Plasma auszutreten ’. Hempelmann recognizes only one sort of plasma granule —‘yolk’, originating from the peri-nuclear * Trépfchen’. According to him, the entry of additional spermatozoa into the oocyte is prevented by the formation of a membrane ; the spermatogenesis has been studied by Hempelmann, who described the tripartite “ Nebenkern’ of the spermatid ; his material was not preserved carefully enough to allow of his giving a good account of the formation of the male gamete. Van Gaver and P. Stephan have published two short notes on Saccocirrus papillocercus. In their second paper these observers state, ‘ Notre désaccord fondamental avec Hempelmann a trait a l’époque de la pénétration du spermatozoide ; | pour nous, le spermatozoide arrive dans loocyte dés que celui-ci est différencié en tant qu’oocyte, lorsque sa taille est encore extrémement minime et avant toute formation de vitellus & son intérieur. Nous n’osons pas affirmer que l’action du spermatozoide soit la cause initiale du développement de loocyte, mais nous avons constamment trouvé un de ces éléments dans les ovules en voie d’accroisse- ment et d’élaboration vitelline’. In addition, van Gaver and Stephan believe that polyspermy and assimilation of sperma- tozoa by the cytoplasm take place in Saccocirrus. They find ‘ désintégration des tétes des spermatozoides et, par suite, leur assimilation par l’oocyte’. The articles of van Gaver and Stephan are not illustrated by figures. The latest observer to attack these problems was Paul Buchner (8). He showed that the whole tail of the sperm may occasionally enter an oocyte, and break up to give rise to a number of peripheral droplets, while the head of the sperm remains quiescent. The yolk-formation he describes as taking place by a partial breaking up of the nucleolus, pieces of which wander to the periphery of the nucleus—particles of yolk appearing simultaneously, apparently from the nucleolar GAMETOGENESIS OF SACCOCIRRUS 5 fragments which pass through the nuclear membrane. The droplets from the sperm tail shrink up and leave the periphery of the oocyte quite clear, while the nucleolar fragments grow and multiply and fill the egg with yolk. Buchner is not altogether happy in his work on Saccocirrus oogenesis, and is very cautious with regard to yolk-formation. He ends his research with the statement, ‘ Die Dotterbildung von Sacco- cirrus verdiente wohl eine eingehendere, mit Hilfe aller zu Gebote stehenden Fiarbungen und Reaktionen ausgefiihrte Analyse ’. No observer has hitherto given any account of Golgi body or mitochondria in the oogenesis of Saccocirrus, and this form has up to the present time resisted the efforts aimed at a satis- factory interpretation of the problems which its oogenesis presents. 3. TECHNIQUE AND MATERIAL. Sections of a large collection of Saccocirrus were prepared according to the plans given in my recent paper on technique (13). The Mann-Kopsch and Champy-Kull methods were particularly valuable. All the other techniques in current use were tried. The material used was sent to me from Plymouth, two lots in the month of June, another in July. The worms were cut into pieces after having been killed by being dropped whole into a capsule full of fixative. Great difficulty was experienced in making successful preparations of the Golgi apparatus. All the first trials were unsuccessful, but for some unknown reason all the latter attempts succeeded completely both with Mann-Kopsch, Cajal, and Da Fano methods (18). T have to thank Dr. Allen, F.R.S., of the Plymouth Biological Laboratory, for sending me the material desired by me. Professor EH. 8. Goodrich, F.R.S., of Oxford, kindly placed at my disposal some of his own material of Saccocirrus papillocercus from Naples, and sections cut from this proved very useful. 6 J. BRONTE GATENBY 4, SPERMATOGENESIS. (a) The Spermatogonium of Saccocirrus. The spermatogonium is a small rounded cell with a spherical granular nucleus containing a karyosome. In Pl. 1, fig. 7, is drawn one of a group or rosette of spermatogonia: there was a spindle bridge as is common in such cases (sB). This cell was drawn from a Mann-Kopsch preparation, and the Golgi body (Ga) shows as a number of black rods or batonettes surrounding an archoplasm or centrosphere. In most sperma- togonia the mitochondria may be detected in the form of a cloud lying in the region of the Golgi body, as in PI. 1, fig. 7, at M. (b) The Spermatocyte. Rosettes of spermatogonia grow synchronously to form groups of spermatocytes, and during the growth stages the mitochondria spread out through the cytoplasm, as in PI. 1, fig. 1, me. The Golgi body grows too, and the number of its individual parts (Golgi rods, dictyosomes, or batonettes) also increases, aS Shown at Ga. The nucleus at this period is often reniform. Possibly the most remarkable fact with regard to the spermatogenesis of Saccocirrus is the presence in many, if not all spermatocytes, of a group of true yolk-granules (lipin) quite separate from either mitochondria or Golgi body : in Pl. 1, fig. 1, the yolk-granules are at y, and form a special cell inclusion. By the Champy-Kull method the nucleus stains bluish, the mitochondria and Golgi rods go dark red, the eyto- plasm is yellowish, and the curious collection of yolk-granules stain brownish green in the osmic acid of the Champy’s fluid. In Pl. 1, fig. 1, at yo on the right, is drawn one of the yolk- cells which accompanied this spermatocyte, and the yolk- granules stained the same shade as the yolk-spheres (y) in the spermatocyte. In Pl. 1, fig. 2, is a Kopsch preparation which shows the effect of prolonged osmication. The Golgi body (G@A) has reduced the OsO, heavily, the mitochondria do not show, and the yolk-spherules are at y; but besides all these there GAMETOGENESIS OF SACCOCIRRUS 7 is found a group of perfectly spherical granules at x. These went black-brown in the OsO,; their true nature or origin was not ascertained, but it was thought that ultimately they formed a part of the spermatid, as will be shown later. The number of granules was about twelve in all the cases I could count. (c) The Spermatocyte Divisions. Pl. 1, fig. 4, is a first spermatocyte metaphase. The peculiar yolk-spherules have taken up a position at the middle of the spindle, and in the next stage (PI. 1, fig. 5) the granules have become sorted out into two groups (vy) subequal in size. Hach spermatid (Pl. 2, fig. 8) receives about one-quarter of the number of granules in the spermatocyte. The behaviour of the mitochondria in the divisions is peculiar: they lose their granular state, and during the prophases break down to form threads as in Pl. 1, figs. 4 and 5, in the telophase (PI. 1, fig. 5) ; the threads le chiefly around the equatorial plate. The Golgi rods are difficult to follow through mitosis; at the prophases they lose their staining power, and it is only in certain cases that the cell at metaphase has distinguishable elements (xy in PI. 1, fig. 4) which might be identified as Golgi elements. It must be admitted that no positive evidence has been adduced with regard to the Golgi elements during division of the Saccocirrus spermatocyte. (d) The Newly-formed Spermatid. In Pl. 2, fig. 8, is a newly-formed spermatid. The yolk- spherules (y) are on the right, while the mitochondria surround the nucleus ; the Golgi elements are at Ga, being scattered. In the next stage the mitochondria collect to one side of the cell, in proximity to the nucleus, the Golgi elements lying behind, as in Pl. 2, fig. 9; this cell was drawn from a Kopsch prepara- tion, and in it no yolk-spherules could be identified. At x are what I consider to be two of the granules marked similarly in fig. 2. In nearly all Kopsch or Mann-Kopsch preparations the spermatid cytoplasm is seen to be formed of very coarse reticulum, as shown in PI. 2, figs. 9 and 14. 8 J. BRONTE GATENBY (ec) Spermateleosis. The spermateleosis stages comprising those leading to the metamorphosis of spermatid into spermatozoon are very peculiar. The mitochondria, which in Pl. 2, fig. 9, le grouped behind the nucleus, begin to run together as depicted in fig. 10. Three main centres for this coalescence exist, but here and there a few separate centres exist ; these ultimately joi up with one of the larger centres, till one gets such a stage as in Pl. 2, fig. 11, where three balls of mitochondrial substance are produced (mm), while the remainder of the free mito- chondria are gradually fusmg up. By the stage of fig. 12 the mitochondria have all fused to form three solid spheres generally somewhat unequal in size, and as well, often in staining affinity. In Pl. 2, fig. 14, the three spheres are viewed from below, their unequal size being apparent. Leaving the mitochondria at this stage, the fate of the yolk- spherules and of the Golgi elements may be deseribed; at such a stage as in fig. 10 of Pl. 2 the Golgi elements lie behind the zone of the mitochondria ; as the mitochondria fuse up, the Golgi elements keep behind the most distal mitochondria as in PI. 2, fig. 11, and, finally, when all the mitochondria have fused to form the three spheres, the Golgi elements lie close up behind as in Pl. 2, fig. 12. In the case of the yolk-spherules a somewhat similar change in position has been noted: in Pl. 2, fig. 16, the yolk-spherules are on the right of the cell, but by stages in figs. 11 and 12 they have moved back behind the mitochondrial spheres. In many, but not all, examples there can be observed between the three mitochondrial spheres, a small, often round, often angular grain, as in Pl. 2, fig. 14, at x. In figs. 9, 10, 11, and 12 such bodies are also seen. These bodies, I believe, are derived from the grains marked xin Pl. 1, fig. 2. They seem to form a part of the mitochondrial sheath of the sperm-tail. The spermatid at such a stage as that of fig. 12 now begins to lengthen. In Pl. 2, fig. 13, the three mitochondrial spheres GAMETOGENESIS OF SACCOCIRRUS $ have become drawn out to form a comet-tail body attached to the spermatid nucleus. The latter has begun to undergo the usual changes which may be seen in figs. 9, 10, 11, 12, and 13. In Pl. 2, fig. 16, the tail has further elongated, the yolk- spherules and Golgi elements are drifting down the length of the tail, while the nucleus is losing its reticular arrangement. This lengthening process now goes on till the fully-formed elongate Saccocirrus spermatozoon is produced. Owing to the small size of the cells, and possibly to the unsuitability of the material, no satisfactory description can be given of the formation of the acrosome. This can be seen in Pl. 2, fig. 18, at as. In figs. 9 and 10, at ex, were bodies which might have some connexion with the formation of the acrosome, but of this 1 am unable to speak with certainty. (f) The Fate of the Golgi Apparatus during Spermateleosis. If one examines a bundle of ripe or ripening spermatozoa in material prepared by the Mann-Kopsch method, it will be found that near the end of the tail of each sperm are rounded or angular bodies which are stained by the osmic acid: in Pl. 2, fig. 17, is drawn such a bundle of sperms, the bodies being marked Gax; at a higher magnification, as in fig. 15, the bodies are seen to be very like the Golgi body drawn in the spermatogonium in PI. 1, fig. 7, at aa. I cannot say for certain whether these bodies are derived from the original Golgi rods depicted in PI. 2, figs. 8-13, but 1 should think that they are so derived, and that they have undergone some change during late spermateleosis. If these bodies are an integral part of the spermatozoon and not the degeneration products of the spermateleosis, one might expect to find some sign of them in the spermatozoa within the female : in Pl. 2, fig. 18, is section of the receptaculum seminis ofa female, and drawn to the same scale as fig. 17 above. The same bodies are to be seen at the tails of the ripe sperma- tozoa. 10 J. BRONTE GATENBY 5. Tue Fare or tHe Tar or THE SPERMATOZOON DURING ENTRY. Buchner (8) showed that the sperm-tail sometimes entered the egg, sometimes not. I have been able to confirm these observations. In Pl. 3, fig. 25, is a typical oocyte to illustrate this: the sperm-head (sp) is wrapped around the oocyte nucleus, while the remains of the tail of the sperm are, in this section, seen as four irregular chromophile bodies at spr. In Pl. 3, fig. 22, in the lower oocyte, the sperm-head is at sp, while the tail is cut across as two irregular bodies at spr. In the upper cell of fig. 22 the sperm-tail seems to be partly inside the egg (upper) and partly outside (lower). While it is generally impossible to say whether these irregular masses (which we can positively identify as remains of the sperm-tail) are, or are not, inside the egg cytoplasm, when we examine eggs at a little later stage of growth, it is quite certain that in the majority of cases the sperm-tail fragments have not only entered the egg but have broken up to form a number of spherical, extremely chromophile, bodies at the periphery. In PI. 3, figs. 24, and 23 at spr, the beads are noted all around the periphery of the oocyte. If Mann-Kopsch preparations be examined for this, the beads appear a pale yellow colour as in Pl. 3, figs. 19 and 21, at spr. In some cases it certainly appeared that the number of beads derived from the remains of the tail of the sperm increased in number as the egg grew. ‘This was probably what van Gaver and P. Stephan thought when they believed that the spermatozoa might have something to do with yolk-formation. I do not believe, however, that the beads take part in fertiliza- tion or yolk-formation, either directly or indirectly. Later on they either disappear or become hidden by the formation of clouds of yolk or nucleolar deutoplasm (deseribed below). 6. THE OOGENESIS OF SACCOCIRRUS. The oogenesis has proved the most difficult problem that I had hitherto attacked, and at one time I despaired of ever GAMETOGENESIS OF SACCOCIRRUS 11 unravelling the intricate story of the origin and nature of the complicated granulations of the oocyte cytoplasm. After a year’s work, and the making of a large number of preparations, I feel that this present account is the correct interpretation of the oogenesis. The egg cytoplasm of Saccocirrus contains four kinds of grains or formed bodies: (a) Golgi elements, (b) mitochondria, (c) true yolk, (d) nucleolar extrusions or plastin-deutoplasm. These can all be distinguished one from another by some staining method, as described on p. 18. (a) The Nucleolus during Oogenesis. Both Hempelmann and Buchner noted the peculiar peri- nuclear bodies drawn in PI. 3, fig. 23, nu, and concluded that they were in some way concerned with yolk-formation. Such a marked process as that depicted in this figure is unknown in any other animal; the history of the formation of these extraordinary attachments to the nuclear membrane is not at all easy to make out, and it is only after a study of material fixed in Champy-Kull and stained by Benda’s crystal violet and alizarin that a satisfactory conclusion can be reached. In Champy-Kull-Benda preparations the nucleolus stains a very characteristic orange-brown shade, while the mito- chondria and chromatin are in shades of violet; true yolk (derived from the Golgi apparatus) is stained by the OsO, of the Champy’s fluid. Now in such preparations the nucleolus of the young oocyte is found to be budding off small pieces, as shown in Pl. 3, figs. 24 and 25; these pieces appear to wander to the periphery of the nucleus and to pass through, but to remain plastered upon the outer surface of the membrane, as in PI. 3, fig. 24, nu. Some considerable variation in the exact method of this process is found: in certain cases the pieces broken from the nucleolus are coarse and easily distinguishable, as in Pl. 3, fig. 24, but in some other examples, of which fig. 25 is hardly typical, the broken-off pieces are so small that they are difficult to identify. 12 J. BRONTE GATENBY Occasionally, as in Pl. 3, fig. 22, the nucleolus may be seen to be differentiated at its periphery into a number of small stainable bodies which may represent the beginnings of the parts to be extruded. In Pl. 3, fig. 23, is a cell showimg the appearance of the nucleolar extrusions after staiming in iron haematoxylin or Champy-Kull, while in fig. 21 is a cell treated by Mann-Kopsch and the nucleolar extrusions appear as pale yellowish bodies. From the stages represented by Pl. 3, figs. 21 or 23 onwards, there is generally marked difficulty in ascertaining the exact fate of the nucleolar extrusions. This is due to the fact that just about this period a second process is set into motion ; this consists of an appearance all around the nuclear membrane of a chromophil cloud, which in most preparations obscures the nucleolar extrusions ; an exaggerated example of this is drawn in Pl. 4, fig. 35, from a silver nitrate Da Fano prepara- tion, but the cloud does not show so darkly with Benda or iron alum haematoxylin. At all events there begins at this period a_peri-nuclear activity, which also corresponds with the change of the chromo- plulty of the egg cytoplasm from a primary oxyphilia to a basophilia. Two other occurrences also tend to obscure the peri-nuclear nucleolar bodies at this period: around each body a Clearly-defined vacuole often appears (PI. 4, fig. 29, Niv, from a Mann-Kopsch preparation), and moreover the mito- chondria near the nuclear membrane are now forming actively- growing and dividing clusters. With iron alum haematoxylin it is not possible to make sure as to the fate of the nucleolar extrusions, because these and the mitochondria stain in the same colour. With the Champy-Kull fixation and Benda stain IT have found examples which, I believe, establish as a fact my view that the nucleolar extrusions first lose their connexion with the nuclear membrane and then either pass right away into the cytoplasm or immediately begin to break up into much smaller fragments. As with the mode of appearance itself of the nucleolar extrusions, so also the subsequent behaviour of these bodies is open to a good deal of variation. GAMETOGENESIS OF SACCOCIRRUS a3 With almost a suddenness a large number of nucleolar yolk- bodies appear in a ring surrounding, but some distance away from, the nucleus (PI. 4, fig. 34, is a somewhat later stage). The appearance of this ring of numerous nucleolar yolk- bodies corresponds more or less closely with the spreading out of the basophil peri-nuclear cloud referred to below (p. 17). The next period sees the complete change of the cell from primary oxyphilia to a temporary basophilia: this activity is often shown plainly with the Mann-Kopsch osmium tetroxide method, of which Pl. 4, fig. 29, is an example; the whole appearance of the cell seems to change. Later the peri-nuclear vacuoles disappear, the cytoplasm becomes smooth, and the nucleolar yolk-bodies are the most noticeable element in the egg. In Pl. 4, fig. 80, is an egg fixed for six weeks in formol- Flemming ; the true yolk (derived I believe from the Golgi elements) has gone black with the osmic acid, while the nucleolar yolk-spheres are pale, in this case fuchsinophile, bodies ; neither mitochondria nor Golgi elements appear in this preparation. That the nucleolar deutoplasm or yolk-spheres go on dividing in the egg cytoplasm seems to me a very likely suggestion, but I was unable to prove that such was the case. How then otherwise can we account for the extraordinarily rapid increase of clouds of these alizarin-stamimg granules such as appear in Pl. 4, fig. 834? It seems certain that smaller nucleoli inside the nucleolus keep budding off extra-nuclear fragments (Pl. 4, fig. 34), but this would not account for the arrangement and rapid growth of clouds of granules such as those at Nu in Pl. 4, fig. 34. In Pl. 4, fig. 80, which was drawn from a very clear example where the nucleolar yolk-spheres were large, I could not see any of the latter undergoing binary fission; 1 am therefore disposed to believe that these cytoplasmic nucleoli bud off little pieces just as the larger nucleolus is doing in the cell drawn in Pl. 4, fig. 84; and then these little pieces themselves orow larger. In Pl. 4, fig. 31, is a part of the cytoplasm of a nearly ripe 14 J. BRONTE GATENBY oocyte ; it will be noted that there are now enormous numbers of granules formed, and the majority of these are nucleolar deutoplasm derived from the original nucleolus of the oogonium. (b) The Mitochondria. In the young oogonia I did not find it possible to demonstrate mitochondria, but in all the oocytes just at or after the last stages of the prophases of the heterotypic division, mitochondria are easily identified, especially after proper staining in iron alum haematoxylin. In Pl. 3, fig. 22, are two oocytes showing the fine grey-staining bodies which I have identified as mito- chondria. These show more clearly in Pl. 3, figs. 24 and 25 ; the mitochondria do not appear to have anything to do with the nucleolus. (c) The Golgi Apparatus. The Golgi apparatus (Golgi body or element) was studied by the Cajal, the Da Fano, and the Mann-Kopsch techniques ; of these the Mann-Kopsch technique was the most suitable. In young oocytes the Golgi apparatus consists of an excentric juxta-nuclear mass, as at GA in Pl. 3, fig. 20. This mass really lies around an archoplasm, as in Pl. 3, fig. 24,at ar. In Pl. 4, fig. 88, is an oocyte showing the Golgi apparatus on the right of the nucleus. Now in the youngest oogonia the Golgi body is isolated at one side of the cell, but quite early in the history of the progerminative oogonium it grows rapidly and begins to fragment ; the additional pieces so derived move out into the other regions of the cytoplasm, as has already happened in all the three cells drawn in fig. 20 of Pl. 3. In fig. 33 on the same plate (though the oocyte is drawn much older in so far as the extrusion of nucleolar deutoplasm is concerned) the Golgi body is still fairly isolated, being just in process of fragmentation. In some cases, as pieces break off from the original body, they pass away into the free parts of the eyto- plasm and form remarkable nests or areas of proliferation, as in the cell in Pl. 4, fig. 28, at aa. While in certain cases the fragments of Golgi body scarcely retain their semi-lunar GAMETOGENESIS OF SACCOCIRRUS 15 spherical condition, in other examples this condition is retained, as in Pl. 3, fig. 19, which was a remarkably clear example ; the neighbouring cell in fig. 21 showed this condition less well, while in fig. 83, most of the smaller fragments were of no special shape. In Pl. 4, figs. 28 and 29, the Golgi elements form a fine dust at the periphery of the cell. In all the cases, however, the ultimate result is the same—the apparatus breaks up into hundreds of irregular grains, as in PI. 4, fig. 82, at Ga. If the Saccocirrus is prepared by the Mann-Kopsch method, and the sections mounted in balsam without any previous treatment, the cell-granules of the oocyte appear as in Pl. 4, fig. 31; here we find a confused mass of granules which have become either blackened or browned to different degrees. But if the sections on the slide be treated for several hours in turpentine all but the Golgi granules become decolourized or a light yellow in colour. In PI. 4, fig. 33, the egg-granules were throughout the colour of those in fig. 31, but the slide was treated in turpentine and the colour extracted from everything except the apparatus, which is here seen to be fragmenting and spreading through the cytoplasm. (d) The Formation of Fatty Yolk. In the egg of Saccocirrus which has been centrifuged, a layer of fatty yolk of an oily type collects on the upper pole of the egg (see Text-fig. 1). The characteristics of this yolk are that it goes greenish or brown only after prolonged osmication and is rapidly destroyed by fixatives containing lipoid solvents. Such fatty yolk is quite distinct from both nucleolar deuto- plasm and mitochondria, but it is best shown by Kopsch techniques which demonstrate the Golgi elements so well. In Pl. 4, fig. 30, the fatty yolk is shown black and the nucleolar deutoplasm yellowish grey, after prolonged immersion in formol-Flemming. From the method and time of appearance of the fatty yolk I beheve it is formed from the Golgi bodies, but I admit it is impossible to make a trustworthy statement in such un- favourable material. . 16 J. BRONTE GATENBY (ec) Changes in Chromophility during Oogenesis. Method I.—Fixation in saturated solution of corrosive sublimate, staining in Ehrlich’s haematoxylin and eosin accord- ing to Scott’s directions (55). The nucleolus of the oogonium is amphophil with distinct basophil preponderance, i.e. more blue than reddish purple. The chromatinic reticulum becomes oxyphil after a certain time, and remains so throughout oogenesis ; the oogonial cytoplasm is oxyphil. During oogenesis at the period of the appearance of the peri-nuclear bodies (nucleoli, Pl. 3, fig. 28) the cytoplasm becomes basophil, especially near the nucleus. This basophility persists in a peri-nuclear position for a considerable time and spreads out, but gradually the entire cytoplasm again becomes completely oxyphil. A typical somatic cell (e.g. gut, or epidermal) shows an oxyphil cytoplasm and basophil nucleus. The head of the sperm is basophil, the tail oxyphil. Method II.—The same material stained by eosin and toluidin blue (in this order) offered a new point of view. Somatic nuclei were blue, the sperm-head and most of the epidermis cytoplasm blue also. The body-muscles and the tail of the sperms were red. The oocyte cytoplasm had an oxyphil ground, but the nucleolar deutoplasm was bluish. The nucleolus itself was generally amphophil, and, as in the case of the Ehrlich preparation, had a basophil central core and an ampho- phil cortex with basophil preponderance. In other cases the nucleolus was entirely oxyphil. The peri-nuclear bodies were completely blue. While these results are in themselves of little importance from the point of view of the detection of peri-nuclear ‘chromatin omissions ’, they show very clearly that at the time when the primary oxyphilia is changing to the basophilia there is great new activity in the region of the nucleus ; this activity leads to the formation of new denser cytoplasm, and it will be noted below that there is a correspondence between the pictures given by methods explained above and with formalin silver- nitrate or chrome-osmium techniques. GAMETOGENESIS OF SACCOCIRRUS 17 (f) On Peri-nuclear Activity. In Pl. 8, fig. 25, is drawn a youtig oocyte at a time when the nucleolar deutoplasm is being formed ; this cell was prepared by Da Fano’s cobalt-nitrate-formol-silver-nitrate method. It is remarkable for the fact that it demonstrates very clearly the extraordinary peri-nuclear activity at this stage of oogenesis. It is possibly this material which stains basophil as described on p. 16. In all my silver-nitrate preparations the oocyte at this stage shows this peculiar appearance. The peri-nuclear cloud stains black or grey according as to whether the preparation has or has not been toned, while the nuclear reticulum and nucleolus are only faintly yellowish and may subsequently be stained bright red in safranin. I look upon this cloud as the direct result of active protein metabolism around the nucleus ; the protein is possibly forming under stimuli sent forth from the nucleus. There is no evidence that this cloud is chromatinic, for the Golgi silver-nitrate methods (Golgi, Cajal, Da Fano) do not impregnate chromatin in any cells I have studied. There is, in addition, no evidence of intra-nuclear specks or dust as described by Schaxel for Aricia foetida. Later on this cloud disperses through the whole egg cyto- plasm. 7. Some Cyromoputiuity REACTIONS oF THE Four CATEGORIES OF CYTOPLASMIC GRANULES. The table given on p. 18 summarizes the differences which can be shown to exist between the four categories of cyto- plasmic granules found in the egg of Saccocirrus. Only those methods which best show these differences are mentioned, but besides I used many other fixing and staining techniques (18). The nucleolar yolk-spheres or deutoplasmic elements approach somewhat in their density to the mitochondria and tend to stain rather like them. Besides these methods quoted there are such tests as the use of alcoholic or acidified No. 261 C 18 J. BRONTE GATENBY (acetic) fixatives, which either wash away the fatty yolk or the mitochondria, or both, and leave the nucleolar deutoplasm. Then there are the formalin silver-nitrate methods which stain the Golgi elements. ing : The table below shows the follow- Method 1 constitutes a difference between mitochondria and nucleolus and its derivatives. Method 8 constitutes a difference between nucleolus and its derivatives and true fatty yolk (from Golgi elements). Method 4 constitutes a difference between nucleolus and Golgi apparatus. Method 5 constitutes a difference between Golgi apparatus and mitochondria, as also do Methods 1, 2, and 3. TABLE. Nucleolus and its Golgi is L Method Used. Derivatives. Apparatus. Yolk. | Mitochondria. Champy - Kull fixation, | Yellow-brown | Does not show. Black Violet Benda’s alizarin and crystal violet / Flemming without acetic) Black Does not show Greenish- Black or acid, and iron haema- brown grey toxylin | Formol-Flemming and Reddish | Does not show Black Did not show | Altmann’s acid fuchsin Mann-Kopsch Mann-Kopsch, Altmann Yellowish Reddisk | Black, diffi- cult to de- colourize in turpentine | Black, and as above plainly Black, easy to Did not show decolourize | in turpen- tine Black, and as |Red above 8. CENTRIFUGE EXPERIMENT ON THE OVARIAN OocyTER. Live specimens of Saccocirrus were placed in a tube and centrifuged for twenty minutes at 3,000 to 5,000 revolutions a minute. fixatives of various types. They were afterwards thrown into capsules of The centrifuged egg shows three layers, viz. an upper cap, a clear subcentral zone, and ] GAMETOGENESIS OF SACCOCIRRUS 19 a large lower zone. The upper cap is formed of delicate sranules which I think are fatty yolk and probably of the Golgi elements ; these granules will go yellowish green TrExtT-Fic. 1. bu : : Gis, = re 0 x ae? i eerie rm o*. -0°e 4 t ex?,- 8, “heen Centrifuged oocyte of Saccocirrus. Shows an upper layer of greenish oily yolk (uL), clear middle layer with nucleus (N), lower layer principally nucleolar deutoplasm (NL) with an upper layer mainly mitochondrial (mM). At cu are the cortical lamellae of the egg membrane. (Chrome-osmium and Benda stain.) after prolonged osmication. The middle layer generally shows two zones—an upper just beneath the fatty yolk and staining in crystal violet, or iron haematoxylin, and looking much like thickened cytoplasmic reticula ; then there is the C2 20 J. BRONTE GATENBY large lower area formed of the heavy nucleolar deutoplasm, forming by far the largest separate part of the centrifuged oocyte. These areas are shown in Text-fig. 1. The nucleus generally lies in the middle layer. The mitochondria appear to be lighter than the nucleolar deutoplasm and take up a position in an area above the latter, m in Text-fig. 1, but are also found throughout the lower area. Around the exposed periphery of the egg, the cortical lamellae are beautifully apparent, especially in Benda preparations (cu in Text-fig. 1). It is from these lamellae that the substance of the fertilization membrane is produced. See also lamellae in Pl. 4, fig. 27, on. 9. Cruts FounD IN THE MALE, INTERMEDIATE BETWEEN SPERMATOCYTE AND OocyTE. In the coelom of the male Saccocirrus are found large cells packed with yolk-spheres ; these large cells often fill up all the coelomic space in the mid region of the body, excepting for the areas occupied by developing spermatozoa. Occasionally one finds large isolated cells lymg completely surrounded by and shut in between the large yolk-cells. These isolated cells were once young spermatocytes, which, during erowth of the yolk, have become shut off. That this is so is indicated by an examination of a sufficient number of Saccocirrus males. Now these isolated cells are sometimes remarkable for the fact that they show a rough resemblance to oocytes at the stage of nucleolar extrusion. In PI. 1, fig. 3, is drawn such a cell. The group of yolk-granules is at y, several groups of fine mitochondria are at mM, M, while the nucleus is found to be in the process of extruding large peculiar nucleoli, Nux. In this one section the nucleus showed four pieces being extruded, two other nucleoli inside the nucleus, and one piece on the lower right of the cell already detached from the nucleus ; an examination of the larger nucleoli in fig. 3 shows that they GAMETOGENESIS OF SACCOCIRRUS OF have a stout triangular base quite like the bodies in Pl. 3, fig. 23. When I came to examine my first preparations illustrating the oogenesis of Saccocirrus, I did not immediately notice the fine mitochondria drawn in PI. 3, figs. 22 and 23, and I was temporarily led to believe that the nucleolar extrusions might represent the mitochondria. But before 1 had made more and better preparations, as the result of my experience on this new material, my belief in this view that the nucleol might represent mitochondria was shaken by finding such cells as that in Pl. 1, fig. 8, in which I noted both mitochondria and bodies which, I concluded, represented the extruded nucleoli of the oocyte. With regard to the probable reason for the appearance of nucleolar extrusions in a spermatocyte, I believe that it is due to the fact that such cells are packed away among yolk- cells which bring about conditions simulating the metabolism of the egg-cell. 10. Discussion. (a) General. The oogenesis of Saccocirrus is likely to be typical of several other Annelida, and possibly of Polychaeta such as Chaeto- pterus and Nereis, judging from Lillie’s figures of centrifuged ova of these genera. A graphic representation of the oogenesis of Saccocirrus would be as follows : Oogonium. Full-grown Oocyte. 1. Nucleolus. 1. Nucleolus . : . Nucleolar deutoplasm or nucleolar yolk-spheres. Peels el : {8 Definitive Golgi elements. ee 4. Yolk-spheres (fatty). 8. Mitochondria . 5. Mitochondria. 4, Chromosomes . 6. Chromosomes only. The part of this scheme about which I feel some doubt is the metamorphosis of the Golgi element into a yolk-sphere ; 99 J. BRONTE GATENBY that such takes place in Ascidians and Molluses is now quite certain, but the egg of Saccocirrus does not provide such clear opportunity for study as that of Limnaea or Ascidians. Nevertheless, I believe that I have made sufficiently clear observations on a large amount of material to justify the above interpretations. It is only after the dispersal and breaking up of the Golgi body that true fatty yolk puts im an appear- ance, and in many cases it seems that Golgi elements can be traced step by step as the eggs grow, metamorphosing into fatty yolk. The extruded nucleolar material has nothing to do with this, and I do not think that the mitochondria are concerned in the process. When we take the case of extruded nucleolar material in Saccocirrus it is difficult to understand why the bulk of the formed reserve granules in the egg should be of nucleolar origin. If one studies the cytoplasm of the egg in a number of different examples of oogenesis, one sometimes finds that the nucleolus supplies the bulk of reserve material (Saccocirrus), sometimes the Golgi apparatus (Patella or Limnaea), some- times the mitochondria, as in certain insects. In each of these cases, however, analysis of the entire reserve materials in the egg cytoplasm leads one to a similar conclusion for each example, namely, that reserve material in eggs of invertebrates consists of protein and fat or lipin (or both these). Then, of course, most eges contain glycogen. In some examples, as in the sponge Grantia, the main bulk of reserve material seems to be delicate vacuoles of lipin originating from the ground plasma. (bo) The Preciseness of the Modern Technique for the Cytoplasmic Inclusions and for Chromatin. ° Some observers apparently unacquainted with the finer usages of modern cytological technique have written doubt- fully of the preciseness of such methods. In all the cyto- logical problems that I have attacked the difficulty I have met with les not especially in the discrimination between yolk, Golgi elements, mitochondria, nucleolar deutoplasm, and GAMETOGENESIS OF SACCOCIRRUS 23 slycogen, for this was generally easy, but im the identification of chromatin; the problem was whether basophile chromatic material was chromatinic ; this is the great problem of cytology at the present time. I believe that I may be forgiven for holding an optimistic view with reference to our present and future understanding as to the behaviour of the Golgi elements and mitochondria during oogenesis: I think that the works of Jan Hirschler, Weigl, Nussbaum-Hilarowitz, Rio-Hortega, and my own series of papers on the cytoplasmic inclusions have gone far to shed a clear ight on the subject, but I do at present feel much puzzled over the nuclear phenomenon in oogenesis.+ One is driven onward trying to avoid the pool of Charybdis, formed by the chromosome theorists who will not admit of true chromatinic extra-nuclear extrusions, and the rock of Sceylla, which in my mind is constituted by the fact that it is at times difficult to believe that the so-called extra-nuclear extrusions are not chromatin. This special matter is further discussed below under the heading of ‘ The Supposed Chroma- tinic Nature of Extruded Nucleolar Material ’. In all probability were it not for the ingenious, and one must say believable theories of chromosome workers of Mor- gan’s, Wilson’s, or McClung’s schools, one would have no hesitation in saying that the extra-nuclear extrusions were chromatin, even though they frequently do not stam quite like the chromatin of the ‘ resting’ nucleus. When one takes the case of the secondary nuclei of the Hymenopterous egg it is very difficult to avoid the conclusion that such granules are chromatin. This is a matter to which I have given a good deal of atten- tion. Quite recently I have again gone over my Apanteles material, and I have found an example which shows the chromatin filaments at the diplotene stage of the prophases of the heterotypic division, while the nucleolus is separate and shows buds, some of which have already passed into the cytoplasm to form minute secondary nuclei. 1 Mr. R. J. Ludford’s recent work has gone far to clear up parts of this obscure ground (* Jour. Roy. Mier. Soe.’, 1920-21). 24 J. BRONTE GATENBY While this helps towards a disposal of the view that the chromosomes, at this period at least, are drawn upon to provide material for the formation of secondary nuclei, it does not dispose of the questions as to the nature and origin of the nucleolus which buds off the secondary nuclei. It is possible to recognize several kinds of nucleolar activity in various examples of oogenesis. : ; definite nucleolus. Saccocirrus nucleolus nucleolar deutoplasm. definite nucleolus. Apanteles nucleolus . com nuclei (associated with yolk-formation). definite nucleolus. Grantia nucleolus eda {mitochondria (‘ chromidia ’). In certain other forms it is possible to recognize a process of nucleolar extrusion early in oogenesis, but which appears to lead to nothing (possibly in Patella). The belief held by some observers that nucleolar extrusion may be looked upon as a process whereby the nucleus sends chemical messengers into the cytoplasm imducing growth to begin, is discountenanced, for Saccocirrus at least, by the very apparent fact that nucleolar extrusion is prone to much variation in the point of time and the rate that it takes place— as shown by comparing the sizes of the eggs in Pl. 3, fig. 23, and Pl. 4, figs. 29 and 33. The process is just beginning in the first-mentioned figure, and has finished in the smaller egg in the last-mentioned figure. (c) On the Supposed Chromatinic Nature of Extruded Nucleolar Material. If one fixes the testis or ovary of any animal in Zenker or Petrunkewitsch fluid, and stains in Ehrlich’s or Delafield’s haematoxylin and eosin or Biebrich’s scarlet, it will be noted that during the greater part of the development of the sperm the chromatin stains blue, or basophil; but there are certain periods when what we can only assume to be true chromatin, GAMETOGENESIS OF SACCOCIRRUS 25 as it is morphologically derived from preceding materials which stained like chromatin, will be found to stain oxyphil, or in the red stain. As Bayliss especially has shown clearly in his valuable ‘ Principles of Physiology ’, stainmg depends on a number of more or less obscure factors, and it is probably injudicious to lay too much weight on the results of staiming fixed material. In many of the parasitic Hymenoptera the egg nucleus contains a large heavily-staining nucleolus which buds off fragments, which pass through the nuclear membrane into the egg cytoplasm, where they form what are known as secondary nuclei. With safranin and light green the nucleolus of the true egg nucleus stains red, and the nuclear (chromatinic) network a green colour. In the sponge Grantia the plasmosome of the oocyte partly passes into the egg cytoplasm to form bodies called by Joérgensen and Dendy ‘ chromidia’; I have objected to the use of this term for such nucleolar fragments, both because we do not know that they are chromatinic and also because such ‘nucleolar’ extrusions appear to be identical with the mitochondria. We must face the facts frankly: the chromosome theorists would object to the identification of ‘nucleolar’ extrusions as chromatinic in nature and as derived from the definitive chromosomes. I have shown above that staining tests are not conclusive; several others, and also I myself, have demon- strated that the secondary nuclei are derived from extruded fragments which in the case of such forms as Myrme- cina or Apanteles are, I believe, to be regarded as of nucleolar origin. We find, therefore, that fragments of the nucleolus can form a true nucleus, with nuclear membrane, linin network, and nucleolus. Seiler (55 a) described in Lepidopterous eggs what he has called a chromatin diminution process; the polar body spindle at metaphase is found to carry three groups of granules, the two outer being the chromosomes which have divided and are becoming separated, the middle group of granules being apparently derived from the ends of the chromosomes by a diminution process, well known in the somatic mitoses of 26 J. BRONTE GATENBY the developing Miastor egg (82). Just before his untimely death Professor L. Doncaster was examining this problem, and sent me some of his slides for examination and sugges- tions; all that I could do was to recommend the use of stains such as Auerbach and Pappenheim, and methyl blue eosin. Digestion tests and such other microchemical tests are im- possible when one is working on the minute spindle in a very small ege. It certainly seemed to me that in the slides sent by Professor Doncaster the intermediate bodies were derived from the ends of the chromosomes as in Miastor.t Here again, however, we are faced with the same difficulty with regard to staining test, as I have pointed out with reference to the nucleolus: we are not justified in saying that a substance is chromatin simply because it selects methyl green from the Pappenheim or Auerbach stains; no one would care to say that the head of the spermatid was not chromatin, yet at certain periods it will select the red stain from the Pappenhemn or Auerbach fluid. To my mind it is useless to declare that the head of the sperm at such stages is not true chromatin, but has only changed its chemical nature ; the head of the sperm is derived from chromosomes before it reaches the egg and breaks up into chromosomes when it has penetrated into the ege. The spermatid nucleus takes the red stain from the Pappenheim or Auerbach fluid possibly because the arrange- ment of its surface or internal substance is more favourable to the molecules of the red stam, and unsuited for the absorp- tion of the green stain. The facts of the matter are that we know very little about the relationship between the nucleolus and the chromosomes, both during mitosis and during interkinesis; the same remark apples when we come to the subject of the microchemical nature of the nucleolus. I believe that a good step towards the elucidation of the first-mentioned problem has been taken by H. M. Carleton. This observer has shown that the nucleolus of certain vertebrates contains an argentophil core, or is related more 1 T have often wondered why this work of Prof. Doncaster was not edited and published. GAMETOGENESIS OF SACCOCIRRUS a7 or less closely to a body which under certain conditions becomes densely black in Cajal’s formalin silver-nitrate technique for the Golgi apparatus. During mitosis Carleton has shown that the argentophil core which he calls a nucleolinus (Haeckel) does not lose its individuality but divides, and may be found among the two chromosome groups of the telophase. I have been enabled to go through the preparations made by Carleton and can vouch for the correctness of his description; more- over, I possess preparations of the gut-cells of Saccocirrus, of the follicle-cells of Stenobothrus, and of many tissues in Rana, all-of which show a typical nucleolinus. What is very important is that Carleton has shown that the nucleolinus may be associated with either a ‘karyosome’ or a ‘plasmosome’ type of nucleolus. These remarks will serve to indicate the importance of work carried out on the nucleolinus, especially with Cajal’s formalin silver-nitrate method; Da Fano’s cobalt- nitrate method also serves to bring out the nucleolinus in some forms. Interpreting the work of Carleton on the nucleolinus, and also in the light of Cajal’s figures of various mammalian tissues and my own materials of invertebrates, I believe that the nucleolus, term used generally, might be morphologically independent of the chromosomes during the germ-cell cycle; the nucleolus during interkinesis might exist as a compound body consisting of a core which is argentophile and sometimes chromophile to other stains, and this core might act as the centre for the proliferation of a more extensive body which functions as the plasmosome or karyosome of the * resting’ nucleus; furthermore, during mitosis this outer region pro- hferated from the argentophil core possibly becomes lost, to be reformed in the next interkinesis. How far these suggestions will be found correct is impossible to say at present, but many of the facts we know now point in the direction I have indicated. Moreover, this view would coincide with the already-formed theories of the chromosome worker. The nature of the nucleolus is mainly proteid, maybe even in some cases nucleo-proteid, but its functions appear to be different from those of the chromosomes. ‘The nucleolus, like 28 J. BRONTE GATENBY the chromosome, Golgi element, and mitochondrium, is capable of growth and binary or multiple fission. Buchner, in his paper on the secondary nuclei of parasitic Hymenoptera, among the other conclusions, comes to the two following: accessory nuclei are to be traced back at the beginning, as naked chromatin (sic) granules lying in the cytoplasm. From these granules develop enchylema, nuclear membrane, and linin network, while the granule itself becomes the nucleolus of the accessory nucleus. Buchner has used safranin and light green and iron haematoxylin as stains; he labours under the delusion that what staims in a basic dye must necessarily be chromatin. He states that the chromatin granule which induces the formation of karyolymph, lmin network, and nuclear membrane, later becomes a ‘nucleolus’. Buchner figures the oocyte of Bombus and Myrmecina showing the nucleoli of the head nucleus as red granules (safranin) and a more or less faint chromatin (?) network green (‘licht- eriin’). The accessory nucleus also shows a red nucleolus and a green network. Buchner and others have concluded that the red-staining substance of the head nucleus, which becomes extruded through the nuclear membrane, is chromatin. As I have mentioned before I do not believe that one should lay too much weight on the staiming tests (and Buchner has not tried several of the stains I should like to have seen used), but the poimts which must be emphasized are, firstly, that it is proven that the nucleolus of many hymenopterous imsects does fragment and partly pass into the cytoplasm; and secondly, that these fragments do form secondary nuclei, exactly similar in certain species, to the head or principal nucleus. Call the red-staining body inside the head nucleus what one may, plastin or chromatin, plasmosome or karyosome, it is a fact that fragments of it can give rise to secondary nuclei. There is some temptation to use the facts which have recently been described in parasitic Hymenoptera, and in this paper, with reference to the behaviour of nucleoli, as support for a * binuclearity ’ hypothesis of some kind. In a recent paper GAMETOGENESIS OF SACCOCIRRUS 29 on the giant germ nurse-cells of Testacella (4) I ventured to interpret certain of my results in this manner, and it must be said that the case of the secondary nuclei is very suggestive. There are three possible modes of general interpretation— either the nucleolus represents a second chromatin of some kind, but separate from the chromosomes, or it derives its chromatin from the chromosomes, or there is some cell sub- stance other than chromatin which has the attribute of forming bodies similar to the ordinary nuclei, except for the presence in them of true chromatin. Whether the power of production of a nucleus-like body is to be looked upon as a proof of the chromatinic nature of a granule is unknown. (d) On the Special Part played by the Nucleus during Oogenesis. Recent studies on the cytoplasmic inclusions of the germ- cells have revealed the fact that all such units possess both Golgi elements and mitochondria, and that these two categories of formed elements take a prominent part in the upbuilding of the egg cytoplasm. No one has claimed a nuclear origin for the Golgi body, and in my work I have found a complete Golgi apparatus in the earliest germ-cells which have been studied—in molluses, insects, birds, amphibians, and mammals. The case of the mitochondria is different; several observers have claimed that they have found the mitochondria to originate from the nucleus during early stages of oogenesis or spermatogenesis. I had never seriously believed these accounts, and still doubt most of them; but in my own studies on the sponge Grantia I was led to identify the ‘chromidia’ of Jér- gensen as the representatives of the mitochondria; now Dendy firstly, and then I, have shown that the ‘ chromidia ’ of Jérgensen are nucleolar in origin. I still have some doubts as to whether true mitochondria do not exist in Grantia, but my efforts to demonstrate other granules which might be mitochondrial have so far not met with success; therefore I can but assume tentatively that in the case of Grantia the mitochondria are of nuclear origin, 30 J. BRONTE GATENBY It is important to notice that careful modern work on oogenesis confirms certain previous accounts of the extrusion of nucleolar material into the egg cytoplasm, and puts on a definite basis of truth the claim that the nucleus takes a part in the development of the cytoplasm. All such positive evidence which we possess in this direction applies to the behaviour of the nucleolus, and I do not believe that, we are able to point to any circumstances which would lead us to conclude that the chromosomes take a part, though I think that such is the case. Probably the only significant fact upon which we can fall back les in the formation of flocculent threads and reticula from the chromosomes after the prophases of the heterotypic division, and just before the real inception of the growth period of oogenesis. But this might just as well be interpreted as preparation by the chromo- somes for their own growth by means of substances absorbed from the egg cytoplasm. With regard to this difficult matter of the relationship between nucleus and cytoplasm during oogenesis, I believe that zoologists may be able to ascertain new facts if they develop and use more constantly the various silver-nitrate techniques, which give pictures unobtainable by other methods. (ec) Schaxel’s Chromatin Emission. From time to time in these papers I have referred to Schaxel’s work on chromatin emission in a number of imvertebrates which he has studied. Criticisms which have already been brought forward by me, in conjunction with Woodger, are that Schaxel has not worked at his material by proper methods, and he has not attacked the problem from the point of view of the cytoplasmic inclusions. Furthermore, he has not estab- lished that his granules are chromatin or that they are emitted through the nuclear membrane. With corrosive fixation, &c., and Ehrlich’s haematoxylin, the granules are found to be baso- phil, which probably proves nothing with regard to their microchemical nature. A new phase in the problem of Schaxel’s work was introduced by Miss van Herwerden, who, by treating GAMETOGENESIS OF SACCOCIRRUS 31 Strongyloentrotus eggs in a ‘ nuclease’ procured from spleen and pancreas, succeeded in dissolving away Schaxel’s granules, which did not appear when the eggs were subsequently treated by methods which fixed and stained the granules in eggs not treated by the enzyme solution. This work has been especially referred to by some recent writers, who consider that weight should be attached to Miss van Herwerden’s statements. With certam precautions, which were incomplete, she prepared a proteolytic enzyme from spleen, according to the directions of Sachs (52). Now I submit that her enzyme solution was probably a mixture of several enzymes, ‘ nuclease’ possibly, but also lpolytic enzyme as well. The fact that cell granules disappear under treatment by such a solution proves nothing with reference to their precise chemical nature. These granules were possibly mitochondria whose proteid basis was washed away by some protease, which would cause them to disappear as definite granules—or what is more likely, Miss van Herwerden’s ‘nuclease’ contained a lipoclastic body which swept away the linin content of the mitochondria. Until an expert on enzymes prepares solutions whose con- tents are known and whose reactions towards various organic materials are completely worked out in vitro, until the microchemistry and origin of bodies in question are better understood, then and then only should one place any weight on such work by enzyme action as that of Miss van Herwerden on Schaxel’s ‘chromatin’ granules. It should be noted carefully that Schaxel’s granules do not produce bodies resembling nuclei, as happens in Apanteles, &c.; one should not without good reasons call any haematinophilous body chromatin : even if his granules are extruded from the nucleus, they might just as well be nucleolar as chromatinic; and he might with advantage try other methods. Zoologists should note carefully that an espousal of Schaxel’s views seems to necessitate either the further adoption of a binuclearity hypothesis or the rejection of the chromo- some theory. 32 J. BRONTE GATENBY For if Schaxel’s granules are chromatin, using the word in the sense that they are made of the same sort of material as the chromosomes, either they must have originated from the latter—have been budded off from them—or there must be two kinds of chromatin in the egg nucleus. I cannot see how the adoption of the first alternative will allow one still to hold that the present-day chromosome theory is likely to be true; and the very behaviour of the nucleolus in Apanteles shows that there is a body other than the chromosomes which can produce a nucleus. By placing one’s belief in the second of the two alternatives— in some form of ‘ binuclearity hypothesis ’, one could also make many of Schaxel’s observations fit in with the more theoretical aspect of the question. While my mind is as open as it well could be in view of my own observations, I do not at present feel that Schaxel has attacked the problem in the best way, and I refrain from definitely accepting any of his views till some other observer carefully reinvestigates his claims and uses all the best and latest cytological techniques. Perhaps it should be mentioned that the above remarks do not commit me to the espousal of any ‘ binuclearity hypothesis ’, though I feel that there is some good evidence for such a postu- lation. (f) Centrifuge Experiments in Annelid Develop- ment and what they demonstrate. Tt has been shown in this paper that the major part of the granules of the egg of Saccocirrus is derived from nucleolar material extruded from the nucleus. If these nucleolar extru- sions represent Schaxel’s chromidia or the granules which form the secondary nuclei in parasitic Hymenoptera, and if they are of chromatinic nature, and not merely metaplasm or yolk, one might expect them to play some special part during embryonic development. They might even represent organ-forming materials. But apparently this is not the case: Lille (48) has given GAMETOGENESIS OF SACCOCIRRUS 33 some figures of centrifuged eggs of Chaetopterus and Nereis which lead me to believe that in these animals the egg contains fatty yolk (or oil) and nucleolar deutoplasm as in Saccocirrus. In Chaetopterus he finds the layers in the centrifuged egg to be a grey cap, upper (the ‘ fatty yolk’ of this paper), a clear area in the middle, and a lower layer of ‘ yolk ’ (my ‘ nucleolar deutoplasm*’ and mitochondria); these areas correspond with the layers in the centrifuged Saccocirrus egg (p. 18). Now, speaking of these layers in developing embryos and of formative stuffs in general, Lillie remarks: ‘So far as they (formative stuffs) are to be identified with the visible sub- stances segregated by the centrifuge, it would appear to be indicated by experiments that they can play no specific réle in differentiation, because in centrifuged eggs they may occupy variable positions in the embryo.’ This view coincides with that of Morgan (quoted in my previous paper (17)) and with Miss Beckwith’s study on Hydractinia. Any physiological derangement during the development of centrifuged eggs seems to be due either to mechanical difficulties of massed yolk or to absence of nutriment. It is interesting to note, too, that Morgan came to his conclusion partly as a result of work on Echinoderm eggs, where Schaxel finds an emission of ‘ chromatin ’ granules. (9) The Probable Part played by Mitochondria and Golgi Apparatus in Heredity. Modern cytologists tend to become divided into two groups— those working on the nucleus and those working on the cyto- plasm. Nearly all modern text-books dealing with Heredity and Sex treat exclusively of the part played by the chromo- somes in the mechanism of Heredity, and most observers are satisfied to accept the view that ultimately the nucleus is the seat of the substances which contribute to bring about the phenomena of Heredity. ‘Die Mitochondrien sind die proto- plasmische Vererbungssubstanz ’ is a statement which serves to show us that the chromosome theorist is not alone in this field. In the germ-cell cycle the chromosomes have been No. 261 D 84 J. BRONTE GATENBY shown by Van Beneden, Boveri, Wilson, Morgan, Montgomery, McClung, Doncaster, and many others, to go through certain definite changes, which have been found to correspond with many of the peculiar phenomena of sex and heredity in breeding experiments. The main facts ascertained with regard to the chromosomes are briefly as follows: 1. They are constant in number in any one species. 2. In ordinary cell-division each chromosome is halved so that each moiety is a complete replica of its fellow. 3. In the formation of the germ-cells there is a process whereby the ripe gamete comes to have the halved or haploid number of the chromosomes. 4. The male and female pronuclei in fertilization are practically equivalent, and possess the same number of chromosomes (overlooking the x and y chromosomes). 5. In the formation of the ripe spermatozoon no visible part of the chromatinic substance is rejected. In the cytoplasm of the animal cell it has been shown that two important categories of formed protoplasmic elements exist: namely, mitochondria and Golgi elements. The purpose of this section is to compare and contrast the behaviour of these protoplasmic bodies with the chromosomes of the nucleus. Under the first heading— That the chromosomes are constant in number in any one species ’—we may compare and contrast the Golgi body and mitochondria. While it is not generally possible to gain absolutely explicit evidence by examining the mitochondria in most animals, it is nevertheless true that in some forms the mitochondria are so few and so large that definite counts may be made. As examples I give the follow- ing: (a) In Paludina the typic spermatid may contain from four to seven spheres. Four is the commonest number. These spheres are subequal in size in those spermatids which contain four spheres and in those which contain seven. (b) Wilson (80) has shown the same variation to apply in Centrurus, and Retzius (25) also in a variety of Molluses. (c) It was shown (9) that in Helix aspersa the mitochondria in one spermatocyte GAMETOGENESIS OF SACCOCIRRUS 35 or spermatid were often remarkably different in size and number from those in another example. It is thus clear that the mitochondria are not usually of markedly definite number or size in the germ-cells or somatic cells of any given species. With reference to the Golgi apparatus the same applies. In Helix aspersa (9) and in other Molluses it was shown that the dictyosomes or Golgi batonnets could vary in number considerably. Moreover, examination of preparations of this apparatus in any somatic cells, as well as germ-cells, gives the impression that the Golgi elements are variable to an extreme. The statement—‘ In ordinary cell-division each chromosome is halved ’—may now be used as a basis for comparison and contrast with what occurs in the mitochondria and Golgi apparatus. In many cases it is difficult to get quite complete evidence as to whether a mitochondrium does divide during cell-division, but the general impression one gathers after examining cells in division is that the mitochondria are sorted out whole and haphazardly. In special cases, e.g. Centrurus (80) and Paludina, it is possible that the elongate mitochondria are halved transversely but not longitudinally. In by far the majority of animals it seems tolerably clear that the process of chondriokinesis or distribution of the mitochondria (or chondriosomes) between the daughter-cells is haphazard, and not in any way comparable to the process of karyokinesis. This result has been arrived at by a number of independent workers, and may be taken as established. The Golgi body in the dividing cell consists of rods or granules (dictyosomes) ; in most cases these dictyosomes keep around the zone of the amphiaster, often stuck on the asters them- selves, and, as with the mitochondria, the observer is impressed with the fact that the whole train of events in dictyokinesis, or the distribution of the dictyosomes between the daughter- cells, is extremely haphazard and much less precise than with the process of karyokinesis. That this is so can easily be shown to be the case in the mollusean germ-cell; in the sper- matid of Limax maximus the Golgi apparatus generally D2 36 J. BRONTE GATENBY consists of two dictyosomes (‘ Nebenkern’ batonnets); but in other cases there may be three, and one never finds a spermatid with a single batonnet. It is therefore certain that during dictyokinesis the Golgi elements are not always sorted out equally. That the Golgi rod is divided or halved like the chromosome is unlikely from this evidence, described in detail elsewhere (9a): in Limax agrestis the spermato- cyte has a Golgi apparatus formed of some eight dictyosomes or batonnets. This cell divides twice to give rise to four spermatids, but each of the latter only contains two of the Golgi batonnets ; this shows that in dictyokinesis the batonnet is not divided like the chromosome. With regard to the fact of the maturation of the germ-cells and the reduction of the chromosome number, nothing com- parable can be found in either mitochondria or Golgi elements of germ-cells. In the egg the polar bodies rarely contain mito- chondrial granules or Golgi elements, and never in such quantity as to suggest a special reduction in number. In the case of the male germ-cells the same applies: the first and second maturation divisions (chondriokinesis) in the male are of the same type, and while they bring about a halving, and then a rough quartering of the original number of mito- chondria in the spermatocyte, this process is not of the same nature as the reduction of the chromosomes. The same remark applies to the Golgi elements in dictyokinesis of the male germ-cells during maturation.? In the last stages of gametogenesis in the male no chromosomes are lost: the case of the mitochondria and Golgi apparatus is instructive, for in many Mollusca it has been shown that possibly all the Golgi elements, and much of the mitochondrial matter, are lost during spermateleosis, being sloughed off the tail of the sperm (9, 9a). Such seems to occur with the mito- chondria in Mammalia; Regaud shows that the bead of sloughed off protoplasm of the sperm of rats may contain mitochondria (24), though the main bulk of the granules forms part of the sperm. In other words, the chromatin of the 1 See also Ludford and Gatenby, “ Proc. Roy. Soc.’, vol. 92, 1921. GAMETOGENESIS OF SACCOCIRRUS 37 sperm is the only part which is meticulously guarded during spermatogenesis of all animals. That the male and female pronuclei contain the same number of chromosomes (leaving out the special x or y chromosomes) is a notorious fact. The sperm never contains as many mito- chondrial granules as the egg, and in only one case (Ascaris megalocephala (82)) has it been shown that at the time of fusion of the o7 and ¢ pronuclei, the number of mito- chondria of the o7 gamete are about the same as those of the @. The above comparisons show conclusively that of all the cell elements the chromosome is the only one whose behaviour is precise and coincident with the expected conduct of bodies directly engaged in the processes of heredity, the results of which, as breeding experiments show, are often of previously calculable exactitude. As direct bearers of any important or precise factors of heredity, the Golgi body and mitochondria appear to be ruled out by their inexact and variable behaviour in the germ-cell cycle. The chromosomes, and the chromosomes alone, fulfil the necessary conditions. 11. Summary. Spermatogenesis. 1. The spermatogonium is of the usual type, containing both mitochondria and Golgi apparatus (PI. 1, fig. 7). 2. The spermatocyte contains the same inclusions as the spermatogonium, but in addition there is to be found, in a large number of cases, a group of granules generally lying near the Golgi elements and giving the microchemical reactions of true yolk, i.e. turning greenish yellow in chrome-osmium fixatives, not staining in haematoxylin or acid fuchsin, and generally dissolved out by strong lipoid solvents (PI. 1, fig. 1, y). 3. Nurse-cells often accompany groups of spermatogonia. The nurse-cells contain large quantities of yellowish yolk, as well as fuchsinophil bodies, possibly mitochondrial in nature (Pl. 1, fig. 1, yc). 38 J. BRONTE GATENBY 4. In the spermatoeyte there is another group of granules to be found, especially in Kopsch (OsO,) preparations. These are individually much larger than the four members of the group of yolk-granules, and are about ten to sixteen in number ; they go brownish in OsQ,. ‘These larger granules have been traced into the spermatid, where about three or four are present, and appear to form the sides of the mitochondrial part of the sperm-tail (Pl. 1, fig. 2, x, and Pl. 2, figs. 11 and 12, x). 5. During the spermatocyte division prophases, the group of yolk-granules is found to take up a position near the equator of the spindle (Pl. 1, fig. 4, y), and subsequently becomes divided into two smaller groups in later stages of the division (Pl.1, fig. 5). This process occurs in both maturation divisions, so that each spermatid contains about one-quarter of the yolk-granules of the spermatid. The mitochondria, .as is generally the case during cell-division, become altered in such a way that they form a tangled mass of thread-like bodies, which are subequally sorted out into two portions, one in each daughter-cell (Pl. 1, figs. 4 and 5). The larger group of granules which were thought to form part of the tail-sheath were not found during mitosis. 6. The newly-formed spermatid contains the usual inclusions plus the group of yolk-granules (y in PI. 2, fig. 8). At this stage the sperm-sheath granules are occasionally found, and occur much more often in later stages (Pl. 2, fig. 9, x). 7. The spermateleosis stages, or metamorphosis of spermatid into spermatozoon, are remarkable for the manner of formation of the tail-sheath. The mitochondria become grouped behind the nucleus and around the outgrowing axial filament, while the Golgi elements and yolk-granules take up a position behind the mitochondria (Pl. 2, fig. 9). Tail-sheath granules are usually found in the vicinity (x im Pl. 2, fig. 9). 8. The mitochondria, hitherto single and all approximately equal in size, now begin to run together, like rain-drops, forming groups of larger and smaller granules (PI. 2, fig. 10, mm and m). This process goes on till only three large subequal spheres are left (figs. 11, 12, and 14), and then these spheres begin to elongate GAMETOGENESIS OF SACCOCIRRUS 39 to form the mitochondrial tail-sheath (Pl. 2, fig. 18 and 16). The tail-sheath granules are seen at x in Pl. 2, figs. 11, 12, and 14. 9. During these stages the Golgi elements tend to become thrown downwards along the length of the sperm (PI. 2, fig. 16), and this also occurs with the yolk-granules. In a bunch of ripe sperms within the body of the male Saccocirrus small eranules are always found on the lower region of the sperm- tails, and there seems to be good evidence that such elements are derived from the Golgi apparatus (PI. 2, figs. 17 and 18). If the receptaculum seminis of the female is examined, such sranules are also found on the tails of the sperms (PI. 2, fig. 18, GAX). Fertilization. 10. Saccocirrus is an example of precocious entry of the spermatozoon into the unripe oocyte (PI. 3, fig. 22). The nuclear head of the sperm alone enters the egg completely at first, while the tail remains plastered on the surface of the young oocyte (Pl. 3, fig. 22, head at sp, fragments of tail at spt). It is very difficult therefore to say whether these sperm-tail fragments are or are not inside the oocyte cytoplasm at this period. Later on, however, it is quite easily observed that the elements of the sperm-tail do enter the egg, break up further, and form large numbers of spherical granules (con- secutive stages given in Pl. 3, figs. 21, spr, 22, spr, 24, and 25). 11. In many cases one cannot help believing that these beads, derived from the remains of the sperm-tail, grow in number and in size (ef. Pl. 3, figs. 23 and 25). 12. These beads always remain in the periphery of the egg, but do not seem to take any noticeable part either in the forma- tion of yolk or in any process of fertilization. Careful examina- tion of the periphery of many oocytes reveals the fact that the granules are of two types, one going yellowish in OsO,, the other going black, as shown in Pl. 3, fig. 21. It was thought that the black granules might have something to do with the black granules noted on Pl. 2, fig. 18, Gax, which were con- sidered to be derived from the Golgi apparatus. 40 J. BRONTE GATENBY 13. The peripheral granules of both types later disappear ‘or become hidden by the formation of clouds of nucleolar deutoplasm. Oogenesis. 14. By staining, fixing, and by centrifuge experiments if can be shown that the full-grown oocyte of Saccocirrus con- tains four distinct kinds of formed ‘ yolk ’-granules, i.e. Golgi elements, mitochondria, true yolk, nucleolar extrusions or nucleolar deutoplasm. (See their fixing and staining reactions in a table on p. 18.) 15. Themost numerous and chemically most resistant granules are neither mitochondria nor Golgi elements, but are derived from nucleolar extrusions. At a very early stage the oocyte nucleolus buds off pieces which pass through the nuclear membrane, but at first remain stuck on its outer surface (stages in Pl. 3, figs. 28, 24, and 25, nu). At one stage these nucleolar deriva- tives form an extraordinary picture, being stuck all over the nucleus in the form of pyramidal bodies, whose base adheres to the nuclear membrane (PI. 3, fig. 23). The nucleolar deriva- tives stain intensely in haematoxylin or fuchsin (fig. 23), but only go yellowish in osmic acid (fig. 21). 16. Later on these pyramidal granules lose their connexions with the nuclear membrane, but, remaining quite near, become the centres of numbers of large vacuoles which appear (Pl. 4, fig. 29). Inside these vacuoles the nucleolar granules partially break up, and subsequently, after the absorption of the vacuoles, the granules move out further from the nuclear membrane and form clouds of granules in the egg cytoplasm (Pl. 4, figs. 80 and 84). The marked vacuolar stage in the history of the egg seems to occur with suddenness, and is not discoverable at this period in all eggs of this size (PI. 4, fig. 29). It is just about this stage that great activity is noticeable around the periphery of the nucleus, as shown in Pl. 4, fig. 35, by formol-silver nitrate technique. The nucleolar deutoplasm forms dense clouds of heavy granules throughout the entire egg cytoplasm. 17. If the ovary of Saccocirrus be prepared by a silver GAMETOGENESIS OF SACCOCIRRUS 41 nitrate or osmic acid Golgi-body method, an appearance such ° as shown in PI. 3, fig. 19, is seen. Large numbers of crescent- shaped bodies, such as were noted in the spermatocyte (Pl. 1, figs. 6 and 7), occur throughout the cytoplasm. In younger oocytes the bodies of the Golgi apparatus are densely packed and placed to.one side of the cell (Pl. 3, fig. 20, ca). Such Golgi elements eventually divide rapidly, and spread out, as fine crescents or slightly elongated rods, through the cyto- plasm of the full-grown oocyte (PI. 4, fig. 32, Ga). 18. If the ovary be treated by a chrome-osmium method, and stained in iron alum haematoxylin or acid fuchsin, fine mitochondria become visible (Pl. 3, figs. 23, 24, 25, m). Such mitochondria are difficult and sometimes impossible to see in the youngest oocytes and the oogonia. 19. In chrome-osmium preparations there are also to be seen fine true yolk-spheres, characterized by the fact that they go yellow-green in the fixative and do not stain in haematoxylin or fuchsin. 20. By centrifuging the oocyte, three layers appear, viz. an upper layer formed of true yolk (greenish), a middle clear protoplasm layer, and a lower layer mainly formed of nucleolar deutoplasm, with a mixture of mitochondria. 21. In many oocytes an enigmatic body, much like a secondary nucleus, was noted (Pl. 8, fig. 28, xn, and Pl. 4, fig. 30, xn). 22. The oogonial cytoplasm is oxyphil, and during oogenesis becomes basophil, and then again oxyphil in the full-grown oocyte (p. 16). Intermediate Cells. 23. In Pl. 1, fig. 3, is a cell found in a male Saccocirrus, and it shows characters intermediate between an egg and a spermatocyte (p. 20). Discussion. 24. The above facts are discussed on p. 21, and also the probable part played by mitochondria and Golgi bodies in heredity (p. 33). 42, J. BRONTE GATENBY 12. BiBpLioGRAPHY. 1a. Bauer, E.— Einfiihrung in die experimentelle Vererbungslehre ’, 1914. 1b. Bayliss, W. M.—‘ The Principles of General Physiology.’ Longmans, 1918. 2. Brachet, A.—‘ L’Ciuf et les Facteurs de lOntogenése.’ Doin et Fils, Paris, 1917. 3. Buchner, P.—‘ Die akzessorischen Kerne des Hymenoptereneies ”’, ‘Arch. f. mikr. Anat.’, Bd. xci, 1918. 4, Carleton, H. M.—‘‘ Observations on an Intra-Nucleolar Body in Columnar Epithelium Cells of the Intestine’, “ Quart. Journ. Mier. Sci.’, vol. 64, 1920. 5. Castle, W. E.—‘ The Effect of Selection upon Mendelian Characters Manifested in One Sex only”, ‘ Journ. Exp. Zool.’, vol. 8, no. 2, 1910. 6. Chambers, R.—‘* Changes in Protoplasmic Consistency and their Relation to Cell Division ’’, ‘ Journ. General Phys.’, vol. 11, 1919. Conklin, E. G.—‘* The Oriestatian and Cell-Lineage of the Ascidian Egg”, ‘ Journ. Acad. Sci. Philad.’, vol. 13, 1905. 8. —— ‘The Share of Egg and Sperm in Heredity”, * Proceed. Nat. Acad. Sci. of U.S.A.’, vol. 3, no. 2, 1917. 9a. Correns, C.—* Zur Kenntniss der Rolle von Kern und Plasma bei der Vererbung”’, ‘ Zeit. Abst. Vererb.’, ii. 9b. Danchakoff, V.—* Development of Cell Organs during the first Cleavage of the Sea Urchin Egg ’”’, ‘ Journ. Morph.’, vol. 27. 10. Dendy, A.—** Gametogenesis of Grantia compressa”, ‘ Quart. Journ. Mier. Sci.’, vol. 60, 1915. 11. Dobell, C.—*‘ Chromidia and the Binuclearity Hy pothesis ” . ibid., vol. 53. 12. Dunn, L. C.—** Nucleus and Cytoplasm as Vehicles of Heredity ” ‘Amer. Nat.’, vol. li, 1917. 13. Gatenby, J. Bronté.—** The Modern Technique of Cytology ’’, ‘ Quart. Journ. Micr. Sci.’, vol. 64, 1920. 14. —— “The Cytoplasmic Inclusions of the Germ Cells ”, Part I, ibid., vol. 1. 7 115% Ditto, Part IV, ibid., vol. 63, 1919. 16.———— Ditto: Part VilL ibid: 17. —— Ditto, Part VII, ibid., vol. 64, 1920. 18. —— Ditto, Part VIII, ‘ Journ. Linnean Soc.’, 1920. 19. —— Ditto, Part IX, ‘Quart. Journ. Micr. Sci.’, vol. 65, 1921. 20. Gatenby and Woodger.—** Mitochondria and Golgi Apparatus and the Formation of Yolk ’’, ‘ Journ. Roy. Micr. Soc.’, 1920. 21. 22. 23 24 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36 37. 38. 39. GAMETOGENESIS OF SACCOCIRRUS 45 Gaver, van, et Stephan.—‘‘ Intervention des Spermatozoides dans Povogenése chez Saccocirrus papillocercus (Bobr.) ”’, ‘ Compt. Rend. Soe. Biol.’, vol. 61, Paris, 1906. ———— ‘Apropos de l’ovogenése de Saccocirrus papillocercus (Bobr.) ”’, ibid., vol. 62, 1907. Godlewski, E.—‘‘ Untersuchungen iiber die Bastardierung der Echinoiden- und Crinoidenfamilie”’, ‘ Arch. £, Ent.-Mech.’, 20, 1906. Goodrich, E. 8.—‘‘ Notes on the Nephridia of Dinophilus and of the Larvae of Polygordius, Echiurus, and Phoronis”’, * Quart. Journ. Micr. Sci.’, vol. 54, 1909. —— ‘‘On the Nephridia of the Polychaeta, Part I11”’, ibid., vol. 43, 1900. —— ‘On the Structure and Affinities of Saccocirrus ”’, ibid., vol. 44, 1901. Goérich, W.—‘“ Zur Kenntnis der Spermatogenese bei den Poriferen und Coelenteraten nebst Bemerkungen iiber die Ovogenese der ersteren ’’, ‘ Dissertation Marburg’, 1903. Haberlant, G.—‘ Uber die Beziehungen zwischen Function und Lage des Zellkerns in der Pflanze ’, Jena, 1887. Haeckel, E.—‘‘ Uber die sexuelle Fortpflanzung und das natiiliche System der Schwimme”’, ‘ Jenaische Zeit. f. Med. u. Naturwiss.’, Bd. vi. Hargitt, G.—‘‘ Germ-cells of Coelenterates, Parts I, II, U1, IV, V, and General Conclusions ”’, in the ‘ Journ. Morph.’, 24-33. Hegner, R.—‘‘ Genesis of Organization of the Insect Egg’, ‘ Amer. Nat.’, vol. li, no. 611, 1917. —— ‘The Germ-cell Cycle in Animals’, 1914. Held, Hans.—‘‘ Untersuchungen iiber den Vorgang der Befruchtung ”’, * Arch. f. mikr. Anat.’, Bd. 89, 1916. Hempelmann, F,.—‘*‘ Die Geschlechtsorgane und -zellen von Sacco- cirrus ’’, ‘ Zoologica ’, 67, 1912. ——“Kibildung, Eireifung und Befruchtung bei Saccocirrus”’, * Zool. Anz.’, Bd. 30, 1906. Hirschler, Jan.—‘‘ Uber Plasmastrukturen in den Tunicaten-, Spon- gien-, und Protozoenzellen ’’, ‘ Anat. Anz.’, vol. xlvii, pp. 14-15. —— “Uber die Plasmakomponenten der weiblichen Geschlechts- zellen ’’, ‘ Zeit. £. mikr. Anat.’, Bd. Ixxxix. Jenkinson, J. W.—‘‘ On the Relation between the Structure and the Development of the Centrifuged Egg’’, ‘ Quart. Journ. Micr. Sci.’, vol. 60, 1915. Jorgensen, Max.—* Beitrige zur Kenntnis der EHibildung, Reifung, Befruchtung, und Furchung bei Schwimmen (Syconen)”’, ‘ Arch. {. Zellforsch.’, Bd. iv, 1910. 44 J. BRONTE GATENBY 40. Just, E. E.—‘‘ Fertilization in Platynereis megalops”’, ‘Journ. Morph.’, vol. 26, 1915. 41. Lankester, E. Ray.—‘ A Treatise on Zoology. Part Il. The Porifera and Coelentera.’ Minchin. 42. Lillie, F. R.—‘‘ Studies in the Fertilization of Nereis, 1V”’, “‘ Journ. Exper. Zool.’, vol. xii. 43. —— ‘Polarity and Bilaterality of the Annelid Egg, &c.”, * Biol. Bull.’, 1908-9. 44. Loeb, Jacques.—* The Organism as a Whole.’ 45. Maas, O.—‘ Uber Reifung und Befruchtung bei Spongien ”, ‘ Anat. Anz.’, Bd. 16, 1900. 46. Macbride, E. W.—‘ Text-book of Embryology. Part I. Inverte- brata’. Macmillan, London, 1914. 47. ——‘ Presidential Address to British Association, Section D, 1914.’ 48. Malsen, H. v.—‘‘ Geschlechtsbestimmende Einfliisse und Eibildung von Dinophilus apatris’”’, ‘ Arch. mikr. Anat.’, Bd. 69, 1906. 49. Nussbaum-Hilarowitz, J.—‘‘ Uber das Verhalten des Chondrioms wihrend der Eibildung bei Dytiscus ”’, * Zeit. f. wiss. Zool.’, Bd. exvii. 50. Retzius, G.—‘ Biolog. Untersuch.”’, ‘ Neue Folge’, xii. 51. Rio Hortega, P. Del.—‘** Détails nouveaux sur la structure de lovaire ”’, ‘Trab. del Lab. Invest. Biol. Madrid’, t. xi, 1913. 52. Sachs.—‘‘ Uber die Nucleasewirkung’’, ‘ Zeit. f. physiol. Chemie’, Bd. xlvi, 1905. 58. Schaxel, J.—‘‘ Das Verhalten des Chromatins bei der Eibildung einiger Hydrozoen ”’, ‘ Zool. Jahrb., Abth. Anat.’, Bd. 31, 1911. 54. —— “ Die Geschlechtszellenbildung und normale Entwicklung von Aricia foetida’’, ‘ Zool. Jahrb.’, Bd. 34, 1912. 55a. Scott, 8S. G.—‘‘ On Successive Double Staining for Histological Purposes ’’, ‘ Journ. Path. and Bact.’, vol. xvi, 1912. 55 b. Seiler, J.—‘‘ Das Verhalten der Geschlechtschromosomen bei Lepidopteren ”, &c., * Arch. f. Zellf.’, Bd. 13, 1914. 56. Shull, A. F.—See reference no. 12. 57. Shearer, C.—** The Problem of Sex Determination in Dinophilus gyrociliatus. Part 1. The Sexual Cycle”, *‘ Quart. Journ. Micr. Sci.’, vol. 57, 1912. 58. Van Herwerden, M. A.—‘* Uber die Nucleasewirkung auf tierische Zellen ’’, ‘ Arch. f. Zellf.’, Bd. 10, 1913. 59. Vejdovsky, F., and Mrazek.—** Umbildung des Cytoplasma waihrend der Befruchtung und Zellteilung”’, “Arch. mikr. Anat.’, Bd. 62, 1903. 60. Von Baehr, W. B.—** Uber die Bildung der Sexualzellen bei Sacco- cirrus major”, * Zool, Anz.’, Bd. xliii, 1913-14. 61. Walker, C. E.—* Hereditary Characters and their Modes of Transmis- sion’, London, ed. Arnold, 1910. GAMETOGENESIS OF SACCOCIRRUS 45 62. Weigl, R.—“‘ Vergleichend-zytologische Untersuchungen iiber den Golgi-Kopschen Apparat”, ‘ Bull. ’ Acad. Scien. Cracovie’, 1912. 63. Wilson, E. B.—‘* On Cleavage and Mosaic Work’”’, ‘ Arch. f. Ent.- Mech.’, vol. 3, 1896. 64. —— ‘Experimental Studies in Germinal Localization”, ‘ Journ. Exp. Zool.’, vol. 1, 1904. 65. —— ‘The Cell’, University Biological Series, 1919. 13. EXPLANATION OF PLATES 1-4. IntustrRatTInG Prorrssor J. Bronte GATENBY’s PAPER ON THE GAMETOGENESIS OF SACCOCIRRUS. EXPLANATION OF LETTERING. AS, acrosome. AR, archoplasm or centrosphere. CH, chromosomes. CHO, vitelline membrane. CL, cortical lamellae of egg. Ga, Golgi apparatus (‘ Nebenkern ’), Golgi body or element. Gax, body supposed to be a part of the Golgi apparatus. Gx, body believed to be forming the acrosome. M, mitochondria. MM, macromitosome or forming middle-piece (mito- chondria) of sperm-tail. N, nucleus. NL, nucleolus, or fragments of latter forming nucleolar deutoplasm. NLXx, nucleolar bodies homologous with the true nucleolar deutoplasm of egg. NLV, vacuoles around the nucleolar extrusions. sB, spindle-bridge. spr, sperm-tail. spz, sperma- tozoa. SP, sperm inside young oocyte. v, vacuoles in ground protoplasm of oocyte. x, bodies believed to form a part of the skeleton of the sperm- tail, on each side of the macromitosomal (mitochondrial) spheres. xv, nuclear-like bodies sometimes found in oocytes, possibly secondary nuclei. xy, bodies near asters, possibly part of the Golgi apparatus. Y, yolk-spheres. yc, yolk-cell or nurse-cell. Scale of Figures.—On PI. 1, all figures, except number 6, are drawn to the scale indicated in the middle of the plate. The scale for fig. 6 is near the drawing. On PI. 2, all figures, excepting 17 and 18, are drawn to the scale in the middle of the plate, the scale for figs. 17 and 18 being near by. All figures on PI. 3 are drawn to the scale on the right-hand side. On Pl. 4, all figures, excepting number 34, are drawn to the scale given below fig. 29. Techniques Used.—m.x., Mann-Kopsch osmium tetroxide method. CH.K., Champy-Kull chrome-osmium acid fuchsin toluidin blue and aurantia method. v.F., Da Fano’s cobalt nitrate formol-silver nitrate method. PLateE 1. Fig. 1—Full-grown spermatocyte and part of nurse- or yolk-cell on left. In the spermatocyte the Golgi apparatus (Ga), the mitochondria (m), and the group of yolk-spheres (y) are to be seen, The nurse-cell contains 46 J. BRONTE GATENBY large yolk-spheres (vy) and smaller fuchsinophile bodies, possibly mito- chondrial in nature (M), CH.K. Fig. 2.—Younger spermatocyte from a Mann-Kopsch preparation, show- ing at x a number (about two) of largish spheres, believed to be identical with the same bodies marked x in figs. 9, 10, and 11, and which seem to take some part in the formation of the tail skeleton of the sperm, At y are the yolk-granules, and at Ga the Golgi apparatus ; compare this with fig. 1, in which the apparatus is formed of delicate slightly-curved rods. In fig. 2 the rods are heavily impregnated with OsO,, and possibly owing to a shrinkage of the centrosphere, they have become much more curved. The mitochondria do not show. M.K. Fig. 3.—Cell of the spermatocyte series but showing a modification ; the mitochondria are small, like those of the egg (fig. 24), and peri-nuclear bodies are present at NLX (compare with egg in fig. 23). At y is a group of yolk-spheres. CH.K. Fig. 4.—Metaphase of second spermatocyte division. Note alteration in shape of mitochondria (Mm), which from their previous granular structure (fig. 1) have become filiform, At ¥ the group of yolk-spheres has become grouped near the spindle preparatory to being sorted out into two groups as in the next figure. At xy are bodies supposed to be Golgi elements stuck on the poles of the asters. Fig. 5.—Telophase of second spermatocyte division, the equatorial plate, is forming, and the mitochondria, still filamentous, are grouped near in a special manner, being most numerous near the forming cell-wall. At y are the yolk-spheres, now sorted out into two groups. CH.K. Fig. 6.—Spermatocyte, for comparison with the oocytes in fig. 20. MLK, Seale above, Fig. 7.—Spermatogonium drawn to same scale as spermatocyte in fig. 1. Shows Golgi apparatus consisting of from eight to ten dictyosomes or rods, a spindle-bridge at sB, and the mitochondria at mM, grouped near the centrosphere. M.K., counter-stained in Altmann, PLATE 2. Fig. 8.—Newly-formed spermatid, showing the Golgi apparatus some- what scattered on the right and the mitochondria surrounding the nucleus. The yolk-granules form a compact group at y. The centrosome would be on the right side of the nucleus. CH.K. Fig. 9.—Later spermatid after the outgrowth of the axial filament. The spheres at x are probably of the same nature as those drawn in PI. 1, fig. 2. The mitochondria have now become grouped behind the nucleus at M, while the Golgi elements (ca) and yolk-spheres (y) have drifted to the bottom of the elongating cell. The cytoplasm, as in many Kopsch (OsO,) preparations, is coarsely fibrillar. At Gx is a body believed to be forming the acrosome. M.K., counter-stained in Altmann. GAMETOGENESIS OF SACCOCIRRUS 47 Fig. 10,—Later stage. The mitochondria have begun to run together to form a number of larger spheres (MM), At xX is one of the large granules seen in fig. 9 and in PI. 1, fig. 2, while at Gx is the same body mentioned in the description of fig. 9. The yolk-granules (y) form a fine group to one side of the cell. The nucleus in this cell is still spherical, in this being less advanced than that of fig. 9, which is depressed. CH.K. Fig. 11.—Later spermatid, nucleus now depressed on one side, or cap- shaped. The macromitosomal spheres (MM), ‘ Nebenkern’ of some authors, are larger, not all the same size, and there is still a collection of unused mitochondria at M. The acrosome is seen as a thickened edge of the nucleus at As. Other parts as before, except that notice should be taken of the fact that the Golgi elements (GA) have become drawn up below the macromitosomal spheres. CH.K. Fig. 12.—Later stage, all mitochondria have run into the macro- mitosomal spheres (MM), only two of which are shown. Golgi apparatus still drawn up below the mitochondrial spheres. The cytoplasm is stringy as is often found in osmic-acid preparations. Nucleus further depressed and shrinking in size. M.K., counter-stained in Altmann. Fig. 13.—Forming spermatozoon, showing elongated macromitosome (mm) and other cell inclusions. CH.x. Fig. 14.—Macromitosome or mitochondrial spheres, at the stage of fig. 12, but viewed from below. Note skeletal granules at x, and unequal size of spheres. CH.K. Fig. 15.—Part of tails of fully-formed sperms, at a higher magnification than in fig. 17, to show the bodies marked Gax, which are thought to be Golgi elements. M.K. Fig. 16.—Forming spermatozoon, at a stage later than that drawn in fig. 13. Fig. 17.—Bundle of ripe sperms from coelom of male; refer to fig. 15, M.K. Fig. 18.—Receptaculum seminis of female, to show the presence of the special granules (GAX) on the tails of the spermatozoa, SPZ. M.K. PLATE 3. Oogenesis of Saccocirrus. Figs. 19 and 20.—Four oocytes prepared by the Mann-Kopsch- Altmann method to show Golgi apparatus. The peculiar peripheral granules derived from the sperm-tail (spr) are shown well. At sp is the head of the spermatozoon, and at NL a peri-nuclear thickening marking partly the nucleolar extrusion, and also as well the peri-nuclear activity, which seems to be something apart from nucleolar extrusion (note also Pl. 4, fig. 35). Fig. 21.—A later stage showing Golgi apparatus and advanced nucleolar extrusion, NL. The peripheral granules in such Kopsch preparations appear to be of two sorts—those staining quite black, and those yellowish. 7 2. , 7 he ~ 7 7 7 vil, it Led hy kitty mare = , » r, =, | af , a . 4 = e ‘ =~ q > ~~ ‘6 ‘ t Vol. 66, N.S. Pl. 2. Quart. Journ. Mier. Sct. i aN as BE 2 J.B.G. del. Quart. Journ. Micr. Sci. Vol. 66, N.S. Pl. 38. J.B. G. del. Quart. Jowrn. Micr: Sci. Vol.66,N.S., Pl. F J.B.Gatenby. del. Te RL hal 4 ey Li. 5 7 4 yo : ‘ : ' i + ‘ ty Re bea, +? i i i j . ; 4 e a _ iy ee se o eS aa = a eS ‘Ca = On the Development of the mw he ‘Enteronephric’ type of Nephridial system found in Indian Earthworms of the genus Pheretima.’ By Karm Narayan Bahl, D.Se., D.Phil. (Merton College, Oxford), Reader in Zoology, Lucknow University, India. With Plates 5—7 and 8 Text-figures. CONTENTS. . INTRODUCTORY . Historical THE Cocoon . GENERAL OUTLINE OF THE DaveLonnne OF THE Nieaateay, System IN PHERETIMA . DEVELOPMENT OF THE PRIMARY Heenaiete N EPHRIDIA . . DEVELOPMENT OF THE PRIMARY SEPTAL NEPHRIDIA . DEVELOPMENT OF THE SECONDARY NEPHRIDIA, SEPTAL AND INTEGUMENTARY . DEVELOPMENT OF THE Puenewanse Nee AND THEIR Ducts . COMPARISON WITH THE Deion OF heideesaaiG? AND THE SO-CALLED ‘ PLECTONEPHRIC ’ TYPES oF NEPHRIDIAL SYSTEM . . PHYLOGENY OF THE Groen Anite re Se arene . MATERIAL AND TECHNIQUE . SUMMARY . REFERENCES TO Lie . EXPLANATION OF PLATES 1. InTRODUCTORY. PAGE 100 101 In a previous paper in this journal (1) I described a new type of nephridial system, which is found in Indian earthworms of the genus Pheretima and which I called ‘ enteronephric’. 50 KARM NARAYAN BAHL The essential feature of this system is that the numerous septal and pharyngeal nephridia (all micronephridia) are connected with an elaborate system of ducts, which open, not on the surface of the skin but into the lumen of the intestine and other regions of the gut (buccal cavity and pharynx). These nephridia of the ‘ enteronephric’’ type co-exist im Pheretima with the integumentary nephridia, which are exceedingly numerous on the inside of the body-wall, and open on the surface of the skin through separate nephridiopores, like ordinary Oligochaete nephridia. Although in my paper I referred very briefly to the possible physiological significance of the discharge of excretory fluid into the gut of this worm, I did not enter, for want of embryological data, upon any discussion concerning the morphological significance of the discovery of the ‘ entero- nephric’ type of nephridial system, in relation to the com- monly accepted view, due mainly to Goodrich (9 and 10), that all Oligochaete nephridia ‘ develop centripetally as it were, and quite independently of the coelom and are probably derived from the epiblast ’. While little doubt could be entertaimed, from a study of the disposition of the nephridial system of the adult worm, with regard to the ectodermal origin of the integumentary nephridia, it was difficult to believe that the septal and pharyn- geal nephridia also had a similar origin for two reasons. In the first place, these nephridia have not only no connexion with the body-wall but are connected instead with the inter- segmental septa, which are mesodermal structures ; and they open, through an elaborate system of ducts, presumably mesodermal, into the lumen of the gut, the wall of which is partly mesodermal and partly endodermal. In the second place, the septal nephridia differ from the integumentary ones in that the former possess open ‘ funnels’, which are absent in the latter. Although no solenocytes or ‘ flame-cells ’ have been found on the integumentary nephridia, the presence of a coelomic funnel in one case and its absence in the other might lead one to ascribe a different origin to the two sets of structures. In fact the connexions of the septal nephridia and their ducts DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 51 in the adult worm seemed to negative the ectodermal theory of the Oligochaete nephridium, and to point to a mesodermal origin of these nephridia of the new type. It thus became evident that interesting results would be obtained from a study of the course of development of the nephridial system of this worm, and accordingly I undertook to investigate the problem and the following pages embody the results obtained by me. The work was carried out in the Department of Comparative Anatomy at Oxford, under the general supervision of Professor HK. $. Goodrich, to whom I am very much indebted for the keen interest he has all along taken in my work and for his valuable help and advice. 9. HisToRICAL. The question of the origin of nephridia in Oligochaetes has engaged the attention of many distinguished observers. The early investigators, like Kowalewski (11), regarded the nephri- dium as a tube connecting the coelom with the exterior, and believed that a nephridium arose by a growth of the septal wall of the coelom, that it gave rise to a chain of cells projecting backwards, which eventually fused with the ectoderm and then became hollowed out, so that the whole nephridium is to be looked upon as a ‘ tail’ of the coelom. Moreover, since the first trace of a cavity appears in the region of the funnel and is a prolongation of the body-cavity, the cavity of the nephri- dium might be said to be part of the coelom. Bergh (6) derives the whole nephridium, including the funnel, of Criodrilus and Lumbricus from a single large cell, the ‘ funnel-cell ’, lying close to the epiblast, between each successive pair of solid mesoblastic somites. The origin of this ‘funnel-cell’, from which the whole nephridium develops, has been a matter of considerable dispute. In a later paper on the subject (7) Bergh denies the origin of the ‘ funnel-cell’ from the nephric row in Criodrilus and Lumbricus, and asserts that the funnel and the body of the nephridium have a separate and different origin in Rhynchelmis, the upper lip of the funnel 52 KARM NARAYAN BAHL arising not from the ‘ funnel-cell’ (‘ Trichterzelle ’) but from a peritoneal cell. This view of the mesodermal or the so-called * intraperitoneal ’ origin of nephridia is in strong contrast with that held by Hatschek, Wilson, Meyer, and Vejdovsky (in his later work), which ascribes an ectodermal or a ‘ retroperitoneal’ origin to the main body of the nephridium and traces the ‘ funnel-cell ’ to the primary nephric row. According to this view the * funnel- cell’ arises from the primitive cell-row, or nephric cord, formed by the repeated division of one of the teloblasts on either side. In the earlier stages, this teloblast and the nephric cord to which it gives rise lie on the surface of the embryo; thus the ‘funnel-cells ’ are epiblastic in origin. From the nephric row one cell enlarges and enters into connexion with each successive segment ; these large cells, arranged metamerically outside and between each pair of somites, are the so-called ‘ funnel- cells’. In some worms, like Dendrobaena and Lum- bricus, the ‘ funnel-cells ’ give off the chain of posterior cells whilst separating from the nephric row, thus remaining for some time in connexion with it. In other cases, such as Criodrilus, the ‘ funnel-cells ’ appear to separate first (9). This view of the superficial origin of nephridia was strongly supported by Goodrich’s work (10), in which he showed that in certain Polychaetes (e.g. Nephthys) the nephridia do not open into the coelom at all, but terminate internally in a bunch of solenocytes which project into the coelom. He regarded the nephridium as essentially an ectodermic structure com- parable with the excretory tube of a Nemertine or of a Platy- helminth. According to him the excretory organs of Oligo- chaeta are ‘ true’ nephridia, i.e. tubes originally blind which have acquired secondary communications with the coelom, as distinguished from the ‘ coelomoducts ’, the term he uses for purely mesodermal structures. He points to the co-existence of the genital duct (which is a wide short coelomoduct) and the nephridium in the same somite, in Lumbricus, as evidence that the two structures cannot be homologous with one another (12). DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 53 The question as to which category (epiblastic or mesoblastic) the Oligochaete nephridia belong has recently been attacked by Staff (15), by renewed researches into the mode of their development in Criodrilus. Staff found ‘that in Criodrilus lacuum the mother- cells of the nephridia appear in the ectoderm at the hinder region of the embryo, and here act as teloblasts, giving rise to strings of cells by continuous budding off of smaller cells in front of them, like the mesodermie teloblasts situated internally to them. There are on each side four rows of such ectodermal teloblasts, and the rows of cells to which they give rise become wedged in between the ectoderm and the coelomic mesoderm. ‘The strings of cells destined to give rise to the nephridia are broken into groups, and one group is pushed into each septum which divides one coelomic sac from another. Here each group grows and gives rise to a chain of cells, and this cell-chain becomes hollowed out and forms a tube. Its most internal cell projects into the coelomic cavity between the coelomic cells forming one side of the septum, and forms the greater part of the coelomic funnel of the nephridium. The lower lip of the funnel is constituted by one huge cell belonging to the coelomic wall’ (12). According to Staff, therefore, the nephridia develop. from the ‘ retroperitoneal ’ cell-row, lying lateralwards to ‘ primitive muscle-fibres ’ in the manner that this breaks up in segmentally- arranged cell-groups, which project into the body-cavity and are covered over with the peritoneum. The whole nephridium is really ectodermal. The result of Staff's imvestigation, therefore, is to uphold Goodrich’s view. The earthworm Pheretima (Perichaeta), the develop- ment of which I have studied for the purpose of this paper, has all along been held to possess a branched ‘ plectonephrice ’ nephridial system, a term which has become inapplicable to the system in Pheretima on our further knowledge of it gained recently (1). The development of this latter type of nephridia has been investigated by Beddard (8) in Octo- chaetus multiporus, by Vejdovsky (16) n Mega- 54 KARM NARAYAN BAHL scolides australis, and by Bourne (8) in Mahbenus imperatrix and Perichaeta pellucida. The earthworm Octochaetus (Acanthodrilus) pos- sesses, in the adult condition in the interior of its body, eight tufts of nephridia in each segment, but a much larger number of external orifices for these nephridia. The funnels are present on these nephridia in the hinder region only and not in the anterior region (4). During development, according to Beddard (3), the embryo possesses a paired series of organs in each segment, which, as Vejdovsky thinks, are probably the equiva- lents of the pronephridia of Lumbricus. These paired nephridia of the embryo are, however, provided with well- developed ciliated and functional nephrostomes. Beddard was not able to follow these paired nephridia to the condition obtaining in the adult, and his work is very incom- plete; but he thinks that the nephridia of the embryo are converted into those of the adult, firstly by a temporary cessation of function (?) in a part of the nephridium—the por- tion nearest the funnel—which is produced by the disappear- ance of the lumen, and secondly by the active growth of this part of the nephridium, as well as other parts, and by the formation of a fresh series of apertures to the exterior. Our knowledge of the development of nephridia in the Australian earthworm Megascolides is fairly complete. In the adult condition of this worm the diffuse network of minute excretory tubules is reinforced by the existence of larger paired tubes, one pair to each segment ; and these large paired nephridia appear to be in connexion with the smaller tubes. We have, therefore, both the ‘ meganephric’ and the ‘ plectonephric ’’ systems existing side by side in the same worm. Vejdovsky (16) has found that ‘in this worm also, during development there is to begin with a pair of nephridia to each segment ; these have a funnel, and from the funnel. leads a straight duct not perforate; here and there the cells become larger and finally form loops; these loops ultimately increase in size and become complicated coils, the connective point of the original tube degenerating into a mere strand of : DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 55 connective tissue. The last step is the absolute severance of the connexion. Thus it appears, firstly, that the nephridial system of this worm originates from a pair of pronephridia to each segment; and, secondly, that this becomes broken up into a large number of nephridia, of which one only—the large paired nephridium—retains the funnel ’ (4). The development of nephridiain Mahbenus imperatrix described by Bourne (8) is remarkably similar to that of the nephridia in Megascolides described by Vejdovsky (17). The only difference is that while in the former the funnel is at no stage well developed, is probably never functional, and afterwards entirely degenerates, in the latter the funnel is retained by one pair of nephridia. In fact, the resemblance in the development in the two forms is so great that there is a remarkable similarity between Vejdovsky’s diagram (Pl. 32, fig. 5) showing the development of nephridia in Megascolides and Bourne’s diagram (PI. 5, fig. 39) showing the same in Mahbenus. From the foregoing account of the history of our knowledge of the development of nephridia in earthworms we arrive at three more or less definite broad conclusions. The first is with regard to the fundamental problem of the ultimate origin (ecto- or mesodermal) of the Oligochaete nephridium. As we have seen, there is an overwhelming amount of evidence to show that the nephridia in Oligochaetes are certainly ectodermal. Secondly, in all forms with the so-called ‘ plectonephric ’ system, studied so far, this adult condition is preceded in the embryo by a condition of paired pronephridia in each segment. In the third place, the adult condition of diffuse micronephridia is derived by the breaking up into separate loops of the em- bryonic pair of pronephridia, the original funnel either being retained by one of the nephridia in each segment or degenerating altogether. The present work on the development of nephridia was undertaken to find an answer to the following questions : 1. Are all the three types of nephridia in Pheretima ectodermal in origin ? 56 KARM NARAYAN BAHL 2. If they are ectodermal, how do the septal nephridia with their ducts come to lose all connexion with the body-wall and be associated with the septa and the gut, which are meso- dermal and endodermal structures respectively ? 3. Is the adult condition of nephridia preceded by a ‘ mega- nephric ’ or paired condition in the embryo ? 4. If so, how is the adult condition derived from the em- bryonic condition ? 5. Do facts of development throw any light on the phylogeny of the Oligochaete nephridial system ? 3. THE Cocoon. The egg-capsules or cocoons of Pheretima do not differ in any essential particular of structure from those of Lum- bricus, Allolobophora, or Acanthodrilus, pre- viously described by Vejdovsky (16) and Beddard (8); but I am recording here my observations on the cocoons of this worm to bring out their special characters. I have no observations to offer on the mode of formation of this structure in Pheretima, but I have no reason to doubt that it is formed in much the same way as in all the other genera where cocoon-formation has been carefully studied, and that the clitellum alone is concerned in its pro- duction. ' Although the cocoons vary somewhat in size, they are very much smaller than those of Lumbricus. On an average they are about 1:5 to 2mm. by 1-8 to 2:4mm., i.e. about one-third the size of the cocoons of Lumbricus. The cocoons are light yellow or olivaceous in colour, the empty cases having a clear transparent olive colour. In form they are more or less rounded in shape and give a distinctly swollen appearance, the two ends being drawn out into very short fibrous appendages. . My observations on the time of egg-laying are based on two species of Pheretima, namely P. posthuma and P. rodricensis. The cocoons of the first species were found by me at Allahabad (India) in spring and summer months DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 57 (March to June) out of doors in moist places in the surface layers of the soil in abundance, but during the rains (July and August) they were very rare. My friend Mr. B. K. Das has since informed me from Allahabad that he has been able to collect cocoons of earthworms (not necessarily of Phere- tima) in the months of November, December, January, and February ; and he rightly suspects that egg-laying continues almost throughout the year. Of course the number of cocoons found in the winter months is very small, smce the surface layers of the soil get very dry on account of the prolonged drought, and the worms go deep into the soil and are them- selves difficult to obtain. As regards the cocoons of P. rodricensis,! my observa- tions are based on worms kept in captivity in garden-pots in a hot-house. In order to make sure of the specific identity of my cocoons I kept worms of this species in sterilized earth, to which decaying leaves previously sterilized were added from time to time. From a number of garden-pots containing these worms I could obtain cocoons in any number containing embryos at various stages of development throughout the year. The statement is usually made in text-books that ‘ege-capsules are formed in spring or early summer and the young worms grow mainly durig the summer months. Sometimes large clusters matted together may be found in autumn packed away under clods or in banks where there is a favourable condition of moisture’.2 Wilson (18) says, ‘ ego-laying seems in special cases to continue throughout the year, though it is most active in the spring and summer months. I have found the capsules of Lumbricus foetidus out of doors in nearly every month of the year, but in mid-winter they are only found in decomposing compost- heaps where the temperature is maintained at a tolerably high point ’. From these authorities and from my own observations I am inclined to believe that the time of egg-laying depends 1 T am indebted to Col. J. Stephenson of the University of Edinburgh for identification of this species. 2 Osborn, ‘ Economic Zoology ’, New York, 1908, pp. 110-11. 58 KARM NARAYAN BAHL very largely on the external conditions—temperature, moisture, and the richness of soil. My garden-pots containing the worm were kept quite damp ; the temperature of the hot-house was always about 60° F. and the soil was frequently ‘ manured ’, so to speak ; and it is no wonder, therefore, that under these artificial conditions cocoons were obtained at all times of the year. In nature these conditions are best fulfilled in spring and early summer, and hence we get the greatest activity in egg-laying in these months, although it seems that it does not stop altogether at other times of the year. I have opened hundreds of cocoons of Pheretima and feel justified in considering, as a rule, there is only one embryo in a cocoon. Occasionally one comes across two embryos in a cocoon of a very young age, and only once did I see three embryos in one cocoon. In fig. 24, I have tried to represent three typical stages of the embryo of Pheretima in their natural size (the external segmentation of the body, although complete throughout, cannot be made out with the naked eye and is therefore not represented). The rate of development is very much slower in Phere- tima than in Lumbricus. Wilson (18) found that in laboratory cultures the young worms (Lumbricus) made their escape from the capsule in about two or three weeks. Beddard (8) judges that the shortest timein Acanthodrilus can hardly be less than five or six weeks. In Pheretima the rate is even slower than that in Acanthodrilus, and I cannot put the shortest period at less than eight weeks in this case. Beddard (8) found that the albuminous fluid filling the cocoon in Acanthodrilus, as in Lumbricus rubellus, was milky and opaque while the shell was transparent ; in Phere- tima, however, the albuminous substance of the cocoon is perfectly clear and transparent like its shell, so that under a binocular microscope I could always see, by transmitted light, the embryo inside the cocoon without opening it, and it was thus very convenient to be able to know roughly the size and age of the embryo before opening it. DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 59 Vejdovsky and Beddard speak of two perfectly distinct membranes forming the shell of the cocoon. I have not been able to see these two membranes in the case of Pheretima cocoons, the shell of which seems to me to be single-layered. 4, GENERAL OUTLINE OF THE DEVELOPMENT OF NEPHRIDIA IN PHERETIMA. The three sets of nephridia of Pheretima, namely, the integumentary, the septal, and the pharyngeal arise in the embryo at successive stages of its development. In order to elucidate, therefore, the development of the whole nephridial system consisting of these three distinct series of nephridia and their ducts, it is necessary to examine a large number of embryos of widely different ages. The work is rendered laborious and difficult on account of three facts: firstly, that each type of nephridium develops independently of the other— these several types are not derived one from the other ; secondly, that the nephridia of the three series develop at different ages and in different positions in the embryo; and thirdly, that each series consists of numerous nephridia that go on developing for a long time even after the embryo has left the cocoon. But before going into the details of each stage of nephridial development, I shall provide here an out- line sketch of the development of the elaborate excretory system of this worm. Leaving aside the transitory excretory cells the earliest beginnings of permanent nephridia appear in this worm, as in Lumbricus (18), Rhynchelmis (16), and Crio- drilus (15), as teloblasts lying on the surface of the embryo, ventral to the mesoblastic bands and in front of the meso- dermal pole-cells. While these teloblasts form part of the surface epiblast im very young embryos (300 » long), they soon sink below the surface and come to lie between the definitive ectoderm and the mesoderm. Strings of cells are budded off from and in front of these ectodermal teloblasts, and it is these cell-rows (nephroblasts) that form the material foundation (‘ Anlage ’) from which are derived all the future nephridia. No. 261 F 60 KARM NARAYAN BAHL These strings of nephridial cells aggregate later into groups that lie opposite and a little posterior to the places where the intersegmental septa joim the body-wall. These groups of cells, situated underneath the mesodermal peritoneal membrane (‘somatopleure ’), proliferate to form masses of cells, which, as they grow, begin to project into the coelomic cavity. These constitute the nephridial rudiments. ‘They carry with them in their growth the sheet of peritoneal membrane, which now forms an enveloping sheath over these unformed nephridia (fig. 9). In longitudinal sections of an embryo, about 4 mm. in length, these embryonic nephridia are seen for the most part as solid clup-shaped masses, lying in the anterior part of each coelomic chamber, a pair in each segment of the body except the first two. While the first two segments are devoid of nephridia and the greater part of the embryo possesses solid nephridial masses, some of the anterior segments (seventh and eighth, for example) have fully-formed nephridia with the charac- teristic shape and the intra-cellular canals of the adult organ. In preparations of whole embryos of suitable age, flattened after opening them through the mid-dorsal line, we can see the rudiments of these primary nephridia as elongated masses lying posterior to the septa towards the hind end of the embryo; but, as we examine the segments in front, we get the nephridia in all stages of development in the same embryo, since development proceeds antero-posteriorly. We may note here that these nephridia have no connexion with the septal partitions, and consequently a ‘ septal funnel ’ is never formed at any stage of development of this primary pair of integu- mentary nephridia. At this stage of development (4mm. long) the embryo exhibits a typical meganephric or paired condition like that of the adult Lumbricus, having a pair of ‘ true’ ectodermal nephridia in each segment (Text-fig. 14). This marks the first stage in the development of nephridia in Pheretima, which comprises the developmental history from the first appearance of teloblasts up to the formation of a pair of primary integumentary nephridia in each segment. DEVELOPMENT OF NEPHRIDIA OF PHERE'TIMA 61 In the second stage that follows we have the appearance and development of the primary pair of septal nephridia in each Trext-Figc. 1. Diagrammatic representation of the three stages of development of the nephridial system in Pheretima. A represents a diagrammatic section of an embryo about 4mm. in length, showing the paired condition of nephridia (meganephric stage), B represents a stage at which the embryo has two pairs of nephridia, in each segment, a primary integumentary pair of the first stage and a primary septal pair. C shows the formation of secondary septal and integumentary nephridia. In B and C the intersegmental septum is shown on the left half. b.w., body-wall; g., gut; d.v., dorsal blood-vessel; v.v., ventral blood-vessel ; s.v.v., subneural vessel; m.c., merve-cord ; p.t.m, primary integumental nephridia; p.s.n., primary septal nephridia ; 8.7.n., secondary integumentary nephridia; s.s.n., secondary septal nephridia. segment of the body behind the first fourteen. As these nephridia begin to appear before all the integumentary ones of the first stage have attained to their full development and F 2 62 KARM NARAYAN BAHL size, we have the later development of integumentary nephridia going on side by side with the appearance and growth of septal nephridia, so that we have an overlapping, so to speak, of the first and second stages. The rudiments of septal nephridia appear in two rows, one on each side of the dorsal vessel. The latter in embryos is single anteriorly but double for the creater part of the posterior portion, and the earliest rudiments of the septal nephridia recognizable in whole preparations lie on both sides of this double dorsal vessel (Text-fig. 4). But while the integumentary nephridia vary in their topographical position from segment to segment, lying close to and away from the nerve-cord alternately, the septal ones lie in two straight rows, nearer the mid-dorsal than the mid-ventral line. As their name implies, the septal nephridia develop on the intersegmental septa and, in sections, can be seen to lie just internal to the commissural that connects the dorsal with the subneural blood-vessel. As a septal nephridium develops, the pre-septal portion elongates to form a long narrow tube ending in the funnel, the body of the nephridium, lying in the coelomic cavity behind the septum, develops the limbs and loops of the adult organ, while the terminal duct elongates to run along the septum, parallel and internal to the commissural vessel, to meet its fellow into the supra-intestinal duct mid- dorsally. These pairs of nephridia of the second stage differ from the primary integumentary nephridia in that the former develop on the septal wall and have no connexion with the body-wall from the very beginning, and that they develop a septal funnel. Thus we see that septal nephridia are not derived from integu- mentary ones, and have no connexion with them except that, as will be shown later, both types can be traced to the same source. When the embryo has developed a pair of septal nephridia in each segment we get to the end of the second stage. At this stage the embryo possesses, in each of its typical segments, two pairs of nephridia, an integumentary pair and a septal one, the former opening to the exterior on the body-wall and lying DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 63 alternately dorsally and ventrally, and the latter opening into the supra-intestinal duct and lying dorsally throughout (Text-fig. 1B). The vertical ducts leading from the supra- intestinals to the lumen of the gut at each intersegmentum are also formed at the end of this stage. In the third stage we have the development of secondary nephridia, integumentary and septal. These begin to appear at a rather late period of development of the embryo, when it is almost fully formed and is about to come out of the cocoon. The circlets of setae are completely formed in all the segments of this age, and while rudiments only can be seen of septal and integumentary nephridia towards the posterior end, we find them in various stages of development anteriorly. These secondary nephridia of both types appear indepen- dently of the primary pairs of their segments. The integu- mentary ones appear earlier than the septal, since towards the posterior end we find segments with rudiments of secondary integumentary nephridia but with no traces of secondary septal ones. In both cases these nephridia in their initial stages are lumps of cells having no connexion with the primary nephridia. The septal secondary nephridia appear immediately ventral to the primary pair, and develop very much in the same way as the primary pair, their terminal ducts running dorsalwards on the septa and meeting the ducts of the primary nephridia. Some of these nephridia develop a pre-septal funnel like the primary nephridia, but others, as shown in fig. 15 a, develop the funnel in the same segment in which they lie. In this way we get two kinds of nephridia, one kind with pre-septal funnels and the other with funnels in the same segments as the nephridia. Subsequent pairs of nephridia develop similarly, and we get the formation of the septal canals by the union of the terminal ducts of secondary nephridia with those of the primary ones. The secondary integumentary nephridia first appear on the body-wall as masses of cells lying beneath the somatic layer of the coelomic epithelium. They do not arise strictly in pairs 64 KARM NARAYAN BAHL like the primary nephridia, but have a more or less scattered arrangement. They may appear either on the dorsal or on the ventral side of the primary nephridia (Text-fig. 1 c). They arise in connexion with the setal sacs, and one can very often see a string of cells running from the rudiment of a secondary nephridium in the coelom to the epidermis alongside a setal sac. Trxt-Fic. 2. C.C Diagram showing the formation of nephridial rows of teloblasts and the intersegmental position of large ‘funnel-cells’. m., mouth. For lettering see Text-fig. 3. The pharyngeal nephridia of the fourth, fifth, and the sixth segments make their first appearance in a manner similar to the primary integumentary nephridia, although the former lag behind in development, the integumentary nephridia growing faster than the pharyngeal. Amongst the pharyngeal nephridia themselves the nephridia of the sixth segment develop faster than those of the fifth, and these, in their turn, faster than those of the fourth. The first pair in each of these segments originates from the ectodermal nephridial row, and has a long terminal duct which meets the wall of the pharynx ventro- DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 65 laterally. Secondary pharyngeal nephridia are formed distal to the primary ones as buds on the ducts of the primary pairs of nephridia (fig. 19 and Text-fig. 3). 5. DEVELOPMENT OF THE PRIMARY INTEGUMENTARY NEPHRIDIA. Although I observed short intracellular canals in certain ectodermal cells of very young embryos (gastrulae) in the living condition, and believed these cells to be of the nature of larval excretory cells, I could not definitely locate them in preserved and stained embryos, nor could I examine and arrive at any definite result about these cells in sections. I shall, therefore, confine myself to the development of per- manent nephridia alone. As already indicated, the first set of nephridia to make their appearance in Pheretima are the ectodermal primary nephridia, a pair in each segment. At a stage of develop- ment when the embryo has fifty to fifty-five clearly-defined segments and is about 4mm. long, we can easily see some of the anterior segments (seventh and eighth, for é¢xample) possessing a pair of fully-formed nephridia opening on the surface of the body-wall. Each of these segments, at this stage, resembles in this respect a segment of the adult Lum- bricus, and we may even call this stage of development of the nephridial system of Pheretima the ‘ meganephric ’ stage. The early history of these nephridia is very similar to that described in Lumbricus, Criodrilus, and other worms by previous writers. In an advanced gastrula in which the -— mesodermal bands are well formed and in which cavities are beginning to appear, we can recognize the earliest beginnings of nephridia. On examining such an embryo, when it is still a rounded sphere and has not begun to elongate, as for example the one shown in fig. 1, which is about 140p in diameter, we see the mesoblastic bands diverging from the two large mesodermal pole-cells lying at the future posterior end. These bands he along the two sides of the ventral surface of the embryo, and, on careful focusing, we can also see the begin- 66 KARM NARAYAN BAHL nings of five or six coelomic cavities in each of the two bands. On examining, however, the surface epiblast covering these mesoblastic bands ventrally, we can distinguish four rather large and rounded cells on each side, called the teloblasts. These teloblasts lie a little way in front of the pole-cells: the three ventral ones lie six or seven cells in front of the pole- cells as seen in longitudinal sections (fig. 2), while the fourth, the lateral teloblast, lies a little farther forward than the rest. In these very young embryos (figs. 1-4) the teloblasts form part of the surface epiblast, but can be easily distinguished from the adjacent epiblastic cells both by their larger size and by the fact that their nuclei are free from granules surrounding the nucleolus and thus give an appearance of greater trans- parency as compared with the nuclei of the other cells. Of these four teloblasts on each side the one near the mid-ventral line is the neuroblast, going to form the nerve-cord of the adult, the two lying outside the neuroblast are the nephro- blasts, which go to form the nephridia, while the outermost and dorsal is the lateral teloblast which lies just outside and dorsal to the coelomic sac on each side at this stage of development of the embryo (fig. 4). In a series of transverse sections of an embryo, about 300, in length, from which figs. 8, 4, and 6 are taken, we can follow these four teloblasts forwards as they bud off rows of cells in front. The rows of cells in front of the teloblasts can be followed for a long way in young embryos. Concerning the nephridial cells in con- tinuation with the nephroblasts, we have to note that while the nephroblasts are large cells and occupy the whole thickness of the epiblast, the nephridial cells in front are small and come to he deep in the ectoderm. They can be seen distinctly marked off by a sort of boundary line from the definite epiblast, which is very thin at places where these nephridial cells occur. These cells are thus embedded in the ectoderm, as shown in figs. 8, 4, and 6, but they have not yet formed a separate layer of their own. The next step in the development of the nephridia, which is slower than that of the nerve-cord, is that the nephridial DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 67 teloblasts and the rows of cells in front of them sink beneath the ectoderm and come to form a separate and distinct layer of their own, between the ectoderm on the outside and the mesodermal lining of the coelomic cavities on the inside. In longitudinal sections of young embryos (fig. 5) this layer towards the posterior end gives the appearance of a kind of string of nephridial cells. The large nephridial teloblast together with a row of smaller cells lying in front of it form the definitive nephridial layer. The transition from the previous stage can be well appreciated by comparing the position of the nephroblasts and the nephridial cells in figs. 2, 8, and 4, where they are superficial, with the deeper position occupied by them in figs. 5 and 7. This nephric cord is single-layered in the beginning and remains so for a long time at the posterior end, but its cells soon begin to multiply and proliferate opposite and behind the intersegmental septa which divide one coelomic sac from another, so that we get groups of these nephridial cells situated at intersegmental intervals. The cells of these intersegmental nephridial groups multiply here beneath the peritoneal lining of the coelom, and the cells tend to travel backwards towards the middle of the segment. Some of these nephridial cells push their way into the septa between the two apposing walls of the adjoiing coelomic chambers (figs. 7 and 13). We thus get these intersegmental nephridial masses segregat- ing into two separate groups, one keeping its * retroperitoneal ’ position while shifting backwards and multiplying rapidly, the other consisting of very few cells which make their way into the septa and lie between the two sheets of peritoneum forming the two faces of the septa. In earlier stages—or what amounts to the same thing, in the posterior segments of the embryos—we can see, in longitudinal sections, one nephridial group in each segment lying at the posterior of the two angles formed by the septum with the ventral body-wall (fig. 8). In later stages, or in the more advanced anterior segments, the segregation into two groups becomes quite evident. ‘The group consisting of a few cells caught in between the two septal 68 KARM NARAYAN BAHL sheets, leads, to anticipate matters, to the development of septal nephridia, which we shall speak of in the next part of the paper, while the other group consisting of a number of cells lying beneath the peritoneum and immediately posterior to each intersegmental septum, is the rudiment of the primary pair of integumentary nephridia. We shall now consider the details of development of these integumentary nephridia. This group of cells beneath the peritoneum is the ‘ retro- peritoneal’ group of Meyer (18) and forms the forecast of the whole primary nephridium of the first stage. The cells of this sroup separate away from the septum, divide and proliferate so as to bulge out as solid masses into the coelomic cavities, as shown in fig. 8, c. They carry with them their peritoneal covering which forms a thin sheath round these solid nephridia. The growth is not only vertical but also horizontal, and the nephridial rudiment besides increasing in thickness and pro- jecting into the coelom also extends laterally, so that in a preparation showing the body-wall of an embryo flattened we get a pair of deeply-staining elongated solid masses of cells lying immediately behind each septum as shown in fig. 11,4, and Text-fig. 4. By what steps this elongated ridge lymg behind each coelomic septum develops into an adult nephridium I have shown in fig. 11 (a-r). The earlier stages, in which the nephro- blasts and their derivatives multiply, form masses of cells at septal places which segregate further into two groups, a smaller one, the cells of which push their way into the intersegmental septa, and a larger one, the cells of which move backwards and form the so-called ‘ retroperitoneal’ group of cells, which forms the elongated solid ridge bulging into the coelomic cavity, can all be followed in a few series of longitudinal and transverse sections; but once the nephridial rudiment has reached the size and shape of the elongated mass, shown in fig. 11 a, we can follow its further development best in whole embryos that have been opened in the mid-dorsal line, their endoderm with food-yolk removed, and the remaining portion mounted flat. DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 69 The ridge-shaped mass of nephridial cells grows in the middle, and we soon get a sort of papilla-like protuberance ; this papilla elongates further into a long loop, having its two limbs close together. At this stage, while the two ends of the loop forming the proximal part are attached to the body-wall, the loop itself forming the distal part lies free in the coelomic cavity (fig. 11 c). This loop now elongates further, and, side by side with the elongation of the loop, we find its two limbs getting more and more closely pressed together so as to form one compact lobe. A bend appears, at this stage, towards the base of this lobe, and we now get two more or less distinct divisions of the embryonic nephridium, the one distal to the bend (fig. 11 p) and lying free, and the other proximal to the bend and connected with the body-wall. The distal portion is now a compact structure and goes to form the short straight lobe of the adult nephridium. Although it has visibly lost its double character, we must note that it is really double morpho- logically, having been formed by a close apposition of the two limbs of the loop. The proximal portion of the developing nephridium, the part connecting the bend with the body-wall, is now the seat of further growth. In this portion the double nature of the nephri- dial loop persists for a time (fig. 11 p), but, soon after, the two limbs of the loop at its proximal end, which are really the two opposite ends of the original ridge, come closer together and elongate further. This further elongation of the part proximal to the bend results in the formation of a twist, a little way from the bend, resulting in a condition of the nephridium represented in fig. 11 5. Elongation and twisting go on further until we get the long twisted loop fully formed, with the number of twists characteristic of the adult nephridium. The straight lobe and the twisted loop having been fully formed, the proximal end connecting the nephridium with the body-wall becomes narrow and slender and forms the terminal duct of the nephridium. This duct has meanwhile grown through the thickness of the body-wall, and opens to the exterior in front of the row of setae occurring in the middle line of each segment. 70 KARM NARAYAN BAHL In this fully-formed integumentary nephridium we have to note the absence of either a coelomic funnel or a solenocyte or ‘ flame-cell’. During the course of development, when the two ends of the elongated nephridial ridge come close together, one end develops into the terminal duct opening to the exterior, while the other remains blind and does not develop any struc- ture at all. These nephridia develop an intra-cellular canal and cilia like the septal ones ; and, no doubt, the excretion in their case takes place by means of the diffusion of the coelomic fluid through their permeable walls. 6. DEVELOPMENT OF THE PRIMARY SEPTAL NEPHRIDIA. When the embryo has acquired a pair of integumentary nephridia in each segment—fully developed in the anterior and in various stages of development in the posterior segments —the second set of nephridia, i.e. the septal, make their appear- ance. These form the second pair of nephridia in the body segments of Pheretima (Text-fig. 1B). The first fifteen segments of the embryo do not develop this second set of nephridia, which appear only in segments behind the first fifteen. Unlike the primary integumentary nephridia the septal primary nephridia appear on both sides of the dorsal vessel instead of the nerve-cord. They form two rows, one on each side of the dorsal vessel, at a distance of about 160 » from it in an embryo 9mm. in length. The alternate or scattered arrangement characteristic of the integumentary nephridia of the first stage does not obtain in these septal ones, which occur in two straight rows. These nephridia of the second set have no connexion with the body-wall, but appear from a very early stage, as their name implies, as outgrowths on the intersegmental septa. They can then be first recognized in whole mounts as masses of cells on the septum, projecting on its posterior surface. These nephridial masses on the septa can be recognized with cer- tainty at the earliest in embryos, about 8 to 9 mm. in length, which have been opened in the mid-ventral line, their yolk removed and the rest including the endoderm mounted flat. DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 71 TEXxtT-FIG. 3. A series of three diagrams showing the common origin and develop- ment of the three types of nephridia in Pheretima embryos. In A (left half) the integumentary nephridia are seen pushing themselves into the coelomic chambers while the ‘ funnel-cells ’ are travelling dorsalwards between the adjoining coelomic chambers. The origin of the pharyngeal nephridia and ducts is alsoshown. The right half shows the nephridia at a more advanced stage of development. Septal nephridia are developing between the adjoining peritoneal sheets, while the ducts of the pharyngeal nephridia are formed even before the nephridia themselves are fully formed. In B (on the left) is shown the development of secondary nephridia, while the right half shows more or less the adult condition of nephridia in the worm. 2.¢., nephridial teloblasts ; mph.c., cells of the nephridial row; c.c., coelomic cavities ; int.nph., integumentary nephridia; s.n., septal nephridia ; s.mr., rudiment of a septal nephridia; ph.nph., pharyngeal nephridium ; ph.d., duct of pharyngeal nephridia; g., gut ; s.i.d., Supra-intestinal excretory duct; f.c., funnel-cells. 72 KARM NARAYAN BAHL In such preparations of embryos (Text-fig. 5, part of an embryo about 9mm. long) we can follow these nephridial rudiments antero-posteriorly. The dorsal vessel in embryos of this age is double in the posterior portion, and consists of two lateral vessels lying on the sides of the gut. Anteriorly, TEXT-FIG. 4. Portion of an embryo mounted flat after removal of the gut, showing the relative position of the developing integumentary nephridia in successive segments. 6.w., body-wall; 7.s., interseg- mental septa; 7.c., nerve-cord. however, the two dorsal vessels converge and fuse to form one single vessel, and we find it as such in the mid-dorsal line in the anterior part of the embryo. Following the nephridial masses from the septum 15/16, we can trace them backwards a good way beyond the point where the two converging vessels meet to form the single dorsal vessel (Text-fig. 5). The further development of these septal nephridia can be followed in whole mounts of embryos, 9 to 18 mm. in length, flattened DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 73 after being opened from the ventral and not the dorsal side. As the nephridial mass grows in size we can soon distinguish the two ends of the growing nephridium, as shown in Text- fig. 6. One end grows inwards along the septum towards the dorsal vessel, beneath which it meets its fellow of the other TEXT-FIG. 5. Portion of the whole mount of an embryo 8 mm. in length, showing the rudiments of the septal nephridia on each side of the dorsal blood-vessel. d.v., dorsal blood-vessel, double behind; 7@.n., integumentary nephridia; s.n., septal nephridia; 7%.s., inter- segmental septa. side to form the supra-intestinal excretory duct ; we can call this end the centripetal end. The other end of the nephridial rudiment at this stage is away from the dorsal vessel, and proliferates to form a mass of cells which project in front of the septum to form the beginnings of the ‘funnel’ of the nephridium. ‘This pre-septal portion of the nephridium soon attains to a considerable size, and is a prominent feature of the 74 KARM NARAYAN BAHL septal nephridia at all stages of their development (fig. 12). While these two ends of the nephridium—the ‘ centripetal ’ end and the ‘ funnel’ end—are growing and differentiating, the portion of the nephridial mass between the two ends also srows and forms a papilla-like projection behind the septum (fig.128). This papilla elongates to form a loop, the two limbs of which come close together ; a bend appears, and the portion distal to the bend forms the rudiment of the short straight lobe of the adult nephridium. This stage of the development of the septal nephridium is represented in fig. 12 c. The two ends of the nephridium are attached to the septum while the body of the nephridium, consisting of a newly-formed straight lobe (s.l.), distal to the bend, and a growing region proximal to the bend, between the latter-and the attached ends of the nephridium, lies free in the coelomic cavity. As this growing region elongates (fig. 12 p) the two limbs come close together, and, as a result of elongation, twists appear, which grow to form the spirally twisted loop of the adult nephridium (fig. 12 and Text-fig. 7). We thus get the main body of the nephridium, consisting of the short straight lobe and the long spirally twisted loop, fully formed. Histological differentiation, along with the formation of intracellular canals with cilia lining them at intervals, completes the development of a nephridium. As will be seen by comparing the foregoing account of the development of a septal nephridium with that of an integu- mentary nephridium described in the last section, the successive steps of growth in the two cases are very similar if not identical. The chief difference lies, of course, in the fate of the two ends of the nephridium. In the case of a septal nephridium one end grows out to be pre-septal and is differentiated to form the ‘funnel’, the other end forms the terminal duct which runs along the septum, parallel to the commissural vessel, and joins its fellow to form the supra-intestinal duct ; on the other hand, in the integumentary nephridium the ‘funnel’ end is blind, and the terminal duct opens on the surface of the skin. We have now followed the development of a septal nephri- DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 75 dium from a stage when it consists of a mass of cells on the septum (fig. 12 a) to a stage when it has attained to its adult structure (Text-fig. 7). But in order to assign these septal nephridia to one of the three primary germ-layers we must trace the ultimate origin of this septal mass—the unformed septal_nephridium. We must note that an intersegmental TEXT-FIG. 6. = SBSEED BSR S) wine Loe vk = ee SSeS F See, [ \ aa 2P0°%e SSO DS CDE Sl. Portion of the whole mount of an embryo showing septal nephridia at a more advanced stage of development than those in fig. 13. s.s., setal sacs; other letters as above. septum is morphologically double and results from a coalescence of the two layers of peritoneum covering the two faces of the septum, and that although this double character of a septum is not discernible in sections of an adult worm the two layers of peritoneum can easily be distinguished in longitudinal sections of embryos. The question naturally arises as to whether this nephridial mass arises by a proliferation of one or more cells belonging to the two layers of peritoneum forming the septum, NO, 261 G 76 KARM NARAYAN BAHL like the testes and ovaries, or whether the mass arises by multiplication and growth of one or more cells lying between the two adjoining sheets of a septum. Is the septal nephridium intra-peritoneal or inter-peritoneal ; or, in other words, is it mesodermal or ectodermal ? This is the fundamental morpho- logical question to be answered. We have already noticed that during the course of develop- ment of the primary integumentary nephridia the mass of nephridial cells lying opposite and behind the intersegmental septa, underneath the coelomic epithelium, segregates early on into two groups—one forming the ‘ retroperitoneal’ group of cells and developing into an integumentary nephridium, and the other consisting of a few cells that push their way into the septum between its two layers of peritoneum. This second group, which is directly traceable to the original nephric row and has thus the same source as the integumentary nephridia, is in fact the primordial rudiment of the septal nephridia. In a series of longitudinal sections of an embryo about 6 mm. long, we can trace how a cell from this primordial group travels through the septum to take up its final position in the row of septal nephridia on each side of the dorsal vessel. If we examine, in this series, a septal nephridium on one of the anterior septa—say the twentieth—we find that it lies at a little distance from the dorsal vessel immediately internal to the commissural vessel (fig. 14). This incipient nephridium and the commissural vessel are both situated between the two sheets of the septum, one below the other (fig. 14). As we trace this nephridial rudiment backwards we find that it retains the same relative position with regard both to the dorsal and the commissural vessels. We can thus trace the nephridial rudiment, consisting of a few cells as it lies dorsally on each side of the dorsal vessel on one of the anterior septa, back through successive segments to the posterior end of the worm, where the nephridial rudiment lies ventrally on each side of the nerve-cord, and is just beginning to push its way into the edge of a septum. On examining the sections shown in figs. 18 and 14 two fundamentally important facts come out, DEVELOPMENT OF NEPHRIDIA OF PHERETIMA re The first is the inter-peritoneal situation of the rudiment of a septal nephridium, i.e. in other words, the septal nephridia are not derived from one of the cells belonging to the peritoneal lining of the septa but take their origin from cells lying between the two sheets of peritoneum forming a septum. The second is that these rudiments can be traced directly to the original nephric row. Since the original nephric row is ectodermal in origin we have established the ectodermal origin, in the last analysis, of the septal nephridia. In all descriptions of previous work on the development of nephridia in earthworms, mention is made of a ‘ funnel-cell ’ (‘ Trichterzelle ’), a term which is used in at least two senses. It is either used for a single large cell which is separated off very early from the nephric row and forms the forecast of the whole nephridium, or it is used for the most internal cell of a series or group of cells which go to form the whole nephridium, and, in this case, the ‘ funnel-cell’ gives rise only to the funnel of the nephridium. The term is used in the former sense by Bergh in the case of Criodrilus and Lumbricus (6 and 7), and in the latter sense by Staff in the case of Criodrilus (15). During the development of nephridia in Pheretima also we can distinguish a large cell which is probably the equiva- lent of the ‘funnel-cell’. So far as the development of the integumentary nephridia are concerned, the ‘ retroperitoneal ’ eroup of cells, which give rise to them, contains no ‘ funnel- cell’ in it, nor, as we have seen, do we get a funnel formed in the adult integumentary nephridia. But with regard to the septal nephridia we can distinguish a cell larger than others at almost all stages of their development. In fig. 6, which represents part of a transverse section of the posterior end of a young embryo, we can distinguish one large cell in connexion with the septum on each side. Further, during the passage of this cell to its dorso-lateral position, one cell can always be distinguished by its very large size as compared with the surrounding peritoneal cells. Finally, when the rudiment of the septal nephridium consist of a group of three or four cells lying on the septum on each side of the dorsal vessel, G2 78 KARM NARAYAN BAHL one of the cells of this group is larger than the rest (fig. 14 B), and we can infer that this large cell is the so-called ‘ funnel- cell’. It would seem, therefore, that this large cell, as it pushes itself into the septum, is the forecast of the whole nephridium, and is a ‘funnel-cell’ in the sense im which Bergh uses it. And later this large cell divides and gives off cells smaller in size than itself ; and while these smaller cells go to form the body of the nephridium the large cell develops into the funnel and becomes a ‘ funnel-cell’ in the sense in which Staff uses it. 7. DEVELOPMENT OF THE SECONDARY NEPHRIDIA, SEPTAL AND INTEGUMENTARY. At the end of the second stage of nephridial development, as we have seen, a typical segment of the embryo contains two pairs of nephridia—an integumentary and a septal. Soon after, rudiments of other nephridia, both septal and integu- mentary, begin to appear. -These rudiments of secondary nephridia (all nephridia appearing after the first pair, septal and integumentary, have been grouped together under the term ‘ secondary ’) can be seen both in sections and in whole preparations of embryos of suitable age as deeply-staiming masses of cells on the septa and the body-wall. In order to study the development of these secondary nephridia, two sets of embryos should be selected—the first set, consisting of those embryos which are fully formed and are about to hatch out of their cocoons: these show the secondary nephridia at a fair degree of development ; the second set, consisting of those embryos which are not fully formed and would take some time before they are ready to hatch out : these show secondary nephridia in their very early rudimentary condition. It may be difficult, in the beginning, to distinguish embryos belonging to these two sets, but when one gets familiar with them after opening a number of cocoons, one can always distinguish them with a fair degree of accuracy. A surer method of distin- euishing the embryos of two sets externally is to examine the setal line. In fully-formed embryos the circlets of setae are DEVELOPMENT OF NEPHRIDIA OF PHERETIMA "9 complete, and, on examination of the embryo under low power, we can see the setae ; but im younger embryos, although setal sacs and muscles can be made out in sections, the setae are not yet formed and so cannot be distinguished externally. (a) Secondary Septal Nephridia. The secondary septal nephridia arise very much in the same way as the primary septal pair. They appear ventral to the primary septals and, like the latter, appear in pairs. It is very difficult to say whether this paired origin is maintained throughout the development of all the septal nephridia, but it is certain that the first two secondary nephridia arise in pairs. We may also note that these nephridia appear later than the secondary integumentary nephridia, since in the posterior segments of embryos with well-developed secondary nephridia in their anterior portion, through we can make out the rudiments of secondary integumentary nephridia, the septal ones have not yet been formed. The group of cells forming a very early rudiment lies, as in the case of the primary pair, between the two peritoneal sheets of a septum and is consequently ‘inter-peritoneal’. One of the cells of this group is larger than the rest and corre- sponds, in all probability, to the ‘ funnel-cell’. As regards the original relations of this secondary pair we have to note, in the first place, that their rudiments lie ventral to and at some little distance from the primary nephridia, and, secondly, that until the nephridium is almost fully formed and has developed its long terminal duct there is no connexion between this and the primary nephridium, nor are there any stray cells lying on the septum between these two nephridia. The obvious inference is that the secondary nephridia do not arise by a process of budding or the like from the primary nephridia, but do so de novo at their place of origin. Whence do the rudiments of these nephridia come ? In describing the ultimate origin of the primary septal nephridia we traced their beginnings to a group of cells which pushed their way into each septum, and which, in their turn, 80 KARM NARAYAN BAHL could be traced further to the original nephric row derived from the ectoderm. ‘his group of cells pushing its way into the septum forms the primitive material foundation of all the septal nephridia. ‘The primary pair is formed from one of the cells of this group, travelling dorsally on each side. More cells move into the septa and give rise to the other nephridia (secondary septals). That this does actually happen is shown firstly by the fact that there are always a number of cells lying into the septum at its junction with the body-wall, even after the rudiments of the primary pair of nephridia are well formed dorsally ; and, secondly, by the fact that we very often come across cells lying interperitoneally within the septa at a little distance dorsal and inwards to the group of cells referred to above (the group pushing its way into the septum), these cells having apparently been detached from the fundamental group and being on their way to their final place of settlement and growth. We thus conclude that although the secondary septal nephridia do not originate as buds from the primary ones and are completely independent of them as regards their origin, they can be traced to the same source as the primary nephridia, i.e. the intersegmental group of nephridial cells, which form a store-house, giving origin to the rudiments of all the septal nephridia, primary as well as secondary. Coming now to the later development of the secondary nephridia, the chief pomt of interest is the topographical position of the funnel. The usual position of the funnel is always pre-septal, and we have seen that it is so with regard to the funnels of the primary pair of septal nephridia. But in the adult Pheretima (1) I have described the funnel as lying in the same segment as the rest of the nephridium, and there was thus an incongruity between the two facts of structure. This led me to a close examination of the funnels of the developing nephridia in the embryos, and also to a re-examination of the position of the funnel in the adult worm. While, on the one hand, it came out that all the primary nephridia have a pre-septal funnel, in the secondary nephridia, DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 81 on the other hand, both conditions prevail—the funnel is pre- septal in some cases and post-septal in others. Both conditions are represented in fig. 15 a and Bs. In the adult Pheretima it was found that while a large majority of nephridia have their funnels in the same segment there are some with pre- septal funnels. The statement that all the septal nephridia have funnels in their own segments is therefore not quite universally true, as I thought before. We may note, however, that in the case of those nephridia which have the funnel in the same segment, all that happens is that the ‘ funnel-cell ’ and the cells going to form the body of the nephridium project in the same direction, either pre-septal or post-septal. When the first secondary nephridium is fully formed, its terminal duct running along the septum meets that of the primary nephridium dorsal to it, and, similarly, the ducts of all the succeeding nephridia jom those of the preceding ones, and that is how we get the formation of the septal excretory canal running parallel and internal to the commissural vessel (dorso-sous-nervien, 2). We may note that, like the septal nephridia themselves, the septal excretory canal is also inter- peritoneal. (b) Secondary Integumentary Nephridia. The secondary integumentary nephridia, in thei early rudimentary condition, can be seen in older embryos about to hatch out of the cocoon. In whole mounts, as shown in Text-fig. 7, they can be distinguished as small solid deeply- staining masses on the side of and between the setal sacs. The setal sacs at this stage have not yet developed full-grown setae in them, and rudiments of secondary nephridia in sections (fig. 17) can be made out lying beneath the coelomic epithelium between the imner ends of two adjoining setal sacs. They arise at almost any place on the body-wall, like the primary nephridia ; im some segments they are found near the dorsal vessel, in others on each side of the nerve-cord (fig. 17). Whether the secondary nephridia develop from some of the nephridial cells, lying beneath the somatic peritoneum, that 82 KARM NARAYAN BAHL have been left over from the original nephridial masses, or whether they arise from epidermal cells that become nephridial at the time and migrate mwards, [I cannot be sure. On Trxt-ric. 7. Pin a Portion of a whole mount of an embryo about to hatch out of the cocoon, showing the fully-formed primary septal nephridia and the primary and secondary integumentary nephridia. s.n., septal nephridia; .i.n., primary integumentary nephridia ; sec.t.n., secondary integumentary nephridia ; 7.c., nerve-cord. examining Text-fig. 4 it would seem as if the nephridial sub- stance (cells potentially nephridial) is spread over the whole of the body-wall, and any part of it might become active and form a nephridium, and hence the appearance of nephridia at all sorts of places on the body-wall. If that be the case, and if, as shown in fig. 8, nephridial cells extend on each side of the definite nephridial rudiments, we probably get these DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 83 secondary nephridia formed from the stray nephridial cells lying beneath the peritoneal membrane of the body-wall. But, as in fig. 17, we have to account for a string of cells; which is not always seen, running from the nephridial rudiment to the ectoderm. It may be that it is a secondary formation leading from the nephridial rudiment to form the terminal duct. 8. DEVELOPMENT OF THE PHARYNGEAL NEPHRIDIA AND THEIR Ducts. The development of the pharyngeal nephridia of the fourth, fifth, and sixth segments can be followed in all its stages in the same embryos which show the development of the integu- mentary and septal nephridia. The pharyngeal nephridia appear at the same time as the primary integumentary nephridia, but are rather slower in growth than the latter. At a stage of development when the embryo is 5 to 6mm. long and the integumentary nephridia of some of the segments behind the first six (e.g. seventh, eighth, and ninth) are almost fully formed and have developed their intra-cellular canals, the pharyngeal nephridia are seen as deeply-staining compact masses of cells lying on the body-wall, a pair in each segment, one on each side of the nerve-cord. They develop from the same source as the primary integumentary nephridia, 1. e. from the nephridial cells belonging to the original ectodermal nephric row; but their manner of development is different from the other two types. While in the case of the mtegu- mentary nephridia the terminal duct is very short and appears rather late in development, forming a lumen at the same time with the rest of the nephridium, the ducts of the pharyn- geal nephridia develop very early. In an embryo 5mm. in length the nephridia of the fourth, fifth, and sixth segments are small club-shaped solid masses produced into long solid strings of cells leading anteriorly to the lateral walls of the pharynx (fig. 18). The terminal ducts are thus formed earlier than the bodies of the nephridia themselves. This is still more marked at a later stage in an older embryo in which the 84 KARM NARAYAN BAHL ducts of the pharyngeal nephridia are seen to have acquired a lumen, while the nephridia themselves have not yet developed their adult form and are still solid. These ducts are intra- cellular, but are surrounded by the muscular tissue of the strands passing from the pharynx to the body-wall for part of their length. The usual order of antero-posterior development is reversed in the case of pharyngeal nephridia. In the two stages of development mentioned above, the nephridia of the sixth segment are advanced further in development than those of the fifth, and the latter than those of the fourth segment. There is only one pair of nephridia anterior to the pharyngeal ones, 1.e. the one belonging to the third segment, the first two segments of the embryo beg anephrous. This most anterior pair, although integumentary in character, follows the pharyn- geals in their time and rate of development. The ducts in these early stages are very thin with intra- cellular lumen, and are therefore to be looked upon as the elongated terminal ducts of the primary pairs of pharyngeal nephridia rather than as outgrowths from the walls of the pharynx. Three successive thickenings on the lateral pharyn- geal wall mark the places of entrance of the ducts ito the pharyngeal lumen. It is a remarkable fact that not only do the terminal ducts acquire a lumen before the formation of the canals in the nephridia themselves, but that they also open into the cavity of the pharynx long before the nephridia are able to function at all. The pharyngeal nephridia, like the integumentary ones, do not develop a ‘funnel’, but we have to note that at an early stage when the terminal ducts have been formed and the nephridia are developing their adult structure, the pharyngeal pair come into connexion with the intersegmental septa not in front of but behind them. The dorsal blood-vessel and the lateral oesophageals act as the afferent and the efferent vessels to these nephridia, and the branches of these vessels near their points of origin and entrance into the main vessels he on the septal supports. DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 85 Secondary pharyngeal nephridia arise in a way different from that of the secondary septal and integumentary ones. They do not appear independently of the primary pair but develop as buds on the nephridial ends of the pharyngeal ducts. In fig. 19 are seen three buds in the fifth and two in the sixth segment ; while in the fourth the primary nephridium itself is not fully formed yet. As these buds develop into fully- formed nephridia, their terminal ducts, longer than those of the other types of nephridia, remain continuous with the primary pharyngeal duct, or rather open into it. Thus we get a large number of pharyngeal nephridia forming big tufts and having their terminal ducts opening into these primary ducts. The primary ducts themselves, although originally very narrow and intra-cellular, enlarge and acquire a muscular investment which makes their walls thick and tough as they are in the adult condition. 9. COMPARISON WITH THE DEVELOPMENT OF ‘ MEGANEPHRIC ’ AND THE SO-CALLED ‘PLECTONEPHRIC’ ‘TYPES OF NEPHRIDIA. I have referred in brief to the known facts of development of these two types of nephridia in the historical part of this paper. So far as the ‘meganephric’ type of nephridia are concerned, we can compare them only with the primary integu- mentary nephridia of Pheretima, a pair in each segment. The obvious differences between the meganephridia of Lum - bricus and the primary pair of imtegumentary nephridia in an embryo of Pheretima are the larger size of the former and the presence of a ‘funnel’ in them. In his recent memoir on the development of nephridia in Criodrilus, as already mentioned on p. 58 (15), Staff derives the nephridium from the ‘ funnel-cell’ and the retroperitoneal group of cells behind each septum, both being ultimately derived from the nephridial string of cells between the ectoderm and the meso- derm. In Pheretima, as we have seen, the primary integumentary nephridia develop from the ‘ retroperitoneal ’ cells alone and there is no ‘ funnel-cell’ taking part in their 86 KARM NARAYAN BAHL formation, and that is why they do not develop any funnel at all. The septal nephridia, on the other hand, develop from the intersegmental nephridial mass of cells which possesses a ‘funnel-cell’, and so we get septal nephridia with funnels developing from this source. This meganephric, or rather the paired condition in the embryo, 1s superseded by, and is assimilated into, the adult condition which is ‘ enteronephric’, so far as the septal and pharyngeal nephridia are concerned ; but is ‘ micronephric ’ and diffuse so far as the integumentary nephridia are concerned. It cannot be called ‘ plectonephric’. If the place of opening of the nephridia be taken into consideration we can divide the nephridia into two groups: those that open to the outside on the surface of the body-wall and may be termed ‘ dermo- nephric’ or ‘ exonephric’, and those that open into any part of the gut and may be called ‘ enteronephric’. The former category will include the ordinary *‘ meganephridia ’ of Luin- bricus, as well as the imtegumentary micronephridia of Pheretima and the plecto-nephridia of other worms. The latter term, i.e. enteronephric, will comprise the septal and pharyngeal nephridia of Pheretima, the ‘ pharyngeal tufts ’, the ‘ peptonephridia’ or ‘salivary glands’ of such forms as Megascolides, Periscolex, and Enchytraeids, and the anal nephridia of Octochaetus and Allo- lobophora antipae (14). Comparing the nephridial development of Pheretima with that of the plectonephridia of Megascolides (17) and Mahbenus (8), we have to note that, while in the latter the whole system results from the breaking up and branching of the first pair of nephridia, m Pheretima there is no such branching and breaking up, and all the nephridia, septal and integumentary, appear independently of one another. More- over, the first pair of nephridia in Megascolides and Mahbenus have ‘ funnels ’ which persist in Megascolides but degenerate in Mahbenus, while Pheretima has no ‘funnels’ even on its initial pair of integumentary nephridia. In my previous paper (1), from a study of the structure of DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 87 adult nephridia, I came to the conclusion that each nephridium is a separate and discrete structure, that there is no network of any kind connecting one nephridium with the other, and that it is a mistake to describe the nephridial system of Phere- tima asa ‘ coelomic network ’ or as ‘ plectonephric ’, implying the idea of a reticular connexion between the nephridia. This conclusion is now confirmed by the evidence we have from the embryology of the excretory system of this worm. We now know that each nephridium, integumentary or septal, originates independently of the others, and, therefore, even the embryonic connexion found in Megascolides and Mahbenus is wanting in this worm. 10. PHYLOGENY OF THE O1IGOCHAETE NEPHRIDIAL SYSTEM. Before any developmental facts were known with regard to the excretory organs of Oligochaetes, it was commonly held ‘that the paired nephridia (meganephridia) of most Oligochaeta were formed by reduction from a network such as now exists in Perichaeta and many other genera’ (4). But after the embryology of the excretory system in Octochaetus and Megascolides had been elucidated, and it was found that a meganephridial condition preceded the diffuse condition, this view had to be given up. In speaking of the phylogeny of the system, Beddard (4) says that ‘it does not follow that the diffuse nephridia are the outcome of a branching and specializa- tion of the paired nephridia ; what the developmental facts prove is that both paired and diffuse nephridia are formed out of similar pronephridia ; that in fact both kinds of excretory organs are equally ancient’. This view was definitely formu- lated by Vejdovsky (16), who says, ‘Es hat daher meiner Ansicht nach sowohl das “ Plecto- als Meganephridium ”’ gleiche genetische Bedeutung. Beiden muss ein einfacher, paarig in jedem Segmente sich anlegender Strang—das Pro- nephridium—vorausgehen, aus welchem erst secundiir seitliche Wucherungen entstehen, die sich als zahlreichere oder spir- lichere Nephridiallappen erweisen. In grosser Menge bilden 88 KARM NARAYAN BAHL sich offenbar die Lippchen bei den mit “ Plectonephridien ” versehenen Oligochaeten und in den vorderen Segmenten von Megascolides. In den hinteren Segmenten des genann- ten Riesenregenwurmes reduciren sich die Lippchen an einige erdssere, welche der Lage nach den Schlingen am Nephridium von Lumbricus entsprechen.’ Vejdovsky’s conclusions are based on his researches on Rhynchelmis, in which a definite pronephridial stage precedes the permanent nephridia, and on the development of nephridia in Megascolides, in which a paired pronephridial condition precedes the per- manent plectonephric system. The facts of nephridial development recorded in this paper do not lend themselves to Vejdovsky’s interpretation of the evolution of the excretory system. In the first place, we cannot distinguish in the development of Pheretima a pronephridial stage as distinguished from a stage of permanent nephridia. What we do get is a paired condition in the embryo which goes to form part of the adult system and is not entirely superseded by it. Secondly, the paired nephridia themselves are not transformed, as they are in Megascolides, into the diffuse system of the adult, but numerous nephridia arise indepen- dently to be added to the primary integumentary nephridia. In the third place, the adult Pheretima does not show the condition referred to by Vejdovsky in some other worms (‘ Riesenregenwurmes ’), where the anterior segments have numerous nephridia but the posterior ones show the paired meganephric condition. Since the diffuse and paired forms of the excretory system occur in genera which are so nearly related, Beddard (8) thinks there can be no profound gap between the two kinds of organs. But when one takes into account the fact that in the family Perichaetidae, Pleionogaster possesses nephridia of the diffuse type all opening to the exterior, Megascolex has a pair of large nephridia in each segment in addition to the small scattered nephridia, while Perionyx and Diporo- chaeta have only large paired nephridia, it becomes very difficult to think of and offer an explanation for the intermediate DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 89 evolutionary stages between the condition in Perionyx (exonephric) and that in Pheretima (enteronephric). Although we can derive the diffuse condition of Pleinogaster from Perionyx, through such forms as Megascolex show- ing an intermediate condition, we cannot ignore the fact that the gap between the ‘ exonephric ’ (p. 86) and * enteronephric ’ conditions is very deep indeed. A few facts in the embryology of the nephridial system, however, seem to throw some light on the possible evolution of the enteronephric system. I have already shown (p. 67) that the primary integumentary nephridia have a common source of origin in the nephridial masses lying opposite the intersegmental septa. We have also seen that, while the integumentary nephridium has no ‘ funnel-cell’, it is repre- sented in the rudiment of each septal nephridium, and con- sequently the former lacks and the latter possesses a ‘ funnel ’ in the adult condition. It is possible to suppose that the first great step in the process of evolution of the enteronephric system was the severance of the connexion between the * funnel ’ and the ‘body’ of the nephridium. That this severance has probably taken place in Pheretima is supported by very strong evidence from the embryology of the nephridia of Octochaetus (8), Megascolides (17), and Mahbenus (8). Inall these forms there is a paired meganephric condition in the embryo, and each nephridium is provided with a well- developed funnel. In the transformation of this embryonic into the adult condition the part to degenerate first is always the duct following the funnel. In Octochaetus (8), Beddard found that the change took place by the disappearance of the lumen in the portion nearest the funnel. Vejdovsky (16) has found in Megascolides that the paired embryonic nephridia have a funnel from which leads a straight duct without lumen, and that this duct joms the nephridial loops. During development the connecting part of the original tube (i. e. the straight solid duct) first degenerates into a mere strand of connective tissue and finally breaks up entirely. But the funnel remains and forms part of the large nephridium in the 90 KARM NARAYAN BAHL adult. Bourne (8) found a similar condition in Mahbenus, the funnel, however, in this case degenerating entirely. It is easy to derive the condition in Pheretima from what takes place in Megascolides. In the latter the funnel, with part of the tube following it, develops into the large paired nephridium with funnel, a pair to each segment, while the body of the nephridium gives rise to a network of minute excretory tubules. In Pheretima the separation of the funnel from the body of the nephridium is carried a little farther and takes place early in ontogeny. ‘The result is essentially the same as in Megascolides, i.e. the formation of two kinds of nephridia: in this case the larger septal nephridia with funnels and the smaller integumentary ones without funnels. The evolution has taken place along the same lines in Octochaetus and Mahbenus also, but in these two genera the degeneration and disappearance of the portion following the funnel has likewise affected the funnel which also degenerates, and that is why we get only one type of nephridium without funnel in these two cases, although 3eddard found some nephridia with funnels towards the posterior end of Octochaetus, along with those without funnels. In Pheretima the ‘ funnel-cell’ along with some other nephridial cells separate off early from the main nephridial mass, and while the integumentary nephridia develop at once from the main nephridial mass the development of the septal ones from the ‘ funnel-cell’ comes a little later. We tacitly assume, of course, that the ‘ funnel-cell ’ itself by division is capable of giving rise to the whole nephridium, and this is what actually takes place (cf. figs. 13 and 14). Why the funnel gets separated off from the main body of the nephridium during development in Octochaetus and Mahbenus, and why the separation begins so early in Pheretima—whether it took place in phylogeny before or after the septal nephridia had acquired their openings into the gut—are questions difficult to answer. How does this ‘ funnel-cell’ travel dorsalwards and take up DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 91 a dorso-lateral instead of its usual ventral position? That it travels dorsalwards can be seen easily from figs. 18 and 14, and that this is very unusual can be realized from the fact that TEXT-FIG. 8. A series of four diagrams showing how the commissural vessel elongates with the migration dorsalwards of the dorsal vessel and how the nephridial rudiment on the septum is carried to its dorso-lateral position. d.v., dorsal vessel (double in A, B, and C) ; v.v., ventral vessel; n.c., nerve-cord ; sn.v., subneural vessel ; i.8., intersegmental septum; 7.n., integumentary nephridium ; s.n.r., rudiment of a septal nephridium ; b.w., body-wall. ordinarily the funnel in a nephridium is the most ventral part of it and hes almost next to the ventral nerve-cord. But although away from the nerve-cord, the position of the primary septal nephridia is not very unusual if we bear in mind the NO, 261 H 92 KARM NARAYAN BAHL scattered arrangement of the integumentary nephridia (Text- fig. 8). By comparing the position of the septal and integu- ‘mentary nephridia in Text-figs. 5 and 6 the discrepancy does not seem to come to very much. But that there is a shifting dorsalwards, however little (and in many septa a great deal), admits of no doubt. This can be explained, however, easily, if we take into account the facts of development and growth of the septa and the structures connected with them, e.g. the commissural vessel and the dorsal vessel. In the series of diagrams in Text-fig. 8 I have tried to illustrate the gradual erowth and formation of the commissural which connects the subneural and dorsal vessels. The dorsal vessel is formed by the progressive backward concrescence of the two lateral vessels, which, as a rule, lie at the dorsal edge of the advancing mesoderm bands. The commissural, which appears very early and connects the subneural with the lateral (the semi-dorsal so to speak), les from the very begining between the two sheets of the mesodermal septa. From an examination of figs. 13 and 14 it becomes clear that the ‘ funnel-cell ’ or the rudiment of the septal nephridium is closely associated with the com- missural vessel, lying in the same intraseptal cavity with and ventral to the blood-vessel. It would seem that in the process of migration of the lateral vessels to the mid-dorsal position the commissural vessel must keep pace and travel dorsalwards. Further, in the general growth of this vessel dorsally, the ‘funnel-cell’ or nephridial rudiment closely associated with it also travels dorsalwards, and hence results the dorso-lateral position of the primary septal nephridium. That the commissural vessel and the rudiment of the septal nephridia le in the same hollow of the septum between the two sheets of mesoderm in the embryo is clear from figs. 15 and 14, and is evidence of the close association of these two structures. That there is some morphological relationship is also shown by the fact that, im the adult worm, the septal nephridia are only present on those septa that have a com- missural vessel. In the first fourteen segments of the worm there are no septal nephridia and no commissural vessels either, DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 93 Once the ‘ funnel-cell’ reached its dorso-lateral position, it developed into a nephridium with a funnel (a septal nephri- dium); but it would seem that the terminal duct of the nephridium had, so to speak, lost its original course, having been removed from the body-wall and having been caught in the ‘ tunnel’ of the septum containing the commissural vessel. The terminal duct followed the course of the vessel and travelled towards the mid-dorsal line, where, on meeting its fellow of the other side, it formed the supra-intestinal excretory duct. It would be a case of induced development and growth, stimu- lated by the course and development of the commissural vessel. When we have once got a septal nephridium formed in the dorso-lateral position, its terminal duct would tend to finda way out. But since the way to the body-wall is blocked, . the terminal duct lengthens out and follows the course of the commissural vessel. Examples of this kind of ‘ dependent differentiation ’, a term due to Roux, are found in the experi- ments of Lewis and Spemann. Lewis has shown that a lens will be formed from any patch of ectoderm taken from some other part of the body and grafted over the optic cup during the development of the eye in Amblystoma and some species of frogs. Spemann and Lewis have also found that in the absence of contact between the optic cup and the ectoderm the cornea is not developed, that is to say, the overlying ectoderm does not ‘ clear’ (lose its pigment), it does not thin out, and Descemet’s membrane is not formed. Once the supra-intestinal duct is formed in the mid-dorsal line above gut, the only possible way to discharge the excretory fluid is to have communications with the gut in each segment. This tendency of the nephridial ducts to open into the gut has been noticed in other worms also. Rosa (14) found in one species of Allolobophora (A. antipae) that all the nephridia of the posterior region of the body, instead of opening on the exterior, communicate with a pair of longitudinal canals which open posteriorly into a median vesicle com- municating with the rectum. It would appear that in outline 1 Jenkinson’s ‘ Experimental Embryology ’, pp. 271-7. H 2 94. KARM NARAYAN BAHL e the condition of nephridia in the posterior region of the body of Allolobophora antipae is remarkably similar to that in the whole length of the body of Pheretima. On the formation of communications between the supra-intestinal ducts and the lumen of the gut, the essentials of the ‘ entero- nephric ’ system are completed. The formation of secondary septal and integumentary nephridia is not difficult to explain. The septals are, no doubt, traceable to the same source as the primary nephridia, i.e. to the cell-group making its way into the intersegmental septum. The primary nephridium already has a_ pre-septal funnel, while, of the succeeding secondary ones, most have a post- septal funnel, there are some with a pre-septal funnel. It is possible that the origimal ‘ funnel-cell’ has something to do with the pre-septal or post-septal position of the funnel. If the funnel is formed from the original ‘ funnel-cell’ or a derivative of it, we get a pre-septal funnel ; but if the funnel is formed from one of the other cells which takes up the character of the ‘ funnel-cell ’, a post-septal funnel results. As regards the secondary integumental nephridia, their separation from the primary nephridia and from one another has gone much deeper and further than that shown in the developing nephridia of Megascolides (17) and Mah- benus (8). They are coeval in origin, but not connected im any part of their development. Both the buccal cavity and the pharynx, forming that por- tion of the alimentary canal which lies in the first four segments of the body, belong to the stomodaeum ; and since the latter is morphologically external, the pharyngeal nephridia opening into the stomodaeum may be said to open on the ectoderm. But the facts of development do not help us in understanding the possible course of evolution of these nephridia. We cannot, for example, assume that these nephridia are really integu- mentary, but have come to occupy their present position and relationship on account of the anterior portion of the worm bemg formed into an ‘introvert’ or a stomodaeum. This assumption does not work in giving us the adult structure DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 95 from the hypothetical original condition of these nephridia. The development of secondary pharyngeal nephridia as distal buds on the pharyngeal ducts is remarkable. 11. MareriaL AND TECHNIQUE. The material for this work consisted of cocoons of Phere- tima contaming embryos of various ages. In the Kew Gardens, where I got my supply of these worms from, they are found along with two or three other genera in the soil of the Lily House, and, if the cocoons were collected from that soil, it would be difficult to distinguish the cocoons of one genus from another. Accordingly, to be quite sure of the specific identity of my material, I tried the isolation and culture method, which I briefly describe below. I took common garden soil and sterilized it for two or three hours to kill all organisms in it, specially the cocoons of other worms, eggs of insects, &c. This earth was mixed with sand, and finally I added to this mixture a quantity of decaying leaves which had also been previously sterilized. Four garden pots, the bottom holes of which were closed with corks, were filled with the sterilized soil and about fifteen worms of this species of Pheretima were kept in each pot. These pots were kept in a hot-house with an average temperature of 60° I.; the earth in the pots was kept moist and sterilized decaying leaves were added from time to time. After the first two months I was able to get cocoons in this way almost throughout the year, and was always sure that the cocoons were of Phere- tima alone and of no other worm. Various methods were tried to sift out the cocoons from the earth, but the least troublesome, and therefore the best, is to put a heap of earth in a fine sieve and to stir the earth while keeping the sieve in a bucket of water. The earth passes through the sieve and settles down at the bottom of the bucket, while the cocoons are left in the sieve along with the large pebbles and pieces of stone. The cocoons can be easily found and picked up with a wet paint-brush. 96 KARM NARAYAN BAHL Since the cocoons of Pheretima have a transparent shell, the embryo inside a cocoon can be easily seen under a binocular microscope by transmitted light. It is therefore easy to know the age and size of the embryo before opening the cocoon. Since in a given lot of earth the cocoons are of all ages, we can at once select an embryo of the desired age, provided we have a large number of cocoons. As a rule there is only one embryo in each cocoon, but we sometimes meet with two or even three. The cocoons are opened in salt solution by means of a pair of sharp needles under the bimocular microscope. Very early stages (blastulae and gastrulae) were mounted whole in clove oil after staining with paracarmine. ‘I'wo pieces of hair were placed below the coverglass, which enabled the rounded embryo to be rolled under the coverslip in order that it could be examined from all sides. Embryos of about 4 to 6mm. in length were used both for whole mounts and sections for the study of mtegumentary nephridia. ‘The embryos while in salt solution were always nar- cotized by ether and fixed either with Bouin’s fluid or corrosive- acetic or Petrunkewitsch. The latter solutions were found preferable to Bouin, since this fluid hardens the food-yolk very much and makes it brittle for section-cutting. Serial longitudinal sections of a few embryos of suitable age enables one to follow the development of integumentary nephridia fairly completely. A series of transverse sections, 5 » in thickness, of an embryo about 800 in length, was very useful in following the very early stages of development, e.g. the teloblasts and their development. For the later stages of development of primary nephridia, as represented in Text-fig. 4, embryos fixed in Bouin or corrosive are slit open by means of a sharp needle along the mid-dorsal line, the roll of albuminous material filling the gut is removed, and with a little care the endoderm itself is removed. What is left is the body-wall with the nephridia attached to it. This is stained with paracarmine, flattened out, and mounted whole. Older embryos, about 10 mm. or more in length, are fixed DEVELOPMENT OF NEPHRIDIA OF PHERETIMA a7 in the same way and opened by a mid-ventral incision ; the albumen is removed but not the endoderm: these, when mounted flat, show the septal nephridia in all stages of develop- ment, as represented in figs. 14 and 15. Since they he in two rows, one on each side of the dorsal vessel, it is best to open the embryos from the ventral side and look for the nephridia on each side of the dorsal vessel. In order to avoid any displace- ment of the septal nephridia it is best to leave the endoderm on; but it is necessary to brush carefully the inside of the embryo with a fine camel-hair brush, so that all yolk-material sticking to the inside of the gut is removed and the prepara- tion rendered quite transparent to show the septal nephridia to the best advantage. The most difficult part of the task, however, was to trace the initial stages of development of the septal nephridia—to find out whether the ‘ rudiment’ (‘Anlage’) of the septal nephridia arose as a multiplication of one or more cells belonging to the walls of the adjoining coelomic sacs (the septa), or it was formed by a group of cells between the two contiguous sheets of the intersegmental septa. For this purpose it was necessary to have serial longitudinal sections of the dorsal and dorso- lateral parts of an embryo of suitable age, i.e. one in which one could expect to find the septal nephridia in very early stages of development. The difficulty in getting such sections arises from the fact that the gut in embryos of this age is so enor- mously distended with food-yolk as to squeeze out of existence altogether the coelomic cavity between the body-wall and the gut dorsally. Consequently it becomes impossible to distin- guish, in sections of such embryos, the septa and the nephridial masses on them. Many series of sections were cut of embryos which had been flattened out, after bemg cut open ventrally and the food-yolk removed. In these series, although the coelom could be distinguished in some of the segments, it was obliterated in others, and no accurate and reliable observations could be made. In many of the embryos part of the posterior end was cut off before fixation to allow the food-yolk to ooze out and thus let the wall of the gut shrmk away from the 98 KARM NARAYAN BAHL body-wall, restoring the coelomic cavity on the dorsal side. I did not meet with much success even by this method. At last I was lucky in finding two embryos, one of which was just of the right age (about 6mm.) and in both of which the gut was narrow and the coelomic cavity all round very spacious. A complete series of longitudinal sections of one of them gave me all the stages of development of the septal nephridia, and I was able to establish, firstly, that the septal nephridia develop between the two peritoneal sheets (inter-peritoneal), and, secondly, that they could be traced to their final septal position from their first place of origin in the primary nephric row in the body-wall. 12. SUMMARY. 1. The three kinds of nephridia—integumentary, septal, and pharyngeal—appear at successive stages of development of the embryo; the integumentary preceding the septal and pharyngeal, both of which develop simultaneously. 2. All the three kinds can be traced back to the original row of nephridial cells of ectodermal origm. ‘Thus all the- different nephridia are ultimately derived from one common origin. 3. The primary pair of integumentary nephridia are the first to appear from a ‘retro-peritoneal’ group of cells. The rudiments lack the ‘ funnel-cell ’, and consequently a‘ coelomic ’ funnel is never developed in these nephridia. They open to the exterior on the body-wall. 4, These primary integumentary nephridia do not appear in the same position in successive segments of the embryo, but are irregularly distributed all over the body-wall. 5. The septal primary nephridia can be traced back to a group of nephridial cells, including the ‘ funnel-cell’, which make their way into each septum between its two adjoiing peri- toneal lamellae. ; 6. The primary septal nephridia have always a well-developed pre-septal funnel, and appear along a straight lme on both sides of the dorsal vessel. They appear after the primary DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 99 integumentary pair has reached a fairly advanced stage of development. 7. The secondary nephridia of both the integumentary and septal types are not budded off from the primary nephridia, but the rudiments of all have a common origin and separate early. They resemble the primaries in every respect, except that in the case of the septal secondaries the funnel is either pre-septal or post-septal. 8. The terminal ducts of the primary septal nephridia form the dorsal portions of the septal excretory canals on the septa, and the canals of both sides form the supra-intestinal duct on meeting the mid-dorsal line above the gut. The segmental ductules establishing a communication between the supra- intestinal duct and the lumen of the gut appear soon after the formation of the supra-intestinal ducts. 9. The primary pharyngeal nephridia of the fourth, fifth, and sixth segments develop from a ‘retro-peritoneal’ group of cells like the integumentary ones, and have long ducts reaching the wall of the pharynx. Secondary nephridia are formed as successive buds on the ducts, anterior to the primary nephridia. 10. The possible phylogenetic stages in the evolution of the ‘ enteronephric ’ type of nephridia are as follows: (1) the severance of the connexion between the septal funnel and the body of the nephridium ; (2) migration of the severed portion, i.e. the ‘funnel-cell’, together with some other nephridial cells from a ventral to a lateral position in the embryo ; (8) the growth of this severed portion into a septal nephridium and the acquisition by the latter of an opening into the gut ; (4) the elongation of the terminal ducts of all septal nephridia towards the mid-dorsal line (induced by the course of commissural vessels) and the formation of continuous supra-intestinal ducts. It is problematic whether the severance of the connexion between the funnel and the body of the nephridium took place before or after the connexion of the nephridium with the gut. 100 KARM NARAYAN BAHL 13. RererRENCES To LITERATURE. 1. Bahl, K. N.—‘‘ On a New Type of Nephridia found in Indian Earth- worms of the Genus Pheretima’”’, ‘ Quart. Journ. Mier. Sci.’, vol. 64, 1919. 2. —— ‘On the Blood-Vascular System of the Earthworm Phere- tima, and the Course of Circulation in Earthworms’”’, ibid., vol. 65, 1921. 3. Beddard, F, E.—‘ Researches into the Embryology of the Oligo- chaeta ”’, ibid., vol. 33, 1892. 4, ——‘ A Monograph of the Order Oligochaeta ’, Oxford, 1895. 5. —— ‘The Cambridge Natural History’, vol. ii (Earthworms and Leeches), 1910. 6. Bergh, R. S.—‘‘ Neue Beitrage zur Embryologie der Anneliden”, ‘ Zeit. fiir wiss. Zool.’, Bd. 1, 1890. 7. ——‘*‘Nochmals iiber die Entwicklung der Segmentalorgane ”’, ibid., vol. 66, 1899. 8. Bourne, A. G.—*‘ On certain Points in the Development and Anatomy of some Earthworms ”’, ‘ Quart. Journ. Micr. Sci.’, vol. 36, 1894. 9. Goodrich, E. 8.—‘‘ On the Coelom, Genital Ducts and Nephridia”’, ibid., vol. 37, 1895. 10. —— “On the Nephridia of the Polychaeta”, ibid., vol. 43, 1898. 11. Kowalewski, A.—** Embryologische Studien an Wiirmern und Arthro- poden ”’, ‘ Mém. de l’Acad. Imp. Se. St. Pétersbourg ’, ser. i, t. xiv, 1871. 12. MacBride, E. W.—‘ Text-book of Embryology ’, vol. i, 1914. 13. Meyer, E.—‘‘ Studien ttber den Kérperbau der Anneliden”, ‘IV: a. d. Zool. Stat. zu Neapel ’, viii, 1888. 14. Rosa, D.—‘‘Sui nephridii con sbocco intestinale comune del- P Allolobophora antipae ”, * Archivio Zool. Napoli’, vol. 3, 1906. 15. Staff, F.—‘‘ Organogenetische Untersuchungen iiber Criodrilus lacuum”’, ‘ Arb. aus dem Zool. Inst. Wien’, vol. 18, 1910. 16. Vejdovsky, F.— Entwicklungsgeschichtliche Untersuchungen *, Prag, 1888-92. 17. —— ‘Zur Entwicklungsgeschichte des Nephridial-Apparates von Megascolides australis’, * Arch. Mikr. Anat.’, Bd. xl, 1892. 18. Wilson, E. B.—‘* The Embryology of the Earthworm ’”’, ‘ Journ, of Morphology °*, vol. 3, 1889. DEVELOPMENT OF NEPHRIDIA OF PHERETIMA Tol EXPLANATION OF PLATES 5-7. Illustrating Dr. K. N. Bahl’s paper *“ On the Development of the ‘‘ Enteronephric ” type of Nephridial System found in Indian Earthworms of the genus Pheretima’. Fig. 1—Whole mount of a very young embryo (146y x 130,) seen in a ventro-lateral aspect, showing the mesodermal pole-cell M.p.c., the neural and nephridial teloblasts N.N.T., and the mesoderm band m.b. x 500. Fig. 2.—Longitudinal section of an embryo 245y in length, showing the relative position of the mesodermal pole-cell and one of the teloblasts. The ectodermal origin of the teloblast is clearly indicated. 1, the neural teloblast ; c.c., coelomic cavities in the mesoderm; arch., archenteric cavity ; end., endodermal cells with yolk-granules. » 500. Fig. 3.—Ventral portion of a transverse section from the posterior end of an embryo 300, long, showing the origin of the teloblasts in the ecto- derm. The teloblasts are still embedded in the ectodermal layer, but are sharply marked off from the definitive ectodermal cells. On the right are seen two teloblasts, the ventral one being the neural and the lateral one the nephridial. On the left are seen three cells still embedded in the ectoderm which lie in front of the teloblasts and have been budded off from them. x.t., neural teloblast ; np.t., nephridial teloblast ; c.t., cells budded off from teloblasts ; m.s., hollow mesodermal somites with coelomic cavities, c.c.; ect., ectodermal cell. x cir. 620. Fig. 4.—Ventral portion of the section just in front of the one shown in fig. 3, showing the gradual demarcation of the neural and nephridial cells from the definitive ectoderm. Jat.t., lateral teloblast ; 7n.c., neural cells ; nph.c., nephridial cells. » cir. 664. Fig. 5.—Ventral portion of a longitudinal section through the hind end of an embryo about 3 mm. long, showing the nephridial string of cells breaking into the cell-masses destined to give rise to the nephridia. M.P.C., position of mesodermal pole-cell; mph.t., nephridial teloblast ; nph.l., nephridial layer of cells; m.l., layer of mesoderm with coelomic cavities ; ect., ectoderm. 545. Fig. 6.—Transverse sections of an embryo 300, long, from which figs. 3 and 4 have been drawn, showing the ‘funnel-cells’ in section between the two adjoining coelomic cavities, i.e. in the septa. .c., nerve- cord ; nph.c., groups of nephridial cells ; f.c., funnel-cells ; ect., ectoderm. x 400. Fig. 7.—Portion of a longitudinal section of an embryo, showing the nephridial groups of cells pushing into the intersegmental septa, Other letters as before. 640. 102 KARM NARAYAN BAHL Fig. 8.—Five consecutive longitudinal sections of a series taken from the hinder portion of an embryo about 4mm. long, showing the develop- ment of primary integumentary nephridia, and their relations with the intersegmental septa. nph., nephridia (integumentary) ; nph.c., nephridial cells pushing their way into the septa; 7.s., intersegmental septum ; c.c., coelomic cavity ; ect., ectoderm. Fig. 9.—A portion of a longitudinal section of the same embryo as in fig. 8, showing the developing integumentary nephridia. Letters as in fig. 8. 1,200. Fig. 10.—Portion of a longitudinal section of an embryo, showing the semblance of a ‘funnel’, the only series in which a ‘funnel’ is seen. f., pseudo-funnel. 1,160. Fig. 11 (A-r).—A series of diagrams showing the developing stages of an integumentary nephridium, taken from whole mounts of embryos with endoderm and albumen removed. Nephridia lie posterior to the septa. i.s., intersegmental septa; s.l., short straight lobe of the nephridium ; t.1., the twisted loop. x cir. 630. Fig. 12 (aA—n).—A series of diagrams from whole mounts showing the exact mode of development of a primary septal nephridium. /., the pre-septal funnel; s.1., straight lobe; ¢.1., twisted loop; t.d., terminal dust ; c.v., commissural vessel; 7.s., intersegmental septum. 530. Fig. 13.—Longitudinal sections of the posterior end of an embryo about 6 mm. in length, showing how the ‘ funnel-cell’ travels from its original position in the nephridial layer of the body-wall to its final position on the septa on each side of the dorsal vessel. 1-6 are sections of septa from the same series showing the change in position. c.v., commissural vessel; (f.c., funnel-cell; v.lat.; ventro-lateral branches of the ventral vessel ; c.c, coelomic cavities. x cir. 1,250. Fig. 14 (a—p).—A series of septa in longitudinal sections showing the early development of a septal nephridium ventral and close to the com- missural vessel and between the two sheets of mesoderm forming a septum. b.w., body-wall; c.v., commissural vessel ; nph.r., nephridial rudiment. x cir. 1,250. Fig. 15.—Two intersegmental septa with nephridia on them from a whole mount of an embryo about to hatch out of the cocoon, showing the develop- ing secondary nephridia of both kinds, one with pre-septal funnel and the other with post-septal funnel. psn., primary septal nephridium ; sec.s.n., secondary septal nephridium with post-septal funnel ; sec.s.n’., secondary septal nephridium with a pre-septal funnel. x 450. Fig. 16.—Part of a transverse section of an embryo, showing a rudiment of a secondary septal nephridium. c.v., commissural vessel; b.w., body- wall; n.c., nerve-cord; d.v., dorsal vessel; other letters as in fig. 15. x 330. Fig. 17.—Portions of two consecutive sections of an embryo, showing DEVELOPMENT OF NEPHRIDIA OF PHERETIMA 1038 the development of secondary integumentary nephridia. sec.7.n., secondary integumentary nephridia; ep., epidermis; circ.m., layer of circular muscle-fibres ; Jong.m., layer of longitudinal muscles; s.s., setal sacs. « 440, Fig. 18.—Longitudinal section of the anterior portion of an embryo about 4:5 mm. in length reconstructed from serial sections, showing the early development of pharyngeal nephridia and their ducts (both being solid at this stage). ph,, ph,, ph,, pharyngeal nephridia of the fourth, fifth, and sixth segments with their respective ducts; pha., pharynx ; i.g., integumentary nephridia of the seventh segment ; 6.c., buccal cavity ; c.g., cerebral ganglion; ph.gl., pharyngeal gland-cell. Fig. 19.—Longitudinal section of the anterior portion of an embryo about 6 mm. in length, showing the primary pharyngeal nephridia and the buds of the secondary ones arising on the ducts. d.v., dorsal vessel ; lat.oe.v., lateral oesophageal vessel. Fig. 20.—Three embryos of typical ages of Pheretima rodri- censis, shown in natural size. A is an embryo about 4mm. in length, from sections of which figs. 5, 8, 9, and 22 are taken. B is an embryo about 6 mm. in length, from sections of which figs. 17, 18, and 23 are taken. Whole mounts shown in figs. 11, 12, 13, 14, and parts of 15 are taken from embryos of about this age. C is an embryo about to hatch out of the cocoon about 17 mm. in length. Figs. 16, 19, 20, and 21 are taken from whole mounts or sections of embryos of this age. Nat. size. uve ad Te. rit arr’ vel TOs RMA? FI Weasel was y tie it iehiali) «00 eile VAT gnes «Ve Tate Wr Wt A ‘ Lilahay) tm, eu Sey hori” i aie uf [a rier oe ; ; wh Lik ee 7 erie ie? Oe f ‘ier We 1 be “aie sige Wf the, AD ae mn iva : ma 2 : ia’ n iw ne iia | ne 1 3 e ——Te Ss \ Bahl del. Quant. Fun Mir S6.V0H. 68 NS Pt. 8. Huth, London. Ch QO <> RRO AYouolo ooo SIO 7 Bahl del, Quart FourndMior' SA.VA. O6.NS§. Ft. 6 7290. AVIV EVAL TIT tit 7? slay Ae pi ae Huth, London. Cant’ vv" - Tae y Poa ee Te ~~ if 7 7 Th - ; i : Behl del. SEC. Quart, FurnMior Sti. VO. 66,1) ti 1 ! if ! r A Huth, London The Occurrence of Situs inversus among artificially-reared Echinoid Larvae. By Hiroshi Ohshima, Assistant Professor in the Department of Agriculture, Kyushiu Imperial University, Fukuoka, Japan. With 3 Text-figures. CoNTENTS. PAGE 1. INTRODUCTION. ; . 105 2. DESCRIPTIONS OF THE eee WITH Teen rma : LOT 3. RESULTS OF THE EXPERIMENTS E ’ : é , . 113 4. CHANGES WHICH MAY POSSIBLY HAVE TAKEN PLACE DURING EARLIER STAGES s : ; i 3 a lily 5. VARIATIONS AMONG DOUBLE- Tianadanil LARVAE AND OTHER ABNORMALITIES } , ; AAC) 6. CONSIDERATIONS ON THE Gneina AND ST aTOnnnse CONCERNED AND THE FACTORS CONCERNED IN THEIR DEVELOPMENT . 128 7, PROBABLE MECHANISM WHEREBY ABNORMALITIES ARE PRO- DUCED . : : 5 UBD 8. EXTERNAL wacions AS Ca USES OF Wee eee : d yll43 9. SUMMARY AND CONCLUSION . ; : j é : a 45 10. LireRATURE CITED : s : j : : : . 146 11. APPENDIX ; : : : E ; 3 : 5 4s 1. INTRODUCTION. A REMARKABLE case, where a hydrocoele and its associated structure had developed only on the right side instead of on the left side of the body, as in normal specimens, came under my notice among the artificially reared larvae of Echinus miliaris. 106 HIROSHI OHSHIMA Cases of situs inversus viscerum are not very rare in nature, and are frequently met with under artificial conditions. Spemann (29, pp. 400-14), in his most interesting experi- mental studies on Triton larvae, has made an exhaustive survey on cases of situs inversus. According to him (p. 401) the cases may be classified into two categories, though the distinction between these two may not be clean-cut. The one comprises such cases where an ‘ inverting’ factor affects an individual very early in its ontogeny, it may be even before fertilization, so that the ‘ microstructure’ of the egg under- goes a change at once and completely. Those Gasteropods with reversed spiral belong to this category. Conklin (3, p. 585) suggested as its cause the reversal of the polarity in the egg. To the second belong those cases where that factor acts much later in the embryonic development, a little while previous to the time when any visible asymmetry of organization occurs. It affects only a single but decisive part, and in consequence of the abnormal development of that part all the other adjoiming organs will assume the inverse situs. ‘There are many interest- ing instances of this: thus, for example, a chick embryo heated on its left side (Dareste, Warynski and Fol), a Triton embryo with a portion of the medullary plate cut out and replaced in the inverted position (Spemann), an egg or embryo which has been constricted along its median plane partially or completely so as to give rise to either a double monster or twins (Spemann; compare Bateson, 2, p. 560, and Morrill, 18, p. 267), and two halves of embryos with different rate of growth grafted together (Spemann) can likewise produce the situs inversus. Cases of such partial situs inversus have also been interpreted in a most satisfactory manner, as has also the striking fact that generally the abnor- mality is exhibited by the right-hand members of double monsters of Triton (and trout) and by the right-hand member of twin Triton larvae. Turning now to the case of the reversed Ee hinus larvae, I have tried to propose tentatively an interpretation. This SITUS INVERSUS IN ECHINOIDS 107 case is, it seems to me, more or less related to, but distinct in some respects from, the above-mentioned second category (see p. 141). The idea came to my mind after the experiments had come to an end, and it needs further test with special reference to this question. The experiments were made during the early summer of this year (1920) in the Zoological Department, Imperial College of Science and Technology, London. It is my pleasant duty to tender my hearty thanks to Professor E. W. MacBride for his kind supervision and unceasing encouragement through- out the time during which the work has been carried out. The writing of the manuscript was done in the Natural History Department of the British Museum. My cordial sratitude is also due to Sir Sidney F. Harmer, Director of the Department, for his kind permission to work there and to use the library. 2. DESCRIPTIONS OF THE LARVAE WITH INVERSE SITUS. It must at the outset be stated with regret that the descrip- tions of internal structures as here given are founded on a very few specimens which I could preserve and section. As will be seen in the table (p. 115) the total number of reversed larvae I found was more than 150, but with the hope of getting as many metamorphosed young as possible I did not kill and preserve many of them. The observation on the early stage when the right hydrocoele makes its appearance, i.e. the earliest visible sign of the abnormality, is also lacking. About half a dozen metamorphosed young were obtained, but all the rest died off gradually without affording me any oppor- tunity of following the internal changes which had taken place. External Characters.—KHight larvae with the inverse situs were first found on May 31, when they were eleven days old. The ‘larval’ body was quite normal both in size and shape: two pairs of larval arms, post-oral and antero-lateral, both symmetrical and fairly long; postero-dorsal arms still very short ; the posterior part of the body-rod beginning to degenerate, with its club-shaped end separated from the rest No. 261 I . 108 HIROSHI OHSHIMA and lying near the hind end of the body. Both the ventral and dorsal epaulettes were already separated from the ciliary bands, the anterior transverse part of the latter showing a peculiar twist which indicated the future position of the paired pre-oral arms. A hydrocoele, stone-canal, and amniotic invagination were all situated on the right side, whilst no such organs were found on the left side. No special attention was paid to such a slight asymmetrical distortion in shape of the stomach as was often noticed by Runnstrém in some abnormal larvae (24, 25). Similar larvae were found later to be fairly numerous, and were transferred to a separate jar where they were allowed to develop further. There was found no difference in the rate of growth between normal larvae and these abnormal ones. When fully grown (T'ext-fig. 1) the abnormal larva possessed four pairs of well-developed arms, a large echinus-rudiment (rd) on the right side, from which five primary tentacles often pro- truded and moved actively. Whether a pair of pedicellariae really appeared on the left side as is the case with Strongylo- centrotus (see p. 136) I cannot assert at present, though it seems to me to be highly probable. As to those paired ceal- careous structures which appeared on the left side, as seen in the text-figure (sp,), I am almost certain that they were groups of spines.t The unpaired spine which should appear in normal cases at the hind end, a little to the right of the median line, was here found shifted to the left side (sp). No less than half a dozen of such abnormal larvae passed metamorphosis when a month old. As to the external feature of these young sea-urchins one can find no difference from 1 While dealing with the living larvae I thought without the slightest doubt that the paired calcareous structures always found on the left side were really pedicellariae. Text-fig. 1, which is the only drawing made of this stage from life and the only evidence now available, shows that they are situated inside the loop of the ciliary band. This position coincides precisely with that of the groups of spines as described by Runnstrém (27, pp. 21-2, figs. 21-3). In this particular specimen at least there were present no true pedicellariae (see p. 138). SITUS INVERSUS IN ECHINOIDS 109 TExtT-FIc. 1. SP; Full-grown larva of Echinus miliaris with inverse situs. Ventral view. x75. an, anus ; cd, constriction between larval oesophagus and stomach ; ep, ventral and dorsal epaulettes ; mo, larval mouth ; py, constric- tion between larval stomach and intestine ; rd, echinus-rudiment formed on the right side; sp,, rudiment of posterior unpaired spine situated a little on the left to the median line; sp,, a pair of groups of spines formed on the left side. === faves (a4 "ei I 2 110 HIROSHI] OHSHIMA normal young. All the sets of primary unpaired and first- paired tentacles, pedicellariae, pointed and square-ended spines, were formed precisely as in the young which had metamorphosed from normal larvae. It was hoped that they would develop further to the stage when the asymmetrical arrangement of the organs, above all the peculiar coil of the intestine, would be more pronounced. Unfortunately, however, they were all lost after ten days, probably being destroyed by a tiny Gasteropod which had been carelessly put into the jar together with some Corallinae. Internal Structures.—so far as the internal anatomy of the larva is concerned, the following short account is all we can learn. The transverse sections of the larva (Text- fig. 2) are exactly the mirror-images of those of normal larvae, so that one cannot distinguish them from sections of a normal larva mounted upside down. An eleven-day-old larva has the pore-canal (pe) still distinctly opening on the right side of the mid-dorsal line (dp), a madreporic vesicle (mv) lying close to the canal, situated at its median side but without any com- munication whatever with it. The canal then leads to the thin-walled axial sinus (ax) which lies close to the oesophagus (oe). The stone-canal (st) connects the axial sinus and the hydrocoele just as in the normal case. The hydrocoele (hy) situated on the right side of the stomach (qg) has just begun to produce lobes, and an amniotic invagination (am) has already appeared. No traces of hydrocoele, stone- and pore- canals were found on the left side. From want of material it is not known on which side of the posterior coelom the genital stolon would be formed. Thus, with doubtful exception of the pedicellariae and genital stolon, the internal organs as well as the external characters showed perfectly the inverse situs im every detail, so far as I could examine. With regard to the pedicellariae and genital stolon I refrain from expressing a definite opmion. We may expect to find some aberrant types as might be suggested from further descriptions of double-hydrocoele larvae. Similar cases previously known.—So far as I am SITUS INVERSUS IN ECHINOIDS 111 aware, the similar cases among Echinoid larvae have only been recorded twice by Runnstrém. He described two such larvae of Strongylocentrotus lividus reared at Monaco. TEXT-FIG. 2. dp mv pe Transverse sections of an eleven-day-old reversed larva of Echinus Miliaris. x300. am, amniotic invagination; ax, axial sinus; dp, dorsal pore ; ep, ciliary epaulette; g, stomach; hy, hydrocoele ; in, intes- tine ; mv, madreporic vesicle ; oe, oesophagus ; pc, pore-canal ; rpc, right posterior coelom ; sé, stone-canal. [Case A.] Runnstrém, 1912 (23), pp. 2-3, ‘no. 1’; 1918 (26), pp. 419-20. Left : no hydrocoele developed, but instead of it a ventral primary pedicellaria was formed. 119 HIROSHI OHSHIMA Right: echinus-rudiment well developed. Dorsal pore remaining at its original position on the right of the mid-dorsal line, no shifting towards the latter taking place. [Case B.] Runnstrom, 1912 (28), pp. 7-10, ‘no. 5’; 1918 (26), pp. 420-4, Taf. xiv, figs. 12-16. Left: hydrocoele not formed, anterior coclom remaining rudimentary. Ammiotic invagination formed as only a shallow depression, and afterwards disappearing. Pedicellariae present, the dorsal one being already formed while the ventral one was indicated by accumulated cells. Right : anterior coelom consisted of two portions, one being the axial sinus communicating with stone-canal and the other representing the madreporic vesicle (* pulsating organ’), which displayed infrequent and irregular pulsation. The dorsal pore was not at first formed, though an ectodermal groove indicated it. After-some days a pore opened anew, and the madreporic vesicle began to pulsate more frequently and regularly than before. The stone-canal became split into two parts, one short and still communicating with the hydrocoele whilst the other was longer and opened freely into the coelom.? These two then degenerated and a new stone-canal appeared, so that the hydrocoele regained its communication with the exterior. ‘The hydrocoele produced two diverticula, one of the ordinary size and the other much larger. The amniotic invagina- tion was not formed, but instead of it there were two ectodermal pits. These the author at first (28) interpreted as rudiments of the primary pedicellariae, but afterwards (26) corrected his former view and called them ‘spime invaginations’ (see p. 138). In other classes of Echinoderms, Auriculariae with the hydrocoele on the right side only were noticed by Miller many years ago (19, pp. 101, 109, Taf. v, fig. 1). From com- munications with Dr. Th. Mortensen I have learnt that he found among the larvae of Ophionotus hexactis two specimens which had a right hydrocoele only. I here 1*Cavité générale’ and ‘ K6rperhdhle’ in the original descriptions. Under these terms the posterior coelom is probably meant. SITUS INVERSUS IN ECHINOIDS 113 express my thanks to Dr. Mortensen for the kind permis- sion to note this discovery, which he has not yet published. 3. Resutts or THE EXPERIMENTS. The purpose of our experiments made under the direction of Professor Mac Bride was to carry out a further test of the influence of high salinity on the production of double hydrocoeles (15, pp. 334-7, 341). Fresh specimens of Echinus miliaris were sent from Plymouth, ripe males and females were then selected from among them, and the eges were fertilized. For detailed descriptions of the method we adopted I refer to MacBride’s paper (15, pp. 326-9). Only a few details need be added here (see table on p. 115). ‘Outside water’ of Plymouth (1, p. 372) was always used in starting the culture, viz. the eggs were fertilized in it and then kept for a day in finger-bowls filled with clean ‘ outside water ’ (‘ finger-bowl period’). One-day-old larvae with pyramidal body and a pair of rudimentary post-oral arms were then transferred to Breffit jars, which had been filled with ‘ outside water ’ supplied with some Nitzschia (° Breffit-jar period ’). Then some of them were treated for several days with ‘ hyper- tonic’ sea-water, which had been synthetically prepared according to Allen and Nelson (1, pp. 369-71), and the salinity increased roughly to 3-7 per cent. while others were left untreated as controls. When about a fortnight old, the larvae were put into plunger jars, which had been filled with synthetic sea-water of normal salinity mixed with a small quantity of ‘outside water’ (‘ plunger-jar period’). The results of five more or less successful cultures are here shown in the table. They were offsprings of three different parents : culture nos. 1 and 4 belonging to the first, nos. 6 and 9 to the second, and no. 11 to the third. The larvae with the inverse situs were first discovered among no. 4, on May 31. ‘Those fifty-four abnormal larvae of this culture were then kept separate in a Breffit jar. On June 19, thirty days after fertiliza- tion, some few among the normal ones of this culture were found just metamorphosed into tiny young sea-urchins, while 114 HIROSHI OHSHIMA one of those fifty-four abnormal larvae also metamorphosed on the same day. Within ten days afterwards 127 normal larvae and six abnormal ones had metamorphosed to young sea-urchins from this culture. MacBride (11, p. 294) got the larvae of Ecechinus esculentus to metamorphose in forty-two to fifty days after fertilization, while Allen and Nelson (1, pp. 420-1) found the earliest metamorphosed young of EH. acutus forty-two days after fertilization, of K. esculentus, forty-eight to sixty-eight days, and of K. miliaris, thirty-eight days. As compared with these records of regular sea-urchins our case was much quicker in development. On the other hand, the culture no. 11 and others from the same parents suffered from want of food seriously after the first week of their development, and when examined on September 3 they were, though seventy-six days old, all very far from metamorphosis, the ‘ larval’ body fully developed, but the echinus-rudiment, if present, being very small. The culture no. 6, for some unknown cause, gave poor results. Most of the larvae died off very quickly, and the survivors showed various irregularities in shape. The food supply was generally good during the first week or so, but afterwards in most cases it could not be continuous, and became unavoidably very irregular, owing to the unsuccess- ful culture of Nitazsehia. Now, from among the ‘treated’ larvae (nos. 4 and 9), which number 784 in all, there were found 88 inverse (11-2 per cent.) and 6 doubles (0-8 per cent.). In ‘ controls’ (nos. 1, 6, and 11), on the other hand, from among 646 larvae, there appeared 69 inverse (10-7 per cent.) and 13 doubles (2 per cent.). This shows clearly that there is no noticeable difference in the rate of producing abnormalities between these two differently treated lots. We shall discuss this question later on (p. 143). The results of Professor MacBride’s experiments of producing the double hydrocoele (15) may here be cited briefly. 1914 (pp. 334-5). The larvae three or four days old were treated for ten or eleven days with ‘ hypertonic’ sea-water ‘p ‘OU TOF UIAIS 127) WIT) qUareYIp YONUI Utes you pp asezUe010d 943 G pPUe ‘9 ‘[ ‘SOU UT “pazUMOD you seA\ oIN4[ND YOR UI VAIL] YONS Jo Joqunu yorxs oy} F ‘ou ul 4daoxy] ,; *sisoydiowejzom passed woyy Jo amos [yun srel 4WFeIg 0} porl1oysuRIy Io ‘arofjeq sev aef sues 94} OUT YOR ynd Jeq4Ie Sulaq ‘laqqyany satje ydey orem Ady], ‘suumyoo Surutofpe oy] ul se popioo9er pue pouTMexXS o10M BVAIL] OY} USM 9UIT} OY} SUBOT ported sTyy Jo pus eT, 7 | | ; *, TayeM apisyno, | (%F:3) | (%G&81) | ‘(sAep ZZ) | urqdoy ys ‘(sfep | *(Aep T) F | G | 99T € ‘ydog-¢gz ounp | g)ezoung-ogeunf | OZ euNf-GT 9UNL | *, oIZMOD, II “Aueur ApIIe 7 | | 146 Se-2e Ae | | ‘sXep aay 10j 1070 | | | | | -was , oTuoqred Aq , he OSal es One (Co0tL) S| | ‘(Aep [) |q714 poqeery “(sAep “MOT | 9 | FE | Oct | ¢g ounf—f oump | ET) F ounp-z Avy ‘0d | *, peveery , 6 | | | | * 1d}@M apisyno , | (%oL-9) (%oe-8) | | ‘(sXep g) | urqdoy [1198 ‘(shep ‘(Avp 1) “MOT 8 I | 08 | L ounp-F ounp | ET) pF ounp—z Av ao Avw-1z Ae | *Jomuog, | 9g | | ‘Le Av-ee Av ‘sAep dal 10J 1092 | | -eas , otuoqrod AY , | | (%8-F) | | (%&91) | q7uM pozvory‘(shep | | 9T 0 #g | FEE | ‘od=s« TT) 1 eung-71z Ae ‘od | *, peyvory , ee | * 1oyeM opIsyno , | | | (%z-0) (%z-0r) | | (step g) — urqdoy [ys ‘(sep (Sep 1) | “MOT | T OF | OLF L ounp-T oung [T) T oung-Tg Avy Tz Aey-0c Av *, foryUOD , I ie | *(, 4aqom -(, wagon ‘apaooouphiy =» “a]20004phyY sens asuaaun ‘pourmnxa | Piso, fo hyywonb ‘(unbag eryos apisyno , ur day “ainyno fore es AU. yun anasny aan) fo ee “ZAIN U0 Buypoaf’) ‘wouneymiaf foo fo sume ee anaLD] Yn ena fy aquny | soqunu mo, | 2mM-mes ompyhis | Honig wnl-nffoug | ayn yw buruurbaq) | jouorsraosg 24" 2fo wquny § fo wqung fo buysisuos sayon) powag jnoq-iaburg | | | 1 powagd iol-sabunyg : Asal “SINGWIYGdIXY JO SLIOSAaY AHL ONIMOHS ATAV], 116 HIROSHI OHSHIMA which had been prepared by evaporating. A right hydrocoele appeared but the amniotic invagination failed to appear, and the larvae refused to develop further. 1915 (p. 335). From among the larvae treated as above the most promising ones were isolated and fed on abundant Nitzschia. One larva produced a five-lobed hydrocoele on the right side. 1916 (p. 335). In both groups, those kept throughout in ‘hypertonic’ sea-water and those put back in normal sea- water, after being treated for one to three days, were found some larvae with an unmistakable right hydrocoele provided with tive tentacles. 1917 (pp. 835-7). ‘ Hypertonic ’ sea-water was prepared this time by adding common salt to sea-water. The fourth-day larvae were transferred to ‘ hypertonic * sea-water and allowed to remain in it for six days, after which period they were again put back in normal sea-water. The larvae with double hydrocoeles were about 2 per cent. in one jar, while at least 5 per cent. were in the other. Amongst hundreds of controls there was found only one specimen which had a double hydrocoele. The result obtamed in 1919 was so similar to that of the fore- going year that he thought it unnecessary to publish anything about it. Before further discussimg the causes and processes of forma- tion of the abnormalities, let us stop for a moment to consider some questions which may naturally arise in the reader’s mind. These are the questions of fundamental importance: (1) Is_ not the writer’s discovery due to an error of observation ? (2) Is not the occurrence of such abnormal larvae also common in nature for this particular species—at least in a particular season and at a particular place? (8) Is not the scantmess of records due to negligence on the part of previous observers ? (4) Is not the so-called ‘ abnormal’ condition hereditary ? (1) It is rather incredibly frequent to find that even careful observers make an error in the use of the so-called endless screw of the fine adjustment of some microscopes so as to confound the upper surface of the object with the under SITUS INVERSUS IN ECHINOIDS U7 surface, for instance, with the result that a minute spiral structure may be taken as turned in a wrong direction. In my case it will be quite sufficient to state that as the larvae were fairly large objects under the microscope, I used to focus by means of the coarse adjustment while examining them with respect to the symmetry relations. (2) It is now impossible to compare our culture with the larvae belonging to the same species which might have been found in plankton near Plymouth in the early summer of the same year (1920). One may suppose that if quite a number of naturally-developed larvae were examined carefully there might also be found some such abnormal forms. I think one may safely say, however, that at least the occurrence of this abnormality in so high a percentage as more than 10 per cent. is really due to artificial conditions. (3) In view of the fact that in our cultures such larvae with inverse situs were eight times as numerous as the doubles (157: 19), I cannot help doubting that the previous workers, who were fortunate enough to discover a few double-hydrocoele specimens from among hundreds of larvae, would have over- looked those inverse forms which might have been more frequent. It is very desirable to know if situs inversus occurs also fairly frequently in other species of sea-urchins when artificially reared. (4) As stated above, the five lots of cultures shown in the table were obtained from three different parents. It is highly improbable that such a remarkable case, if inheritable, was found in at least three individuals out of seventy sea-urchins (more than 4 per cent.) which had been sent from Plymouth. From all these considerations I am driven to conclude that the occurrence of the abnormality is true, and can even be fairly frequent among artificially-reared larvae. 4, CHANGES WHICH MAY POSSIBLY HAVE TAKEN PLACE DURING EARLIER STAGES. One of the most remarkable and well-known cases of situs inversus among animals is that of the snails with sinistral 118 HIROSHI OHSHIMA shells. In some genera and species it is a normal character, while in others it is regarded as abnormal. As is well known, the sign of the reversal goes as far back as the segmenting egg, which shows its spiral cleavage in the direction contrary to that found in the eggs which will give rise to normal dextral snails. Conklin (8, p. 585) tried to interpret the phenomenon by assuming the reversal of the polarity in the egg, which change might have taken place in its very early stage. This hypothesis, though still lacking any satisfactory experimental evidence, is very simple and admirable ; and besides this we have as yet no other explanation. There is no reason to deny that a state similar to that occurring among sinistral Gasteropods may occur also among Kchinoderms. But can we not find in our cases of E¢hinus larvae any other interpretation which is more plausible and more probable than this ? The Kchinoderm egg has been known to be ‘ equipotent ’, or, in other words, the distribution of the organ-forming substances becomes established much later than in the eggs of most other groups. We owe to Runnstrém our know- ledge of this question. In his series of experiments with Strongylocentrotus lividus (24 pp. 533-44, Text- figs. 7 a, 10) he showed that in this species embryos developing from half-eggs assumed normal characters later than did similar embryos of Echinus microtuberculatus and Sphaerechinus granularis. The larva developed ‘probably’ from the right half of the egg of Strongylo- centrotus has its skeleton more strongly developed on the right side than on the left, and, moreover, the coelomic sac appeared only on the right side. Another of his experiments (28, pp. 471-3, Text-figs. 16 a, b) shows that when an early gastrula of Solaster sp. had been constricted along its median line, in the double monster so produced, no hydrocoele formed ; but a dorsal pore appeared on its left side instead of on the right, forming a mirror-image of the dorsal pore of the left half. He thus confirms what Driesch observed in some few double monsters of Echinus microtuberculatus SITUS INVERSUS IN ECHINOIDS 119 in 1906 (4, p. 765). These results, considered in connexion with Spemann’s Triton twins and double monsters and also with Morrill’s double monsters of the trout referred to in a foregoing page (p. 106), lead us to expect that if successfully reared we might get an inverse larva from the right half of the egg in these Echinoderms also. Indeed, Spemann suggested this idea at the end of his work (29, p. 413). I may, however, only mention that our inverse larvae were all of normal size, and that there can be no doubt as to their having been developed from whole unseparated eggs. Gemmill’s information of several cases of twin larvae of Luidia sarsi (6) is not uninteresting in this respect. Eggs of early cleavage stages were sent from Plymouth to Glasgow, and, according to him, the long-continued shaking during the transportation might have caused the blastomeres to dissociate and such twins resulted. His figures, especially of those ‘ side-by-side ’ doubles (Pl. ii, fig. 18; Pl. ii, figs. 19, 21), clearly show that there is no perceptible difference in structure between the two halves developed from partially-separated blastomeres, nor is there any sign in the right half of assuming a mirror-image of the left. We cannot, however, help doubting whether separation really took place during the long-continued shaking. Judging from the haphazard relative positions of the halves and from apparent differences in age between them in some cases, one may naturally suspect that the conditions observed resulted from fusion of two individuals. It is desirable to learn how the left side of a member will affect the right side of the other in artificially-grafted larvae. Results of both chemical (Goldfarb) and mechanical (Runnstrém) grafting of the eges or embryos are unfortunately inadequate to solve the present problem. 5. VARIATIONS AMONG DouBLE-HyDROCOELE LARVAE AND OTHER ABNORMALITIES. Our attention will naturally turn to the double-hydrocoele larvae which appeared in cultures associated with the reversed larvae. To try to find if any relation exists between these 120 HIROSHI OHSHIMA two kinds of abnormalities we may first examine those known cases of double-hydrocoele and other abnormal larvae, and then consider the behaviour of individual organs and the interrelations to be found between them. I. Hydrocoeles formed on both sides. (a) Right hydrocoele and its associated structures more or less incomplete. [Case 1.] Strongylocentrotus lividus. Runn- strém, 1912 (28), pp. 3-5, ‘no. 2’; 1918 (26), pp. 417-18, Taf. xin, figs. 8a, b. Reared at Monaco. Left: anterior coelom large, divided into three regions : first, the ampulla to which the stone-canal opens ; second, the main body of the axial sinus extending transversely to the right and communicating with the third region, the madreporic vesicle. The last-named vesicle exhibited no pulsating move- ment. Pore-canal and dorsal pore lacking. Stone-canal and hydrocoele well developed, the latter produced into five lobes. Amniotic invagination deeper than normal. Right: anterior coelom smaller than that of the left, with pore-canal given out towards the epidermis, without, however, an opening to the exterior. Stone-canal showing a sign of degeneration, its anterior end beginning to be absorbed. Hydrocoele smaller than that of the left side. Amniotic invagination did not form on this side. Posterior coelom produced into an anterior process, which probably corresponds with genital stolon. [Case 2.] Strongylocentrotus lividus. Runn- strédm, 1912 (28), pp. 5-7, ‘no. 3’; 1918 (26), pp. 413-14, Taf. xin, fig. 4. Reared at Monaco. Left : anterior coelom large, consisting of two regions, one on the left, connected with stone-canal, the other on the right, corresponding with madreporic vesicle. The latter became later separated from the former, and was not seen pulsating. Pore-canal absent. Stone-canal and hydrocoele well developed, the latter produced into five lobes. Amniotic invagination formed but remaining totally undifferentiated, SITUS INVERSUS IN ECHINOIDS 121 Right : anterior coelom smaller than that of the left side. A vesicle was seen to be produced from it, which latter the author interpreted with some doubt as the hydrocoele. All these parts were seen beginning to degenerate. Amniotic invagina- tion formed very late, but soon disappeared. Pedicellariae not formed. The posterior coelom produced an anterior diverticulum, probably representing the genital stolon. [Case 3.] Strongylocentrotus lividus. v. Ubisch, 1913 (80), pp. 440-8, Text-fig. u. Reared at Naples (Gies- brecht). Left : axial sinus well developed with the pore-canal which opened externally by a dorsal pore. Madreporic vesicle (‘ dorsal sae ’) large, lying close to the axial sinus, but no com- munication between them existing at all. Echinus-rudiment fairly advanced. Genital stolon developed. Right: axial sinus smaller than that of the left side, pore-canal only represented by a knob from wall of the former, and a fibrous tissue connecting the epidermis with this knob. EKchinus-rudiment less advanced than that of the left side. [Case 4.] Strongylocentrotus lividus. Runn- strom, 1918 (26), pp. 418-19, Taf. xii, figs. 9,10. Reared at Monaco. Left : axial sinus, stone-canal, and hydrocoele all developed normally. Pore-canal and dorsal pore present, the latter later shifted its position towards the median line. Amniotic in- vagination formed. Right : axial sinus well developed, with pore-canal and dorsal pore. The latter hke its left fellow changed its position later towards the median line. Stone-canal formed later, its slight expanded posterior end representing the hydrocoele. No amniotic invagination formed. _ [Case 5.]| Echinus miliaris. MacBride, 1918 (15), p. 347, Pl. v, fig. 9. Reared in London. Left : axial sinus fused with that of the right side and com- municated with the exterior ‘through a single pore-canal, Kehinus-rudiment large. 122 HIROSHI OHSHIMA Right : Echinus-rudiment smaller than the left one. No pedicellariae formed. [Case 6.] Echinus miliaris. MacBride, 1918 (15), pp. 339, 348, 347, Pl. vi, fig. 11. Reared in London. Left: axial sinus provided with a pore-canal and dorsal pore. Lobed hydrocoele and amniotic invagination developed normally. Right: axial sinus with a pore-canal and dorsal pore. Hydrocoele smaller than that of the other side and no lobes were formed. Amniotic invagination absent. Two pedicel- Jariae developed. [Case 7.] Eechinus miliaris. MacBride, 1918 (15), pp. 338, 339, 348, Pl. viii, figs. 18, 19. Reared in London. Left : axial sinus fused with that of the right side. Madre- porie vesicle situated between the compound axial sinus and the gut. Pore-canal and dorsal pore single. Echinus-rudiment well-developed. Stone-canal double, probably formed by the splitting of the strmg which had connected the hydrocoele bud with the anterior coelom. Right : echinus-rudiment developed but, judging from the figures, it was smaller than the left one. [Case 8.] Echinus miliaris. Culture 9, ‘treated’. The larva was fifteen days old when found and killed. Left : axial sinus well developed with a pore-canal and dorsal pore. Madreporic vesicle rather rudimentary, but distinetly seen lying close to the pore-canal. Stone-canal normal, leading to the slightly-lobed hydrocoele. Amniotic invagination formed. Right : axial sinus and pore-canal rudimentary leaving no visible lumen. Dorsal pore absent. No madreporic vesicle found on this side. Hydrocoele as large as that of the left side, but simply vesicular in shape. Stone-canal has in its posterior part a distinct lumen and its calibre is as thick as its left fellow ; but the canal passes into a solid cell-string as it goes dorsad towards the vestigial axial sinus. No amniotic invagination formed. (b) Left hydrocoele and its associated structure more or less incomplete. Such is found very rarely, and hitherto any SITUS INVERSUS IN ECHINOIDS 1238 definitely-recorded case belonging to this category is lacking. By the kind permission of Professor Mac Bride I examined his preparations and found among twenty whole mounts only a single specimen with the right echinus-rudiment larger than that of the left side. [Case 9.] TEXT-FIG. 3. Transverse sections of a fifteen-day-old double-hydrocoele larva of Echinus miliaris, in which the water-vascular system of the left side has begun to degenerate. 300. am, amniotic invagination ; ax, axial sinus; dp, dp’, dorsal pores ; g, stomach; hy, hy’, hydrocoeles; mv, madreporic vesicle ; oe, oesophagus ; pc, pore-canal; st, st’, stone-canals. [Case 10.] Echinus miliaris. Culture 9, ‘treated’. This larva was found to have double hydrocoele when fifteen days old, and then killed and examined by means of sections (Text-fig. 3). Left: Axial sinus quite reduced, being represented by NO, 261 K 124 HIROSHI OHSHIMA a solid thickening at the dorsal end of the stone-canal, while the latter also has no visible lumen (st’). Pore-canal repre- sented by a solid cell-mass (pe’). Hydrocoele simple and vesicular (hy’). No amniotic invagination. Right : anterior coelom well developed (ax), with its external communication through a pore-canal (pc) and dorsal pore (dp). Madreporic vesicle with fairly distinct lumen (mv), with its wall in contact with the axial sinus, but no communication whatever existing between them. Stone-canal (st) formed normally, and hydrocoele (hy) well developed, being provided with five lobes. Amniotic invagination formed (am). [Case 11.] Echinus miliaris. Culture 9, ‘ treated’. This was also fifteen days old when found and killed. Left: no trace of axial sinus and pore-canal to be found. Stone-canal ending blindly at the anterior end, while posteriorly it opens to a small widened cavity of the hydrocoele. Amniotic invagination formed but small. Right: axial simus, pore-canal, dorsal pore- and stone- canal all well developed. Madreporic vesicle lying close to the pore-canal. Hydrocoele lobed. Amniotic invagination a little smaller than the left one. (c) The hydrocoele and its associated structures of both sides equal in their state of development or nearly so. [Case 12.] Spatangoid pluteus collected at Messina. Metschnikoff, 1884 (17), p. 64. Echinus-rudiments with ambulacral feet and spines ‘ quite equally ’ developed on both sides. [Case 13.] Mellita pentapora. Grave, 1911 (9), pp. 35-46, Text-figs. 1-3. Reared at Beaufort, N.C. Axial sinuses, hydrocoeles, and amniotic invaginations of both sides very nearly equal in development and symmetrically arranged. ‘I'wo pore-canals opened by a common dorsal pore situated on the mid-dorsal line. [Case 14.] Echinus miliaris. MacBride, 1911 (14), pp. 237-41, Pl. xxiv, fig. 1. Reared in London. Here also axial sinuses, hydrocoeles, and amniotic invagina- tions were found almost exactly in the same state on both sides. SITUS INVERSUS IN ECHINOIDS 125 There was only a single dorsal pore, and no trace of madreporic vesicle was found. [Case 15.] Echinus esculentus. MacBride, 1911 (14), pp. 241-4, Pl. xxiv, figs. 2-4. Reared at Plymouth (de Morgan). Echinus-rudiments fully developed when examined and drawn by the observer, the larva then being fifty-five days old. The echinus-rudiment of the right side- very slightly smaller and less advanced, than the left one. ‘Two pedicellariae developed on the right side and a third appeared at the posterior end. [Case 16.] Strongylocentrotus lividus. v.Ubisch, 1913 (80), pp. 440-3, Text-fig. 7, Taf. vin, fig. 26. Reared at Naples (Giesbrecht). Axial sinus well developed on both sides and almost the same in size, each beset with a pore-canal, which opened to the exterior separately by dorsal pores. Mid-dorsally-situated madreporic vesicle communicated through a narrow canal with the right axial simus. Echinus-rudiments and_ stone- canals on both sides equally well developed. [Case 17.] Echinus miliaris. MacBride, 1918 (15), pp. 338, 347, Pl. v, fig. 8. Reared in London. EKchinus-rudiments on both sides of almost equal size. Dorsal pores two, and no pedicellariae. [Case 18.] Echinus miliaris. MacBride, 1918 (18), pp. 3389, 347, Pl. vi, fig. 10. Reared in London. Both echinus-rudiments nearly equal in size, the right one slightly smaller. Dorsal pore single. A single pedicellaria formed on the right side. [Case 19.] Echinus miliaris. Culture 1, ‘ control’. The larva when found was eighteen days old, and had a lobed hydrocoele, stone-canal, and amniotic invagination developed almost exactly in the same state on each side. ‘Two pore-canals opened separately through a respective dorsal pore. The larva has been carefully fed on sufficient food, and the echinus- rudiments on both sides developed at equal rate. The two dorsal pores retained as before their side-by-side positions. K 2 126 HIROSHI OHSHIMA No asymmetry in shape of the stomach was to be found. The unpaired spine appeared at the hind end on the median line. The larva lived for forty-six days and was at the end of that period very near to metamorphose, but was missed suddenly, and hence no further information on the internal structures could be obtained. [Case 20.] Echinus miliaris. Culture 9, ‘ treated’. The larva was found and killed when it was fifteen days old. The flattened axial sinus, pore-canal, stone-canal, and lobed hydrocoele developed nearly symmetrically on each side. Only on the right side the pore-canal ended in a solid cell-mass, and no dorsal pore opened. Amniotic invaginations formed on both sides, the right one being smaller than that of the left side. No madreporic vesicle found. (d) Two hydrocoeles formed on one side. [Case 21.] Echinus miliaris. MacBride, 1918 (15), p. 839. Reared in London. There were formed two hydrocoeles on the right side due to the splitting of the hydrocoele bud which had been formed at the hinder end of the anterior coelom. One of them normally developed and had associated with it an amniotic invagination. The other, smaller and situated posteriorly to it, also possessed well-developed lobes. There was, however, no amniotic invagination for the smaller hydrocoele. Il. Hydrocoele formed on the left side only as in normal larvae, but some abnormalities found in other associated structures. (a) Amniotic invaginations on both sides. [Case 22.] Strongylocentrotus lividus. Runn- strém, 1912 (28), p. 7, ‘no. 4’; 1918 (26), pp. 414-15, Taf. xin, figs. 5,6. Reared at Monaco. Left: anterior coelom extending along dorsal side to the right to form a canal, which had no external opening. It seems, however, that an opening existed in an earlier stage. Stone-canal ending blindly at the posterior end. Hydrocoele isolated, with five lobes, and a blind canal sent towards the SITUS INVERSUS IN ECHINOIDS 127 stone-canal. ‘The author suggests that this hydrocoele may probably have differentiated from the posterior coelom of the left side. Posterior coelom was absent at first, but appeared later. Amniotic invagination formed. Right: anterior coelom posteriorly situated, being very elongated and developed much more strongly than normal. Despite the absence of hydrocoele on this side a small and shallow amniotic invagination appeared later. [Case 23.] Strongylocentrotus lividus. Runn- strom, 1918 (26), pp. 415-17, Taf. xii, figs. 7a, b. Reared at Monaco. Left : anterior coelom undifferentiated and devoid of external communication. Pore-canal ‘ post-generated ’ and dorsal pore opened on the mid-dorsal line. Hydrocoele at first remained undifferentiated, but later, when amniotic invagination appeared, it began to develop again. Right: hydrocoele not formed. Amniotic invagination appeared later, but soon degenerated. (b) No amniotic invagination formed. [Case 24.] Strongylocentrotus lividus. Runn- strém, 1918 (26), p. 424, Taf. xiv, figs. 17a, b. Reared at Monaco. The left anterior coelom represented by a widened end of the stone-canal, and a short wide pore-canal opening externally. Hydrocoele provided with four lobes, but owing to the absence of amniotic invagination its development was abnormal. One of the primary tentacles gave out a small branch which corre- sponds to one of the paired tentacles. (c) Pore-canal and madreporic vesicle doubled. [Case 25.) Strongylocentrotus lividus. Runn- strém, 1918 (26), p. 419, Taf. xii, fig. 11. Reared at Monaco. Both axial sinuses beset each with a pore-canal opening to the exterior by a dorsal pore. A pair of vesicular organs lying each near the pore-canal of each side were identified without doubt by the author as madreporic vesicles. The one on the left side acquired later a communication with the left axial sinus, while the other on the right side began to degenerate. 128 HIROSHI OHSHIMA Ill. Hydrocoele absent from both sides, [Case 26.] Strongylocentrotus lividus. Runn- strom, 1918 (26), p. 424, Taf. xiv, fig. 18. Reared at Monaco, Amniotic invagination failed to be formed on the left side ; and, instead of it and at the place where the former should in normal case be formed, a calcareous spine appeared. [Case 27.] Echinus miliaris. MacBride, 1918 (15), pp. 389-40, Pl. vi, figs. 12-14; Pl. x, figs. 22-8. Reared in London. Under this heading more than one specimen will be described together. Anterior coelom on neither side enlarged so as to form an axial sinus. On neither side was a hydrocoele discovered, nor was there any vestige of a stone-canal or au dorsal pore. Only in an exceptional case was there found a dorsal pore. In one case a madreporic vesicle was seen and figured (fig. 22). A group of pointed spines developed on each side within the loop of the ciliated band and another spine was found situated dorsal to this loop on both sides. 6. CONSIDERATIONS ON THE ORGANS AND STRUCTURES CONCERNED AND THE FACTORS CONCERNED IN THEIR DEVELOPMENT. (2) Anterior Coelom.—this is formed separately on each side pinched off from the posterior coelom, the left one being earlier in its formation than the right fellow (see MacBride, 11, p. 298). Sometimes the two anterior coeloms unite to form a single sac on the dorsal side of the larval oesophagus (Cases 5, 7). The left one is connected with the pore- and stone-canals and remains as a distinct sac, called the axial sinus, while the right one normally remains as a simple sac and very often degenerates later. (b) Madreporic Vesicle.—tThis is a minute round sac normally found a little on the left side lying close to the pore- canal, and often is stated to exhibit a rhythmic pulsation. Mac Bride (11, p. 299) discoveredin Echinus esculentus that this vesicle was derived from the right anterior coelom, SITUS INVERSUS IN ECHINOIDS 129 at first as a solid thickened end of a string of cells given out from the posterior end of this coelom. Later (16, pp. 261-2) he confirmed this in Eehinocardium cordatum, in which species the vesicle in question is unusually large. Runnstrém found a pair of madreporie vesicles in a larva of Strongylocentrotus lividus (Case 25), and, moreover, according to him, the one on the left side became later connected with the axial sinus of the same side. Perhaps other instances of the presence of a communication between the vesicle and one of the axial sinuses (Cases B, 1, 2, 16) may also be due to a secondary change. In the Case 2 the vesicle is seen later again separated from the coelom. Often this vesicle is absent (Cases 14, 20). v. Ubisch (80, p. 448) is of the opinion that the madreporic vesicle was not possessed by the ancestor of the sea-urchins, but that it represented the only remnant of the degenerated right anterior coelom having assumed a new but unknown function in the course of phylogenetic develop- ment. And, further, according to him, when the right anterior coelom made its unusual development the highly-differentiated and functioning madreporic vesicle could not be affected thereby and both of them existed side by side. In the reversed and also in some double-hydrocoele larvae (Cases 10, 11) the madreporic vesicle was found on the right side, close to the right pore-canal. In the case where two such vesicles are present (Case 25) the right one may be the homo- logue of this. In neither case is its origin made clear. From want of sufficient material and from our ignorance of its func- tion any definite statement will be premature. (c) Pore-canal and Dorsal Pore.—The primary dorsal pore is formed from the left coelomic sac to communicate with the exterior before the latter becomes divided into the anterior and posterior coeloms. In the course of further development of the larva the pore shifts from its original position on the left side towards the mid-dorsal line. This shifting is preceded by the formation of a transverse groove of the ectoderm. Probably in connexion with this shifting process it is often the case that the canal gets temporarily or permanently obliterated (Cases B, 130 HIROSHI OHSHIMA 1, 2, 20, 22, 283). The cause is unknown to us; still, [ think there is hardly any doubt as to its being due to artificial conditions. ‘loo large a number of diatoms or bacteria in the vessel in which the larvae have been kept may cause this. Shortly afterwards the pore and canal can regenerate (Cases B, 93) and the revived development of the whole water-vascular system follows. In other instances no second pore was formed, and degeneration of the system soon set im (Cases 1, 8, 10, 11). The presence of the right pore-canal side by side with the left is a constant and normal character in the larva of Mellita pentapora (Grave, 9, p. 42; and also his former paper, 1902, p. 58). The same is not common in Echinus miliaris; still, it has been recorded by MacBride in a larva which was otherwise quite normal (15, p. 339). Although the presence of two pore-canals is a very common occurrence among double-hydrocoele larvae (Cases 4, 6, 16, 17, 19, 20) it seems by no means to be a necessarily associated feature. In starfish the occurrence of the double dorsal pore has never been seen even among double-hydrocoele larvae (Gemmill, 5, p. 230; 7, p. 31; 8, p. 62). To such an important difference found between these two classes let us return later (p. 142). According to Runnstr6ém the forma- tion of the dorsal pore and pore-canal seems to be a self- differentiation (25, p. 301). (d) Stone-canal.—This is the part which at first connected the hydrocoele bud with the main body of the anterior coelom. This canal is sometimes found doubled, being caused from either its defective origin (Case 7) or abnormal regeneration (Case B). When degeneration takes place, probably due to the lack of communication with the exterior, it begins from that end which is adjacent to the axial sinus (Cases 1, 8, 11). (e) Hydrocoele.—lIt is a well-known fact that the mght coelomic sac has in normal larvae the potentiality of producing a sac which is homologous with the left hydrocoele. Such a special organ-forming substance seems to be located especially at the place where the coelomic sac has to divide later into the anterior and posterior coeloms. We see from MacBride’s SITUS INVERSUS IN ECHINOIDS 131 work on Ophiothrix fragilis (18, pp. 578, 586) that this sac, homologous with the left hydrocoele, exhibits varying degrees of development among normal larvae, and in a few extreme cases it gives rise to a five-lobed hydrocoele (PI. xxxvi, fig. 54; compare further those double-hydrocoele Ophioplutei described by Miller and Metschnikoff). Whether this unusual development of the right hydrocoele is to be regarded as a case of atavism or as another kind of variation is a matter of choice. MacBride (14, pp. 240, 244) is of opinion that the free-swimming ancestor of the Echino- derm had a pair of hydrocoeles, equally developed on each side, the right one has, however, become atrophied as soon as the free-swimming habit was given up. The appearance in some abnormal larvae of a right hydrocoele is an atavistic feature. But, according to him, the appearance and further completion of the associated structures, such as amniotic invagination, set of spines and dental sacs, derived from the ectoderm and mesoderm respectively cannot be accounted for by atavism, because it is quite impossible to endow the ancestor with such a double set of highly-developed spines and Aristotle’s lanterns. Therefore, he introduced the idea of the internal secretion, in that the abnormally-developed right hydrocoele must have given off some stimulating substances which caused both ectoderm and a part of the posterior coelom to respond, with the result that there appeared a second set of spines and dental sacs. He further discussed this theory in his second paper on the double hydrocoele (15, pp. 341-5). Some months earlier than the first of these papers Grave (9, p. 48) discussed the same idea and made the objec- tion ‘ that such an explanation presupposes that the series of structures in question was present and in some way related in the normal development of the ancestral echinoderm, a supposition for which there is no basis in observed fact ’. Now, we may find no great difficulty im assuming that such stimulating power of the left hydrocoele has been acquired since the disappearance of the right hydrocoele, as v. Ubisch (80, p. 444) remarked in reply to Grave’s objection. It 132 HIROSHI OHSHIMA is necessary, however, to introduce another supposition to understand how the right hydrocoele in our abnormal case acquired that power of stimulating other tissues, which power was not possessed by the right hydrocoele of the ancestor. In short, even if we accept the view that the Echinoderm ancestor possessed a double hydrocoele, it seems to me that the atavistic interpretation has to encounter with such a diffi- culty as stated above. The development of a right hydrocoele to such an unusual degree may then safely be regarded as a case of homoeotic variation. The examples of this kind of variation given by Bateson (2, pp. 721-35) should be classified at least into two different groups. One group contains the cases characterized by the appearance on one side of a wholly new structure, which is quite unknown in the animal’s phylogenetic history, whereas a mirror-image of it is normally present on the other side. Gemmill’s ‘ primary’ homoeosis (8, p. 71) seems to be this. A tadpole of Pelobates fuscus with a second spiracle on the right side is an example, and if Runnstrém’s view is accepted the appearance by self-differentiation of an amniotic invagination on the right side of the sea-urchin larva would be another. The second group comprises those cases where, in obviously paired organs, one member, which is normally vestigial, develops in certain circumstances to the same degree as its fellow. A double-tusked narwahl is the best illustration of this kind. Gemmill’s term ‘ secondary ’ homoeosis perhaps denotes the same phenomenon. I feel very doubtful whether the case of our double hydrocoele should be placed under this latter category or under the first. The paired origin of the front teeth in the narwahl is quite obvious, while the presence of a pair of well-developed hydrocoeles in the Echinoderm ancestor will not be accepted unanimously by all zoologists. I do not believe that the development of a double hydrocoele has ‘resulted in a larval organization better adapted to the conditions under which the existence of the pluteus is led’, as Grave (9, p. 45) states in his discussion on the homoeosis. SITUS INVERSUS IN ECHINOIDS 133 We need not explain the cause of homoeosis in this way only. * The chance by which the double hydrocoele is induced to develop seems to be quite unusual, as I will try to show presently. It is not at all a result of adaptation. In his famous experiments on Alpheus, .Przibram showed that if a large claw of this Crustacean is amputated a small claw will appear at the spot, whilst the small claw of the other side, which was not operated upon, will become a large claw. This phenomenon he calls ‘ compensatory hypertypy ’. For more detailed information I refer to his later paper (22). A similar but slightly different idea can be applied in the case of double hydrocoeles. The right hydro- coele might have arisen as a result of compensatory hypertypy caused by the arrested state of development in the left hydro- coele. ‘The differences from the case with Alpheus are that (a) the presence of a rudimentary right hydrocoele is not a normal feature, but no doubt the right anterior coelom has a potentiality of producing it, while the small claw of Alpheus is present constantly and quite functional, and (b) the left hydrocoele has not yet been fully developed but arrested in its early stage of development, while the large claw of Alpheus was removed after it had reached the full-grown state. With these differences kept in mind we may use Przibram’s term in our case as well. According to Runnstrém (25, p. 305) the further dif- ferentiation of the hydrocoele, left or right as the case may be, depends largely on the formation of an amniotic invagination. There was, however, an exceptional case (Case 24). Besides, from lack of a corresponding amniotic invagination and from obliteration of the dorsal pore, the hydrocoele and its associated structures will degenerate from hunger (Runnstr6ém, 265, p. 265-5; MacBride, 15, pp. 339, 340). The presence of two hydrocoeles on one side was noticed by MacBride (Case 21), and interpreted as being due to the splitting of the hydrocoele bud. Another curious abnormality was described by Runnstrém (Case 22). There are, accord- ing to this observer, two possibilities as to the cause of such 134 HIROSHI OHSHIMA an isolated hydrocoele: (a) it may have been separated from the end of the stone-canal, or (b) the posterior coelom may have * given rise to it under the influence of the amniotic invagination. From the absence of posterior coelom, though one appeared afterwards, he thinks the latter more probable. In one of Runnstrém’s larvae of inverse situs (Case B) we see another extraordinary feature in the right hydrocoele (28, p. 9; 26, p. 423). The hydrocoele was three-lobed, and close to it there were two curious structures. One was a round closed vesicle, the origin of which the author could not ascertain. The other was an ectodermal groove running nearly parallel to the stone-canal and lined with very actively-moving cilia. This groove at last became separated from the ectoderm, and together with the above-stated closed vesicle, united with the hydrocoele, remaining as a larger lobe of the latter. Runnstr6ém is of the opinion that in those pathological cases a hydrocoele or a part of it can be formed both from posterior coelom and ectoderm. (f) Amniotic Invagination.—This is formed some days later than the appearance of the hydrocoele. It seems to me highly probable that this structure is homologous with the stomodaeal invagination of Holothurians. As early as 1906 MacBride (12, p. 615) pointed out that the larval stomodaeum of Holothurians reminds one ‘of the amniotic cavity in the Echinopluteus ’. This idea has since found another support in the fact that in Cucumaria the stomodaeal invagination is formed to the left of the mid-ventral line, as was first discovered by Newth (20, p. 634, Pl. 1, fig. 6) and after- wards confirmed by the writer (21, pp. 379, 384, Pl. v, figs. 5 and 6). It is therefore quite improbable that the ancestral Kehinoid had a pair of amniotic invaginations. Mac Bride (15, p. 343) never found in any single instance an amniotic invagination formed where no hydrocoele existed, and con- firmed his former view (14, pp. 240-1) that the undifferentiated ectoderm can give rise to an amniotic invagination only under the influence of the hydrocoele. Runnstrém’s view is diametrically opposed to this. He has shown us several — SITUS INVERSUS IN ECHINOIDS 135 instances where an ectodermal invagination was formed at a place under which no hydrocoele had been developed (Cases B, 22, 23, and also 25, p. 271). He further made experi- ments to prove his view that the formation of the amniotic invagination is a self-differentiation and is not formed from stimulus of an underlying hydrocoele. He could produce a new amniotic invagination in a larva of Echinus miliaris from which the echinus-rudiment had been removed (27, pp. 9-11). In another of his experiments an amniotic invagina- tion was seen to appear in each of two pieces of a larva where the normally-formed invagination had not been included ; thus in this larva three amniotic invaginations in all were formed (pp. 13-14). It may be mentioned that in all of his cases the ectodermal invagination was very small and lined with flat epithelial cells. In another place he states (25, p. 302) that the invaginated ectoderm forms cylindrical cells only at the place where the hydrocoele wall comes to be in contact, while in the other part the cells remain flat. I myself understand by the term amniotic or ‘ echinid ’ invagination an ectodermal pit whose epithelial cells are from its first appearance high and cylindrical, even when fairly apart from the hydrocoele (Text-figs. 2, p, and 3, c, am). In this sense I cannot help doubting whether all of Runnstrém’s structures deserve the name amniotic invaginations. He admits that the further development and differentiation of the amniotic invagination is conditioned by the presence of a hydrocoele, and that without it the former degenerates (25, p. 305). It is of interest to see that he pointed out that the role of an amniotic invagina- tion could be played, to a less extent, by other ectodermal invaginations, such as that which he termed ‘ spine invagina- tion’ (Case B). According to him, if there was no amniotic invagination formed the stone-canal stopped developing when it had reached its normal length, and later gradual degenera- tion set in of the whole water-vascular system. But, as for the larva in question, the ‘ spine invaginations ’ were situated further back than normally an amniotic invagination is placed, and the stone-canal did not stop at the normal length, 136 HIROSHI OHSHIMA but continued to lengthen until the hydrocoele reached those invaginations (26, p. 421). As to the nature of these invagina- tions let us examine again (p. 138), (g) Posterior Coelom and Genital Stolon.—The anteriorly-prolonged end of the left posterior coelom shares the formation of the echinus-rudiment (MacBride, 11, pp. 304-5). This change takes place also on the right side in abnormal larvae where a right hydrocoele developed. In the normal case the genital stolon makes its appearance shortly before metamorphosis from the wall of the left posterior coelom (Mac Bride, 11, p. 309). How its right fellow behaves in abnormal larvae is still an open question. Runnstrém inclines to think that in two of his double-hydrocoele larvae (Cases 1 and 2) a rudiment of genital stolon was formed from the right posterior coelom. v. Ubisch (80, p. 445) concludes that the doubleness is not extended to all organs as shown from the fact that in his older double-hydrocoele larva (Case 3) the genital stolon was seen formed only on the left side. This conclusion cannot pass unchallenged because in this larva the right echinus-rudiment was much less advanced than the left, and also because the structure in question is not distinct until the larva reaches the height of its growth. (hk) Pedicellariae——In normal Echinus larvae there appear a pair of pedicellariae on the right side, one being dorsal to the loop of the ciliary band, the other ventral to the same. In some imperfectly symmetrical double-hydrocoele larvae one or both of them appear on the right side only (Cases 6, 15, 18) or on both sides of the larva (MacBride, 15, p. 343). According to Runnstr6ém the reversed larvae of Strongylocentrotus had pedicellariae appearing on the left side (Cases A and B), and I am inclined to believe that it 1s also the case with our Echinus, though unfortunately any positive evidence is lacking at present. In the complete absence of hydrocoele from both sides no true pedicellariae appear (Case 27). Thus the relation between the pedicellariae and echinus-rudiment (or hydrocoele) is somewhat compli- cated. Probably the echinus-rudiment calls forth the forma- SITUS INVERSUS IN ECHINOIDS 137 tion of pedicellariae on the opposite side. It seems to me that they are not inhibitory to each other on the same side, because they can co-exist side by side. The fact, however, that in most of the double-hydrocoele larvae the pedicellariae are not formed may simply be due to lack of sufficient material, or that the echinus-rudiment, being more vigorous in develop- ment than the pedicellariae, wins the competition. Mac Bride assumes that ‘the influences emanating from a hydrocoele not only tend to inhibit the formation of pedicellariae on the same side but to determine their formation on the opposite side of the larva’ (15, p. 348), and that the hydrocoele can act as such even in its early stage. Thus, the fact that an echinus-rudiment and a pedicellaria or two can co-exist on the same side is explained by him in the following manner : ‘Tf we assume that in these larvae the growth of both hydro- coeles has been arrested at an early stage, but after the stage at which the stimulus to form pedicellariae on the opposite side had already gone forth from them, and that then, after the formation of these organs on both sides had been deter- mined, further nourishment became available and the left hydrocoele developed further, the structure of such larvae can be explained’ (pp. 343-4). Runnstrém’s case that some starved larvae, which had no hydrocoele, developed a pair of pedicellariae (25, pp. 269-70, Text-figs. 33-5) is now very difficult to understand. It is doubtful whether the hydrocoele was really absent in those larvae. (1) Spine.—The larva of Echinus miliaris produces, when fairly grown, a rudiment of a spine at the hind end a little towards the right from the median line. This gives rise, as do some others which develop later, to a square-ended spine on the future abactinal side. This rudiment is found situated a little on the left side in reversed larvae (Text-fig. 1, sp,), and in most of the double-hydrocoele larvae, in an almost median position. Such a different position of this spine is undoubtedly correlated with the different behaviour of the echinus-rudiment. Characteristic are the spines which develop in the larvae devoid of hydrocoele (Case 27). As already 138 HIROSHI OHSHIMA stated there is a group of pointed spines and a solitary one on each side of the larva. Runnstrém found such a spine only on the left side (Case 26). I am much inclined to think that from want of regulating influence of the hydrocoele the rudiments of pedicellariae were developed in an aberrant way into some of those peculiar spines. Runnstréom (27, pp. 21-2, figs. 21-3) discovered in the normal larva of Kchinus miliaris a pair of small ecto- dermal invaginations formed inside the loop of the ciliary band on the right side. In each of these invaginations 2-3 spines belong to the Basalia 3 and 5 are later formed. He called the former ‘spine invaginations’ (26, p. 420). Spines un- doubtedly identical with these have been seen by me on the left side of one of the reversed larvae (Text-fig. 1, sp,). In Strongylocentrotus lividus these structures do not appear normally, still Runnstr6ém identified the pair of pits found in an abnormal larva with them (Case B). These may be an abnormal amniotic invagination divided into two. His descriptions and figures (28, p. 8; 26, pp. 420-1, Taf. xiv, fig. 18) are not quite satisfactory enough to sub- stantiate his refusal to look on them as modified amniotic invaginations. (j) Gut.—We know really nothing about the change the gut undergoes in accordance with the formation of the double echinus-rudiment or situs inversus. Normally the definitive stomodaeum appears at the centre of the floor of the epmeural space (MacBride, 11, p. 307), and the rudiment of the oesophagus, as an outgrowth from the left wall of the stomach meeting the stomodaeum, appears later (p. 310). The adult mouth breaks through some days after metamorphosis, and the anus is formed still later (pp. 311-12). Runnstrém (24, pp. 544-52) found in the larvae developed from the eggs which had been treated with potassium-free sea-water some asym- metrical distortions in the larval stomach and formation of a new oesophagus on the left. He interpreted the phenomenon as the formation of the definitive oesophagus precociously indicated. It is quite conceivable that in the course of the SITUS INVERSUS IN ECHINOIDS 139 development of an echinus-rudiment, no matter on which side of the larva it may lie, the hydrocoele, working together with other ectodermal and mesodermal tissues, can induce this new structure to appear and thus the actinal part of the young sea-urchin be completed. 7. PRoBABLE MECHANISM WHEREBY ABNORMALITIES ARE PRODUCED. From those observed facts above considered the following conditions seem to concern the production of abnormalities of the hydrocoele and its associated structures. 1. Obliteration of the pore-canal. This seems to be a cause of the arrest of the further development of the water-vascular system and then a quick degeneration of the whole system follows. 2. Activation of the right anterior coelom of its latent poten- tialities of producing a hydrocoele, to compensate the degenerat- ing left hydrocoele. 3. Regeneration of the pore-canal or fusion of the two axial sinuses. Both afford the left hydrocoele a renewed com- munication with the exterior, and the further development and differentiation of the water-vascular system thereby take place. 4. Development of a right amniotic invagination and the peculiar change of the anterior prolongation of the right posterior coelom. These changes seem to have been evoked by the stimulus of the unusual right hydrocoele. These three elements working together give rise to an echinus-rudiment. From these data, if adequately combined, the following changes are quite possible. Let us start from a young normal larva, in which hydrocoele, axial sinus, pore-canal, and dorsal pore are all formed on the left side. An amniotic invagination may already be formed on the left side. The right anterior coelom may have a pore- canal. Now, the dorsal pore of the left side becomes obliterated, which fact is followed by the arrest of development and further NO. 261 L 140 HIROSHI OHSHIMA degeneration of the left water-vascular system. ‘Two courses are here open: A. The right anterior coelom begins its unusual development to produce a right hydrocoele, which acquires communication with the exterior through a pore-canal. B. The right anterior coelom does not become active either from very weak disposition of the right anterior coelom or, more probably, from want of sufficient nutrition. The result is the total absence of hydrocoele from both sides. The further fate of larvae in which the course of events has been that indicated by A will be one of the following three : 1. Appearance of a new dorsal pore on the left side which revives the power of the left hydrocoele to develop further. If well fed the hydrocoele on each side will continue to develop side by side so as to give rise to a double-hydrocoele larva. 2. Axial sinuses of both sides come in contact with each other and then unite, thus making the left hydrocoele regain its communication with the exterior and enabling it to develop further. The result is also a double hydrocoele. 3. No reappearance of a second dorsal pore nor fusion of the axial sinuses takes place. The left water-vascular system will then degenerate quickly, while the right one will develop like the normal left. A larva with situs inversus is the result. In both the courses of events indicated by 1 and 2 the following three conditions may possibly arise, according to the different stages at which the right hydrocoele had arrived, when the recovery of the left hydrocoele took place : (a) The recovery of the left hydrocoele takes place before the right hydrocoele attains a size equal to the left. The period during which the hydrocoele is deprived of communication with the exterior is very short. Under such a condition the result is a larva whose left hydrocoele or echinus-rudiment is larger or more advanced than that of the right side. This is very frequently met with among double-hydrocoele larvae. (b) The left hydrocoele recovers at the time when the right one attained a size about equal to it. The larva developed under such a condition has two hydrocoeles or echinus-rudiments i ee SITUS INVERSUS IN ECHINOIDS 141 equal in size. Such a case is less frequently met with than the former. (c) The left hydrocoele recovers late when the right one is in a more advanced state than it. The period during which the hydrocoele is deprived of communication with the exterior is here very long. The result is that the larva has the left hydrocoele or echinus-rudiment smaller than the right. Usually the hydrocoele and its associated structures cannot remain unchanged for so long a time after being deprived of its external communication. This case is therefore met with very rarely. The above may not be the only ways of reaching the respec- tive results, but probably are the commonest. Many modifica- tions are naturally conceivable: for instance, the right dorsal pore may be obliterated in its turn, which causes the degenera- tion of the whole water-vascular system of the right side and thus a normal larva will result secondarily (see Case 8). Let us now compare this interpretation of the occurrence of the inverse situs in Echinus larvae with Spemann’s case of Triton larvae (29, p. 407). Though equally caused by a ‘defective’ development of a single organ—alimentary canal in Triton and hydrocoele in Eehinus—tfurther results in which the other organs become affected are different in these two cases. Instead of displacement of other adjoin- ing organs, the arrest in development of the left hydrocoele causes a new hydrocoele to appear on the other side and also a new set of associated structures as a consequence. The normal left hydrocoele can, if it regains its opportunity of further development, produce another echinus-rudiment, so as to give rise to a double-hydrocoele larva. Any parallel of such a feature is very improbable in Triton larvae. There is no reason to expect that the above is equally applicable to the formation of double hydrocoele of other classes of Hchinoderms. Conditions may be totally different. Let us, for instance, take the case of the double-hydrocoele larvae of starfishes. Normally in most species of starfishes the paired coelomic vesicles grow forwards, and their anterior L 2 142 HIROSHI OHSHIMA ends meet and unite in front of the larval mouth. The presence of two dorsal pores is very common, but the right one gradually atrophies (Gemmill, 5, p. 231), and still the right coelomic vesicle retains its communication with the exterior through the, left dorsal pore. The hydrocoele becomes later differen- tiated from the middle portion of the spacious left coelomic sac. In the case of the double hydrocoele the right one is likewise formed from the middle portion of the right coelomic sac. Among the double-hydrocoele larvae of Porania pulvillus and Asterias rubens Gemmill found no case of the presence of double dorsal pores, in all instances the left pore only being present (5, p. 230; 7, p. 43; 8, pp. 62, 69). Thus it is evident that the obliteration of a dorsal pore has hardly any influence on the further development of the hydrocoele on the same side. Under such a different con- dition I suppose that the occurrence among starfish larvae of the situs inversus as we find in Echinoid larvae will be extremely unusual. Gemmill tried to explain the cause of the double hydrocoele chiefly by the supposition that, owing to the over-fed condition of the larva, its stomach becomes expanded and globular, so that the ventral horn of the left posterior coelom tends to fail to unite with the right middle coelomic region. The latter region, being thus left isolated from the posterior coeloms, produces a right hydrocoele (5, p. 244; 8, pp. 54-5). This interpretation in its turn cannot hold true in the case of those double-hydrocoele Echinoid and Ophiuroid larvae, in which no such extension of the left posterior coelom takes place normally (MacBride, 15, p. 826). The discovery by MacBride (10, pp. 368-70) of a double-hydrocoele larva in Asterina gibbosa, in which species the egg is heavily laden with yolk, is a serious objection to the hypothesis of excessive food. One feature is, however, certainly common in the double-hydrocoele larvae of the three different classes: namely, the temporary arrest in the development of the left hydrocoele in some way or other in an early stage. And this occurs more frequently under artificial conditions than in nature, SITUS INVERSUS IN ECHINOIDS 148 With regard to the occurrence of the reversed Auriculariae, as discovered by Miller (19, pp. 101, 109, Taf. v, fig. 1), the attempt to interpret the phenomenon by virtue of the compensatory hypertypy is nearly hopeless. It is a widely- accepted fact that in Holothurians the right anterior coelom does not exist at any stage throughout life, whilst the hydro- coele is differentiated even before the coelomic sac divides into right and left halves (posterior coeloms). It is not easy to imagine that the right posterior coelom could ever produce a hydrocoele, when the normal hydrocoele happened to be arrested in its development. If this cannot be the case we must regard it as a result either of the change of polarity in the egg (according to Conklin, 8) or of twin formation (of Spemann’s sense, 29). 8. ExternAat Factors as Causts or ABNORMALITIES. From series of his experiments MacBride (15) came to the conclusion that the chief cause producing double-hydrocoele larvae of Echinus was the increased salinity of the water used for culture. Unfortunately, as I have pointed out in a foregoing page (p. 114), the result of our experiments of this year was quite different from our expectations. As shown in the table the number of double-hydrocoele larvae was greater in ‘ controls’ than in ‘ treated’, i.e. 2 per cent. and 0-8 per cent. respectively. As the double hydrocoele and situs inversus start, I believe, under the same condition, the figures of reversed larvae may also be used in this connexion. The occurrence of the reversed larvae was practically equal in both ‘controls ’ and ‘ treated ’, i.e. 10-7 per cent. and 11-2 per cent., the difference being within the range of probable error. Let us now turn to examine whether artificial synthetic sea-water had anything to do with the production of abnor- malities. Culture 11 came into contact with the synthetic sea-water when the larvae were four days old, Culture 1 when twelve days old, and Culture 6 when fourteen days old. They were examined and counted seventy-two, six, and three days afterwards respectively. Though it is unsafe to draw any 144 HIROSHI OHSHIMA decided conclusion from such few cases and numbers one can hardly see any effect of the synthetic sea-water on the production of doubles or reversed if allowed to avt earlier in one culture than in others. One might reasonably expect that the artificial treatment of the egg and sperm might have caused some disturbance from the normal development of the larva. This is of course quite possible, but I may only mention that it is curious to see that among such material as the sea-urchin egg so commonly used for study and demonstration in embryological work only a very few cases of the abnormalities in question have been noticed. One of the most important factors which differ more or less from the conditions in nature is the food supply. The method of feeding marine larvae on diatom cultures, through which many different forms of pelagic larvae have been successfully reared, is relatively a recent introduction. The result is very often over-feeding. In an over-fed larva hypertrophy and other disturbances in growth is quite conceivable. From uneven distribution of food in the culture jar and from a different state in the health of larvae, over-fed and under-fed individuals may arise withm one and the same jar. The obliteration of the normally-formed left dorsal pore, which seems to me a direct cause of the production of the double hydrocoele and situs inversus of the Echinus larvae, may be associated with the excess of diatoms and other minute organisms in the jar. Whether it is physiological or mechanical it is hard to decide at present. Runnstrém (25, pp. 321-2) found that the larvae of Strongylocentrotus showed the degeneration of organs when over-fed on yolk. The echinus-rudiment was above all the most sensitive to the treatment and degenerated com- pletely. Undigested yolk granules were found migrating everywhere, even scleroblasts were laden with them and the absorption of calcareous bodies followed. The effect of over- feeding on diatoms will naturally be very different from this. Though somewhat difficult to coutrol (MacBride, 15, SITUS INVERSUS IN ECHINOIDS 145 p. 388) it is desirable to experiment on the effect of different amounts of diatom-food upon the development of the larval organs. The effects of hunger were observed both by Runnstrém (25, pp. 254-821) and MacBride (15, pp. 339-40). The difference between the results of these two observers 1s remark- able. In every instance of Runnstrém’s larvae showed extreme degeneration of skeletons, while in MacBride’s case the larval arms were almost normal, owing to the well- developed state of the skeletons, but the hydrocoele degenerated and peculiar spines formed. Besides the differences in degree and duration of hunger, the stage at which the larvae were treated, &c., there must be still other complicated factors which caused such different results. For those starved larvae bacterial infection is no doubt another important cause of abnormal development (25, pp. 278-4). Grave (9, p. 36) remarked that among the larvae of Mellita only those well fed developed the echinus-rudiment. As to the effect of other chemical and physical environments upon the development of the sea-urchin larvae we have those valuable results obtained by Vernon, Tennent, and others. But we know hardly anything with regard to the changes of coelomic vesicles and hydrocoele treated specially. 9. SUMMARY AND CONCLUSION. 1. Under artificial conditions more than 10 per cent. of the larvae of Echinus miliaris exhibited the situs inversus. 9. So far as I could examine, the internal as well as external structures of such abnormal larvae were mirror-images of those of the normal larva. 3. The young sea-urchins metamorphosed from such inverse larvae showed no abnormal features externally. 4. The manner in which such abnormal larvae departed from the normal development seems to be analogous to that in the case of ‘compensatory hypertypy’ in the claws of Alpheus. 5. In an early stage of the normally-developing larva it 146 HIROSHI OHSHIMA happens sometimes that the left dorsal pore becomes obliterated. This seems to be associated with the shifting of the pore towards the mid-dorsal line. The hydrocoele, thus deprived of its communication with the exterior, ceases to develop and then degeneration of the whole water-vascular system sets in. 6. The right anterior coelom, on the other hand, is now evoked to realize its latent potentiality of producing a hydrocoele (homoeosis). The degenerating left hydrocoele gives place to a newly-appearing right hydrocoele. 7. The right hydrocoele stimulates its adjoining tissues to give rise together to an echinus-rudiment. 8. The external factor or factors which cause the obliteration of the dorsal pore could not be found. This probably is con- nected with the presence of too much diatom-food and other micro-organisms in the culture jar. 9. If a new dorsal pore is formed on the left side before the degeneration of the left hydrocoele sets in, the developing power of the latter will thereby be revived. If sufficiently fed a double-hydrocoele larva will result under such a condition. 10. If, while the left hydrocoele is arrested in its develop- ment and then degenerates, the right anterior coelom fails to develop a new hydrocoele presumably from want of sufficient food, a larva devoid of hydrocoele will result. ZooLoGicaL DEPARTMENT, IMPERIAL COLLEGE OF SCIENCE AND TECHNOLOGY, LONDON. December 16, 1920. 10. Literature CITED. 4. Allen, E. J., and E. W. Nelson (1910).—‘‘ On the Artificial Culture of Marine Plankton Organisms ”’, ‘ Quart. Journ. Mier. Sci.’, vol. 55, no. 218; and also ‘ Journ. Mar. Biol. Ass.’, vol. 8, no. 5. 2. Bateson, W. (1894).—‘ Materials for the Study of Variation, treated with Special Regard to Discontinuity in the Origin of Species’, London. 3. Conklin, E. G. (1903).—‘‘ The Cause of Inverse Symmetry ”’, ‘ Anat. Anz.’, Bd. 23, Nr. 23. 4. Driesch, H. (1906).—‘‘ Studien zur Entwicklungsphysiologie der Bilateralitat ’’, ‘ Arch. Entw.-Mech.’, Bd. 21, Heft 4. 11 SITUS INVERSUS IN ECHINOIDS 147 . Gemmill, J. F. (1914).—‘‘ The Development and certain Points in the Adult Structure of the Starfish Asterias rubens, L.’, * Phil. Trans. Roy. Soc.’, Ser. B, vol. 205. —— (1915).—‘‘ Twin Gastrulae and Bipinnariae of Luidia sarsi, Diiben and Koren ”’, ‘ Journ. Mar. Biol. Ass.’, vol. 10, no. 4. —— (1916).—‘“‘ The Larva of the Starfish Porania pulvillus (O. F. M.),” ‘ Quart. Journ. Micr. Sci.’, vol. 61, no, 241. . —— (1916).—‘‘ Double Hydrocoele in the Development and Meta- 12. 13 14. ° 15. 16 17 18 19. 20. 21. 22 morphosis of the Larva of Asterias rubens, L.’’, ibid. Grave, C. (1911).—‘*‘ Metamerism of the Echinoid Pluteus ”, ‘ Johns Hopkins Univ. Cir.’, no. 232. MacBride, E. W. (1896).—‘* The Development of Asterina gibbosa”’, ‘Quart. Journ. Micr. Sci.’, vol. 38, no, 151. —— (1903).—“‘ The Development of Echinus esculentus, together with some Points in the Development of E. miliaris and E. acutus”, ‘ Phil. Trans. Roy. Soc.’, Ser. B, vol. 195. —— (1906).—‘‘ Echinodermata ”’, “ The Cambridge Natural History ’, vol. 1. —— (1907).—‘‘ The Development of Ophiothrix fragilis”, ‘Quart. Journ. Micr. Sci.’, vol. 51, no. 204. —— (1911).—‘‘ The Abnormal Plutei of Echinus, and the Light which they throw on the Factors in the Normal Development of Echinus”, ibid., vol. 57, no. 226. —— (1918).—‘ The Artificial Production of Echinoderm Larvae with two Water-Vascular Systems, and also of Larvae devoid of a Water-Vascular System ’’, ‘ Proc. Roy. Soc.’, Ser. B, vol. 90. —— (1918).—“‘The Development of Echinocardium cor- datum. Part Il. The Development of the Internal Organs ’’, * Quart. Journ. Micr. Sci.’, vol. 63, part 2. Metschnikoff, E. (1884).—‘‘ Embryologische Mittheilungen ber Echinodermen ”’, ‘ Zool. Anz.’, Jrg. 7, Nr. 159. Morrill, C. V. (1919).—‘‘ Symmetry Reversal and Mirror Imaging in Monstrous Trout and a Comparison with similar Conditions in Human Double Monsters ”’, ‘ Anat. Rec.’, vol. 16, no. 4. Miller, J. (1850).—‘‘ Uber die Larven und die Metamorphose der Echinodermen. II. Abhandl.’”’, ‘Phys. Abhandl. k. Akad. Wiss. Berlin ’, 1848. Newth, H. G. (1916).—“* The Early Development of Cucumaria: Preliminary Account ”’, ‘ Proc. Zool. Soc. London ’, 1916. Ohshima, H. (1918).—‘‘ Notes on the Development of Cucumaria echinata’”’, ‘ Annot. Zool. Japon.’, vol. 9, part 4. Przibram, H. (1905).—“‘ Die ‘ Heterochelie ’ bei decapoden Crustaceen (zugleich: Experimentelle Studien iiber Regeneration. LI. Mitt.)”’, ‘ Arch. Entw.-Mech.’, Bd. 19, Heft 2. 148 HIROSHI OHSHIMA 23. Runnstrém, J. (1912).—‘‘ Quelques observations sur la variation et la corrélation chez la larve de VOursin”’, * Bull. Inst. Océan. Monaco’, no. 247. 24, —— (1914).—‘‘ Analytische Studien iiber die Seeigelentwicklung. I. Mitt.’’, ‘ Arch. Entw.-Mech.’, Bd. 40, Heft 4. 25. —— (1917).—Ditto. ILI. Mitt., ibid., Bd. 43, Heft 3. 26. —— (1918).—Ditto. IV. Mitt., ibid., Heft 4. 27. —— (1917-18).—“* Zur Entwicklungsmechanik der Larve von Par- echinus miliaris”, ‘ Bergens Mus. Aarb.’, Nr. 14. 28. —— (1920).—‘* Entwicklungsmechanische Studien an Henricia sanguinolenta Forbes und Solaster spec.”’, * Arch. Entw.- Mech.’, Bd. 46, Heft 2-3. Spemann, H., und H. Falkenberg (1919)—‘* Uber asymmetrische Entwicklung und Situs inversus viscerum bei Zwillingen und Doppelbildungen ”’, ‘ Arch. Entw.-Mech.’, Bd. 45, Heft 3. Ubisch, L. v. (1913).—‘ Die Entwicklung von Strongylocen- trotus lividus (Echinus microtuberculatus, Arbacia pustulosa)’’, ‘ Zs. wiss. Zool.’, Bd. 106, Heft 3. 29 30 11. APPENDIX. To the examples of the reversed larvae (p. 112) a case in the starfish Cribrella oculata as figured and described by A. T. Masterman is to be added (see ‘ Trans. Roy. Soc. Edin- burgh’, vol. 40, 1902). My thanks are due to Professor J. F. Gemmill, who has kindly called my attention to this paper. Compare also: Gemmill, ‘“ Notes on the Develop- ment of the Starfishes Asterias glacialis O.F.M.; Cribrella oculata (Linck) Forbes; Solaster endeca (Retzius) Forbes; Stichaster roseus (QO. F. M.) Sars’, ‘London Proc. Zool. Soc.’, 1916, p. 557. With regard to the reversed larvae of Ophionotus hexactis a description can now be found in the following work: Th. Mortensen, ‘Studies on the Development and Larval Forms of Echinoderms ’, Copenhagen, 1921 (p. 182). Ho Princeton, N. J. December 28, 1921. a ee Ser = SITUS INVERSUS IN ECHINOIDS 149 Nore spy Prorrssor FE. W. MacBriprt on Mr. Hirosut Ox- SHIMA’S Parser on ‘ THE OCCURRENCE OF SITUS INVER- SUS AMONG ARTIFICIALLY-REARED EH CHINOID LARVAE’. The most interesting paper by my friend and pupil Dr. Oh- shima, which appears in this number of the ‘ Quarterly Journal of Microscopical Science’, calls for some comment from me. Dr. Ohshima refers to a paper published by me in the ‘ Proceed- ings of the Royal Society ’ in which I described a method for inducing the formation of a second (right) hydrocoele in EKchinoid larvae by stimulating the larva at a critical period of its growth by exposure to hypertonic sea-water. Dr. Ohshima states that an attempt which he made to repeat this experiment in my laboratory in 1920 resulted in failure. Nevertheless certain larvae with two hydrocoeles turned up, and he gives a different explanation of the cause for their appear- ance. I am convinced that the explanation which Dr. Oh- shima gives is the right one to account for the phenomena which he observed in 1920; but I wish to emphasize the fact that his and my explanations agree in one most important particular, viz. we both feel convinced that the right anterior coelom of an Echinoid larva has the innate constitutional power of developing a right hydrocoele. This power I account for on the hypothesis that Echinoderms are derived from a free- swimming ancestor provided with sets of tentacles on the right and left sides of the body. Dr. Ohshima’s explanation is that it is a case of ‘homoeosis’, but to use this term of Dr. Bateson seems to me to be merely restating the difficulty in other language without offermg any explanation at all. The fact that when the right hydrocoele does appear it appears in similar form to that exhibited by the left, and not in the condition in which the original right hydrocoele must have been when it was functional, is in my judgement to be accounted for by the assumption that the modifications which the left hydrocoele subsequently underwent have been pushed backwards in development according to the principle of tachygenesis till they now affect the earliest differentiated 150 HIROSHI OHSHIMA organ-forming substance out of which the hydrocoele arises— and that this organ-forming substance gives rise to both hydrocoeles. The results which I obtained in 1917 I was able to obtain under precisely similar conditions in 1919. Dr. Ohshima’s failure to obtain them in 1920 may, I think, be attributed to several causes. I stated that for success several conditions were necessary, one of which was a vigorous culture of the diatom Nitzschia. For some unknown reason this was exces- sively difficult to obtain in 1920. Again and again our cultures died off and the larvae were checked in their development. Dr. Ohshima obtained a few ‘ doubles’ both in the control and the ‘treated’ culture which were started in May, and a few more doubles in the control culture started in June. But the May cultures were not obtained from satisfactory females: they were obtained from masses of eggs in which only a small pro- portion developed, and they could not be described as vigorous cultures or likely to show a proper reaction to stimulation. The June cultures were vigorous, but, owing to the failure of the Nitzschia culture, the ‘treated’ culture died off com- pletely, and the ‘ control’ culture was for weeks in a condition of checked and stunted growth and only recovered later when the Nitzschia finally re-established itself. Ina word Dr. Ohshima obtained his specimens with a right hydrocoele through the checking of the growth of the normal left one by starvation, whilst I obtained mine in 1917 and 1919 by stimulating the larvae in their early growth by the action of hypertonic sea- water. E. W. MacBripe. The Behaviour of the Golgi bodies during nuclear division, with special reference to Amitosis in Dytiscus marginalis.' By Reginald James Ludford, B.Se. (Lond.), Department of Zoology and Comparative Anatomy, University College, London. With 4 Text-figures. CONTENTS. PAGE 1. INTRODUCTION ; : : ‘ : : ‘ 5 poet 2. Previous WORK . : ‘ ‘ : : ‘ E Pella? 3. Tue Gotat Bopy DURING AMITOSIS IN THE FoLLicLE CELLS OF THE Ovary OF DyTIScUS : : ; ‘ : > Lae: 4, DISCUSSION . : : ; , ; : : : 5 SYS 5, CONCLUSION . i : : : : ’ : : = AlSy/ 6. BIBLIOGRAPHY : c - P P : ; : 17 1. INTRODUCTION. Driorvoxrnssis, or division of the Golgi bodies during cell division, has recently been worked out by Professor J. Bronté Gatenby and the present writer in the male germ-cells of several animals (5). It was found that the distribution of the Golgi elements, or dictyosomes, to both halves of a dividing cell was a very haphazard process and was unaccompanied by any splitting of the dictyosomes such as occurs in the case of the chromosomes. In the cricket Stenobothrus, the dictyosomes 1 Part of the materials used in this research was purchased by a Govern- ment Grant of the Royal Society, for which I express my thanks, 152 REGINALD JAMES LUDFORD become scattered in the cytoplasm of the spermatogonium before cell division takes place, and they remain in this condi- tion during meiosis, so that approximately a half become contained within each of the newly-formed cells. A different process, however, occurs in the Mammals, Mus and Cavia, and in the Molluses, Helix and Limnaea. The Golgi bodies in the spermatogonia of these types consist of a number of dictyosomes arranged around the archoplasm, inside which is the centrosome. As this latter organ divides, preparatory to the formation of the spindle, its two constituent parts separate and carry with them to both ends of the cell, approxi- mately half of the archoplasm, still with the dictyosomes attached. During late prophase, the dictyosomes become temporarily detached from the archoplasm and _ scattered throughout the cell, and then at the late telophase they collect together again around the archoplasm. The examples of dictyokimesis described in our previous paper were those which occurred concurrently with meiotic nuclear division. Professor Gatenby suggested to me the desirability of investigating the behaviour of the Golgi body during amitotic nuclear division, and in the present paper is described the behaviour of the apparatus during amitosis in the follicle cells of the ovary of the beetle, Dytiseus mar- ginalis. 2. Previous Work. Deinecka (1) has described dictyokinesis in the dividing epithelial cells of Descemet’s membrane and connective-tissue cells of the cornea, during both mitotic and amitotic nuclear division. He found that the Golgi body surrounded the archoplasm, and during mitosis divided into two parts so that each daughter-cell received a ‘ Netzapparat ’, as he calls it, but that in amitosis there is no division of the centrosome and no change in the Golgi body. These observations are quoted by Mackln (8) in support of his own conclusions derived from a study of nuclear division in cells of tissue cultures of the heart of the embryo chick, that amitosis involved a GOLGI BODIES IN DYTISCUS 153 merely division of the nucleus, and not of the cytoplasm. He observed that binucleate and polynucleate cells were formed as the result of amitotic nuclear division. During such division the centrosome and archoplasm remained unchanged. ‘The archoplasm could be seen in the living cell, but not the centro- some; but the latter was to be seen in fixed preparations stained with iron haematoxylin as two small black bodies embedded in the archoplasm. Mitochondria were visible in the living cells, but not the Golgi body. Macklin’s observations are of special interest in that they substantiate materially the evidence that has been accumu- lated against the view upheld by writers such as Meves (9), that mitosis can occur in an amitotically-formed nucleus. Binucleate cells were observed by Macklin to undergo mitosis, but in this process the nuclei which had been formed by amitotic division came together and their chromatin masses fused to form the chromosomes which underwent the usual stages of mitotic division. It is therefore concluded that amitosis does not imply division of the cytoplasm but only fission of the nucleus. 3. Tur Goter Bopy puRING AMITOSIS IN THE FOLLICLE CELLS OF THE Ovary oF DytTiIscus. In common with most insects Dytiscus has an ovary laterally disposed on either side of the abdomen. Hach ovary is com- posed of a number of tubules containing a single row of odcytes in all stages of development, the most mature being at the distal end. The odcytes are surrounded by the follicle cells, and between each odcyte is a group of nurse or nutritive cells whose function is to provide nourishment for the developing odcytes. At the proximal end of the ovuliferous tubules is a mass of undifferentiated cells from which arise three types of cells, viz. odcytes, nutritive cells representing modified odcytes, and the follicle cells in which the behaviour of the Golgi body during amitosis was studied. It was found rather difficult to impregnate the body in the follicle cells of insects, After a number of unsuccessful attempts 154 REGINALD JAMES LUDFORD the best results were obtained by adopting Cajal’s method (4) with shghtly longer fixation than he describes. Ovaries of Dytiscus were fixed for about eight hours in Cajal’s standard fixative, and after rapid washing in distilled water were left in silver nitrate solution for three days. Prepared slides were stained by safranin or carmine. The Golgi body then appears as a number of black granules in a pink-coloured cytoplasm, the mitochondria when visible are usually brown in colour, and the nucleoli of the cells are red. TExtT-Fias. 1-4. Follicle cells of the ovary of Dytiscus marginalis. The upper cell-wall is in each case in contact with the odcyte wall. ge, elements of the Golgi body (Dictyosomes) ; m, mitochondria ; n, nuclear membrane ; xl, nucleolus (plasmosome). In the text-figures are shown the various stages of amitosis. At fig. 1 is seen the so-called ‘ resting stage’ of the cell. It will be observed that there is a single nucleolus within the nucleus, and scattered through the cytoplasm are the darkly- impregnated elements of the Golgi body, while the mito- chondria are more or less evenly distributed in the cell. In the stage shown in fig. 2 the nucleus has elongated and the nucleolus is dividing into two. The Golgi elements still remain scattered in the cytoplasm, but it will be noticed they show a tendency to lie near the nuclear membrane—a tendency which is apparent in the other figures. At a later stage, as shown in fig. 3, the GOLGI BODIES IN DYTISCUS 155 two parts of the nucleolus have separated and the nucleus is greatly constricted, but the dictyosomes are still irregularly scattered: while in the cell shown in fig. 4, when the two parts of the nucleus are completely separated, the dictyosomes are still irregularly disposed around them. It will be noticed that the nucleolus appears to play quite an important part in this process: its division seeming to initiate the division of the nucleus. This process has been verified by observations on material prepared by fixation in corrosive acetic and Bouin, and stained with Mann’s methyl blue eosin (7). In such preparations the nucleolus stains oxyphil, and is apparently of the nature of a plasmosome. Its appearance is the same in both kinds of preparations. This type of amitosis, originally described by Remak, in which the nucleolus appears to play an important part, has been found by recent workers to be exceptional rather than typical, and Macklin, observing amitosis in living cells, says ‘ the division of the nucleolus has no direct relationship to nuclear division. It may, however, have to do with the size of the nuclear portions ’ (8). The extent to which the dictyosomes are distributed in the resting follicle cell is subject to variation. In some cases, evidently owing to the large size of the nucleus in comparison with the width of the cell, the elements of the apparatus are crowded together towards its outer wall and appear in rare cases to be attached to an archoplasm ; but this does not occur when the cell is dividing amitotically, and in no case has the separation of two distinct groups of dictyosomes, as occurs in mitosis, been observed. 4. Discussion. Gatenby has suggested (8) that the scattermg of the Golgi elements during o6genesis is a means whereby it is able to exert a maximum formative influence upon the cytoplasm, as well as prepare for even distribution in the cells of the seg- menting ovum. In a previous paper (6) I have described how the dispersing dictyosomes influence the formation of NO, 261 M 156 REGINALD JAMES LUDFORD yolk in the odcyte of the Mollusc Patella. It would seem possible, therefore, that the spreading out of the apparatus in the follicle cells of an insect might be related to the high degree of metabolism existing in such cells. Chun, quoted by Nakahara, regarded the division of the nucleus in amitosis as a means of increasing the nuclear surface as an aid to metabolic interchange between nucleus and cytoplasm; while Flemming pointed out that amitosis was especially associated with intense secretive and assimilatory activity, but he considers such cells as being on the way to degeneration (2). Recent work has shown that fragmentation of the nucleus does occur in pathological growths, in cells subject to faulty nutritive conditions, and in tissue cultures which have been left unattended for some time. Such frag- mentation is regarded by Macklin (8) as an altogether different phenomenon from amitosis, but in the past there is no doubt that there has been confusion between the two. Nakahara, who has made an investigation into the subject of amitosis in adipose cells of insects and an extensive survey of the literature of the subject, concluded that ‘ amitosis, occurring in secreting or reserve forming cells and in other cells of similar activity, may be for the purpose of securing an increase of the nuclear surface to meet the physiological necessity due to the active metabolic interchanges between the nucleus and cytoplasm. Apparently it is not a method of cell multiplication nor a sign of degeneration or senescence of cells, but, whenever it occurs, it seems to indicate an intense activity in the vegetative functions of the cell’ (11). It is altogether in accordance with our present knowledge of the Golgi apparatus to assume that in such cells, as for example the follicle cells of insects’ ovaries, the dictyosomes scattered in the cytoplasm would play a by no means unimportant part in the lipoid metabolism. In conclusion, I have to acknowledge my indebtedness to Professor J. Bronté Gatenby, of Trmity College, Dublin, for his kindness in reading through the manuscript of this paper. GOLGI BODIES IN DYTISCUS 157 5. CONCLUSION. We may now recognize the following modes of behaviour of the Golgi bodies during nuclear division : (1) During karyokinesis the Golgi bodies may either, (a) remain scattered in the cytoplasm and be approxi- mately shared out amongst the two newly-forming cells, e.g. male germ-cells of Stenobothrus (8) ; (b) divide into two masses surrounding the separating centrosomes and thus pass into each cell, e.g. (i) during meiosis in the male germ-cells of the Molluscs, Helix and Limnaea, and the Mammals, Mus and Cavia (5) ; (ii) during mitosis in the epithelial cells of Descemet’s membrane and connective-tissue cells of the cornea (1). (2) During amitosis either they (a) remain as a number of elements or dictyosomes arranged around the archoplasm, e.g. in mammalian epithelium (1), or they (b) become irregularly scattered throughout the cytoplasm, as described in this paper in the follicle cells of insects’ ovaries. It is suggested that these differences are related to different conditions of metabolism existing in cells exhibiting these phenomena. 6. BIBLIOGRAPHY. 1. Deinecka, D.—‘‘ Der Netzapparat von Golgi in einigen Epithel- und Bindegewebszellen wihrend der Ruhe und wahrend der Teilung derselben ’’, ‘ Anat. Anz.’, 1912. 2. Flemming, W.-—“ Attraktionssphiiren u. Centralkérper in Gewebs u. Wanderzellen ’’, ‘ Anat. Anz.’, 1891. 2a. —— “Entwicklung und Stand der Kenntnisse tiber Amitose”’, ‘Merkel und Bonnet’s Ergebnisse d. Anat. u. Entwick.’, 1892. 3. Gatenby, J. Bronté.—‘‘ Cytoplasmic Inclusions of the Germ-cells, Part V. The Gametogenesis and Early Development of Limnaea stagnalis, with special reference to the Golgi Apparatus and Mitochondria ”’, ‘ Quart. Journ. Micr. Sci.’, vol. 68. 4, —— “Identification of Intracellular Structures”, ‘Journ. Roy. Micr. Soc.’, 1919. M 2 158 REGINALD JAMES LUDFORD 5. aa: Ludford, R. J., and Gatenby, J. B.—‘“‘ Dictyokinesis in Germ-cells, or the Distribution of the Golgi Apparatus during Cell Division’’, ‘Proc. Roy. Soc.’, vol. 92, 1921. . Ludford, R. J.—‘‘ Contributions to the Study of the Oogenesis of Patella”, ‘ Journ. Roy. Micr. Soc.’, 1920. —— “The Behaviour of the Nucleolus during Oogenesis, with special reference to the Molluse Patella ”’, ibid., 1921. Macklin, C. C.—‘‘ Binucleate Cells in Tissue Cultures ’’, ‘ Contribu- tions to Embryology ’, no. 13, Carnegie Institution of Washington, 1902. Meves, F.— Uber amitotische Kernteilung in den Spermatogonien des Salamanders, und das Verhalten der Attraktionssphiren bei derselben ”’, ‘ Anat. Anz.’, 1891. —— ‘Uber die Entwicklung der minnlichen Geschlechtszellen von Salamandra ”’, ‘ Arch. fiir mikr. Anat.’, 1896. Nakahara, W.—“ Studies of Amitosis: Its Physiological Relation in the Adipose Cells of Insects, and its Probable Significance ”’, *‘ Journ. of Morph.’, 1918. On the anatomy and affinities of Paludes- trina ventrosa, Montague. By Guy C. Robson, B.A. (Published by permission of the Trustees of the British Museum.) With 12 Text-figures. CoNTENTS. PAGE 1. INTRODUCTION, TECHNIQUE, ETC. . F . : ‘ - 159 2. STRUCTURE . F é : : : : ; ; - 160 3. HABITS, ETC. . 2 p ; : : A A : ae) TKS 4. AFFINITIES . t ‘ : : : j 2 : . 182 5. SUMMARY . F ? 5 ; 5 - : 3 . 184 6. BIBLIOGRAPHY A 2 F z 3 : e 3 - 85 1. INTRODUCTION. THE Prosobranch molluse which is described in this paper is a small insignificant animal found upon plants or bottom débris in the brackish water of creeks and tidal ditches in various parts of England, Wales, and Ireland. It is also found in the upper waters of estuaries. Jeffreys (9) records it from ‘ the sea coasts of Sweden, France, and Portugal, as well as of Algeria ’, though such cases are open to a great deal of doubt. It has been selected for study for several reasons. In the first instance there is very urgent need for more informa- tion about the Taenioglossate Prosobranchs to which group the Paludestrinidae are referred. In the second place, though some substantial knowledge is available upon the classification and structure of the Paludestrinidae, the euryhaline habits of 160 GUY C. ROBSON some species of the genus and the general tendency in the group to show a transition from a marine to a fresh-water habit render them a peculiarly interesting group and worth an intensive study. In the last place the recent discovery of Partheno- genesis in Paludestrina jenkinsi (Boycott, 4) makes a closer study of the kindred species necessary. Our knowledge of the European Paludestrinidae includes good accounts of the anatomy and histology of Bythinella dunkeri (Bregenzer, 6) and Vitrella quenstedtii (Seibold, 20), and more incomplete descriptions of part of the anatomy of P. ulvae (Henking, 10) and P. jenkinsi (Robson, 16). In spite of this amount of work a good deal remains to be cleared up as to the structure of these animals. The material employed was obtained from tidal ditches at Leigh-on-Sea, Essex. It was fixed in Bouin’s solution after the shell had been carefully cracked away so as to expose the columella. Reconstruction models of various organs were made. A rapid method, which may be capable of improve- ment, was devised, in which the usual plates were made up of modelling-clay mixed with varying proportions of glycerine and water and rolled out on pieces of thin paper. This mixture can be graded to give a harder medium than Plasticine, and is therefore more suited to making models of such parts as contain delicate ducts, nerves, &c. The paper, if cut larger than the plate, allows of rapid handling and can be cut away after the plate is in position. The surfaces and edges of the plates can be easily painted with water-colours. The author is indebted to the late Dr. W. G. Ridewood for suggestions as to reconstruction models, and to Professor Paul Pelseneer for information upon the general morphology of the ‘Taenioglossa. 2. STRUCTURE. 1. The Alimentary System. The oral cavity (Text-fig. 1) is usually deep and narrow. Ventrally it exhibits a pair of lateral diverticula which are sometimes forked. In general it agrees with that of other ANATOMY OF PALUDESTRINA 161 members of the genus. It is lined by a cuticle which is fairly thin ventrally, but becomes thicker dorsally. This cuticle is secreted by a columnar epithelium which is continuous with that of the lips and adjacent parts. Hach cell of this epithelium contains an elongate, deeply-staining mass of secretion which occupies the major portion of the cell and usually obscures the nucleus. These secretion-masses are especially well developed where the cuticle is deepest ; and in these areas the TErxt-Fia. 1. Transverse section through the mouth. c¢, cuticle; m, mandibles ; se, secretory epithelium. whole epithelium is characterized by a mass of extra-cellular pigment in the shape of very small, subcircular granules. In the upper portion of the mouth is found a pair of man- dibles. These consist of a number (minimum 13, maximum 20) of columnar pieces of specialized cuticle, each secreted by a single cell of the basement epithelium, as Seibold found in Vitrella (20). That the secretion-masses are intimately con- cerned in the formation of these is shown by the fact that plates are often continuous with the former. The mandibular plates stain very sharply with eosin, the rest of the cuticle being more or less unaffected by the stain. There is sometimes present, in addition to the mandibles, a specialized piece of 162 GUY C. ROBSON cuticle just below the mandibles on each side. There is also usually a median dorsal projection, dagger-shaped in transverse section. This also stains sharply with eosin, but less intensely than the mandibles. Behind this projection are to be found at about the same level in the mouth on either side two glandular patches of unknown function which run backwards to the origin of the salivary ducts. Posteriorly to the mandibles the mouth expands laterally and is flattened dorso-ventrally over the lingual cartilages. TEXtT-FIG. 2. P. ventrosa. Transverse section through buccal bulb. c, oral cartilage; cm, circular muscles; Im, longitudinal muscles ; bm, basal membrane of radula; 7, radula ; sg, salivary gland. In this area it shows in transverse section three main divisions— a median, unpaired cavity with a thin roof, dorso-lateral expansions with ciliated and glandular walls into which the salivary ducts open, and ventro-lateral expansions which dip down beside the cartilages. These have a cuticular lining. The cilia of the dorso-lateral cavities no doubt serve to circulate the saliva. In P. ulvae Henking considers their function consists in driving the food particles backwards. In Vitrella and Bythinella the cilia are continued on to the roof of the median portion. The lingual cartilages (Text-fig. 2) correspond to those found in other genera of the family, and in general to the ANATOMY OF PALUDESTRINA 168 excellent description given by Bregenzer for Bythinella. They are two somewhat piriform bodies, united dorsally and in the median line. They are rather flattened dorso-ventrally. Posteriorly they diverge somewhat. Bregenzer states that in Bythinella they are each divided into a *‘ Haupt-Knorpel’ and a ‘ Knorpelspange’, and regards the difference between Paludestrina and Bythinella in this respect as of taxo- nomic value. Certainly no such division is apparent in Palu- destrina ulvae and ventrosa. Lateral expansions (‘ Fliigel’) observed in Bythinella occur in P. ventrosa as well. The tissue of the cartilages is composed of irregular polygonal cells with relatively small nuclei and a large amount of granular pigment. It should be noted that Rougemont (18) states that in Hydrobia sp. the cartilages are capable of movement upon each other, if pressed downwards. In Paludestrina, Vitrella, and Bythinella, this is, of course, quite possible; as they are no doubt very elastic and have a certain amount of ‘ play ’ on each other. The radula has been described and figured by Woodward (24). The salivary glands are two in number. In P. ulvae Henking describes two pairs with separate openings. Those of P. ventrosa appear to correspond with Henking’s first (larger) pair. They run as far back as do the second pair of P. ulvae and sometimes cross over in a_ similar fashion. The lateral diverticula of the pharynx disappear posteriorly, and the oesophagus develops a fresh series of diverticula in the form of deep longitudinal furrows. Behind the cerebral com- missure the oesophagus in most cases shows a tripartite arrange- ment as in Bythinella. Itis cilated almost up to its distal extremity. The stomach (Text-fig. 3) is an irregularly- shaped organ with a forked appearance exteriorly due to the fact that the intestine and style-sac leave the stomach parallel to each other from its anterior end. The oesophagus and 164 GUY ©. ROBSON hepatic duct open into the stomach at its upper (posterior) end. The pyloric part of the intestine and the style sac are in open communication with each other, as in P. jenkinsi (Robson, 16), by means of a longitudinal slit for a considerable way. Bregenzer does not refer to this as occurring in TExtT-FIG. 3. Diagrammatic reconstruction of stomach to show relationship of crystalline style sac. p, pylorus; 7, intestine; sf, style sac; l, hepatic duct ; 0, oesophagus. Bythinella. No mention is made by Seibold of a style sac in Vitrella, and, as his account is painstaking and thorough, we have to assume that the sac is absent. This is a very curious fact and one to which we shall return later. The oesophagus and hepatic duct open into the stomach fairly closely together. In this region the stomach epithelium is ciliated, the ciliated area being continued downwards and ANATOMY OF PALUDESTRINA 165 forwards into the pylorus. On the side of the stomach opposite to the oesophageal and hepatic apertures the epithelium gives rise to a dense cuticle which occupies the major part of the posterior part of the stomach but diminishes anteriorly. Vitrella and Bythinella apparently differ conspicu- ously in the lining of the stomach. In the latter form only a small part of the stomach is ciliated, while the contrary is true of Vitrella. P. ventrosa is more or less inter- mediate between the two in this respect. The base of the epithelium which secretes the cuticle is, as in the case of the oral cuticle, rendered conspicuous by a layer of densely-staining granules. The stomach is crossed by numerous ridges of which the most constant and most conspicu- ous is a large and strongly-developed one lying transversely in the cavity above the hepatic duct. Grooves with specialized cuticle are found in the neighbourhood of the latter. The style sac is blunt externally and rather thimble-shaped. In transverse section it is circular and exhibits on the side towards the pylorus a groove of characteristic structure. The latter corresponds in its histological features to the similar structure in Bythinella. Bregenzer has offered no explana- tion of the function of this groove. Unfortunately the structure and relationship of the style itself cannot be demonstrated in fixed material. I am under the impression, however, that, in the living animal, the style is not loose in its sac but attached. If that is the case it may be secreted in the groove. The rest of the sac is simple, being composed of a thick ciliated epithelium. The cilia are very dense and much longer than the cells. No one familiar with the recent work on the style sac in Lamellibranchia can fail to be struck with the similarity between the structure here described and that figured by Nelson (14) for Lampsilis anodontoides. In both forms the pyloric part of the intestine communicates by a narrow slit with the style sac, the walls of which are composed of a single layer of columnar ciliated cells. In Lampsilis the resem- blance to Paludestrina is still more emphasized by the 166 GUY C. ROBSON presence of a large mass of deeply-staining cells near the slit. In several figures given by Nelson and Edmondson (7) we find a lateral groove like that of Bythinella and Paludes- trina. The similarity between the crystalline style of Gastro- pods and Lamellibranchs has been commented upon by various authors, and a short discussion may be found in Moore (12). The pylorus is ciliated and passes gradually into the intestine proper. This follows the usual course. It is ciliated almost to its extremity; a well-marked typhlosole is found extending for some distance down the intestine. This is apparently absent nm Vitrella and Bythinella. The hepato-pancreas occupies the apical whorls as usual, and consists of branching finger-like processes. The duct is very short and fairly wide. No definite differentiation of the liver-cells into granular- and ferment-cells with different stainnmg and contents could be made out. Vacuoles with inclusions are seen in the plasma of the liver-cells. The rectum, when viewed transversely, exhibits a number of longitudinal folds. It runs forward in the roof of the pallial cavity projecting from the latter and ultimately becomes free for a short distance. 2. The Nervous System (Text-fig. 4). The only description of the nervous system of Paludes- trina is that of Henking, which is insufficient and leaves a good deal to be desired. The cerebral ganglia of P. ventrosa are long and rather pointed anteriorly. The cerebro-pedal commissure is normal though very short. The buccal commissure calls for some comment as it is extremely short and thus unlike the elongate form found in this and allied families. The buccal ganglia are closely applied to the anterior end of the cerebral ganglia. In one or two cases very short connectives were found ; but such instances are rare. Henking does not refer to the buccal commissure as such in P. ulvae; but from his description and figure the connectives would appear to be as short as in P. ventrosa, ANATOMY OF PALUDESTRINA 167 As in allied forms the cerebro-pleural connectives are absent, the ganglia being practically contiguous on each side. The right pleural ganglion is larger than the left. The pleuro-pedal connectives are short and closely applied to the cerebro-pedals. The pedal ganglia are rather triangular in transverse section. Posteriorly they bear a pair of otocysts. TEXT-FIG. 4. C Central nervous system (diagrammatic). 6, buccal ganglion ; c, cerebral ganglion ; lp, rp, pleural ganglion ; pd, pedal ganglion (l); pp, propodial ganglia ; sb, subintestinal ganglion ; sp, supra- intestinal ganglion; on, osphradial nerve; ¢, tentacular nerve and ganglion; pdn, parapodial nerve. The visceral commissure is of the shortened type and resembles that seen in Vitrella and Bythinella, though it agrees with the former in the amount of abbreviation, the supra-oesophageal pleural connective not being so shortened as in the latter. The supra-oesophageal portion follows the oesophagus more or less closely to the abdominal ganglion 168 GUY C. ROBSON which is situated in the columella region between the kidney and the reproductive organs. The subintestinal portion passes over the surface of the columella muscle to the abdominal ganglion. Three main nerves are given off from the anterior end of the cerebral ganglia—a large tentacular nerve with a tentacular ganglion and two oral and labial nerves. No separate optic nerve was found, the innervation of the eye being achieved by optic fibres of the tentacular nerve. It may be recalled that Vayssiere found that in Truncatella (28) the tentacular nerve is apparently responsible for the oculo-motor innervation. From the left pleural ganglion are given off two nerves of pallial distribution. Each pedal ganglion gives off three main roots—anterior, ventral, and postero-lateral. ‘The first-named always bear at a short distance large and well-defined propodial ganglia. ‘The ventral pair is very stout and sometimes bears ganglia. The postero-lateral pair is very slender and, as in Bythinella, sometimes bears a diminutive ganglion. No certain trace of a metapodial commissure was found. It is absent in P. ulvae and present both in Vitrella and Bythinella. From the supra-intestinal ganglion a nerve passes upwards over the roof of the pallial cavity to the osphradium. Henking’s description of the visceral commissure and its prolongations in P. ulvae isin need of correction. Bregenzer, who admits this, is inclined to make taxonomic capital out of his statement that two connectives from the pedal ganglion are found joining the ‘ Oberschlundganglion ’, and infers from this that the pleural and cerebral ganglia are completely fused. If this is correct it is difficult to account for Henking’s ‘ acces- sorische ’ and ‘ unpaares’” ganglia unless we call these respec- tively supra-intestinal, subintestinal and abdominal (!) It seems far more sensible to assume that Henking made a mistake about the point of insertion of the pedal connectives, to call his ‘1. accessorische Ganglion’ the left pleural and his ‘ unpaares Ganglion ’ the subintestinal. ————————— ANATOMY OF PALUDESTRINA 169 3. Sense Organs. The statocysts are in close proximity to the posterior surface of the pedal ganglia. The auditory nerve is very diffi- cult to trace, and in a large number of preparations I have only succeeded in finding one in which it is clearly seen. It runs backward from the statocyst in close proximity to the cerebro-pedal connective and ultimately becomes indistin- euishable from the latter. The cysts contain each a single moderately-sized otolith. In certain other Taenioglossa, e.g. Valvata (Bernard, 1), Melania and Paludina (Pelseneer, 15), numerous small otoconia replace the single otolith of Paludestrina, Vitrella, and Bythinella. This diversity, which con- trasts with the remarkable constancy in the number of otoliths found in Teleostean fishes, might well supply a subject for independent study both from the taxonomic and the physio- logical point of view. The cysts are formed of an external layer of very thin epithelium covering an internal layer of irregularly-shaped cells. ‘These are flattened and resemble rather those figured by Bernard (1) for Valvata. In Valvata Bernard found no cilia, while Garnault (8) observed that in Cyclostoma they are very sparsely developed. In Bythinella and Vitrella they are not referred to. In P. ventrosa it is almost certain they are absent. There is quite definitely no ordinary ciliated layer and nothing more on the interior surface than occasional vague clumps of unrecognizable tissue. How the otoliths in this case stimulate the sensory layer is therefore uncertain unless the latter has a general tactile sensibility. In Valvata Bernard (loc. cit.) observed a sort of network formed of prolongations from the membrane of the lining cells. The eyes do not call for special attention. They resemble those figured by Henking for P. ulvae in general form, though they differ in the closer approximation of the inner and outer cornea. In Bythinella the space between the two corneas is considerable and filled with connective tissue. In ad 170 GUY C. ROBSON the same form Bregenzer notes no special differentiation of the outer cornea from the adjacent epithelium of the tentacle base. This is not the case in P. ventrosa, in which the external cornea is always noticeably thinner. The osphradium, asin Vitrella and Bythinella, is a simple ridge-like elevation on the left-hand side of the gill close to the junction of the roof of the pallial cavity with its floor. There are no foliations such as occur in some other Taenioglossa. The laterally-disposed pigment cells contain brownish pigment granules. The ciliated and sensory layer overlies a large and elongate osphradial ganglion. 4. The Respiratory and Circulatory System. (1) Pericardium and Heart (Text-fig. 5). The pericardium lies on the posterior side of the body- whorl in a superficial position covered only by the body- epithelium. It is placed at the posterior end of the pallial cavity on the left-hand side, and is roughly bounded by the kidney and the extremity of the style sac. No trace of a reno-pericardial orifice could be found. Seibold was unable to find one in Vitrella; nor is it referred to by Bregenzer for Bythinella. It occurs in both Cyclo- stoma (8) and Valvata (1). The auricle was never found in an expanded condition, so that its general structure cannot be defined. The ventricle is, as usual, thick-walled and muscular, though a thinner-walled portion of varying extent is invariably to be seen. (2) Vascular System. The descriptions of this system in other Taenioglossa are very unsatisfactory. The accounts given are usually incom- plete, and frequently omit some portion from consideration. It may also be pointed out that on one point at least two of the most up-to-date treatises on the Mollusca are at variance. Lang (11, p. 322) says, ‘ Bald 6ffnen sich die Arterien ... in arterielle Sinusse. Unter diesen verdient besonders der grosse Kopfsinus . ..’ Pelseneer (15), who alludes to the ANATOMY OF PALUDESTRINA 171 cavities in which the arteries terminate as ‘ interorganic lacunae’, says (p. 100), ‘ the venous blood is collected into an anterior or cephalopedal sinus’, &c. I share Professor Pelseneer’s view that the large cephalopedal cavity is venous. The following outline, which is by no means complete and owing to the size and contractility of the animal is not founded TExtT-FIG. 5. Heart, pericardium, and kidney in section. a, auricle ; ¥, ventricle ; k, kidney ; p, portal vein; pc, pericardium. upon injections, may serve to enlarge our knowledge of this system in the Taenioglossa. There appears to be a large general venous sinus of which the chief components (cephalic, pedal, and visceral) are in communication with each other. The two first-named are in open communication posteriorly, but are separated anteriorly by a horizontal septum. Both Henking and Bregenzer speak as though the latter completely separated the two cavities in P.ulvae and Bythinella. In none of my series, however, is this the case. NO, 261 N 172 GUY C. ROBSON From the general venous sinus the blood passes (a) to the renal portal system by more or less clearly-defined vessels, (b) to the gills through the rectal simus. From the kidney the blood passes into the portal vein (q.v.). The complete course of the latter is not easy to trace. In one or two cases it was found entering the auricle close to the root of the pulmonary vein, though in other cases its junction is not so clearly seen and it might even open into the pulmonary vein itself. It is possible, however, that some of the blood in the kidney may find its way into the gill directly, as the rectal sinus was adjacent to the kidney on part of its course. The rectal sinus proper appears to be cut off from the other venous sinuses, but to be in communication with them by means of a loose lacunar system. From the rectal sinus afferent vessels run to the gill and pass along the base of the gill-lamellae (q.v.). The arterialized blood is carried from the gill by the pulmonary vein. It leaves the ventricle by the aorta, which divides into anterior and © posterior branches. The first-named runs forward along the wall of the pericardium sending out branches in its course. The posterior branch passes between the stomach and intestine and probably enters a lacuna in this position. From the lacuna is given off among other vessels a clearly-marked genital artery which can be traced backwards to finer branches distributed to the various processes of the gonad. (3) Respiratory System. The gill (Text-fig. 6) is monopectinate and composed of broadly-triangular plates hanging freely in the pallial cavity. On the rectal side the extremity of each plate is free for a short distance. The histology and structure correspond in general with that of Bythinella, but Bregenzer does not indicate the relation of the gill to the adjacent parts of the circulatory system. On the whole it would seem that in P. ventrosa the apical portion of the lamellae is more elongate and triangular in transverse section than in other forms. It should be ANATOMY OF PALUDESTRINA 173 noted that in this form and Bythinella the cilia are not distributed all over the lamellae as in Vitrella but are concentrated at the apex. Hach plate shows _longi- tudinal folds towards its basal part, becoming flatter where they jom up with the efferent vessel. In Bythinella, Vitrella, and Paludestrina jenkinsi the lamellae are flat and unfolded. This folding seems rather difficult to explain. Were it not for the fact that similar folding occurs in P. ulvae (Henking) I would be inclined to think TEXT-FIG. 6. Gills in section. av, afferent vessel; ev, efferent vessel; pv, pul- monary vein. that it might be due to shrinkage arising from excessive contraction of the transverse muscles in each filament. But in addition to the occurrence of similar folds in P. ulvae there is the fact that, if it were due to shrinkage, one would expect such contraction to take effect over all the lamella, which is not the case. If this ultimately proves to be an invariable character of these two species of Paludestrina it may very well be correlated with their brackish-water and marine habitat. It should also be pointed out that unless the text-figures G and H in Bregenzer’s paper are diagram- matic the afferent vessels and blood-spaces are much smaller in P. ventrosa than in Bythinella. N 2 174 GUY C. ROBSON From the rectal sinus blood passes by rather irregular and inconstant lacunar spaces to the roof of the pallial cavity, and finally into more definite and constant lacunae at the base of each gill-lamella. Thence it passes to the afferent vessels, and from these through the blood-spaces in the lamellae to the efferent vessels at the apex of the plates. The efferent vessels of all the lamellae are united on the left-hand margin of the gill by the pulmonary vein. On the left-hand side the afferent vessels apparently lose themselves in lacunae. On the rectal side the efferent vessels come to an end in the free portion of each lamella. It is probable that blood is brought from the left-hand side of the mantle cavity in a wide irregular lacunar system which ultimately debouches into the same sub-lamellar spaces as the blood from the rectal sinus. 4. Renal System (Text-figs. 5 and 7). The kidney lies between the body-wall, the pallial cavity, and the pericardium, sending ramifications among some of the other organs as in P. jenkinsi. It communicates with the pallial cavity at its own anterior end by a ciliated aperture furnished with sphincter and dilator muscles. In general the kidney is a thin-walled cavity with a lining of special secretory vacuolated cells as figured and described by Bregenzer (6, p. 248). Anteriorly, however, there is a special area between the body-wall and the pericardium characterized by a compact stroma of connective tissue with blood cavities communicating with the portal vem. This is the ‘ blood-gland’ of various authors (cf. Simroth, 21, p. 564). It is presentin Bythinella but absent in Paludina, Cyclostoma, and other forms (Simroth, loc. cit.). It is not specifically described in Vitrella, and it is not clear from Seibold’s description if it actually occurs. With regard to the epithelium covering this gland on the renal side I do not find the condition described by Bregenzer. The lining is usually a flat epithelium with flattened or roundish nuclei (fig. 7). I have never found the peculiar epithelium fizured by Bregenzer (loc. cit., Pl. xvi, fig. 15). ANATOMY OF PALUDESTRINA 175 5. Reproductive System. (1) Female Organs (Text-figs. 8, 9, 10). The ovary lies as usual between the liver and columella muscle. It consists of branched tubular follicles. The material employed for this study was all collected and preserved during May and June when apparently the odgonia were approaching maturity, but were not being shed in any number. A few TEXT-FIG. 7. Renal aperture and ‘ blood-gland’ in section. ck, cavity of kidney ; b, blood-gland ; r.o., renal orifice ; s, secretory epithelium. were found in the oviduct, and in a small number of cases spermatozoa were found in the receptaculum seminis. It may be therefore considered that, speaking generally, the material examined represented a stage coincident with the beginning of the breeding season. A great diversity of cellular elements was found in the ovary. The following types were invariably distinguished (‘T'ext-fig. 12): (a) Ripe odgonia distinguished by their large size, large 176 GUY C. ROBSON yolk content, usually with a clear slightly-granular nucleus and a deeply-staining nucleolus. (b) Ovarian cells only distinguished from (a) by the less intense staining of the yolk and their smaller size. (c) Small cells of various sizes, free or attached to the epithe- lium of the follicles with darkly-stainmg cytoplasm, clear nucleus, and dark nucleolus. TExt-FIGs. 8-9. Fig. 8.—Female genitalia. a, ac- Fig. 9.—Section of oviducal cessory gland; 0, oviduct; og, gland. m, outer muscular oviducal gland ; rs, receptaculum layer ; g, gland. Seminis ; v, vagina. (d) Cells of the germinal epithelium in various stages, either very small and irregular or enlarged and approximating to (¢). The germinal epithelium was never found in the regular columnar condition seen in Bregenzer’s figure U; and it is sometimes very difficult to interpret, beg full of masses of deeply-staining material of irregular disposition and uncertain nature and often flattened out by the pressure of the ripening odgonia. The various types of ovarian cell with all the inter- mediate stages are frequently met with in one and the same follicle, and the gradual transition seems to indicate the ANATOMY OF PALUDESTRINA iT development of one main type of cell from the germinal epithelium, viz. odgonia. No traces of nurse-cells could be found. The oviduct follows the usual course down the columellar region accompanied by the genital artery. It ultimately becomes thick-walled and convoluted. It gives off in succession a receptaculum seminis and an oviducal gland, Trext-Fic. 10. agi SS og Cee, Pome e001 TOUR OGUI Ir Portion of accessory ? glands in section. 6, darkly-staining area ; 1, purple-staining area ; c, cavity continuous with vagina. and opens into the vagina close to the entrance of the accessory glands of the latter. In its upper course its walls consist of a single layer of flattish epithelial cells. In the neighbourhood of the kidney its walls are formed of deeper and more columnar cells which contain at their apices (i.e. towards the lumen) a darkish secretion. They are ciliated and covered by an external layer. The receptaculum seminis is rather club-shaped and has a short duct. It is surrounded by a muscular layer. The cells are columnar with basal nucleus and their structure 178 GUY C. ROBSON seems to indicate a glandular nature. Very occasionally spermatozoa were found in the receptaculum aggregated imto small subcircular clumps. (2) The Oviducal and Accessory Glands. Some excuse is perhaps required for cumbering nomenclature with an additional obscurity. The appendage (Text-fig. 9) borne upon the oviduct just below the receptaculum seminis is called by Seibold the ‘Anhangsdriise des Receptaculum seminis’’ and by Bregenzer the ‘ Hiweissdriise’. The latter’s figures are not a sufficient indication whether structurally the organs are similar in Bythinella and P. ventrosa. Devoting our attention to the latter we find the ‘ oviducal gland’, as I prefer to call it, to be covered by a strongly- developed muscular sheath with circular muscle fibres. In general form it is an irregular-shaped gland with a short duct. Internally it is very much folded. The cells of its inner layer when not loaded with secretion are tall and narrow. ‘The nuclei are basal, and, when the cells are full of secretion, they become driven close up against the basal membrane. There are not very many accounts of the albumen gland in Gastropoda. But from those available we can safely assume that we are hardly warranted in calling this structure in Paludestrina by that name. In Valvata (1) on the one hand and Physa (22) on the other we see radically different types of ‘albumen gland’, and we can identify this form with neither.1. Until more is known of this structure in Gastropoda, and particularly in Prosobranchia, it is perhaps better to avoid a too positive terminology. The vagina is a narrow slit-like cavity surrounded by a large accessory glandular mass. It is thin-walled and ciliated internally. The glandular mass is very interesting but difficult to interpret. Previous authors of recent work upon Taenioglossa do not discuss it at any length, though Seibold poited out that differences of staining could be observed init. Subject to cer- tain qualifications, we may state that this mass is divisible most 1 Cf. Slugocka’s Pl. iv, fig. 20. ANATOMY OF PALUDESTRINA 179 frequently into two parts which occupy more or less opposite sides of the vagina and, where they meet, show a certain amount of transition in their structure. One portion is usually stamed in haematoxylin and eosin a vivid light purple in which the pink tinge predominates. It consists of two kinds of cells. A layer of ciliated, cubical cells lines the cavity of the gland. Some of them are drawn out into irregular, elongate extensions with which are asso- ciated other rather elongate cells. These form irregular digitiform glandular masses. A distinct lumen is seen in these masses (‘Text-fig. 10). It 1s uncertain how they pass their secretion to the exterior, as I have never observed a com- munication between the lumen and the exterior. The second area usually stains a deep purplish blue with the same stain. Seen in its most characteristic form it is composed of the same columnar ciliated cells and an inner glandular mass. The latter is more compact, the nuclei of the constituent cells are fewer and often arranged at the periphery of rudely quadrate masses. One is tempted to conclude that this second portion only represents another stage of the condition observed in the first described part, and that in the one the cells are full of secretion and tend to obscure a structure like that described in the first case. In the compact portion it is very hard to make out cell outlines, and certainly nothing like the digitiform glandular processes can be seen. It is, on the other hand, very certain that in certain areas transitional masses are to be found. I am inclined on the whole to consider that there are two functionally distinct portions of this gland mass, though inter- mediate stages are found. A comparison may be made with the rather similar structure of the accessory glands (odtype, shell- gland, &c.) of Neritacea which have been described by Bourne (8). The cavity of the vagina is continuous with those of the glands. (3) Male Organs (Text-fig. 11). The testis consists of a number of branching follicular tubes and in general plan resembles the ovary. Only one kind of 180 GUY CG. ROBSON spermatozoon was found, viz. the ‘typical’. The definitive stage of the latter, which is found in the vas deferens and the receptaculum seminis of the female, exhibits an elongate conical ‘head’, a usually well-developed acrosome, an acute apical portion, no discernible middle-piece, and an elongate tail. ‘The precise length of the latter could not be very satisfactorily ascertamed, but it is apparently very much longer than that of Bythinella, in which the tail is between twice and thrice as large as the head. In P. ulvae and P. taylori (Robson MS.) the tail is relatively enormous. One of the constant features of spermatogenesis is the occur- Trxt-Fies. 11-12. Fig. 11.—Section of penis. a, free Fig.12.—Transversesectionthrough portion of ‘ appendage’ ; vd, vas ovarian follicle. deferens. rence in the spermatids of an arrangement of the chromatin of the nucleus in bent rods or half-hoops at the periphery of the nucleus. My friend Dr. J. B. Gatenby has pointed out to me the rather similar concentration of chromatin at the posterior part of the nucleus in the spermatid of Murex trunculus recorded by Schitz (19). I am also indebted to Dr. Gatenby for pointing out to me the frequent occurrence of abnormal stages of spermatocytes, though of course, as has been stated above, the spermatozoa are monotypic. The vas deferens is thin walled during the first part of its course. It passes down the columellar region and in the neighbourhood of the kidney gives rise to a large glandular swelling, the prostate. The latter has plicate walls im- ANATOMY OF PALUDESTRINA 181 teriorly, lined with columnar ciliated cells with more or less basal nuclei. The rest of the structure of this gland, which stains violet with haematoxylin and eosin, is not unlike that of the lighter-staining portion of the accessory gland of the female. Below the prostate the vas deferens becomes smaller, thick- walled, and ciliated. It eventually runs just below the epidermis in the floor of the pallial cavity to the penis, which it traverses up to its apex. The penis is single in contrast with the remarkable complexity of Bythinella and Bythinia (Moquin Tandon, 18), in which a flagellum and a second branch occurs. It therefore exhibits the condi- tion seen in Cyclostoma (Garnault) and Vitrella. In P. ulvae the penis is quite simple according to Henking, while drawings made from the livmg animal by my friend Dr. H. Quick also show no accessory structures upon the male organ. The intromittent portion in P. ventrosa is long and pointed. 3. HABITS, ETC. A preliminary attempt has been made (Robson, 17) to analyse the ecological conditions under which P. ventrosa is found. But a great deal remains to be done upon this subject as well as upon the distribution and ecology of the plants associated with it and upon which it may be presumed to depend. Though a more definitely brackish-water form than P. ulvae, the case worked out at Leigh-on-Sea demonstrated a greater adaptability and tolerance on the part of P. ven- trosa. If, as we may rightly assume, the British Paludes- trinidae show a progressive tendency to become adapted to fresh-water, P. ventrosa represents an intermediate stage of adaptation, but exhibits the tendency in its initial rather than its later stages. Little can be said upon the more intimate habits of this animal. It is usually found upon some water- plant, but quite frequently upon mud or bottom débris. In several examples from Leigh, Wakering Wick, and elsewhere, the stomach contained a variety of diatoms and a few foramuini- 182 GUY C. ROBSON fera. The rest of the contents were usually too much digested to enable their nature to be made out. No remains of plant fibre, &c., was ever found. I am inclined to think that it browses upon the microfauna and microflora of the plants upon which it lives, and that it does not actually chew the leaves of the latter. 4. AFFINITIES. (a) I cannot agree with Bregenzer’s verdict upon the imme- diate relationships of Paludestrina (6, p. 276). According to her, the latter genus is separated into a group distinct from Bythinella and Vitrella upon the following characters : (1) Fusion of the cerebral and pleural ganglia. (2) The possession of two pairs of salivary glands. (3) Reduction of the ‘ Knorpelspange’ of the lingual cartilages. (4) Brackish-water habitat. Of these characters the first is open to question. In P. ven- trosa there is no more fusion of the ganglia in question than in Bythinella, while in P. ulvae we have seen (p. 168) that Henking’s statements are open to question. In the second place, only one pair of salivary glands is foundin P. ventrosa. As to the third character it is scarcely worth anything as the ‘ Knorpelspange’ is absentin Vitrella! Lastly, Paludes- trina is not restricted to brackish water, at least as far as England is concerned. As the result of a scrutiny of the characters available for taxonomic purposes we might with equal justification select the simple penis and longer super- intestinal part of the visceral commissure in order to unite Paludestrina with Vitrella as against Bythinella, or the crystalline style and certain features of the radula to unite Paludestrina and Bythinella against Vitrella. In any case I venture to think that an animal’s taxonomic position cannot be summarily decided in this fashion. Until we have objective evidence as to the taxonomic value of the various characters such groupings as those discussed above are of little value. On the whole we can safely consider these ANATOMY OF PALUDESTRINA 188 genera as referable to the same family ; but I feel that we require another technique for deciding their closer affinities than a mere inspection and assorting of characters. A character such as the absence of the crystalline style in Vitrella would+appear in the first instance to be profoundly important. But we do not know the precise significance of its absence. As an alternative to a close study of genetics, evolution, habits, and ecology in relation to structure which alone can give us a sound taxonomic method, the only procedure that could be suggested would be a complete enumeration of characters and a grouping based upon agreement or disagreement in a large number of structures. This method would be crude, but it would be better than an arbitrary selection of a few characters. In the present case I have distinguished a total of twenty-one important characters. The agreement or disagreement of the three genera in question is indicated as follows : Paludestrina=Bythinella alone in 5 3 =Vitrelta alone in a | 2s =Bythinella and Vitrella in 4 21 se =neither in 8 relationship uncertain in 9] (b) Though it would be beyond the scope of this paper to offer a criticism of the present arrangement of the Taenioglossa, we may nevertheless attempt to define the position of the Paludestrinidae with regard to some of the main tendencies of Prosobranch morphology. The Paludestrinidae represent a stage in the abbreviation of the nervous system which involves the pleural-intestinal portions, and is seen in its extreme condition in Bithynia and Valvyata in which the sub- and supra-intestinal ganglia are either fused or closely approximated to the pleural ganglia. In Melania and Cerithium this condition of close approximation is seen on one side only, the ganglia being separated on the other side. In Paludestrina and Bythinella they are slightly separated on both sides, while in Littorina and Paludina they are widely separated. 184 GUY CG. ROBSON Paludestrina agrees with Pterocera, Tiphobia, Lithoglyphus, and a few others in possessing a crystalline style. We may assume, however, that this is without phylo- genetic significance within the group. Another interesting tendency which may not be of phylo- genetic importance is the possession of a single otolith in the Paludestrinidae. It shares this character with Littorina, Trunecatella, some Melanias, and Natica. On the other hand, Paludina, Ampullaria, Valvata, Cyclo- phorus, and others have multiple otoconia. A blood-gland is absent from the kidney of Paludina, Valvata, Ceri- thium, &e., and is found in Littorina and in the Paludes- trinidae. Finally, while possessing a simple osphradium, Paludes- trina exhibits a definite osphradial ganglion—a stage appar- ently more advanced than such forms as Littorina and Bithynia, im which (Bernard, 2) no osphradial ganglion is found. 5. SUMMARY. 1. Paludestrina ventrosa_ possesses the general Taenioglossate organization. 2. It represents a genus of Paludestrinidae equivalent to Bythinella and Vitrella. 8. It is peculiar within the family as possessing : i. Folded gills ; ii. A slit connecting the style sac through nearly all its length with the intestine ; in. A typhlosole ; iv. A non-ciliated roof to the median part of the pharynx. It represents an intermediate stage in the acquirement of the fresh-water mode of life, bemg essentially a brackish-water form with a fairly well-marked euryhaline tendency. 4. Several structures not fully described by previous authors are discussed in this paper (e.g. the accessory female and circu- latory organs), and it is not certain in what form these struc- tures occur in other Taenioglossa, Ee 2 ————————EE—— ANATOMY OF PALUDESTRINA 185 5. Within the Order Taenioglossa, Paludestrina is referable to the group which possesses : pe aS ee a a a Onan — & bd = oO 20. bo = 22. Do bo = CO (1) a brevicommissurate visceral commissure ; (2) a single otolith ; (3) an osphradium with basal ganglion ; (4) a renal-portal system and blood-gland ; (5) an ‘oviducal’ gland immediately adjacent to the receptaculum seminis. 6. BIBLIOGRAPHY. . Bernard, F.—‘ Bull. Sci. France et Belgique ’, xxii, 1890. —— ‘Ann. Sci. Nat. Zool.’ (7), ix, 1890. Bourne, G. C.—‘ Proc. Zool. Soc.’, 62, 1908. Boycott, A. E.—‘ Journ. Conchology ’, xvi, no. 2, 1919. Bouvier, E.—‘ Ann. Sci. Nat., Zool.’ (7), iii, 1887. . Bregenzer, Aloys.—‘ Zool. Jahrb.’, 39, Heft 2, 1916. Edmondson, C. H.—‘ Journ. Exp. Zool.’, 30, 1920. . Garnault, P.—‘ Actes Soc. Linn. Bordeaux ’, xli, 1887. Gwynn Jeffreys, J.—‘ British Conchology ’, 1, 1862. Henking, H.—‘ Ber. Naturf. Ges. Freiburg’, viii, 1894. . Lang, A.—‘ Lehrbuch d. Vergl. Anatomie (Mollusca) ’, 1900. Moore, J. E. S.—‘* Quart. Journ. Micr, Sci.’, 41 (N.S.), 1899. . Moquin Tandon, A.—‘ Hist. Nat. Moll. France ’, Paris, 1855. Nelson, T. C.—‘* Journ. Morphology ’, 31, 1918. . Pelseneer, P.— Lankester’s Treatise on Zoology (Mollusca) ’, 1906. . Robson, G. C.—‘ Ann. Mag. Nat. Hist.’ (9), 5, 1920. Thid., 6, 1920. . Rougemont, Ph.—‘* Etude de la faune des eaux privées de lumiére ’, Neuchatel, 1876. . Schitz, Vim Arch. Zool. Exp.’, 59, fase. 2, 1920. Seibold, W.—‘ Jahreshefte Ver. Vat. Naturk. Wiirtemberg ’, 60, 1904. . Simroth, H.—‘ Bronn’s Klassen und Ordnungen des Thierreichs’, Bd. 3, 1902. Slugocka, M.— Revue Suisse Zool.’, 21, 1913. . Vayssiére, A.—‘* Journ. Conchyliol.’, 33, 1885. 24. Woodward, B. B.—‘ Ann. Mag. Nat. Hist.’, ix, 1892, > & aN f PLE Ye ryPan i} 4 f © F j i ’ : 7H OH Aavea wl oetal, ole ives voc toe pay eh ORT? £1 i 56 Y i ye hy ol gate =Qa8 j a iga. ag ~) ¥ The Gastric Mucosa. By Robert K. S. Lim. (From the Department of Physiology, Edinburgh University.) With Plate 8 and 1 Text-figure. INTRODUCTION THE gastric mucous membrane is described as being disposed in three regions, known as the cardiac, fundic, and pyloric. These regions, although distinguished from one another by definite microscopic characters, yet merge gradually the one into the other, so as to present no well-defined lines of demarca- tion. The actual extent of each region varies in different animals. It has not been sufficiently recognized, however, that the cardiac and pyloric areas are very small, especially in the carnivora. In the cat, the (microscopic) pyloric region is a narrow zone, extending for not more than 35 mm. from the junction of pylorus and duodenuin; it may not even correspond in extent with the so-called pyloric antrum. In view of this fact it is possible to doubt the exactness with which pure pyloric pouches can be isolated either by the Heidenhain (18) or Pavlov (21) technique. What is known regarding the functions of the different regions of the stomach is not compatible with their differences in structure. And current descriptions of the cells forming the gastric glands are by no means uniform, much confusion bemg due to the fact that the histological descriptions vary according to the method of fixation and staining employed and the species of animal investigated. The following description of the histology of the gastric glands is divided into three parts. In Part I only the gastric mucosa of the cat will be described, since the observations NO, 262 oO 188 ROBERT K. 8S. LIM I have made upon its stomach are more complete than in the other cases. Other animals, both adult and foetal, have, however, also been investigated, and the special features of some of the cells of their glands are described in Parts II and IIT. PART I. THE GASTRIC MUCOSA OF THE CAT. The cats were killed both while fasting and at various intervals after a meal. They were usually fed cn boiled fish, milk, and bread, but some were put on a meat and milk diet. In all twenty-five animals have been examined. HistoLoaicaL ‘TECHNIQUE. For microscopical purposes the animals were killed either by carbon monoxide or chloroform. ‘The stomach was then examined fresh or was prepared for sections. For the fresh preparations a piece of the mucous membrane was either scraped off and teased in Ringer or serum, or the fresh tissue was frozen in a little serum and cut up with a micro- tome. The fresh sections, however, gave no more information than those obtained after fixation, so that this method was discontinued. For permanent preparations the fixatives used were Zenker, Altmann’s fluid, osmic acid 1 per cent. and formol (either neutral 20 per cent. or acid 10 per cent.). When Zenker or formol was employed the stomach was slightly. distended with the fixative and suspended in the same solution for the period necessary for penetration. It was then cut into suitable pieces, which were either placed in gum or carried through in the usual way into paraffin. For some preparations pieces of fresh stomach were pinned out on a cork and immersed in the fixing reagent; this was the chief method when using osmic acid solutions, but a few pieces were fixed in osmi¢ without stretching. The stains employed were alcoholic eosin and methylene blue (16), haematoxylin and eosin, van Gieson, iron haema- THE GASTRIC MUCOSA 189 toxylin (Heidenhain), Mallory (24), or polychrome methylene blue. It has not been thought necessary to give details of the application of the above stains ; they may be found in the refer- ences indicated. In the description which follows acid formol fixation is implied, although the observations recorded have been corroborated by other methods. Where a notable differ- ence occurs the special fixative concerned is mentioned. Tur Mucous MEMBRANE AS A WHOLE. It is not mtended to describe the naked-eye appearances. Suffice it to say that with a lens (Sprott Boyd (28)) differences may be noted between the duct orifices of the pyloric region and those of the remainder of the stomach. In the former the mucous membrane is thicker and the ducts wider, longer, and more funnel-shaped than in the latter. The gland-tubes are simple, but may branch slightly towards their blind ends. Several gland-tubes are usually served by a common duct. Only in the part of the pyloric canal close to the duodenum do the glands become markedly racemose, but the glands adjacent to the oesophagus may also take on a racemose character. A gastric gland-tube may be described as consisting, besides the duct, of a superficial part, which is the portion of the gland-tube immediately below the duct, and a deep part composed of the remaining portion of the gland. THE CONNECTIVE TISSUE. Between the glands hes the supporting connective tissue Gnterglandular tissue) which contains plain muscle-fibres arranged vertically, blood-vessels, lymphatics, and nerves. In addition to these there are three kinds of cells in the tissue : (1) Finely Granular Branched Connective- tissue Cells.—These stain a deep magenta with poly- chrome methylene blue and a purplish blue with alcoholic eosin and methylene blue. They form by far the most numerous yariety and are more numerous in the stomach than in other O02 190 ROBERT K. 8S. LIM portions of the digestive tract. This has already been noted by Cade (5). (2) Finely Granular Oxyphil Leucocytes.—These are sometimes massed together in groups: more generally they are scattered throughout the mucosa. (3) Coarsely Granular Eosinophil Cells (PI. 8, fig. 5, g). These are present in least numbers. They oceur mainly near the surface, and may be found between the cells lining the duct of the gland as well as in the interglandular tissue. The eosinophil granules or globules vary considerably both in number and size, some being as much as 2-3p in diameter. They stain with iron haematoxylin, which does not colour the oxyphil granules of leucocytes ; they are thus not unlike the cells of Paneth of the small intestine. All three types may be found in the interglandular tissue of other animals, e.g. dog, pig, and rabbit. The interglandular tissue is more abundant at the cardiac and pyloric ends of the stomach than in the middle of the fundic region. Here the connective tissue is more plentiful immediately under the surface epithelium. The mucosa rests on a thick condensation (membrane of Zeissl, stratum compactum of Oppel (Text-fig. 1, a, se)) of white fibrous tissue, immediately underneath which hes the muscularis mucosae. This membrane-like condensation is of interest as it is not common to all animals, e.g. it 1s absent in man, pig, and rabbit, but is present in cat and rat. Further, it is non-elastic and separates the muscle-fibres within the inter- glandular tissue from the muscularis mucosae. It is_per- forated by vessels, and the plain muscle-fibres reach the mucosa by the same communications. THe SuRFACE EPprrHELIUM. This epithelium includes the cells covering the surface and those lining the ducts. These cells are essentially of one type. Those on the surface are columnar, becoming shorter and more cubical as they are traced into the ducts. A corresponding change may also be noted in the nucleus, which is elongated eo} bead THE GASTRIC MUCOSA 1 on the surface but almost spherical within the ducts (see evs, fig. 5). The cytoplasm is finely granular in the fresh and certain fixed (neutral formol, osmic) specimens, and may be differen- TExt-FIc. 1. B. High power (x 400). p, gland lined with peptic cells; m, gland = mm, lined with mucoid cells ; oxyntic : : - cells occur in the parietal parts of A. Low power (x 75). sc, stratum com- both glands, pactum ; mm, muscularis mucosae. Glands from the middle of the anterior surface of the stomach. Cat 10; 24 hrs.; acid formol; haematoxylin and eosin. (Photograph.) tiated into two parts (Ellenberger and Scheunert (11) by staining methods, viz. an outer goblet-shaped part, which is clear but tinted red in haematoxylin and eosin preparations, stained blue by Mallory and a pale blue by polychrome methylene blue, and an inner part consisting of the remainder 192 ROBERT K. 8. LIM of the cell, in which the nucleus is situated, and which is stained of a reddish colour by Mallory. The surface cells show a larger goblet part than the duct-cells (see Pl. 8, figs. 3 and 5). During active digestion this part diminishes in size, but in both fasting and feeding animals cells in which the goblet part is defined but not stamed may be seen. This presumably indicates that the cells in question have discharged their contents and have not had time to supply the part with new material (granules). With regard to the mode of attachment of the cells to one another I have sometimes observed the intercellular bridges described by Carlier (5a). These, however, are only apparent when the cells appear unduly vacuolated. In sections tan- gential to the surface [ have seen no indications of bridges. The surface epithelium is continuous with the epithelium of the gland-tubes, the transitional cells losing their goblet portions and staining a uniform bluish colour with Mallory. The transition is short (see Pl. 8, fig. 5, ¢). Tse Carprac REGION. The junction of the oesophagus and the stomach is well defined in the cat, the stratified epithelium of the former stop- ping abruptly and being replaced by the columnar epithelium of the latter. At this junction a lymph follicle may sometimes be seen, but there is more frequently a large vesicle or cavity lined by one or two layers of cubical cells. The cardiac region (when present) is extremely narrow, measuring about 2-3 mm. from the cardio-oesophageal junction to the nearest group of parietal or oxyntic cells. It meludes only cells of one type (cardiac cells) unmixed with others. Beyond this there is a boundary zone extending for ancther 3mm., which contains both oxyntic and cardiac cells. Frequently there is no definable cardiac area; oxyntic cells are found at the junction itself and only the ‘ boundary zone ’ is present. Beyond the boundary zone another type of cell, characteristic of the fundus, is met with ; this may be regarded as the cardiac limit of the fundic region. THE GASTRIC MUCOSA 193 The glands of the cardiac region consist of relatively simple tubes, with short ducts and somewhat wide lumina. In most animals they are fairly numerous, in others only a few such glands are to be found near the oesophagus. ‘They are lined by a single layer of columnar or cubical epithelium, which appears granular in the fresh condition. In sections, however, sranules are absent and a fine reticulum is seen in its place. The reticulum is irregularly distributed throughout the cell, and is stained blue by alcoholic eosin and methylene blue, pale magenta by polychrome methylene blue, and blue with Mallory (Pl. 8, fig. 1). In some cases a reticulum which stains reddish with Mallory is present in addition to the above finer reticulum which stains blue. Haematoxylin hardly stains the ‘ blue’ reticulum at all, nor does it tint the spaces between the reticulum. In the case of the other stains just mentioned the spaces are coloured im the same way as the reticulum, but more faintly. The nucleus is irregularly rounded or ovoid and is invariably situated towards the base of the cell. In a fasting animal the cell is more columnar and the nucleus less flattened than in an animal which has been fed. On the whole, however, there is little change to be noted in these cells. No compound tubular glands such as have been described by Hllenberger (10), Edelmann (8), Schaffer (27), and others in various animals are present in the cat, nor are any structures resembling crypts of Lieberkthn met with; this also applies to other regions of the cat’s stomach. The simple tubular glands of the cardiac region were first described by Schafer and Williams (26) in the kangaroo, and with their description those of the cardiac region agree. It will be shown later that the cardiac cells do not constitute a special type, but form a variety of mucoid cells, a term which is explained elsewhere. Tur PytLoric REGION. The pyloric region is considerably larger than the cardiac in area, although smaller than is generally supposed. It extends 194 ROBERT K. §. LIM for about 15mm. from the pyloro-duodenal junction along the greater curvature and about 12-15 mm. along the lesser curvature. Beyond these limits small oxyntic cells make their appearance, and about 20 mm. further full-sized oxyntic and peptic cells are met with in large numbers. Here lies the pyloric limit of the fundic region. With regard to the general features of the pyloric glands, they have long and wide ducts and become more racemose and exhibit more interglandular tissue near the intestine. Lymph-follicles are numerous in this region of the stomach, several being invariably present at the pylorus itself. At the pyloro-duodenal junction the pyloric glands pass through the muscularis mucosae, which is here incomplete, and become Brunner’s glands of the duodenum. ‘The lumen of the glands is large, and this, along with their racemose character, serves to distinguish the pyloric glands from those of the cardia which they otherwise resemble. The glands of the pyloric region are lined by a single layer of cells, which are columnar or cubical in shape and irregularly reticulated (Pl. 8, fig. 4). They are stained in the same way as the cardiac cell, the whole cytoplasm appearing blue with methylene blue combinations and with Mallory, pale magenta with polychrome methylene blue, and colourless with haema- toxylin. As is the case with the cardiac cell, the basal portion of the cell may in some animals be occupied by a second reticulum which stains red with Mallory. This may be seen in fasting and fed animals, but more often in the latter con- dition. The nuclei are irregularly rounded and _ situated basally. During activity the cell becomes shorter, indicating a discharge of its contents, and the nucleus appears more spherical, i.e. less compressed. The similarity between the cardiac and pyloric glands has been noted by many observers (Cobelli (6), Ebstein (7%), Schaffer, Stohr (29), and others). Bensley (2, 8) compares the pyloric cells with the cells lining the ‘neck’ of the fundie gland as well as with the cardiac cells. On the other hand, Heidenhain (18), Langley and Sewall (15), Kranenberg (28), THE GASTRIC MUCOSA 195 and all later writers believe that they are fundamentally the same as the ‘chief’ cells of the fundus. It will be shown later that there can be no doubt regarding their difference from the ‘ chief’ cells, and their resemblance to the cardiac gland-cell is too close not to regard them as identical in structure if not in function. THE GLANDS or THE IuNDUs. Histologically, the portion of the stomach between the cardiac and pyloric regions just described has a uniform structure. The glands of this intermediate area are generally known as the glands of the fundus, though they might be more appropriately termed the glands of the body of the stomach. The general form and arrangement of the fundic glands have already been noted. ‘hey are simple tubes with short ducts, and as the glands are closely packed together there is little interglandular tissue. Three kinds of cell occur in the glands of this region, although hitherto, with the exception of Bensley (1) and Cade (5), histologists have recognized only two, namely ‘central’ or ‘ chief’ and ‘ parietal’ or ‘ superadded ’ cells. (1) Peptic Cells.—These are usually known as ‘ chief’ cells; they are quite distinct from a second type of central cell which are intermingled with them, and are described later as mucoid cells. Peptic cells occur throughout the lower or deep half of the gland-tube, although it is comparatively uncommon to find this part of the tube lined wholly by such cells. They look somewhat columnar in shape in section, but when isolated are polyhedral. The cytoplasm contains granules in the fresh state (Langley and Sewall) ; these are irregular in size. On examination in saline, weak acids, or alcohol, the granules tend to increase in size and become less distinct. Finally they disappear, apparently by passing into solution. A few granules always remain unatfected. In fixed preparations, whether formol, Zenker, or osmic, the granules are replaced by a coarse but regular reticulum (PI. 8, / 196 ROBERT K. 8. LIM figs. 2 and 8, p, xxi). Nevertheless, with both formol and osmic, a few granules may be preserved; this is especially the case after osmic fixation (PI. 8, fig. 3, p, xx, mm). The regularity of the reticulum suggests that the extra-granular cytoplasm is coagulated before the granules are dissolved out. The reticulum may therefore be taken as a rough index of the amount and size of the granules contained in the cell. With regard to their reaction to various dyes, both the reticulum and the granules become intensely stained blue with alcoholic eosin and methylene blue, deep purplish blue with polychrome methylene blue, and brownish violet with Mallory. They are only lightly stamed by haematoxylin, but more strongly so by the iron haematoxylin method. The nucleus is irregularly ovoid or rounded ; it varies in shape and position according to the activity of the cell. Functional changes are easily noted in these cells. In the fasting condition the nucleus is found towards the base of the cell and the cytoplasm is reticulated throughout. After a period of activity, i.e. during digestion, the cell gradually shrinks, and the nucleus becomes larger and occupies a more central position. ‘leased preparations seem to show that the granules are on the whole larger in the fresh condition, while in fixed specimens the meshes of the reticulum are wider. Ergasto- plasmic fibres occur at the base of the cell, while the reticula- tions (granules) diminish near the lumen of the gland. In some cases (five to six hours after a large meal) half of the cell may be occupied by ergastoplasmic fibres. These fibres stain in the same way as the reticulum, although more definitely than it (Pl. 8, fig. 3, p, mm, and fig. 5, p). Langley was the first to demonstrate the diminution of granwles durimg activity ; he also stated that the cells become clearer at their bases. Later Bensley, Zimmermann (80), and Theohari (28) showed that the basal clear zone is occupied by ergastoplasmic fibres (prozymogen of Macallum (19)). These observations I can confirm in the cat. The swelling of the granules during diges- tion appears to be a stage in the conversion of zymogen into soluble ferment and occurs more rapidly than the formation THE GASTRIC MUCOSA 197 of new granules. Hence the diminished reticulated area, and the absence of any increase in the size of the cell, contrary to Heidenhain’s observation. (2) Mucoid Cells.—This other type of central cell has somewhat finer granules, and when fixed they are replaced by a fine reticulum (perhaps a precipitate) (PI. 8, fig. 3, m, xx, xx1b). No granules ever remain intact after fixation. In the fresh condition these granules are more rapidly dissolved by reagents than those of the peptic cells ; this, perhaps, partly explains the entire absence of granules after fixation. Mucoid cells occur mainly in the superficial half of the gland-tube, but are interspaced among the coarser reticulated peptic cells towards the deeper part, and may be found throughout the whole gland-tube. In places a portion of a gland may be lined entirely by these cells. In form they are roughly globular, but variations in shape occur according to their position and fit in the tubule (PI. 8, fig. 5, m). Their staining reactions render them distinctive. They are coloured a pale blue by alcoholic eosin and methylene blue, a pale magenta by polychrome methylene blue, and a deep blue by Mallory ; as is the case with the fasting peptic cells, they are unaffected by haematoxylin. When a definite reticu- lum is present it stains blue with Mallory, but in some of the cells the basal portion takes on a brownish or even reddish tinge. When there is no reticulum the precipitate-like material invariably stains blue. The nucleus is small and compressed against the base of the cell: it is generally deeply stained. Changes during digestion consist in the cell becoming first larger and later smaller and staiming less heavily with Mallory, while the nucleus appears to be a little more prominent. Mucoid cells are most marked in the boundary zones, where they are continuous with the cardiac cells on the one side and the pyloric cells on the other. (3) Oxyntic Cells.—In the cat these cells are mostly found wedged in between the central cells with a corner abutting on the lumen; nevertheless, they le sufficiently far outwards to be termed parietal cells. They are most numerous 198 ROBERT K. 8. LIM in the superficial half of the gland, and may form the sole lining of a portion of the gland-tube. They may be found even between the columnar cclls of the gland-ducts. In shape (judging from vertical and transverse sections) they are roughly pyramidal, but there are many variations from ovoid to crescentic. Unlike the peptic and mucoid cells the granules of the oxyntic cells are very fine, and are not readily attacked by reagents. ‘They are fixed by all the methods employed ; with osmic acid those situated immediately underneath the membrane of the cell may be demonstrated to be lipoid in character. Similar observations have been made by Béhm and Davidoff (4) in the rat. The staiming reactions of the oxyntic cell-granules are as follows: red with alcoholic eosin and methylene blue, haematoxylin and eosin and Mallory ; pale blue with polychrome methylene blue ; and dark brown with osmic. The nucleus is spherical and usually central. Occasionally it is excentric or there may be two nuclei within the same cell. A number of the cats examined showed the presence of parasitic spirochaetes (lim (18)). ‘These organisms were sumetimes found within oxyntic cells in what appeared to be a single dilated canaliculus, continuity with the lumen of the gland being demonstrated. Otherwise there was no histo- logical disturbance. Vacuoles may often be seen within the oxyntic cells of all animals. With regard to functional changes, oxyntic cells appear on the whole to become larger (Heidenhain) during digestion and their granules more easily distinguished, being less closely packed together and probably fewer in number. ‘The differ- ence, however, is not marked, and may be partly due to shrinkage of the central cells. It ought to be noted that oxyntic cells occur throughout the whole stomach, bemg absent only some 3 mim. from the oesophagus and about 15 mm. from the pyloro-duodenal junction. The oxyntic cells of the pyloric boundary zone are somewhat small in size and are situated mainly in the super- ficial portion of the gland; they are probably primitive in THE GASTRIC MUCOSA 199 character. These have already been described in other animals (Stohr (29), Trinkler (23), Nussbaum (20)). Nussbaum, how- ever, does not consider these smaller cells to be the same as oxyntic cells. GENERAL CONSIDERATIONS. These observations show firstly that the term ‘chief’ or ‘central’ cells is inadequate, since there are two types differing widely from each other. Secondly that the cells of the cardiac and pyloric regions are similar in structure and of the same characteristics as the mucoid cells of the fundus. Thirdly that the fundus is the all-important region of the stomach from the point of view of the secretion of gastric juice, the other two regions being small by comparison and containing no recognizable zymogen-secreting cells. Let us first consider the characters of the two types of central cells. We have seen that the peptic cell is granular (or reticu- lated) and that after a period of activity the granules diminish in number and are replaced at the base of the cell by ergasto- plasmic fibres. In the case of the mucoid cell the cytoplasm is also granular (when fresh), but functional changes do not cause any alteration in its architecture. ‘The nucleus of the peptic cell at rest is irregularly rounded or ovoid, and is applied against the basement membrane, but during digestion is more regular in outline and frees itself from the base so far as to occupy a more central position. The mucoid cell-nucleus, on the other hand, is not markedly changed either in shape or position. There are also the differences in staining reactions. The peptic cell is coloured in an entirely different manner from that of the mucoid cell (compare m and p, Pl. 8, fig. 3). This difference is manifested not with one staining method alone but with several, although Mallory’s is the best for the purpose. Both types cf central cell may be seen in man, dog, and rabbit (and also in the frog) ; they are prebably common to all mammals. There can thus be no doubt regarding the separate existence of these two types of cells. Hdinger’s theory that all the 2.00 ROBERT K. 8. LIM varieties of cells found in the stomach are functional modifica- tions of one type is untenable. It is impossible to reconcile this view with the differences in structure and reactions in both fasting and feeding animals. Heidenhain (18) long ago observed that some chief cells stain more readily with aniline blue than others —and referred this to functional changes. This was later confirmed by Green- wood (12) in the pig’s stomach ; she suggested that the ‘ clear ’ cells might be mucus cells, thus anticipating the results of two subsequent observers. Both Bensley and Cade have distin- euished two types of central cells (older observers from Edinger (9) and Pilliet (22) downwards have found various modifications of the central cells but not separate types), which appear to be similar to the peptic and mucoid varieties described here. Bensley was the first to note that the cells of the ‘neck’ region of the fundic glands stain in the same manner as mucus-secreting cells; these cells he termed ‘ in- dulinophilous mucous cells’. Cade confirmed Bensley’s finding with indulin and called them ‘ cellules principales du col’. In the cat the neck region is lined by oxyntic and transitional cells, i.e. cells which have almost lost the division of the cytoplasm into two zones so characteristic of the surface mucous cells (see PI. 8, fig. 5, t). It is the portion of the gland below the neck, therefore, that is lined chiefly by mucoid cells (see Pl. 8, fig. 5). Bensley (2) does state, however, that an occasional ‘ indulinophilous cell’ may be found among the central (peptic) cells of the deeper part of the gland, and from an examination of his figures (Pl. 8. fig. 6) it is clear that the neck region he describes includes the superficial portion of the eland. ‘To him credit is due for their discovery, although a more definite description and wider distribution of the mucoid cells must now be recognized. ‘Mucoid’ cells are described in only two text-books in English, Schafer’s ‘Essentials of Histology’ (26), and the American edition of Béhm and Davidoff, translated by Huber (4). Of continental works I can only find a mention in Prenant, Bouin, and Maillard (28), who have an excellent diagram in THE GASTRIC MUCOSA 901 their text-book of Histology showing typical mucoid cells— which they hesitatingly label ‘ cellules principales muqueuses ? ’ to illustrate the mucus cells of Bensley. It is evident. therefrom, that hitherto the distribution and even the existence of mucoid cells have scarcely been recognized. The names ‘ peptic’ and ‘mucoid’ have been chosen for chvious reasons. The structure of the peptic cell is charac- teristically that of a zymogen-secreting cell, and by the term ‘chief’ or ‘ central’ this cell was meant, so that there is no need to dispute its function. The term mucoid is applied because the cell resembles other mucus-secreting cells, but it is not identical either with the mucus-secreting cells lining the surface or with the goblet cells of the mtestine (compare cells m and sin PI. 8, fig. 3 ; also see Lim (17)). We may next consider the relation between the cardiac, pyloric, and mucoid cells. We have seen that there is little or no difference structurally between the two former (cardiac and pyloric) cells, and that the mucoid cells resemble them in most respects except position. They are staimed in the same way, and their structural characters are very similar both during rest and activity. The cardiac and pyloric cells show in some aninals a reddish basal reticulum ; this may or may not constitute a difference, although it is to be noted that the reticulum is more frequently absent than present. Lastly, they are continuous with each other, for cardiac cells can be traced into the fundus in the form of mucoid cells: the same applies to pyloric cells. ‘he close resemblance which thus exists between these three types (they are all obviously mucoid) presumes a similarity im their functions. The striking differences in structure between the peptic and pyloric cells have been quite missed by all the workers on pyloric pouches, and it is possible that their histological examination was inadequate to ensure the purity of the pouches which they made. but apart from this the suggestion that pepsin is secreted by cells which are not typical of the zymogen-secreting type calls for a closer investigation inte the origin of the secretion of the pyloric pouches, 202 ROBERT K. §S. LIM SuMMARY. The gastric mucous membrane is principally formed by relatively simple tubular glands which become more complex near the orifices of the viscus, especially near the pylorus. The glands are lined by one or more kinds of cells ; the following types may be recognized : 1. Surface mucus-secreting cells, which include the cells lining the surface and the gland-ducts leading therefrom. 2. Mucoid cells, of which there are two closely allied groups, V1Z. : (a) The cardiac and pyloric cells which form the sole lining of the glands within about 0-2 mm. and 15 mm. of the oesophageal orifices respectively. (b) The mucoid cells proper, which occur in the large inter- vening region (fundus) where they are intermingled with the peptic and oxyntic cells; they chiefly occupy the superficial or upper part of the gland-tube. 8. Peptic cells, which are found (often in conjunction with mucoid cells) within the deep part of the gland; both peptic and mucoid cells were formerly described as ‘ chief ’ or ‘ central’ cells. 4. Oxyntie cells, which chiefly oceupy the upper portion of the gland where they are found between the mucoid cells ; in the deeper portion of the gland they take up a parietal position. The interglandular tissue contains basiphil connective-tissue cells, oxyphil leucocytes, and a few cells with large eosinophil clobules. LITERATURE. 1. Bensley.—‘ Proc. Canad. Inst.’, 1897, i. 11. 2. —— ‘ Quart. Journ. Micr. Sci.’, 1898, xli. 361. 3. ‘ Amer. Journ, Anat.’, 1902, ii. 105. 4, Bohm and Davidoff.—‘ Text-book of Histology ’, 1914, 2nd edition, transl. by G. C. Huber, Philadelphia. 5. Cade.—‘ Arch. d’anat. micr.’, 1901, iv. 1. 5a, Carlier, quoted by Schafer.—‘ Text-book of Microscopic Anatomy ’, 1912, 528. London, THE GASTRIC MUCOSA 903 6. Cobelli. Wiener Sitzungsb.’, 1866, lii. 250. 7. Ebstein.—‘ Arch. f, mikr, Anat.’, 1870, vi. 515. 8. Edelmann.—‘ Deutsch. Zeit. f. Tiermed.’, 1889, xv. 165. 9. Edinger.—‘ Arch. f. mikr. Anat.’, 1880, xvii. 193. 10. Ellenberger—‘ Handb. d. vergleich. Histol. u. Physiol. d. Haus- saugetiere, 1890. Berlin. 11. Ellenberger u. Scheunert.—Ibid., 1910. 12. Greenwood.—‘ Journ. Physiol.’, 1884, v. 195. 18. Heidenhain, R.—‘ Arch. f. mikr. Anat.’, 1870, vi. 368; Hermann’s *Handb. d. Physiol.’, 1883, v. 91. 14, —— ‘ Pfliiger’s Arch.’, 1878, xviii. 169. 15. Langley and Sewall.— Journ. Physiol.’, 1879, 281. 16. Lim.—‘ Quart. Journ. Micr. Sci.’, 1919, Ixiii. 541. 17. ‘Proc, Physiol. Soc., Journ. Physiol.’, 1920, iii. 18. —— ‘ Parasitology ’, 1920, xii. 108. 19. Macallum, A. B.—‘ Trans. Canad. Inst.’, 1890, i. 20. Nussbaum.—‘ Arch. f. mikr. Anat.’, 1877, xiii. 721. 21. Pavlov.— The Work of the Digestive Glands’, 1910, 2nd edition, transl, by W. H. Thompson, London. 22. Pilliet.—‘ Journ. d. l’anat. norm. et path.’, 1887, v. 23. Prenant, Bouin, et Maillard.—‘ Traité d’Histologie ’, 1904, 112, 793. 24. Schafer.—‘ Essentials of Histology ’, 1920, 11th edition, appendix. 25. —— Ibid., 339. 26. Schafer and Williams.—‘ Proc. Zool. Soc.’, 1876, pt. 1, 165. 27. Schaffer.—‘ Wiener Sitzungsb.’, 1897, cvi. 353. 28. Sprott Boyd.—‘ Inaugural Dissert.’, Edinburgh, 1836. 29. Stohr.—‘® Arch. f. mikr. Anat.’, 1882, xx. 221. 30. Zimmermann,—Ibid., 1898, lii. 546. PART II, THE GASTRIC MUCOID CELLS OF FOETAL AND NEW-BORN ANIMALS. The stomachs of two litters of new-born and of one foetal cat have been examined, and in addition those of three still- born children and one four months’ human foetus. ‘Ihe method employed was acid formol fixation ; the staining was effected with either Mallory’s stain or Heidenhain’s iron haematoxylin. Cat. Ina foetus of about six weeks the stomach exhibits a simple lining of columnar epithelium, which is entirely NO, 262 P 204 ROBERT K. 8. LIM devoid of a superficial mucous portion. The cytoplasm stains reddish with Mallory. Only a few invaginations represent the primitive gland-tubes. At birth short simple gland-tubes are present. They are lined by oxyntic and mucoid cells. Some of the latter are wholly, others are only partially, mucoid, having a portion of non-mucoid (red-staining with Mallory) cytoplasm within the basal half of the cell. The surface cells are similar to those of the adult. One week after birth the glands are larger and the oxyntic cells more prominent. Mucoid cells are present in large numbers ; a few developing peptic (?) cells are visible. These show no mucoid reaction ; they are coloured principally by the red and brown dyes in Mallory’s mixture. The pylorus is now becoming defined ; it contains only mucoid cells. Three weeks after birth the peptic, mucoid, and oxyntic cells are all plainly evident; the appearance of the mucous membrane now approximates that of the adult. HuMAN. In a foetus of about four months the stomach is lined by a mucous membrane of the simple type, bearing only short gland-tubes. ‘These are formed partly by mucoid and partly by red-staining non-mucoid cells; oxyntic cells are as yet absent. The junction between the stomach and the duodenum is sharply marked off by the pyloric sphincter, but the mucous membrane does not show a corresponding division. The pyloric portion of the stomach for some distance from the actual muscular junction contains both goblet and columnar cells with striated borders. The glands are wholly mucoid. At birth peptic and oxyntic cells are fully developed: the elands are much longer than at four months and altogether more like the adult. CONCLUSIONS. It is quite clear that the gastric glands are in the first instance formed of non-mucoid, red-staining cells. Later these THE GASTRIC MUCOSA 905 cells become mucoid in character throughout the whole stomach. The next type to differentiate is the oxyntic, and at a later stage still comes the peptic. Peptic cells are present in the human foetus at birth, but in the cat do not appear until between the second and third week after birth. ‘This difference may give an important clue to the function of the fundic mucoid cells, for it has been observed that the new-born human stomach contains pepsin while the stomach of the new-born cat contains none, and does not exhibit a ferment until the third week after birth ((Ham- marsten 1874, Zweifel 1874, Morrigia 1876) quoted by Moore (2), Sewall (8)). Obviously pepsin is not secreted by the mucoid cells. These cells are essentially primitive, or at least less specialized than either the peptic or oxyntic. Cade arrives at a parallel conclusion from an entirely different point of view (1). He found that oxyntic cells disappear and peptic cells lose their granules in the vicinity of gastero-enterostomy openings, and all the cells appear mucoid in character. He thus inferred that the altered conditions had caused the specialized cells to revert to the more primitive mucoid cells. In cats I have been able to confirm Cade’s observation completely. Thus while the mucoid cells are undoubtedly a definite variety of the gastric gland-cells they are closely allied to the peptic cells to which they give rise in early and perhaps in later life. LITERATURE. 1. Cade.—‘ Arch. d’anat. micr.’, 1901, iv. 1. 2. Moore.—‘ Schafer’s Text-book of Physiology ’, 1898, i. 330. 3, Sewall.—‘ Journ. Physiol.’, 1878, i. 321. PART IIT. THE GASTRIC MUCOID CELLS IN MAN, DOG, RABBIT, AND FROG. The gastric mucous membrane of several species of animal has been examined in order to compare the histolegical features and the distribution of the mucoid-reacting cells in each P 2 206 ROBERT K. 8S. LIM species, and to determine the general relationship which exists between the mucoid group and the peptic cells of the fundus. The technique employed is similar to that referred to in Part I. The material was invariably obtained from the newly-killed or from the living anaesthetized animal. Human material came partly from the operation table, partly from the post- mortem. Acid formol fixation and Mallory’s and Heidenhain’s methods of staining were the routine procedures. THE Mucorp CELis oF THE FUNDUS. Human.—In man mucoid cells are abundantly present. They have the same characteristics as those of the cat except that their cytoplasm is more homogeneous and stains a lighter blue with Mallory. Their distribution is somewhat different ; they form the entire central lining of rather less than the super- ficial two-thirds of the secreting tubule—hence their regular cubical outline. This portion of the tubule is thinner than the deeper portion which (with rare exceptions) contains typical peptic and oxyntic cells. A few tubules are lined throughout their whole extent by mucoid cells. There is not the same amount of intermingling between the mucoid and peptic cells as in the cat, and thus the mucoid portion of the tubule is more easily defined, especially since it is narrower than the peptic portion. Dog.—The mucoid cells of the dog are intermediate in appearance between those of man and the cat. In some individuals the cytoplasm is almost homogeneous and stains lightly with Mallory ; in others it is more reticular and stains heavily as in the cat. This may be due to functional changes. The distribution of the cells, however, shows fewer mucoid cells in each tubule, i.e. they line less than the superficial half; nor do the mucoid and peptic cells intermingle to any great extent. The widening of the calibre of the deep portion of the tubule occurs gradually as in the cat, but nevertheless the mucoid and peptic portions are sharply marked off from each other. Rabbit.—The mucoid cells of the rabbit stain faintly blue with Mallory and are nearly homogeneous ; they appear like those of man, They are not easily made out since they are THE GASTRIG MUCOSA 207 hidden by the numerous overlapping oxyntic cells. This seems to be a very characteristic feature in the rabbit and accounts for the shape of the cells being very irregular. ‘These cells occupy the superficial three-fourths of the tubule, but there is a good deal of intermingling with peptic cells. The deep portion of the tubule rarely shows mucoid cells. This is best shown in iron-haematoxylin-stained sections of the actively secreting stomach, the presence of the overlapping oxyntic cells making it difficult to examine the more centrally situated cells. In the above preparations the peptic cells alone are clearly stained on account of the marked development in them of ergastoplasmic fibres. The mucoid cells are left unstained by iron haematoxylin. The proportion of mucoid to peptic elements in each tubule varies in different parts of the fundus ; from two-thirds to four-fifths of the whole tubule may be mainly mucoid. Frog (Rana temporaria).—lIn the frog’s stomach only oxyntic and mucoid cells are to be seen. ‘The latter have a clear cytoplasm which stains a faint blue with Mallory. ‘They are found in the superficial third of the gland-tube and rarely extend to the deeper parts. THe Carpiac AND Pytoric Mucoip CELLS. The cells forming the cardiac and pyloric glands are so similar in appearance and staiming reactions that they may be grouped together for consideration. They differ from the mucoid cells of the fundus in their regular shape and in some- times exhibiting a red-staining reticulum with Mallory. The extent of the cardiac and pyloric zones along the two curvatures of the stomach have been measured and are set forth below. | Cardiac | Cardiacand | Pyloric | Pyloricand | Animal. Cells. | Oxyntic Cells. Cells, Oxyntic Cells, Curvatures. Cat 0-4 mm. 3mm. | ld5mm. 20 mm. Greater 0-3 mm. Stmma.! ') 122 5hnam, 20-5 mm, Lesser | Dog i 2mm. 20 mm. 40 mm.! | Greater a= 3mm, 25 mm. 45 mm.? Lesser Rabbit | 0-1 mm. 2mm. | 35mm. | 2mm. | Greater 0-2 mm. 2-3mm.,. | 40mm. | 2-3mm. | Lesser 1 Oxyntie cells small and primitive. 208 ROBERT K. 8S. LIM Human.—'The pyloric cells of man resemble those of the cat in every respect except that they are longer and stain more lightly. Sufficient material was not available from which measurements of the cardiac and pyloric regions could be made, Dog.—There are no pure cardiac glands in the dog. Oxyntic cells may be found at the cardio-oesophageal junction along both curvatures, while peptic cells are present within 2-3 mm. of the junction. In this small zone the cells are longer but otherwise show the same features as those of the cat. Race- mose glands are very constantly present ; they extend from the oesophagus into the cardia under the muscularis mucosae. Their acini are mucous with a few serous crescents here and there. They are thus not to be considered as cardiac glands, but as part of the salivary apparatus which occurs abundantly in the mucosa of the oesophagus. The pyloric region extends for about 40 mm. along the greater curvature and 45 mm. along the lesser. The boundary zone bearing full-sized oxyntic cells and pyloric cells occupies only about 2mm., but small (primitive) oxyntic cells may be observed especially at the neck of the glands within 20-5 mm. of the pylorus. The cells, like the cardiac group, resemble those of the cat—the red-staining reticulum being more constantly present; this is best seen in those near the duodenum. Rabbit.—There are few cardiac glands corresponding to those seen in the cat. These usually occur along the lesser curvature, occupying a small zone of about 2 mm. distal to the oesophagus. Along the greater curvature and sometimes along both curvatures oxyntic cells may be found night up to the cardio-oesophageal junction. When the cardiac glands are present the cells which form them are not typical. They only show a faint mucoid reaction near the surface ; elsewhere the cytoplasm is both granular and reticular, and stains reddish with Mallory.. The condition appears to be an exaggeration of the ‘red reticulum’ seen in the cat and other animals. In addition to this peculiarity glands of the racemose type are also met with under the muscularis mucosae. They extend THE GASTRIC MUCOSA 209 (along the lesser curvature) for only a very short distance (about 3mm.). The acini are mainly serous, a few being mucous; the cells liming the terminal ducts have granules in striae and have centrally-placed nuclei. True mucoid and peptic elements are present beyond the cardiac area described above, the former forming a boundary zone of about 3-4 mm. with the oyxntic cells before the latter are met with. The pyloric region is somewhat larger than that of the cat, since oxyntic cells are only seen about 35-40 mm. from the duodenum (see table, p. 207). There is almost no boundary zone ; the peptic cells appearing a few millimetres beyond the oxyntic. The gland-cells are more mucoid than those of the cardia, but like these show a well-marked non-mucoid basal area. Langley (5) described the cells of the rabbit’s fundus along the greater curvature as being finely granular and similar in appearance to the pyloric cells, while the cells of the remainder of the fundus are coarsely granular. I have not been able to make out this distinction, but perhaps Langley took the superficial mucoid cells to be the only kind of central cell and failed to see the peptic (coarsely granular) cells in the deepest part of the mucous membrane. Frog.—tThere are no true cardiac glands in the frog; the peptic cells merely stop short at the end of the oesophagus while mucoid and oxyntic cells make their appearance. The pyloric region extends about 3-4mm. from the duodenum ; its gland-cells are not different from the mucoid cells of the fundus. GENERAL CONCLUSIONS. The results of this investigation confirm those of Bensley (1) and more especially those of Cade (2), who has examined all the species dealt with here. They show that the fundic mucoid cells vary slightly in appearance in different animals, and that their distribution in the tubule is roughly about the superficial half. From the study of new-born cats it is found that the peptic cell arises from cells of the mucoid type. This 210 ROBERT K. 8. LIM is also probably true for animals other than the cat, since the peptic cells are invariably found in the deep or blind end of the tubule, which may be considered to have developed last. This encourages the view that the mucoid cell gives rise to the peptic cell, without suggesting that the latter is merely a fune- tional phase of the former. Mucoid and peptic cells are undoubtedly different functionally and structurally. In this connexion it is noteworthy that mitoses have never been observed in peptic cells, while they have been seen in mucoid and more frequently in oxyntic cells. In short the mucoid cell is a stage in the genesis of the peptic cell. Transitions from the one state to the other are difficult to demonstrate, but con- sidering the differences, slight though they may be, which oceur in the mucoid cells of the same and of different animals, and especially the occurrence of the basal ‘ red-staining ’, the gap in the genesis of the peptic cell is perhaps partially filled. Looked at in this light, the observation of Cade on the retro- gression of the peptic cells in the vicinity of gastero-enterostomy openings (see Cade (2) and Part II of this paper) may be translated as the inhibition of peptic cell-formation and the arrest of its genesis in the mucoid stage. Utilizing the above hypothesis, the cells of the cardiac and pyloric glands may be regarded as cells which have been prevented from attaining full development by the conditions existing at the orifices of the stomach. The relationship between the various gastric cells may therefore be classified as follows. The mucoid cell of the fundus forms the lowest functional type, for it apparently does not secrete pepsin. The cardiac and pyloric cells are a little more advanced, since Klemensiewicz (4) and Heidenham (8) have shown that the pyloric region secretes a proteolytic ferment. Structurally these cells show the basal * red-staining ’ more constantly (especially in the rabbit) than the mucoid cell, and this may be taken as indicating a certain degree of zymogen formation. The cardiac cells may not function exactly as the pyloric cells do, but they are at least cells of the same develop- mental order, and they constitute such a small element in the THE GASTRIC MUCOSA 911 animals under consideration that they probably have no physiological significance. The peptic and oxyntic cells are the most highly specialized. LITERATURE. 1. Bensley.—‘ Quart. Journ. Micr. Sci.’, 1898, xli. 361. 2. Cade.—‘ Arch. d’anat. micr.’, 1907, iv. 1. 3. Heidenhain.—‘ Pfliiger’s Arch.’, 1878, xviii. 169. 4, Klemensiewicz.—‘ Jahresb. ii. d. Fort. d. Tierchem.’, 1875, v. 162. 5. Langley and Sewall.— Journ. Physiol.’, 1879, ii. 281. ‘The expenses of this work were defrayed by grants from the Carnegie Trust and from the Earl of Moray Fund for the promotion of research in the University of Edinburgh. EXPLANATION OF PLATE 8. (All are from the cat.) Fig. 1.—A cross section of a gland-tube from the cardiac end of the stomach along the lesser curvature, about 1mm, from the oesophagus. Animal killed fourteen hours after last meal. Acid formol fixation ; stained with Mallory. Fig. 2.—A cross section of a gland-tube from the cardiac end of the stomach along the lesser curvature, about 6 mm. from the oesophagus. From the same preparation as fig. 1. m, mucoid; p, peptic; 0, oxyntic. These cells are in the resting condition, Fig. 3.—Cells from the glands of the middle region of the stomach. s, Surface mucus-secreting cells ; m, p, as in fig. 2. xx, Cat 20; 14 hrs.; Altmann’s fluid; Mallory. The peptic cell on the left is somewhat homogeneous (granules intact), while the cell on the right shows the more usual reticulated appearance. Note the cytoplasm of the mucoid cell. I, Cat 1; 1 hr.; acid formol; alcoholic eosin and methylene blue. The granules in the peptic cell are imperfectly preserved ; the cytoplasm stains intensely in a blotchy manner. Note the almost homogeneous appearance of the mucoid cell. m1, Cat 3; 6 hrs.; acid formol; very dilute polychrome methylene blue. The peptic cell here shows well-marked ergastoplasmic fibres and zymogen granules, and is in striking contrast with the mucoid cell. xxIa, Cat 21; 24 hrs.; acid formol; iron haematoxylin. 912 ROBERT K. 8. LIM xxtb, same tissue; Mallory. The peptic cells show a well-marked reticulated appearance. Note that the mucoid cells also show a reticulum. Fig. 4.—Cross section of a pyloric gland from the lesser curvature about 19mm. from the pyloro-duodenal junction. Cat. 21; 24 hrs.; acid formol ; iron haematoxylin. The sparse reticulum which can be seen here stains red with Mallory. Fig. 5.—A longitudinal section of a gland-tube from about the middle of the greater curvature. Cat 3; 6 hrs.; acid formol; Mallory. 8, m, p, 0, a8 in figs. 2 and 3; 1, transitional cells; g, cell containing large eosinophil globules, This drawing gives an idea of the distribution of the various cells which compose a gastric gland-tube in the fundie region. The cells are in an exhausted condition. Compare the mucoid cells with the peptic, and also this figure with fig. 2. ——— Quart. Journ. Micr. Sci. Vol. 66, N.S. Pl. 8 eee? KS. Lim del es? " ~ ¥ 6 . * » fae tt aa ee dé | os te eee yt ee [ai anil On the Labral Glands of a Cladoceran (Simo- cephalus vetulus), with a description of its mode of feeding. By H. Graham Cannon, B.A., Demonstrator in Zoology, Imperial College of Science, South Kensington. With Plates 9, 10 and 2 Text-figures. Lrypie (12) in 1860 was the first worker to point out that the possession of labral glands is common to all the Cladocera. In 1846, however, Schédler (16) had observed that in the labrum of Acanthocercus there exist paired glandular bodies; he states, ‘Im sog. Labrum (des Acanthocercus) glauben wir ein paar rundliche, fast nierenformige Conglomerate als driisige Korper (vielleicht als Speicheldriisen, glandulae salivales) ansprechen zu miissen’. Claus (8) in 1876 mentioned these glands in his work on the anatomy of Daphnids, and later Cunnington (5) in 1903 described them in Simo- cephalus sima (Simocephalus vetulus). Among the other Phyllopoda, Claus (4) in 1886 mentions and figures the glands in Branchipus and Artemia. Refer- ring to the labrum he states: ‘endlich in dem terminalen Theil die grossen als Speicheldriisen gedeuteten Driisenzellen, deren Ausfuhrgangsd6ffnung und Drisenstructur auf Quer- schnitten leicht zu constatieren sind’. Sars (15) states that these glands exist in Limnadia and Limnetis, and in his figures of other Phyllopoda large cells are indicated in the interior of the labrum. With regard to the anatomy Claus (8) was the first to give a description in any detail, but apart from this the only deserip- tion at all complete is due to Cunnington in his description of the glands in Simocephalus sima. Claus considered that 214 H. GRAHAM CANNON the glands could be separated into two groups, the first group lying under the brain and over the oesophagus and the second group consisting of very large cells lying nearer the tip of the labrum. ‘The first group sent out a long thin efferent duct which, after making many twists, allowed the exit of the secretion in front of the mouth. Cunnington’s description differs essentially from this in that he could not observe a duct from the first group but did observe an efferent duct from the second group. Cunnington also distinguishes two groups of cells—a proximal group of several small cells and a distal group of large cells. ‘The proximal group, he states, lie close against the chitinous cuticle and are obviously modified epidermal cells and possibly act as replacement cells, taking the place of cells in the distal group when these lose their secretory power. ‘The latter group usually consists of four cells only and these are placed one behind the other, the most extreme possessing a duct opening on the inner side of the labrum. ‘They have characteristic nuclei, which are shaped like a hollow bowl and thus appear circular or semi-circular in section, ‘The secretion is formed in the neighbourhood of the nuclei in the form of little drops which fuse to larger drops or rods or bands and pass to the exterior. Cunnington suggests that the duct of the extreme cell of the distal group acts as a common duct for the whole group. Meruobs. For Simocephalus vetulus the best fixative was found to be cold saturated sublimate in distilled water. his gave excellent fixation and did not produce distortion as did most other fixatives. Good results were also obtaimed with a mixture of equal parts of saturated sublimate in distilled water and 1 per cent. osmic acid. This mixture, a modifica- tion of Mann’s fixative, was allowed to act for about an hour. In comparing Simocephalus with other Cladocera, it was found that for Daphnia the best results were obtained with sublimate acetic acid, while for Graptolebris and Camp- tocercus Carnoy gave the best fixation. A young Chiro- LABRAL GLANDS OF SIMOCHPHALUS 915 cephalus metanauplius was fixed in cold saturated sublimate and was found to be very well fixed. Ehrlich’s haematoxylin was used considerably for staining. Iron haematoxylin gave too intense a stain for the gland-cells. The best differential stain, however, was obtained by .using Mallory’s triple method for connective tissue. The fixed material was embedded direct into paraffin and cut Su. On THE ANATOMY OF THE LABRAL GLAND. The two groups of gland-cells, as described by Cunnington, were found to be very distinct and will be described separately, but before doing so the extent and position of the labrum must be stated. The labrum, or upper lip, is an immediate prolonga- tion backwards of the ventro-posterior part of the head, passing ventrally to the two laterally-working mandibles and ending under the maxillae which are immediately behind the mandibles. When viewed from the ventral side it may be described as dagger shaped, but its contour is peculiar and reference must be made to Text-fig. 1, which is a ventral view of the animal as it is seen resting normally in a watch-glass, and to Pl. 10, fig. 13 which is a diagrammatic lateral view of the animal. Anteriorly the labrum is marked off from the dorsal part. of the head by a groove on each side (PI. 9, fig. 8) which extends forward to the level of the nauplius eye and then expands dorsally into the bay from which arises the second antenna. In the living animal the labral glands can be seen indistinctly in the anterior part of the labrum and are of a pale-yellow colour, as was observed by Leydig (12). Proximal Group.—this consists of two laterally placed groups of epidermal cells which almost meet in the mid-ventral line between the first antennae. Hach group commences just in front and close to the ganglion of the nerve to the first antenna (Pl. 10, fig. 13), and extends postero-dorsally over a lozenge-shaped area lining the lateral cuticle of the labrum as far back as the labral nerve (Pls. 9 and 10, figs. 1-4 and 13). Each group consists of about twenty cells, and the nuclei of these 916 H., GRAHAM CANNON vary in size, being smallest at the base of the first antennae and largest about the centre of the group. ‘The nuclei are usually about 20, long, but the smallest are never less than half this Trxt-Fic. 1. wf E Za YZ Vib ioe. —— Still) ) Semi-diagrammatic ventral view of Simocephalus vetulus. The thick dotted lines ending in arrow heads on the animal’s left side indicate the direction and extent of the normal move- ment of the appendages figured on that side.) adg, anterior pair of distal gland-cells ; ant 1, first antenna ; ant 2, second antenna ; cd, connexion between anterior and posterior pairs of distal gland-cells ; en 3, proximal endite of third trunk-limb ; ex 2, exopodite of second trunk-limb; gn 2, gnatho-base of second trunk-limb; 1, & 2, & 3, branchiae of first, second, and third trunk-limbs respectively ; l, labrum ; mdb, mandible ; mx, maxilla; pdg, posterior pair of distal gland-cells; pg, proximal gland; dl, #12, #3, first, second, and third trunk-limbs respectively. length. For comparison it may be stated that the length of the nuclei of nerve-cells or of musele-cells, which are of very uniform size and oval shape, is 44. Thus the volume of these large gland-cell nuclei must be many times. at least twenty —— LABRAL GLANDS OF SIMOCEPHALUS mt WF times, that of the nucleus of a nerve- or muscle-cell. The chromatin in these nuclei is distributed fairly evenly in small clumps (Pl. 10, fig. 9), and there is a conspicuous oval nucleolus which stains red with Mallory’s stain. The cell outlines are not distinct, but where one would expect the cell boundaries to be there are accumulations of large clear vacuoles (PI. 10, fig. 9), undoubtedly the secretory product of these cells. In the peripheral cells of this group the cytoplasm is not very vacuolated, the vacuoles bemg very markedly intercellular ; but more centrally and towards the anterior end the whole of the cytoplasm of the cells is full of small vacuoles while the larger vacuoles lie in between the cells. In this region the proximal group is seen to be attached to the distal group of eland-cells (Pls. 9 and 10, figs. 3 and 9). The proximal group is supplied by a small branch of the nerve to antenna 1 which comes off very near to the brain. ‘There is no efferent duct from the proximal group as described by Claus. The Distal Group.—The distal glands (Text-fig. 1) on each side consist of five cells, four gland-cells and a duct-cell. The gland-cells are arranged in two pairs situated anteriorly and posteriorly, connected with each other—the hinder pair embracing the duct-cell. The anterior pair of cells are in direct connexion with the posterior side of the nerve to the first antenna at a point a little further from the brain than the branch to the proximal group (Pls. 9 and 10, figs. 2 and 13), and there is a conspicuous group of nerve-cells in the nerve in this region (PI. 9, fig. 2). Laterally, as stated above, these cells are connected with the proximal group, and at this pomt the vacuolated cytoplasm of the proximal gland-cells is seen to be continuous with that of the distal gland-cells, the vacuoles passing freely from one group to the other (Pl. 10, fig. 9). The peripheral cytoplasm, except at this point of juncture, is denser than that in the interior of the cells, and is free from vacuoles of secretion (Pl. 10, fig. 9). There is no distinct division between these two cells, but in between the two nuclei there is a confused mass of vacuoles. 918 H. GRAHAM CANNON Centrally these vacuoles coalesce and form an irregularly flat, ill-defined reservoir (Pl. 10, fig. 9). The vacuoles are not very transparent, and in passing from the proximal glands to these two cells of the distal glands one can see the vacuoles becoming more opaque. The nuclei are not cup-shaped as Cunnington (5) stated to be the case generally with the nuclei of the distal glands, but are roughly spheroidal (PI. 10, fig. 9). Their diameter is not usually so great as the length of the largest nuclei in the proximal sroup, but there is probably not much difference between the volumes of these nuclei. There are larger clumps of chromatin in the nuclei than in those of the proximal group, and also the nucleoli, which stain red with Mallory’s stain, are about twice as large. But there is also a diffuse scattermg of chromatin all through the nucleus which gives it a much darker appearance in a stained preparation. These two anterior cells of the distal group are connected by an attenuated process with the two posterior gland-cells (Text-fig. 1; Pl. 10, fig. 13). The reservoir in the anterior pair is not continuous as a duct through this drawn-out con- nexion, but vacuoles are to be seen here, so that presumably the secretion can pass from the anterior to the posterior pair of cells. This connexion is always attached to the dilata- tores oesophagi (Pl 9, fig. 4), and its middle point is a little posterior to the labral nerve loop (Pls. 9 and 10, figs. 4 and 13). The nuclei of the posterior pair of gland-cells are cup-shaped, as Cunnington states. Most of the nucleus forms a thin lamella but there is usually a swelling in the region of the nucleolus (Pl. 10, fig. 12). This is large and usually flat and shows the same staining reactions as the nucleoli of the other gland-cells. The chromatin is gathered together in clumps as shown in Pl. 10, fig. 10, but a more irregular clumping as shown in Pl. 10, fig. 12, is more characteristic. The cytoplasm is pervaded with vacuoles of secretion which are opaque to varying degrees, and these are very conspicuous in the hemispherical recesses formed by the nuclei. As before, LABRAL GLANDS OF SIMOCEPHALUS 919 there is no distinet division between these two cells, but the nuclei are placed with their concave sides facing towards each other and in between the two is a very conspicuous and clearly- detined reservoir (Pl. 10, figs. 10 and 12). This is apparently formed of a flat plate of transparent coalescing vacuoles of the secretion produced by the gland-cells. Neither of these cells possesses an efferent duct as figured by Cunnington, but posteriorly they embrace a separate duct- cell (Pls. 9 and 10, figs. 6,11, and 12). This cell has the form of a tube opening to the exterior at its posterior end and anteriorly opening into the reservoir of secretion. The lumen of this tube is often flat (Pl. 10, fig.11) especially at its posterior end. The nucleus of this duct-cell stains very lightly and is small compared _ with that of a gland-cell, although it is slightly larger than that of a nerve- or muscle-cell. The cytoplasm stains very lightly and is not vacuolated. In sublimate material there is in the secretion reservoir a granular coagulum which stains faintly blue with Mallory’s stain, while in the lumen of the duct-cell it stains red. Pre- sumably the cytoplasm of the duct-cell alters the constitution of the secretion in some way, so that its staiing reaction when fixed is changed. A section through the duct-cell at its anterior part shows the secretion in contact with the walls of the tube staining red, while that more centrally placed, which has not yet been acted upon by the duct-cell, still stains blue. The external apertures of the duct-cells form two small slits on the side of the labrum near its tip (PI. 9, fig. 7) where the latter is com- pressed laterally. They are situated a little towards the dorsal surface of the labrum and are ventral to about mid-way between the mandibles and maxillae. In other Daphnids studied it was not found possible to obtain preparations sufficiently well fixed on which to base critical considerations, but it is evident that the same ground-plan underlay all the cases studied. In Chirocephalus, however, the results obtained are very good and agreed comparatively well with Claus’s (4) figure for Branchipus. The proximal group is very scattered and ill defined. Its cells do not all line NO. 262 Q 22.0 H. GRAHAM CANNON the chitinous cuticle, but, however, they are connected with the gland-cells of the distal group and loosely fill the anterior part of the large labrum. ‘The distal group is represented by three pairs of gland-cells—two placed laterally and one medially -—slightly nearer the tip of the labrum. The nuclei of these cells are very large but not cup-shaped. In each pair of cells is a secretion reservoir which opens into the lumen of a very conspicuous duct-cell just as in Simocephalus vetulus. On THE MANNER OF FEEDING. Simocephalus vetulus feeds on small particles and planktonic organisms contained in a current of water which it maintains over its mouth appendages. In observing the animal it is usually on its back as figured in Text-fig. 1, but in describing the method of feeding, to avoid confusion, the animal will be assumed to be dorsal side uppermost. The valves of the carapace form an incomplete tube about the posterior part of the animal, this tube being effectively completed by the hairs along the ventral edges of the carapace (Text-fig. 1). Posteriorly the tube is open to the exterior and anteriorly it expands at each side of the labrum into the bays from which arise the second antennae. Further, this tube is incompletely divided into a dorso-lateral chamber, which includes the brood-pouch and in which are the branchiae, and a median ventral food passage. ‘The latter is bounded dorsally by a well-marked food groove (PI. 9, figs. 6, 7, and 8) which runs along the ventral side of the trunk. Ventral to it are the hairs along the edges of the carapace while laterally are the trunk limbs. The current of water carrying the food passes in at the bases of the second antenna, and so passes close to the first antenna on which are situated, according to Scourfield (17), the supposed olfactory organs. and passes out at the postero-ventral angle of the carapace in the neigh- bourhood of the anus. The appendages chiefly responsible for maintaining the food- stream are the first, third, and fourth trunk-limbs. Calman (2) states that the third and fourth pairs of trunk-limbs ‘ are LABRAL GLANDS OF SIMOCEPHALUS 291 characterized by the development of the proximal endite with its comb-like row of setae’. These endites are placed almost vertically with their setae pointing upwards into the food eroove. They diverge slightly from behind forwards and in passing upwards towards the trunk they slope inwards. They move in and out laterally. From the fact that they are nearest together at their posterior end the outward movement sucks in the water from before backwards. Since also they are not placed vertically but are slightly further apart at their proximal end than they are at the end of the comb of setae in the food eroove, the outward movement, in all probability, causes a small backwash in a forward direction in the food groove. Although the food current is produced mainly by the third and fourth trunk-limbs the first also plays an important part. The shape and arrangement of the first trunk-limbs can best be seen from Text-fig. 1. Its setae form a curved shield over, that is, ventral to, the second trunk-limb. In its normal move- ment it synchronizes with the other trunk-limbs but is not in the same phase. It commences its backward stroke just after the other limbs begin to beat outwards. The outer part of the limb moves in an are of a circle with the tip of the labrum as centre (T'ext-fig. 1) so that those setae which le against the side of the labrum scarcely move at all. The two limbs together thus form a funnel-like entrance to the food passage down the centre of which projects the labrum. The reason for the retarded lateral movement of this pair of limbs is not at all certain, but in all probability it is to secure a passage of water over the branchiae. The second trunk-limbs are peculiar in possessing a large and specialized processus maxillaris or gnathobase. Their exopodites or outer branches lie over, that is ventral to, the succeeding limbs, and their function is probably merely to assist by their oar-like movements in maintaining the food- stream. The gnathobases point inwards and are beset with setae which point inwards and almost meet in the middle line. On each gnathobase there are ten setae. The posterior seven point backwards while the anterior three point forwards. They Q 2 299, H. GRAHAM CANNON are numbered in Text-fig. 2 from behind forwards. No. 1 is very long, reaching back to the hind end of the body, and is beset with long hairs. No. 2 is much shorter and ends in a small hook, and possesses a comb-like row of minute closely- set hairs over a little more than half its length on one side. Nos. 3, 5, 6, and 7 really form a series. They are short stout setae ending in a brush-like tuft of hairs. No. 4 differs slightly from them in being shorter but terminating in a long thick TEXT-FIG, 2. Gnathobase of second trunk-limb of Simocephalus vetulus (for explanation see text). hair projecting beyond the rest. Lilljeborg (18) does not figure this difference. Nos. 8, 9, and 10 have the form of for- wardly projecting combs. No. 8 is usually bent at an angle at about its middle point, while Nos. 9 and 10 are curved. The hairs on No. 8, which only occur on its distal half, are very fine and regular, and are about twice as closely set as those on Nos. 9 and 10, which occur along the whole length of the setae. During the movement of the second trunk-limb out- wards and forward the gnathobase also moves upwards so that its three anterior setae comb the side of the tip of the labrum. When the limb is in its most forward position these three setae have passed across the labrum on to the maxillae. Each maxilla is armed with three setae which do not point dorsally as figured by Cunnington (5), but point forwards along LABRAL GLANDS OF SIMOCEPHAILUS 293 the food groove (PI. 10, fig. 18, and Pl. 9 figs. 5 to 8). Hach seta is beset with a double row of hairs on its inner side. During the movement of the maxillae the setae move backwards and forwards and in their forward movement move inwards, so that the hairs of the opposite setae meet in the food groove. At the hinder margins of the biting surfaces of the mandibles there are blunt spines (Pls. 9 and 10, figs. 6 and 18), while the anterior parts are scored with vertical serrated ridges. The two mandibles, which are never symmetrically apposed to one another, appear to work like two cog-wheels fitting into one another and thus crush the food and at the same time force it forwards into the beginning of the oesophagus, up which it rapidly passes by peristalsis. The mechanism of the method of feeding is as follows: food particles in the food-stream, drawn in by the action of the united movements of the trunk-limbs, are diverted towards the median groove along the side of the labrum, by the first trunk-limbs. At the tip of the labrum they are caught by the anterior setae of the gnathobases of the second trunk-limbs and brushed dorsally into the food groove above the tip of the labrum and between the maxillae. The brush-like setae of the snathobase are in all probability the main agents in bringing this about. The more anteriorly-placed comb-like setae which brush the side of the labrum also assist in collecting the food on to the maxillae, but their chief function seems to be to brush the secretion of the labral glands on to the food as it collects between the maxillae. Hardy and MacDougall (8) state that when the food is swallowed it consists of particles—‘ which are glued together by some sticky substance’. It is suggested that this sticky substance is the secretion of the labral glands. The food which collects as a bolus between the two maxillae is now and again pushed forwards by the movements of the appendages on to the mandibles. PI. 10, fig. 18, shows how the hairs on the setae of the maxillae point forwards, and Pl. 9, figs. 6, 7, and 8, show how the hairs of the adjacent setae fit together and so make an admirable broom for sweeping a bolus forwards on to the mandibles. A movement of the maxillae 294 H. GRAHAM CANNON is always followed immediately by a movement of the mandibles, but the latter rotate many times without any movement of the maxillae, so that probably the maxillae push forwards a large bolus on to the mandibles and these gradually pass it into the oesaphagus. Hardy and MacDougall (8), referrmg to Daphnia, which is no doubt essentially similar in its feeding to Simocephalus vetulus, state that food particles are carried over the mouth by a current of water and ‘ many of them adhere to the sticky surfaces of the mouth appendages’, and that these adherent particles are formed into a bolus by the movements of the appendages. ‘To observe the method of feeding these workers fed the Daphnids on milk, yolk of egg, and carmine. When the animals are fed on any of these substances they always become dirty, the particles adhering all over their bodies. With the former two substances they become greasy and break through and adhere to the surface of the water. It is thought that this is merely due to the presence of an abnormally large quantity of food. In the normal animal, feeding on its normal food, no particles are to be seen adherent to the appendages. If the animal is at all moribund it soon becomes covered with adherent particles. If the animal be fed on milk—a drop of milk is carefully placed at the bottom of a watch-glass containing the water in which the water-fleas are swimming—the regular movement of the appendages is often stopped while the setae of the first trunk-limb are combed over the lateral surface of the labrum to remove any milk adhering to it. Also by this method of feeding a large amount of fatty drops collect im the food groove posterior to the maxillae. These are in all probability drawn there by the backwash previously mentioned that must pass out along this groove. When this accumulation of food becomes too great the labrum is raised by its levator muscle—which runs from the base of the labrum to the covering of the brain— the trunk is flexed forwards, and, with the caudal furea, the accumulation is lifted out of the food groove and, by the extension of the body, removed to the exterior. bt i Or LABRAL GLANDS OF SIMOCEPHALUS On tue FuNGTION oF THE LABRAL GLANDS. Some preliminary experiments of staining Simocephalus intra vitam suggested that further investigations might elucidate the functions of the labral glands. ‘The experiments which were accordingly carried out did not prove of much use in the direction expected, but were interesting and will be described here. Vischel (6) describes experiments on intra vitam staiming using, among other stains, alizarm, neutral red, Bismarck brown, Nile blue sulphate and hydrochloride. In repeating his experiments using the stains named, the only stains with which successful results were obtained were neutral red and Bismarck brown. It may be mentioned that these two stains were Griibler’s chemicals while the others were not. In Fischel’s figure of Daphnia magna staimed intra vitam with neutral red, there are figured two large red patches in that region where the labrum should be drawn which probably represent the labral glands. He states that these glands are always to be found faintly stained in animals stained intra vitam with neutral red. In adult Simo- cephalus vetulus the most conspicuously-stained organs in such animals are the labral glands and the body which Vischel describes as a gland of unknown nature, which has since been shown by Langhans (10) to be the end-sac of the shell gland, and both these stain intensely. In the labral glands both proximal and distal groups stain, but the duct-cell remains unstained. The connexion between the anterior and posterior pairs of cells of the distal group appears very distinctly, and was at first thought to be a distinct duct. In the gland-cells there appear accumulations of an intensely staining material—these accumulations being often as large as the nuclei of the cells. The reservoir of secretion which can be seen in the living animal remains unstained. Fischel maintains that the stainmg with neutral red is not due to the staining of passive metabolic products but to the staining of preformed elements in the protoplasm. In support 226 H. GRAHAM CANNON of this he states: ‘ist einmal die Granularfirbung eingetreten, so bleibt sie auch konstant, das Bild derselben indert sich in keiner Weise, wie lange auch die Tiere beobachtet werden mogen. Und was ebenso wichtig ist, firbt man eime gréssere Anzahl von Tieren, so weisen Zellen der gleichen Art stets auch die gleiche Granulierungsart auf’. In the experiments on Simocephalus vetulus no such constancy was observed in the labral glands. While these remained stained they did not continually present the same appearance ; moreover, not only did the glands of different imdividuals stain differently, but the glands of the different sides of the same individual stained differently, which is what one would expect from the mobile, vacuolated nature of the protoplasm constituting the labral glands. Howeyer, quite apart from this case, this con- stancy in the appearance of a cell stamed intra vitam with neutral red does not agree with the fact that by such staiming methods the mitochrondria are stained (Gatenby (7)). Lewis and Lewis (11) have shown that not only do mitochondria continually change their shape but also are continually shifting their position. If specimens are fixed in sublimate after staming intra vitam with neutral red and dehydrated rapidly some of the stain remains in the specimen. If they are now embedded and sectioned, on mounting the ribbon the stam can be seen in patches in the labral gland, and the position and shape of these can be drawn with reference to the contour of the glands. If now the wax is removed and the sections brought down to water the remaining stain is washed out. Staimimg now with an aqueous solution of thionin there appear dark bodies in the section staining an intense violet, almost black, and these patches agree with those stained by the neutral red. In sections of the animals which have not been stained with neutral red but which had been similarly fixed and stamed in thionin, these very conspicuous dark bodies do not occur, and it seems safe to assume that they are formed by the action of the neutral red on the animal. Weak solutions of neutral red apparently always have a harm- LABRAL GLANDS OF SIMOCEPHALUS DA ful effect on Simocephalus vetulus. No individual of Simocephalus exspinosus was found to survive a weak solution longer than twelve hours. In Simocephalus vetulus the movements of the limbs is always retarded when the animals have been in such a solution for about twelve hours. Advantage was taken of this fact to study the move- ments of the limbs during feeding. Usually, even if the stained individuals are removed to pure water, they survive only a few days. Sometimes, however, with young individuals they sur- vive and completely lose all effects of the stam. No adults have been obtained to survive long the effects of the stain, but among these adults the stain often shows signs of disap- pearing and yet the labral glands always remain as con- spicuously stained as at first. It was thought from these results that the labral glands might be partly the agents causing the disappearance of the neutral red. However, in the well-known experiment of feeding Daphnids on carmine, while the end-sac of the shell gland is stained by the carmine there is never any trace in the labral glands. This experiment was also repeated with neutral red, Bismarck brown, Nile blue sulphate and hydrochloride, using a filtered mixture of the stain with milk to feed, but there was no indication as to where the stain was excreted. Apparently with neutral red and Bismarck brown the staining effect is not produced through the gut but the stain acts directly through the cuticle. Thus young embryos in the brood-pouch stain just as markedly as their parent. Both these stains show a great affinity for yolk. Individuals with nearly fully- developed embryos in the brood-pouch were stained in neutral red for twenty-four hours. Those imdividuals were then selected which had given birth to their brood, but had not yet laid their next batch of eggs, and these had deeply-stained ovaries. These were returned to fresh water. The eggs which were subsequently laid were stained deep red. As these developed the stain was seen to be confined chiefly to the yolk. In most cases the adults died before giving birth to the young but in a few cases the young were born, but the adults never 228 H. GRAHAM CANNON survived the succeeding eedysis. ‘lhe young in these cases showed stain chiefly in the tips of the first and second antennae and in the ‘ Haftorgan’. By the third instar all traces of the stain had disappeared. These experiments show a similarity to those of Sitowski (18) on Tineola biselliela. This worker fed these caterpillars on food stained with Sudan red, and their fat became stained red, giving them a red appearance. ‘The eggs laid were also stained red while the animals hatching from them showed signs of a slight red coloration. ‘hey are also most probably similar to a certain experi- ment of Agar (1) on Simocephalus vetulus. In Agar’s experiment he fed the Daphnids on a food which pro- duced in them a curious abnormality, which consisted in a change from the normal, of the curvature of the valves of the carapace. On removing the abnormal individuals to normal conditions the abnormality disappeared in a few generations, and up to this point the result is analogous to Sitowski’s results and to the experiment recorded here. How- ever, Agar states that not only did the abnormality disappear, but in the third generation of the offspring there was a ‘ very decided reaction ’—the valves of the carapace not only came back to their normal position but overshot the mark and became more curved in the opposite direction. This is stated to be due to the overproduction of an anti-body antagonistic in its effects to the substance causing the abnor- mality. This occurrence of a reaction was supported by a table of ratios representing the transmission of the abnor- mality, and about this table Agar says that, by itself, * it cannot be said to give unequivocal evidence, especially when the high degree of imaccuracy in the original measurements is considered ’, but that this table bore a ‘ striking resemblance ’ to a second table representing the transmission of another abnormality which was based on much more accurate measure- ments and on a much greater number of individuals. But, even supposing that this latter table accurately represents the course of the second experiment, the value of the resemblance LABRAL GLANDS OF SIMOCEPHALUS 229 between the two tables as a basis on which to postulate similarity between the two sets of experiments, of which the tables are representative, depends solely on the accuracy of the first table. While it is admitted that the second table is no doubt comparatively accurate, it must be emphasized that in the first table referrmg to the abnormal curvature of the carapace, the presence of a reaction in 3 generation was indicated by an increase among only forty-seven individuals, of 6 per cent. over a normal ratio—and in measuring this ratio an error could be made of as much as 20 per cent.—the average error according to a table quoted by Agar to show this inaccuracy is roughly 10 per cent. It appears very uncertain, on such data, to make the definite statement that there was a ‘very decided reaction’. The repetition of these results of Agar was abandoned because, in the individuals used, the inaccuracy of the measurement of the ratio which indicated the extent of the abnormality was even more marked than in those used by Agar. It may be mentioned here that Agar merely stated that the food producing the abnormality was a ‘ culture of protophyta grown in a mixture of cowdung, soot, and water’. It was found that the abnormality can be produced by feeding a culture containing no other protozoon than a species of Chlamydomonas. Also, contrary to Agar’s finding, it was not found possible to produce the abnormality in Simo- cephalus exspinosus. In the experiments recorded here there seems no evidence as to how the neutral red disappeared. As Agar suggests for cases of parallel induction, it may have disappeared by mere dilution caused by the increase in the bulk of the protoplasm without a corresponding increase in the amount of the stain. Partly the stained matter may be oxidized or changed in some way into a colourless material which may or may not be excreted ultimately. These experiments on intra vitam staining were carried out before the mechanism of feeding was closely studied. ‘The latter investigation made it obvious that, as already stated, 230 H. GRAHAM CANNON the food is entangled in some substance before it reaches the mouth, and from the disposition of the appendages and their method of working it seems most probable that this substance is produced by the labral glands. This brings Simocephalus vetulus into the same category of feeders as those Gastro- poda, Pelecypoda, Protochordata, and Branchiopoda whose method of feeding is described by Orton (14), in which the prehension of the food is brought about by the secretion of some food-entangling substance. The nature of this substance in Simocephalus is, however, peculiar. Sections of the elands were stained according to the method recently deseribed by Keilin (9) using thionin as a metachromatic stain for mucin. The nuclei of the gland-cells stained blue while the cytoplasm was purplish, as would occur in a mucous gland, but the secretion filing the reservoir not only did not stain red, as it would do if it contained mucin, but showed a pale-blue tint. Bismarck brown also left the contents of the reservoir unstained. If, then, the metachromasy of thionin is used as a definite method for the detection of mucin, the labral glands of Simo- cephalus vetulus must not be described as mucous glands. Irom the quotation from Schdédler’s work on Acanthocereus at the beginning of this paper it will be seen that he suggested that the labral glands were possibly salivary glands. Claus does not discuss their function but merely states: ‘ Die grossen Zellen der Oberlippe .. . betrachte ich als Lippendriisen ’. Cunnington, discussing the physiological significance of the secretion from the labral glands, states that the fact that the secretion flows out in front of the mouth suggests that the gland is a salivary gland. ‘The term ‘ salivary gland’, derived as it is from vertebrate and more especially human anatomy, is now unfortunately used in a variety of senses in the different groups of animals. In some groups a certain physiological sense is implied, while in others the term is used only in a topographical sense. Among the Arthropoda, it is not possible to find, from the physiological sense, a character common to all the secretions of their salivary glands, while from a morphological standpoint, LABRAL GLANDS OF SIMOCEPHALUS 931 owing to the very uncertain homologies of the various mouth parts in the different classes, it is not advisable to base a defini- tion of salivary gland in this group on such considerations. Hence the term salivary gland should not be extended still further to include the labral glands of Simocephalus vetulus. SoutH KENSINGTON. August 1921. SUMMARY. 1. The labral glands consist of a proximal and a distal group of gland-cells. 2. The proximal group consists on each side of about twenty cells. The cells possess large flat nuclei and their secretion collects as intercellular vacuoles. 3. The distal glands which are in connexion with the proximal eroups consist on each side of five cells-—four gland-cells and a duct-cell. The anterior pair of gland-cells possess large spheroidal nuclei between which is an ill-defined reservoir of secretion. The posterior pair have cup-shaped nuclei between which is a very definite reservoir of secretion. 4, The duct-cell is in the form of a hollow tube, one end open- ing to the exterior near the tip of the labrum and the other end opening into the reservoir of secretion between the nuclei of the posterior pair of distal gland-cells. The duct-cells act as ducts for the whole of the labral glands, the secretion passing as vacuoles from cell to cell. 5. The duct-cell alters the reaction of the secretion before passing it to the exterior. 6. Food particles carried in the stream which is maintained by the trunk-hmbs through the carapace are abstracted by the gnathobases of the second trunk-limbs. 7. There are ten setae on the gnathobase of the second trunk-limb, the anterior three of which are comb-like and brush the secretion of the labral glands on to the food particles as they collect between the maxillae. 939 H. GRAHAM CANNON 8. The setae of the maxillae are directed anteriorly, and by their action pass the food on to the mandibles at the entrance of the oesophagus. 9, The labral glands stain very markedly intra vitam with neutral red and Bismarck brown. ‘There is no evidence that this effect is due to the staining of the preformed structures in the protoplasm. 10. Females stained intra vitam with neutral red, when removed to fresh water will lay red eggs from which young will hatch which are also stained. ‘he stain disappears from these during growth. 11. Agar’s experiments on the transmission of an abnormality produced by a certain food are criticized. This abnormality can be produced by feeding Simocephalus vetulus with Chlamydomonas. 12. The secretion of the gland contains no mucin. EXPLANATION OF PLATES 9 AND 10. DESCRIPTION OF FIGURES. All figures are from Camera lucida drawings. ligs. 1-8 were drawn using a Zeiss D. objective and are at a magnification of 222 diameters. Iigs. 9-12 were made with a Zeiss apochro- matic N.A. 1.4, 2mm. objective and compensating ocular 8. The maenitication is about 860. List oF ABBREVIATIONS. a.l.g. external opening of duct of labral gland; br, brain; c¢.c. cireum- oesophageal commissure; d.c. duct-cell; d, duct of labral gland; d.o. dilatores oesophagi; e.c. epidermal cell; /.g. food groove; ga 1, ganglion of Antennarius 1; g.pg. group of nerve-cells in Antennarius | at root of branch to proximal group ; J, labrum ; /.d.c. lumen of duct-cell ; Lg. lateral groove dividing labrum from dorsal part of head; 1./. labral nerve-loop; md, mandible ; mg, mid-gut; mx, maxillae ; 7.a. 1. Antennarius 1; 7.@. 2’. Antennarius 2 major; 2.a.2”. Antennarius 2 minor; 2.md. mandible nerve ; .ma. nerve to maxilla; 2.n.e. nerve to nauplius eye; 7.pg. nerve to proximal group of gland-cells ; n/, nucleolus; nu.p.g. nucleus of proximal LABRAL GLANDS OF SIMOCEPHALUS 2338 gland-cell; nu.a.d. nucleus of cell of anterior pair of distal gland-cells ; nu.p.d. nucleus of cell of posterior pair of distal gland-cells ; nw.d.c. nucleus of duct-cell ; oe, oesophagus; o0.g. olfactory ganglion; p.c. peripheral layer of cytoplasm of anterior pair of distal gland-cells; r, reservoir of secretion ; s.m. setae of maxilla; v, vacuoles of secretion. Figs. 1-8 form a series of transverse sections through the labrum and adjacent parts of an adult specimen of Simocephalus vetulus. The position of these sections is indicated in fig. 13 by the series of vertical lines numbered 1-8 at their upper ends. Figs. 9, 10, and 11, are drawings at a higher magnification of parts of the same sections that are represented in figs. 3, 5, and 6 respectively. The orientation of figs. 9-11 with respect to the plate is the same as it is in figs. 3, 5, and 6. In figs. 1-8 and in fig. 13 the proximal glands are shaded thus ////////////// and the distal glands are shaded thus \\\\\\\\\\\\\\. Fig. 1.—Section cutting the most anterior part of the proximal gland at the level of the nerve to the first antenna. Fig. 2.—Section through anterior end of the distal gland. This section passes through the nerve to the proximal gland. Fig. 3.—Section through the connexion between the proximal and distal groups of gland-cells. Fig. 4.—Section through the attenuated connexion between the anterior and posterior pairs of gland-cells of the distal gland. This section passes through the commencement of one of the mandibles and includes the labral nerve-loop. Fig. 5.—Section through the posterior pair of cells of the distal gland. This section includes the tips of the hairs on the setae of the maxillae. Fig. 6.—Section through the duct-cell. Fig. 7.—Section through the external opening of the duct of the labral gland, Fig. 8.—Section through the maxillae and the tip of the labrum. Fig. 9.—See fig. 3. Fig. 10.—See fig. 5. Fig. 11.—See fig. 6. Fig. 12.—Horizontal section, the position of which is indicated in fig. 13. It passes through the aperture of the duct of the labral gland. Fig. 13.—Somewhat diagrammatic figure of a side view of Simo- cephalus vetulus. Of the appendages behind the antennae only the right mandible and maxilla are represented. 934 H. GRAHAM CANNON i; 10. 11. 12. 13. 14. 15. 16. at. 18. BIBLIOGRAPHY. Agar, W. E. (1913).—‘‘ Transmission of Environmental Effects from Parent to Offspring in Simocephalus vetulus”, ‘ Phil. Trans. Roy. Soc. Lond.’, 203. . Calman, W. T. (1909).—‘ A Treatise on Zoology’, ed. by E, Ray Lankester, Part VII, Fasc. 3, ‘‘ Crustacea’. London. . Claus, C. (1876).—‘‘ Zur Kenntnis der Organisation und des feineren Baues der Daphniden und verwandter Cladoceren ”’, ‘ Zeit. f. wiss. Zool.’, 27. (1886).—‘‘ Untersuchungen iiber die Organisation und Entwicklung von Branchipus und Artemia ”’, “ Arb, a. d. Zool. Inst. Wien’, 6. . Cunnington, W. A. (1903).—‘* Studien an einer Daphnide Simo- cephalus eima”, ‘ Zeit. f. Naturwiss. Jena’, 37. . Fischel, A. (1908).—‘‘ Untersuchungen iiber vitale Farbung an Siiss- wassertieren, insbesondere bei Cladoceren’’, ‘ Int. Rev. der ges. Hydrobiologie und Hydrographie ’, 1. . Gatenby, J. B. (1919).—‘“ The Identification of Intracellular Strue- tures’, ‘ Journ. Roy. Mier. Soc.’ . Hardy, W. B., and MacDougall, W. (1894).—‘‘ On the structure and function of the Alimentary Canal of Daphnia”, ‘ Proc. Camb. Phil. Soe,’, 8, il. . Keilin, D. (1920).—‘‘ On the Pharyngeal or Salivary Gland of the Earthworm ”’, ‘ Quart. Journ. Micr. Sci.’, 65, i. Langhans, V. V. (1909).—‘ Eine rudimentire Antennendriise bei Cladoceren als Ergebnis der Vitalfiirbungsmethode”’, ‘ Int, Rev. der ges. Hydrobiologie und Hydrographie ’, 2. Lewis, M. R., and Lewis, W. H. (1914).—** Mitochondria (and other cytoplasmic structures) in tissue cultures ’’, ‘ Amer. Journ. Anat.’, a7: Leydig, F. (1860).—‘ Naturgeschichte der Daphniden ’, Tubingen. Lilljeborg, W. (1900).—‘‘ Cladocera Sueciae *, “Nova Acta reg. Soc. Sci. Upsala ’, 19. Orton, J. H. (1914).—‘‘ On Ciliary Mechanisms in Brachiopods and some Polychaetes, &c.”’, “ Journ. Marine Biol. Assoc, U.K.’, 10. 2. Sars, G. O. (1896).—‘ Fauna Norwegiae’, vol. i, “‘ Phyllocarida and Phyllopoda”’. Christiania. Schédler, J. E. (1846).—‘‘ Ueber Acanthocercus rigidus, &c.”, ‘ Arch. fiir Naturgesch.’, 12. Scourfield, D. J. (1905).—‘* Die. sogen. ‘ Riechstabchen’ der Clado- ceren ”’, ‘ Pléner Forschungsberichte ’, 12. Sitowski, L. (1905).—‘ Bull. de Acad. des Sci. de Cracovie ’. H..G. Cannon, del. Quart. Journ. Micr. Sci. Vol. 66, N.S., Pl. 9 Fig. 3 Fia. 4 ie} ? . a Be a } AV OS No‘ is% ae? ete pe . Rind fr 6 ff an) 4 ae = - . Quart. Journ. Micr. Sci. Vol. 66, N.S., Pl. 10 H. G. Cannon, del. ot. ft the Rh Ata > 7 >= * Se Ma 54g a “wh att nae w ee Yee i it +4 ver ™~ § 7 oe a ee -s r rd 7 oo ry & y a -_ LU star Fr ‘a 4! «¢ ' << ) ® ne Ps y ys > ol GT ry Surface Tension and Cell-Division. By J. Gray, M.A., Fellow of King’s College, Cambridge. With 9 Text-figures. Tue series of changes which a dividing cell exhibits has long suggested to biologists that surface tension plays a dominating role in the process of cleavage. Without exception, theories based on this suggestion have postulated regions of differential tension on the cell-surface ; the surface tension at the equator of the cell has been held to be either higher or lower than that at the polar regions of the cell. Such theories have proved of but little value as a means of further investigation, since there is no apparent means of determining how such a state of affairs could arise, nor is there any apparent differentiation in the microscopical structure near the equator of the cell- surface. The evidence here presented suggests that regions of dif- ferential surface tension are unnecessary assumptions, and that cell-division does not take place owing to a change in surface tension at the cell-surface, but owing to a force inside the cell which operates against the surface tension. It is the equilibrium between this force and the normal surface tension which deter- mines the shape of the dividing cell. The fertilized eggs of Echinus miliaris form very satisfactory material for a study of cell-division, since the protoplasmic surface of the egg is in direct contact with an aqueous medium and because the actual process of cell-division can readily be followed under the low powers of the microscope. The normal egg is spherical, and if it be crushed or broken the resultant portions show no tendency to mix with the water, but rapidly acquire a more or less spherical shape. From this we may infer that the protoplasm of the egg resembles that R2 236 J. GRAY of many other cells, in that it is essentially a hquid which is immiscible with water. Further, an overwhelming body of evidence is available to show that the protoplasmic surface contains a lipoid or oily phase. ‘To what extent is the form of the egg dependent upon those conditions which determine the form of inert drops of oil surrounded by water ? Consider the simple case of oil-drops in water. The drops are all spherical owing to the existence of a foree—usually called surface tension—acting at the oil/water interface. In any such system the amount of free energy will tend to reach a minimum, and since the volume of oil presents a minimum amount of surface wher the drop is spherical it is obvious that TEXT-FIG. OV S4 Form of oil-drops in (a) acid water (b-d) increasing amounts of alkali in water. the position of stability is reached when the drops are round. The higher the surface tension the more rapidly is the spherical form assumed, and the more resistant is the form of the drop to external or internal disturbance. Now, it has been shown that the surface tension at an oil/water interface is materially affected by the hydrogen-ion concentration of the water. Hydrogen ions raise the surface tension ; hydroxyl ions lower it. The effect of such changes is a marked alteration in the form of the oil. In an acid solution the oil-drop remains perfectly spherical and is not readily deformed by external forces. In an alkaline solution, however, the drop becomes very irregular in shape and is readily deformed. In highly alkaline solutions the sur- face tension may actually become ‘ negative ’, and the condition of stability is reached by the splitting up of the drop to form an emulsion. ford SURFACE TENSION AND CELL-DIVISION 237 To what extent the drop of protoplasm responds to similar changes in its environment is seen from the following figures. Tt will be seen that in an alkaline medium the capacity of the egg (like that of an oil-drop), to retain a spherical form is lost. The outline of the egg becomes distorted by the production of numerous blunt irregular processes, just as is the surface of the oil. On returning such eggs to normal sea-water the spherical form is gradually reformed; in acid sea-water the spherical form is quickly regained. In some cases the recovery TEXtT-FIG. 2. EN ’ oy 4 <7 b c > Co o>) 5 / ° a mo) Oe re ae ae e f a, Egg in normal sea-water, P;, 7-9; b, egg in acid sea-water, P;, 6:0; c-e, successive stages in alkaline sea-water, P, 9°6; f, same egg transferred from alkaline to acid water. a of the spherical form causes the protoplasm to divide com- pletely into two or more parts. These parts are always spherical or eliptical. The nucleated fragment alone divides to form normal blastomeres. Saponin possesses the power of lowering the surface tension at an oil/water interface, and produces similar changes in the form of the egg to those produced by alkalis. It seems reasonable to conclude that the form of the undivided egg is determined, at least in part, by the surface tension at the egg-surface. Before proceeding to consider the part played by surface tension during the process of division, it is necessary to draw attention to certain facts in connexion with normal cleavage. The first division in the egg of Echinus miliaris takes 238 J. GRAY place (at 15°C.) about fifty mimutes after fertilization, and takes roughly three minutes. The process of cleavage is shown diagrammatically below. It is important to note that during cleavage there is a progres- sive increase in the length of the main axis of the egg; this is just as distinctive as the production of the shorter axis which is brought about by the development of the cleavage furrow. Text-ria. 3. JOG CO CO Og Stages in normal cell-division of egg of Echinus miliaris. As the egg elongates so the polar regions become more and more convex, while the equatorial region becomes more and more concave. In making a comparison between the dividing egg and an oil/water system it is convenient to consider the fusion of two oil-drops rather than the division of a single drop into two equal parts. When two oil-drops fuse it is obvious (Text-fig. 4) that a reversal of the process would approximate very closely to the process of cell-division. Now, in fusing together the amount of free energy at the surface of the oil is reduced in the ratio of 5: 4, so that when the single drop is mechanically shaken into two equal parts, SURFACE TENSION AND CELL-DIVISION 939 or when the egg divides into two equal blastomeres, it is necessary to provide the surface of the two systems with free energy. Hence, during cell-division the egg must do work in order to provide free surface energy. Before proceeding further with this argument, let us consider the effect of altering the surface tension at the protoplasmic/ water interface during the actual process of cleavage. In order to do this, eggs at different stages of cleavage are transferred TEXT-FIG. 4. a : : ; b c d Stages in the fusion of two oil-drops. to acid and alkaline sea-water. In the latter case division occurs quite normally; in the former case striking effects are produced. The effects of acid sea-water may be summarized as follows : (i) The cleavage furrow is entirely lost, and in the early stages of cleavage the egg tends to become spherical in form. (1) In the total abolition of the cleavage furrow there is distinct evidence that the egg is elongated along the main axis of the astral figure, so that in the latter stages of cleavage the form of the egg is that of a well-marked cylinder with hemispherical ends. (ui) The effect of acid sea-water is entirely reversible. If an egg which has been taken from normal to acid sea-water be 240 J. GRAY replaced in normal sea-water, it very rapidly returns to the stage of cleavage which it had reached prior to being placed in Text-Fia. 5. I Il Il Normal sea-water. Acid sea-water. Normal sea-water, we uh tae ; | a ees b G a (S Effect of acid sea-water on the form of a dividing egg. acid sea-water: it then proceeds to complete the division at the normal—relatively slow—rate. The distinction between the two phases of recovery from the acid is most marked. SURFAOE TENSION AND CELL-DIVISION 241 There can be but little doubt that the explanation of these facts is as follows: the acid sea-water raises the surface tension at the egg-surface, and tends to make the egg regain its spherical shape. Owing, however, to some force, which elongates the main axis of the egg, equilibrium is reached (during the latter stages of cleavage) when the egg is cylindrical and not spherical.’ It will be noted that the increase in length of the main axis, which was noticed in normal cleavage, is much more obvious in acid sea-water owing to the abolition of the cleavage furrow. It has been shown that when the egg is removed from acid to normal sea-water the cleavage furrow reappears at once. The amount to which the cleavage furrow develops depends entirely on the equilibrium between (a) the surface tension at the egg-surface and (b) some other force within the egg. In acid sea-water the surface tension is high and equilibrium is reached when the egg is a cylinder with hemispherical ends ; in normal sea-water equilibrium is reached with a well-marked cleavage furrow. Whereas the effect of a change in surface tension is very rapidly reflected in the form of the cleavage furrow, it is also clear that the elongation of the main axis is the active process whereby free energy is supplied to the egg- surface and allows the furrow to form under normal conditions. This process is stopped in an acid solution (like so many other physiological processes) and is resumed on return to normal sea-water. The rate at which this force acts is entirely indepen- dent of the experimental rate of change of the surface tension of the egg-surface. That the elongation of the egg axis is the active process involved is shown from the experiment of Plateau.” If a drop of oil be placed between two metal rings A and B so as to form a complete cylinder (Text-fig 6, a), and if the rings be now moved apart, then when the distance of A to B becomes 1 The relative surfaces which enclose an equal volume of protoplasm are (i) Sphere : i : 100 (ii) Cylinder with hemispherical ends . 105 (111) Two spheres (each half vol. of i) . 126 * Statique des liquides ’, vol. ii. 949, J. GRAY greater than 2 of the diameter of the rings (‘lext-fig 6, b), the form of the drop changes in the same way as that of a dividing ege. The further A is moved away from b the more convex do the ends of the cylinder become, and the more marked is the development of the ‘cleavage’ furrow; finally the drop is resolved into two separate parts. There is, however, one respect in which the protoplasmic system differs from that of an oil-drop. When a drop of oil is divided into two—as in Plateau’s experiment—it is a simple matter to reverse the process and reform a single drop of oil. TEXxtT-FIG. 6. (a) (b) Form of completely divided egg in (a) normal sea-water. (b) Acid sea-water. In the case of the living cell this does not occur. It would seem that this is due to the existence of a surface layer (Traube membrane) which is automatically formed when protoplasm comes into contact with water, and that the blastomeres fail to fuse with each other just as oil-drops fail to fuse together if they are shaken in smaller drops in the presence of a soap or any similar substance which can form a condensation layer at the surface of the oil. The conclusion reached is that division of the cell is brought about by the elongation of one axis of the cell, and that the cleavage furrow results as an equilibrium between this process and the normal surface tension at the cell-surface. It need SURFACE TENSION AND CELL-DIVISION 9438 hardly be mentioned that the existence of any form of mechani- cal membrane, or the presence of elements (e.g. other cells) capable of exerting a pressure in any particular direction will materially alter the system which is under discussion. It can hardly be doubted that the elongation of the cell-axis is associated in some way with the elongation of the astral TExt-FIG. 7. figure. Since the latter process goes on continuously during mitosis and does not begin with the elongation of the whole cell, it follows that the elongation of the cell is probably the result and not the cause of the elongation of the astral figure. If this conclusion is correct the cell can be divided into three parts ; a central cylindrical portion ABDC (Text-fig. 8) which is tending to increase the length of its shorter axis, and two convex ends to this cylinder AEB and CFD. Until the ratio AC AB approaches 4 the form of the cell will not change, but as 9.44 J. GRAY this figure is approached the sides AC and BD will begin to flatten ; as soon as is )2, a definite cleavage furrow will AB result between A and C and between B and D. At the same time the convexity of the surfaces AEB and CFD will increase. If the change in length of the axis AC is dependent upon a change in the distance of one centrosome from the other, then one would expect to find a similar relationship between this latter distance and the sector AB. The following measure- ments appear to show that in a variety of cells the cell begins to be deformed in appearance at the equatorial region when the distance between the asters is about -67 of the sector of that part of the cell lying between the asters, and that the cleavage furrow becomes well marked when the ratio has reached the value ‘8 or -9. Ratio Stage of | Length of Axes, Kv T ype. Authority. | Division. xy | AB | == Multicilia lacus-|Lauterborn | 2g | 1:15 2-0 58 tris | gee | Diplogaster longi- | Ziegler 328 115 | 1-9 ‘60 || Average cauda (2 celled) | | 330 ; &Y Diplogaster longi- | Ziegler ena | 1:7 2-65 “64> Sit aE cauda (4 celled) fide fo es Acanthocystis Schaudinn | 6.6 5 2 | 1-45 2-10 -69 Ascaris Boveri |" SESE 2-2 2-9 ‘16 | Multicilia lacus- Lauterborn | (| 1°85 2-15 86 | tris Og ' Diplogaster longi- Ziegler zo 1-27 1-64 7 i | cauda , : Spe i ee Average Diplogaster longi- Ziegler ae |] 1:36 7 3278 “80 | XY cauda FE=3 | | Ap 789 Diplogaster longi- Ziegler ae | 1:15 1-38 83 || cauda (4 celled) 24 | | Acanthocystis Schaudinn 1-80 2-00 “98 TT} Ascaris Boveri | 2°65 | 2:9 9] Prior to any visible change in the appearance of the cell the ratio AB 3S progressively increasing from the beginning of the formation of the astral figure ; it then passes through the critical value at which the furrow appears and continues to increase until complete cell-division is effected. SURFACE TENSION AND CELL-DIVISION 245 It is somewhat rash, perhaps, to press too far the analogy of the astral figure to the rmgs in Plateau’s experiment. It is obvious that the mechanical model is a limited one, and that subsequent work may show that regions of differential viscosity such as are suggested by Chambers’s! work may be found to be involved. Yet in the particular case of a cell which possesses a structure curiously fitted to play the part of Plateau’s rings, division is found to proceed on just those lines demanded by the above analysis. Text-fig. 9 shows the division TEXT-FIG. 9. Division of Cole ps hirtus (after Doflein). The undivided form gives rise to two daughter individuals by passing through stage B. of the protozoon Coleps hirtus. The body is covered by four hard skeletal plates P,—-P,. When division occurs the distal ends of the cell remain fixed in form owing to the existing plates, and between plates P. and P, the cell becomes elongated in exactly the form demanded by the hypothesis put forward. In the stage of division illustrated a well-marked cleavage furrow has formed and the ‘ Plateau’ ratio is about -8. SUMMARY. Cell-division may be accounted for by the movement of the two asters away from each other. The appearance of the cleavage furrow is due to an equilibrium between the effect of this movement on the protoplasm and the surface tension at the cell-surface. The behaviour of the cell under these forces is precisely similar to a drop of oil subjected to similar condi- tions. There is no necessity to postulate regions of differential surface tension at the poles or equator of the cell. 1 * Journ. Exp. Zool.’, vol. 23, p. 483, 1917. NO. 262 Ss 4 _ oat PPT Te wT re ee - wf & = 7 q 7 cd td a2 2 , ay 2 = ee au @ — =4 7 4 . - @ ‘< - ul : Ay — 7 ( bo . rd hh EW ah, AO ek ga + eee *y lj : ie ' mod te, ‘ wens Paste 1st = er taie 7 oy eure. 2 OP wired oben a it ETE. kk = m V« ata Clk eerany Le a le ee ae if. GATM x etal) Fh 1) re ; er mes PS ee i ree: ys iF ir - hls leita) Bre) x. t oA eek oe ate J ei eee ’ 7 9 4 . - \4 ’ é <. ~¥TT ‘ ® big (he ft » st te “ Aik 4 » — f * . 4 pie “i rd - - : ey Sr See r — +s aoe ei Wie 1. » ' ’ ' \ ane ¢ a) 5 eo = fie ‘ tin On the Classification of Actiniaria. Part [11.—Definitions connected with the forms dealt with in Part. Els By T. A. Stephenson, M.Se. University College of Wales, Aberystwyth. CONTENTS. \ PAGE |, DEFINITIONS : 3 ; ‘ : 3 : : . 247 2. APPENDIX : } : : ; ; ‘ : ; 2 eu 3. List oF LITERATURE - : : : - . F a old 4. INDEX TO GENERA DEALT WITH IN Parts II ann III . 2 _ Sallec 1. Derinitions. N.B.—In the following pages only a necessary minimum of synonymy is given ; species-lists are not necessarily exhaustive ; and dubious forms are often omitted. Except where stated explicitly, all reasons for the classification here used will be found in Parts J and IT. Sub-class ZoaANTHACTINIARIA, Van Bened. (As used by Bourne, 1916, pp. 514-15 = DopEcacorRaLiia. Carlgr., Bronn’s ‘ Thierreich *, 1908.) Order DoDECACTINIARIA. (As used by Bourne, 1916, p. 515.) Sub-order ACTINIARIA. Tribe 1. Protrantunar, Carler. Founded by Carlgren in 1891: used here in its original and narrow sense, as covermg Gonactinia and Protanthea NO. 262 T 248 T, A. STEPHENSON (and probably Oractis) only, not in the wider sense of Carl- gren’s later work. Actiniaria with or without a definite base, but without basilar muscles. The body is smooth. There is a sheet of longitudinal muscle-fibres in the ectoderm of body-wall and actinopharynx as well as of dise and tentacles, and the body-wall ectoderm at least has also spirocysts. Sphincter absent or weak diffuse endodermal. Tentacles few or more numerous, simple. In normal animals only eight mesenteries are perfect, and these are the analogues of the eight macrocnemes of Edwardsia. The mesenterial muscle is weak, often hardly more developed than the body-wall muscle, and not forming a very definite retractor usually. The number of mesenteries present beyond the eight protocnemes varies, but the eight are not sharply marked off from the others as macrocnemes, although they have a certain predominance, especially in Oractis. Four imperfect mesenteries pair with the lateral proto- cnemes, and the rest form a secondary cycle or cycles. The distribution of gonad and filament may affect the protoecnemes only (Oractis), or the protocnemes and their lateral partners (Gonactinia), or the whole of cycle 2 as well (Protanthea). The filaments have no ciliated tracts. There are no well-marked siphonoglyphes. Family 1. Gonacrinrpar, Carlgr., 1893. Used here sensu stricto for Gonactinia, Protan- thea, and Oractis only. With the characters of the tribe Protantheae. GonacTINiA, Sars, 1851, p. 142. jonactiniidae with the gonads confined to the lateral protoenemes normally, whereas the filaments are found not only on these but also on their partners and on the directives. There is a definite base. Repro- duction often asexual. Species : G. prolifera, Sars, 1835, p. 3. (See Carlgren, 1893. p. 31.) ProtantHEaA, Carlgr., 1891. Gonactiniidae with gonads and filaments on the mesenteries of cycles 1 and 2, and beyond these cycles small mesenteries devoid of appendages and confined to the uppermost parts of the body. There is a definite base. Species : P. simplex, Carlgr., 1891, p. 81; 1893 p. 24. —— a —— — CLASSIFICATION OF ACTINIARIA 249 Oractis, MeM., 1893, p. 138. Gonactiniidae (?) in which only the eight protocnemes are fertile and filamented. There is no definite base, and there are only ten tentacles. Species : O. diomedeae, MeM., 1893, p. 138. This genus is not yet very fully known, but is probably referable to this family. Tribe 2%. PryvcHopDACTHAE, mihi. Containing the family Ptychodactidae only. Actiniaria with a definite base, which may rather merge into the column, but without basilar muscles. Body-wall smooth or with vertical rows of hollow outgrowths ; in structure, however, similar to the tentacles, with ectodermal muscle-sheet and at least usually spiro- cysts. Sphincter absent or weak diffuse endodermal. Tentacles few or more numerous, simple. Actinopharynx either quite rudimentary and reduced to a narrow band, or else quite well developed and provided with siphonoglyphes and ectodermal muscle-fibres. Six to twelve or more pairs of mesenteries perfect. Musculature of mesenteries weak, hardly forming retractors. The free borders of the mesenteries (or their representatives if the mesenteries fuse below) are occupied by filament above and gonad below, not both together. The filaments have no ciliated tracts, but those of the imperfect mesenteries end up above in a curious structure like a bisected funnel, unusual in form and make-up. Family 1. Prycnopactipar, Appelléf, 1893. dee also Carlgren, 1911, p. 12, &e. With the characters of the tribe Ptychodacteae. Genera: Ptychodactis, Dactylanthus Prycnopactis, Appellof, 1893, p. 4. Ptychodactidae with about 100 tentacles or more. Smocth body. No sphincter. Actinopharynx rudimentary, reduced to a narrow band just inside the lip, which is produced at certain points into lappets for the attachment of larger mesenteries. Mesenteries irregularly arranged ; primaries and usually secondaries perfect. Species : P. patula, Appelléf, 1893, p. 4. (See also Carlgren, 1911.) Aus 250 T, A. STEPHENSON DactyLtantuus, Carlgr., 1911, p. 2. Ptychodactidae with twenty-four tentacles, Body with twenty-four vertical rows of hollow outgrowths or vesicles, corresponding to the twenty-four regular endocoels and exocoels. Sphincter very weak diffuse. Actinopharynx quite well developed, with two siphonoglyphes and with curious pockets between the insertions of some of the mesen- teries. Twelve pairs of mesenteries, six pairs or all of them perfect, all fusing together down below in the gonad region, in such a way that the gonads occupy no longer the now non-existent free edge of the mesentery, but the region nearest to the point of fusion. Species : D. antareticus, Clubb, 1908, p. 5. (See Carlgren, 1911.) Tribe 3. NynantrHRraAg, Carler. Used here in a different sense than that of Carlgren. so that it excludes Edwardsiaria, Corallimorphidae, and Discosomidae, but includes Boloceroides and the Endocoelactids. Actiniaria with or without a definite base, with or without basilar muscles. Body-wall smooth or with verrucae or outgrowths of one sort or another. The presence of a sheet of ectodermal muscle in body-wall or actinopharynx is exceptional, occurring sporadically, and sometimes reduced to a vestige such as ectodermal muscle in the siphonoglyphes. Spirocysts in body-wall ectoderm are also exceptional save in Endo- coelactaria. A sphincter may or may not be present, and if present may be weak or strong, endodermal or mesogloeal. Tentacles few or many, simple or complex, their longitudinal musculature ectodermal or mesogloeal. Siphonoglyphes are typically present. The mesenterial filaments have ciliated tracts. Pairs of perfect mesenteries are present save in abnormal cases, and usually at least six pairs, often more. Six is a fundamental number for arrangement of parts, but there are a good many deviations. Mesenterial musculature does not often exhibit so low. * a grade as in Gonactiniidae, Ptychodactidae, and many Madreporaria— often it is highly developed, very definitely marked off retractors being formed—cases of weakness are usually sporadic and secondary rather than universal and inherent. Sub-tribe 1. AtTHENARIA, Carlgr. Used here as covering Haleampids and [lyanthids but not dwardsians. Nynantheae representing those forms which being the outcome of « Haleampa-like ancestor have retained more similarity to that CLASSIFICATION OF AC'TINIARIA 251 ancestor than most forms, and live a more or less burrowing life. Size variable, predominating shape vermiform, this being attained in greater or less degree in different cases ; the diameter of the body in some forms, or at least in some states, bearing a fair proportion to its length. There is no adherent base. the aboral erid being a physa, which does sometimes adhere to small objects. There is little or no sphincter, but if present it may be endodermal or mesogloeal. Cinclides often present. Number of tentacles usually small, even at greatest not passing about forty ; not more than one communicates with each endocoel and exocoel. Number of mesenteries similarly limited, and they are either all macro- cnemes or else a division into macro- and microcnemes is to be found— with an intermediate condition in the case of Peachia. Secondary mesenteries develop in the exocoels of the primaries. Sometimes the larvae seem to be parasitic on medusae. Family 1. Hatcampipan, Andres. Used here m the general sense of Andres, 1883, p. 312. Ilyanthidae as used by Gosse, pro parte. Including Halcampoidinae, Appellof, 1596, p. 13; Hal- campomorphinac, Carler., 1893, p. 88; Haleampinae, Carlegr., 1893, p. 88; Monaulidae, Hertw., 1882, p. 104; Hahanthinae, Kwiet., 1896; and‘ Fenja’ and‘ Aegir’. Athenaria of more or less vermiform shape, with or without suckers or papillae on the body, with or without cuticle or incrustation. Cin- clides may occur in the physa, Tentacles eight to twelve, fourteen, twenty, or more, and with other variations, their longitudinal muscula- ture ectodermal. Sphincter absent, weak mesogloeal, or weak endo- dermal. The mesenteries have as their main feature six pairs of macrocnemes, but there are variations ; the full six pairs may not be developed, or there may be one or two unpaired macrocnemes in addition to them. Microcnemes are sometimes present, their number varying. Genera: Halcampa, Halcampoides, Pentac- tinia, Seytophorus. Hancampa, Gosse, 1858. See Carlgren, 1893, pp. 37-8; IXwietniewski, 1896, p. 585 ; Haddon, 1889, p. 335 ; Carlgren, 1900 (on Pentactinia), p. 1170, &c. ; and Stephenson, 1918 a, pp. 8-10. Halian- thus, Kwiet. Halianthella, Kwiet. ? Halcam- pella, Andres. 252, T. A. STEPHENSON Haleampidae typically worm-like, more or less, but very changeable in form (see Part II, Text-fig. 7, c, D). There is a physa which may or may not be retractile and which has cinclides in it (always 7), The main part of the body, or scapus, may be without suckers, or it may have suckers to which sand adheres, so as to make a more or less dense cover- ing. A clear external separation into capitulum, scapus, and physa is not necessarily present, Some species have no sheath. Sometimes the scapus has solid papillae. Tentacles retractile, eight to twelve or more (e. g. thirty-two), their longitudinal muscle ectodermal. Sphincter weak mesogloeal (see Part I, Text-fig. 1, and Pl. 22, fig. 7). Mesenteries either all macrocnemes, or else divided into macro- and microcnemes. Macro- enemes six pairs (rarely one or two additional unpaired ones) or fewer ; microcnemes if present variable in number (see Part Il, Text-fig. 8). Species : Genotype, H. chrysanthellum, Peach, Johnst., 1847, p. 220. (See also Gosse, 1860, p. 247; Haddon, 1889, p. 335; Walton and Rees, 1913, p. 65; Haddon, 1886, p. 1; Faurot, 1895, p. 127; Stephenson, 1918 a, p. 9; 1920 4, p. 440.) H. duodecimeirrata, Sars, 1851, p. 142; Carlgr., 1893, p. 38. H. arctica, Carlgr., 1893, p. 45. H. limnicola, Annan., 1915, p. 89. H. aspera, Steph., 1918 a, p. 10. H. chilensis, MeM., 1904, p. 223. H. kerguelensis, Studer. 1878, p. 546. (See Kwietniewski, 1896.) H. arenaria, Haddon, 1886, p. 616; 1889, p. 335, (See also Walton and Rees, 1913, p. 66.) And probably others. I have been obliged to transfer H. aspera from Hal- campoides to Haleampa, because on re-examination of some sections of it I find appearances which I take to indicate a mesogloeal sphincter. The reasons for my overlooking it in my original investigation were that there is not much of it, and it was not until I had subsequently examined several other species with insignificant sphincters that I found out exactly where one must look for it in a deeply introverted and some- what twisted specimen such as mine was. I had only trans- verse, and not the more serviceable longitudinal sections of the region where it lies, being further misled by mistaking certain parts of the endodermal circular muscle for a slight endodermal sphincter. There is also perhaps a fourteenth CLASSIFICATION OF ACTINIARIA 253 perfect mesentery, but if so, whether it has a retractor is uncertain, and it is probably asymmetrical, not placed as in Seytophorus. But only more and better material can clear it up. H. arenaria I have left in the genus on the assumption that it has a mesogloeal sphincter, but that remains to be proved. Hatcamporpes, Dan., 1887. See Appelléf, 1896, p. 8, and the references given under Haleampa. Fenja, Dan. Aegir, Dan. Halcam- pomorphe, Carlgr. Halecampa as used by Kwiet- niewski. Halcampella as used by Hertwig. ? Hal- campella, Andres. Halcampidae typically more or less worm-like, not necessarily with a clear distinction into scapus, capitulum, and physa. Cinclides may occur in the physa. Naked or incrusted. Tentacles twelve or more, can be retractile, and with the tentacular longitudinal muscle ectodermal. No mesogloeal sphincter, but there may be a slight endodermal one. Six pairs of macrocnemes. Microcnemes present or not. Species: H. abyssorum, Dan., 1887. H. clavus, Q. and G., 1833, p. 150. (See Hertwig, 1882, p. 92 ; Pax, 1912, p. 310; Appelléf, 1896, pp. 3, 18, &c.; and Haddon, 1889, p. 336.) H. maxima, Hertw., 1888, p. 29. (See Wassilieff, 1908.) H. kerguelensis, Hertw., 1888, p. 28. H. purpurea, Studer., 1878, p. 545. (See Kwietniewski, 1896.) And probably others. H. minuta, Wass., seems to be more like a Haloclava than a Haleampid. It is possible that of the species listed some may be synonyms of others—-it has been suggested that H. clavus is the same as H. purpurea and H. abys- sorum. Halcampella endromitata, Andres, and others, cannot yet be definitely allocated. In Wassilieff’s description of H. maxima there are some remarks which suggest that he had some form other than a Halcampoides to deal with; but they may be due to a couple of critical misprints. 254 T. A. STEPHENSON Prntactinia, Carlgr., 1900, p. 1166. Halcampidae with body which may be long, and physa. Seapus with papillae to which fragments may adhere, Tentacles typically twenty, their longitudinal musculature ectodermal. No sphincter, Ten macrocnemes present—the ‘ Edwardsia eight ’+one couple pairing with the dorsolateral protocnemes. The sixth primary couple repre- sented by two perfect but weak mesenteries. Four pairs of micro- enemes, confined to distal part of body. Species : P. californica, Carlgr., 1900, p. 1166. ScyropHorus, Hertw., 1882. p. 104. Halcampidae with body which may be long, with cuticle developed chiefly on the scapus. The physal end may attach itself. No verrucae. Tentacles fourteen, their longitudinal musculature ectodermal. Mesen- teries fourteen, the usual primary six pairs+ one couple, the individuals of the couple with their retractors facing one pair of directives. All these mesenteries are macroecnemes, but some may be without gonads. No sphincter. The ciliated tracts of the filaments may be peculiar. Species : S. striatus, Hertw., 1882, p. 104. S. antarcticus, Pfeff., 1889, p. 11. (See Carlgren, 1899, p. 7.) Family 2. InyanrTuipar, Gosse. Ilyanthidae as used by Gosse, 1860, pro parte, Le. excluding Ceriantharia, Edwardsiaria, and Haleampids. Athenaria often attaining a fair size, frequently with stout bodies which are often capable of becoming vermiform. Suckers present or absent, Cinclides may occur in the physa or on the scapus. Patches of cuticle are sometimes present. Tentacles simple or capitate, eight, twelve, twenty, or more, up to about forty. Sphincter absent or slight and endodermal. Never fewer than ten pairs of mesenteries in adult animals, the number varying up to about eighteen pairs. They are usually all macrocnemes, even if there is some distinction among them. though in Peachia four at least of the ten pairs are imperfect and without gonad or filament, but they have strong retractors and cannot be called microcnemes. There is often only one siphonoglyphe, and this may bear a specialized upper end or conchula. Genera: [lyanthus, Eloactis, Haloclava, Harenactis, Peachia. Polyopis striata, Hertw.. may just possibly have been a battered member of this family. CLASSIFICATION OF AG'TINIARIA 955 Inyantuus, Forbes, 1840. Ilyanthidae with a physa. Body-form may be thickish, without suckers, but there may be patches of cuticle. Margin of scapus forms a collar with a narrow capitulum above it. Tentacles simple, in 3 cycles, from about 28 to 36. Noconchula., Mesenteries the same in number as the tentacles, all macrocnemes (perfect, with circumscribed retractor, and fila- ment), but not all fertile. Seven tentacles form the primary cycle ; these are held permanently over the mouth, and divide up those of the outer eycles into radiating groups. The arrangement is exactly bilateral, and not radial; one directive tentacle is a primary, the other a secondary. Tentacles of cycle 3 the longest. Species : I. mitchelli, Gosse, 1853, p. 128; 1860, p. 232. ?I. scoticus, Forbes, ‘ Ann. N. H.’, 1. v. 183. The only species to be certainly referred to this genus is I. mitechelli, which I have been able to study alive and anatomically. It is a unique and extraordinary form, and further details about it will, I hope, be shortly forthcoming. It is clear that I. parthenopeus, Andres, 1s something quite different from I. mitehelli, and merits at least a distinct genus, and as it seems to me a distinct family ; see Andresiidae for further detail. Knoactis, Andres, 1883, p. 464. Ilyanthidae with a physa which may adhere slightly. Column without verrucae, its upper margin well marked. Tentacles twenty, capitate, not fully retractile, the outer largest; tentacular longitudinal muscle ectodermal; tentacle-heads especially rich in sting-cells. No con- chula. No sphincter. Mesenteries ten pairs, all macrocnemes: (see Part I, Text-fig. 9), probably all fertile. One siphonoglyphe. Species : E. mazeli, Jourd,, 1880, p. 41. (See Faurot, 1895, p, 152; and Rees, 1913, p. 70.) Perhaps others. Hanocuava, Verr., 1899, p. 41. Ulyanthidae with body which may be long. Base physa-like but capable of adherence to small objects. The body has rows of adhesive suckers above. There is some sort of sphincter. Tentacles twenty, usually clavate. No conchula. Mesenteries ten pairs, all perfect, very muscular, but six pairs are larger and form a primary cycle. 256 T. A. STEPTLENSON Species : H. producta, Stimp., 1856, p. 110, is the genotype, and perhaps, others should also be included. (See Verrill, 1899.) Harenactis, Torrey, 1902. Ilyanthidae with a physa which can flatten into a dise and stick to something. The body may attain great length; it is smooth, but has a vertical row of cinclides in the upper part of it, corresponding to each endocoel and exocoel. No conchula. No sphincter. Twenty-four simple tentacles. One siphonoglyphe. Longitudinal musculature of tentacles ectodermal. Mesenteries all macrocnemes, twelve pairs, only the six primary pairs usually fertile. Species : H. attenuata, Torrey, 1902. Pracuia, Gosse, 1855. Siphonactinia, Dan, and Kor., 1856. Uyanthidae with a physa, in which there are typically cinclides. Column can be thick, or can become vermiform by attenuation (see Part Il, Text-fig. 7, A), without verrucae, Pp Possibly A. erythraea, H. and E., and others may come here. Probably Comactis flagellifera, Dana, is only A. suleata. I have fuset Anemonia and Isactinia and made one definition cover both, because I cannot feel convinced of any real distinction between them. The main point seems to be a slightly different grade of sphincter, but it is not enough for separation in a family like this. Gyrostoma, Kwiet., 1898, p. 424. Paranemonia, Carlgr., 1900, p. 61. Actiniidae with smooth body and a more or less well-marked margin and usually some sort of fosse. There are no acrorhagi, but sometimes the margin is notched. Tentacles simple, their longitudinal musculature ectodermal. Sphincter absent, or diffuse or circumscribed, but not very strong; sometimes with a mesogloeal tendency though actually endodermal. Retractors weak or strong, diffuse to circumscribed diffuse. 268 T, A. STEPHENSON Gonads may appear from cycle 1 onwards. Siphonoglyphes variable in number, there may be several—also directives ; latter may be absent. Species : G. hertwigi, Kwiet., 1898, p. 424. is the genotype. (See Haddon, 1898, p. 420.) G. ramsayi, H. and8., 1893, p. 124; Haddon, 1898, p, 420. G. kwoiam, H. and §., 1893, p. 125; Haddon, 1898, p. 422. G. cinerea, Cont., 1844, p. 183. (See Pax, 1907, p. 36.) G. tristis, Carlgr., 1900, p. 36. G. dubia, Carlgr., 1900, p. 38. G. Stuhlmanni, Carlgr., 1900, p. 39. G. Sancti-thomae, Pax, 1910, p. 177. G. incertum, MeM., 1904, p. 230. G. selkirkii, McM., 1904, p. 227. G. dysancritum, Pax, 1907, p. 48; 1909. G. haddoni, Lager, 1911, p. 229. G. suleatum, Lager, 1911, p. 230. Other species which may be referred with a query to the genus are G. tulearense, Pax, 1909, p. 404; G. inequale, MeM., 1893, p. 149; G. adhaerens, H. and E., 1832, p. 258 (Pax, 1907, p. 51); G. dichogama, Kirk and Stuckey, 1909, p. 384; G. olivacea, Hutton, 1878 (Stuckey, 1909, p. 381); and G. insessa, Gravier, 1918, p. 3. The above list includes forms which have been erroneously placed under Anemonia, and had to be transferred. The two genera have been a good deal confused, and especially when dealing with preserved material it is difficult to be certain about them. I have included Paranemonia as repre- sented by P. cinerea in this genus, because it seems to me to be simply a Gyrostoma without directives, and not worthy of a distinct genus. G. glaucum, Annan., is no Gyrostoma, and needs a new genus (see p. 263). Conpyuactis, D. and M., 1866. Cereactis, Andres, 1880. ?Ilyanthopsis as used by Hertwig for I. longifilis (1888, p. 13). Actiniidae which sometimes reach a large size (C. passiflora may be a foot across). Body with verrucae in the upper part, which may — be well developed, or weak, or practically or even entirely absent ; sometimes they are arranged in vertical rows; fragments may adhere CLASSIFICATION OF ACTINIARIA 269 to them. There are no acrorhagi, but there is a well-marked margin or collar. Tentacles simple, may be long and large, their longitudinal muscle ectodermal or with occasional anastomosis. Sphincter absent or very weak diffuse. Strong retractors. As a rule the mesenteries are all or mostly perfect and fertile, directives may be sterile; more rarely only twelve pairs perfect. Brood-pouches sometimes occur in the females. The tentacles and mesenteries may run in eights or tens, &c., as well as sixes. Species : C. aurantiaca, D. Ch., 1825, p. 438. (See Andres, 1883, p. 455 ; Pax, 1907, p. 22.) C. passiflora, D. and M., 1866, p. 31. (See Pax, 1910, p. 171; MeMurrich, 1889, * Journ. Morph.”) C. georgiana, Pfeff., 1889, p. 15. (See Carlgren, 1899, p. 13.) C. kerguelensis, Studer, 1878, p. 524. (See Pax, 1907, p. 32.) C.erythrosoma, H. and E., 1832, p. 257. (See Pax, 1907, p. 30.) ?C. cruentata, Dana, 1849, Syn. p. 8, pro parte. (See Carlgren, 1899, p. 10; Pax, 1907, p. 26; McMurrich, 1893 and 1904; and Clubb, 1908, p. 2.) There is a certain amount of ambiguity about this genus, or about the way im which it has been understood. Some forms have been split off from it and established under the separate name Parantheopsis, and these (P. cruentata and P. ocellata) seem to have acrorhagi of some sort, and on account of this and their lack of sphincter they stand half-way between Condylactis and Bunodactis. For Condy- lactis is essentially a genus with smooth collar and no acrorhagi; and Bunodactis is wide enough in its limits already, without the inclusion of sphincterless forms. ‘To avoid too great a fusion of genera it is perhaps wisest to retain three : Condylactis for forms with smooth collar and no acrorhagi or appreciable sphincter; Parantheopsis for such as have vertical rews of verrucae and also acrorhagi but little or no sphincter: and Bunodactis for those which have vertical rows of verrucae, usually acrorhagi, and some sort of sphincter— this may, admittedly, be weak, but typically is circumscribed. It seems possible that two distinct species have been described and confused under the name cruentata; the descriptions rather suggest this, and that one of the two isa Condylactis 270 T. A. STEPHENSON and the other a Parantheopsis. MHertwig’s Ilyan- thopsis longifilis is probably C. passiflora. C. hert- wigi, Wass., is no Condylactis. It has, so far as one can tell, good acrorhagi and a weakish circumscribed-diffuse sphincter. If, as stated, it has no verrucae it should go to Anemonia. Ifithas,to Bunodactis. C. parvicornis, Kwiet., does not seem to be a very typical Condylactis, possibly that also is a Parantheopsis or Bunodactis. ParRANTHEOPsIS, MeM., 1904, p. 233. See note under Condylactis. Actiniidae with verrucose body, the verrucae usually above; they are in vertical rows, ending, at least some of them, in acrorhagi (which are not necessarily nematocyst batteries); foreign bodies sometimes adhere to the verrucae. No sphincter or only a trace. All or a good many of the mesenteries perfect, and all may be fertile save directives. Good retractors. Tentacular longitudinal muscle of the simple ten- tacles ectodermal. Tentacles and mesenteries may be hexamerous or octamerous. Species : P. cruentata, Dana, 1849, pro parte. See under Con- dylactis cruentata for references. P. occellata, Les., 1828, p. 79. (See McMurrich, 1904.) Bunopactis, Verr., 1899. Cribrina, Ehr., pro parte; Bunodes, Gosse, 1855 ; Aegeon, Gosse, 1865; Anthopleura, D. and M., 1860; Aulactinia, Verr., 1862; Evactis, Verr., 1868; Bunodella, Verr., 1899; Actinioides, H. and §., 1898. Actiniidae. A large genus of forms which are some of them easily retractile, others of more lax habit and only retractile with diffi- culty if at all. The body has regular vertical rows of verrucae, which are sometimes graded in length, and in size of the individual verrucae, according to the cycles of mesenteries they are connected with, and this may be accompanied by colour distinctions between the verrucae. Foreign bodies are often attached to the verrucae, which may also be somewhat lobed, distally. Acrorhagi are usually developed in connexion with the upper ends of at least some of the rows ; but they may be there or not even in one and the same species. They may be simple, small or large, or decidedly compound. Sphincter variable, never very CLASSIFICATION OF ACTINIARIA 271 powerful, but ranging from weak diffuse to more definite diffuse and slightly or distinctly circumscribed, weak or moderate in development ; the state may vary even within one species, but the circumscribed form is the more usual and may attain a good strength. (Some of the weaker Bunoda:tis-sphincters are illustrated in Part H, Text-figs. 11 , F, and 12, p, £.) Tentacles simple, their longitudinal muscles ectodermal. Retractors often strong, diffuse to circumscribed diffuse or even circum- scribed. Gonads may appear on the older mesenteries, or all mesen- teries may be perfect and fertile, save sometimes the directives. Brood- pouches may occur. Siphonoglyphes and directives variable in number. Symmetry hexameral, octameral, &c., or irregular. Species : The genotype is B. gemmacea, Ellis and Solander, 1786, p. 3. (See Gosse, 1860, p. 190; G. Y. and A. F. Dixon, 1889, p. 321.) The other British species are B. thallia, Gosse, 1854, p. 283 (see Gosse, 1860, p. 195; G. Y. and A. F. Dixon, 1889, p. 310), B. ballii, Cocks, 1849 (see Gosse, 1860, p. 198), and B. alfordi, Gosse, 1865, p. 41. Foreign species numerous. This is a genus somewhat parallel to Sagartia among the Mesomyaria. Its synonymy has been much discussed. ‘he genus Cribrina, like Urticina, seems too vague to be adopted. Against one’s wishes it seems necessary to let the familiar Bunodes lapse, since the name was pre-occupied for a Eurypterid in 1854; and Verrill’s name Bunodactis steps into the breach with several synonyms. There may be some forms which have been wrongly placed under the genus, and their position should be reconsidered if they do not come under the above definition, which is wide enough already. The genera Bunodactis, Anthopleura, and Aulac- tinia have already been fused by Torrey (1906, pp. 47-52), and I fully agree with him that there is no valid way of separat- ing them. I now add Actinioides to the list of synonyms. It has always represented the Bunodes species with the weaker sphincters, but apparently because of its being placed in the Actinidae while Bunodes was placed in the Buno- didae, the similarity was overlooked. It is now evident that the Bunodidae and Actiniidae are one aud the same family (see Part Il, p. 526), and the two genera can no longer be 972, T, A. STEPHENSON separated. The sphincters show a graded series from weak to fairly strong, and from diffuse to more circumscribed, so that it would be difficult to draw a boundary line (ef. Part TI, Text-figs. 11, p, r, and 12, p, n, for some of the weaker ones). And there are no other special points of difference. B.¢api- tata is said to have only six pairs of perfect mesenteries, in which case it should be transferred to Macrodactyla, and then the latter name might have to give way to the earlier name Aulactinia, ip a new sense. It is possible of course for young specimens of a species to have only six pairs of per- fect mesenteries, but if it persists in the adult the foi1m needs exclusion both from Bunodaetis and from the Actinidae too. 'T'HALIA, Gosse, 1858, p. 417; 1860, p. 205. Rhodactinia, Agassiz, 1847, p. 679; 1865, p. 13. Actiniidae. Body sometimes low and broad, attaining however a high cylindrical form in full expansion. Body in some specimens very mobile and changeable in form, in others less so, Column with verrucae which are not usually arranged in definite vertical rows, at any rate in the adult ;. they are sometimes strongly, sometimes very weakly developed, in still others quite absent—and all this within one and the same species, probably. There is a parapet and fosse, but no acrorhagi in the ordinary forms—they occur in certain Antarctic cases though. Tentacles simple, stout, their longitudinal muscle varying from ectodermal to meso- gloeal, the latter being perhaps the more typical condition in the common forms. Strong circumscribed sphincter. Tentacles and mesenteries often in multiples of ten in the common forms, but not invariably. Primary mesenteries may be fertile or sterile. Retractors strong or very strong, diffuse or circumscribed diffuse. All mesenteries may be perfect. (For T. crassicornis, see Part II, Text-figs. 7, B, and 12, B.) Species : , T. crassicornis, O. F. Miller, 1776, p. 231. (T. coriacea, Cuvier, 1797, p. 653.) (See also Gosse, 1860, p. 209; Carlgren, 1893, p. 58; Carlgren, 1902, p. 38; Clubb, 1908, p. 9; MeMur- rich, 1911.) (Bolocera eques, Gosse, 1860, p. 351; Madon- iactis lofotensis, Dan., 1887, p. 47, pro parte.) T. carlgreni, Clubb, 1902, p. 297. T. sulcata, Clubb, 1902, p. 295. Tealia seems to be the famous genus for synonymy- discussions, and I will add as little as possible. I cannot CLASSIFICATION OF ACTINIARIA 278 pretend to go into detail about it, but I venture to support Tealia as the best name to use, even if the legality is doubt- ful—in any case something would be doubtful. Urticina is too ambiguous, and although it probably contained one of our common forms it seems justifiable to reject it in favour of the non-ambiguous Tealia. Rhodactinia, Agassiz, has priority, but although R. davisii seems to be identical with one of the forms more usually called crassicornis or Goriacea, yet the genus is insufficiently described and is not free from ambiguity. Tealia is well defined and a familiar name, and seems clearly to have the advantage. I cannot accept Carlgren’s division of the genus into two distinct genera—Tealia and Rhodactinia—because it cannot (as MeMurrich has already pointed out) be upheld by anything stable. With regard to species under the genus, I do not like to speak with finality ; but my experience with living specimens and study of literature suggests that there is really no valid way of splitting up our British forms into species—I imagine that they are all races or forms of one variable species, but the warts and other things do vary a great deal. This has not always been my opinion, and it is the kind of point about which one is hable to change one’s mind more than once—the other possibility being to regard ow British forms as two species (the extremes are certainly very different from each other in appearance) with intermediate grades. Clubb’s two Antarctic species are certainly distinct from ours, but seem very like each other, and in some ways verge towards Bunodactis—might even be members of that genus, though Tealia-like in build and probably best left where they are. As to the right specific name for the British species, it will perhaps always be a disputed thing, but on the idea that we have only one variable species the name c¢rassicornis has priority over coriacea. I have seen Bolocera eques, Gosse, alive, and it is simply a form of T. crassicornis. It has been suggested that the irregularly arranged verru- cae of Tealia must really be in vertical rows since they O74: T. A. STEPHENSON communicate with mesenterial spaces. But although this is true in a sense, 1t smothers the more important fact that in Tealia the total arrangement of warts on the body-wall as a whole presents an irregular appearance to the eye, whereas in the contrasting Bunodactis the warts run in regular vertical series visible as such, and often with a cyclic arrangemeut corresponding to the cycles of endocoels, and even coloured differently according to cycle in some cases, so that in the end the difference is a marked one. Eptactis, Verr., 1868. Kpigonactis, Verr., 1899, p. 378. Glyphoperidium, Roule, 1909, p. 10. Nor Leiotealia, Hertw., 1882, p. 37—which is probably a synonym of Aureliania; see note under that genus on p. 292. Actiniidae in which the base may be well developed, or somewhat reduced, or may be crack-like through attachment to a spine. Wall smooth, no verrucae or acrorhagi, but well-marked parapet and fosse. There may be young ones adherent to the surface, or even shallow or deep brood-pouches. Tentacular longitudinal muscle ectodermal or with a tendency to anastomosis; disc radial muscle ectodermal or meso- gloeal. Sphincter well developed, circumscribed, varying in strength but usually very strong, sometimes exhibiting a good deal of anasto- mosis of processes. Retractors diffuse, circumscribed diffuse, or even circumscribed, often very strong. Mesenteries may be all perfect ; gonads from cycle one onwards, or not; or on all, save perhaps the directives. (For two Epiactis sphincters see Part II, Text-fig. 12, A and ©.) Species : E. prolifera, Verr., 1868, p. 492. (E. fertilis, Andres, 1883, p. 574.) See also Torrey, 1902, and McMurrich, 1901. E. fecunda, Verr., 1899, p. 378. (E. regularis, Verr., 1899, p. 380.) (2? E. spitzbergensis, Kwiet., 1898, p. 134, pro parte.) E. marsupialis, Carlgr., 1901, p. 482. E. badia, McM., 1893, p. 194. EK. thompsoni, Coughtrey, 1874. (See Stuckey, 1909, p. 370.) I. novo-zealandica, Steph., 1918 4, p. 24. K. ritteri, Torrey, 1902, p. 393. E. bursa, Roule, 1909, p. 11. ae CLASSIFICATION OF ACTINIARIA 275 E. vas, Roule, 1909, p. 13. E. dubia, Wass., 1908, p. 20, may be an Isotealia. Tsornatia, Carlgr., 1899, p. 24. Actiniidae with smooth wall and well-marked margin provided with some sort of acrorhagi; there may be cuticle. Sphincter well developed, circumscribed. Tentacular longitudinal musculature ectoderma}. Retractors fairly developed or strong. Gonads absent on older mesen- teries, or present on all save directives. ° Species : I. antarctica, Carlgr., 1899, p. 25, and probably I. dubia, Wass., 1908, p. 20. PsEUDOPHELLIA, Verr., 1899, p. 376. Actiniidae in which the scapus is covered with a thick soft cuticle, the capitulum smooth and with parapet and fosse. There may be brood-pits in the scapus. Sphincter circumscribed. Strong retractors. Species : P. arctica, Verr., 1868, p. 328; 1899, p. 376. This genus seems to be a sound member of the family, but more details about it are desirable. Botocnra, Gosse, 1860, p. 185. Liponema, Hertw., 1882, p. 129. (See Hertwig, 1888, p. 17; MeMurrich, 1893, pp. 160 and 209; Carlgren, 1899, p. 89; Haddon, 1898, p. 429.) Actiniidae with smooth body (no verrucae or acrorhagi) and un- specialized margin which may be tentaculate or have some sort of fold. Tentacles provided with basal sphincters so that they are deciduous ; they are variable in size and number, sometimes very large, at others very numerous, or again neither unusually large nor numerous. Sphinc- ter diffuse, may b2 well developed, sometimes with one or more of its processes predominating in size over the others, or with a tendency to circumscription. Diffuse retractors. Gonads’ may appear on the older cycles, or these may be sterile. Tentacular longitudinal muscle ecto- dermal. Perfect mesenteries variable in number; sometimes all are perfect. Species : B. tuediae, Johnst., ‘Mag. Nat. Hist.’, v. 163. (See Gosse, 1860, p. 186; Walton, 1908, p. 215; Stephenson, 1918 B, p. 112.) B. longicornis, Carlgr., 1891, p. 241. (See Carlgren, 1893, p. 50; Stephenson, 1918 4, p. 20.) 276 T. A, STEPHENSON B. kerguelensis, Studer, 1879, p. 544. (See Kwietniewski, 1896.) B. multicornis, Verr., 1879, p. 198. (See Carlgren, 1902 (Olga), p. 36.) B. brevicornis, MeM., 1893, p. 158. B. pannosa, McM., 1893, p. 156. B. occidua, MeM., 1893, p. 154. B. multipora, Hertw., 1882, p. 129. (See also references given above for Liponema.) The genus Bolocera should be limited to the above defini- tion and list of species, as regards known forms. B. pollens has now a genus apart on account of its sphincter. B. eques is a Tealia. B. norvegica is of doubtful standing. B. africana is, according to Carlgren (1911. p. 21), a Sagartid wrongly described as a Bolocera. LEIPSICERAS, Steph., 1918 B, p. 112. Actiniidae with smooth wall, no verrucae or acrorhagi. Very strong circumscribed sphincter. Tentacles provided with sphincters, therefore deciduous. Species : L. pollens, MeM., 1898, p. 230. This genus was separated from Bolocera on account of the sphincter. It seemed advisable because of the wide gap between the typically diffuse Bolocera-sphincter and the elongate circumscribed muscle with a mesogloeal axis found in L. pollens; there is a gap here, not a series as in Buno- dactis. Bonocrropsis, MeM., 1904, p. 255. Actiniidae with smooth body and tentaculate margin. The only known species has large tentacles rather like those of a Bolocera, but without tentacle-sphincters ; their longitudinal muscle ectodermal. Sphincter and retractors diffuse. Species : B. platei, McM., 1904, p. 255. Whether there is really any sound distinction between this genus and Gyrostoma it is not easy to decide, but pending further knowledge it is safest to let it stand. The nature of the margin may be quite a good distinguishing point. CLASSIFICATION OF ACTINIARIA 277 Dorrernta, Wass., 1908, p. 13. Actiniidae with smooth body. The only described form has large tentacles rather like those of a Bolocera, and both these and the disc are covered with papillae, plainly visible to the naked eye, and which represent batteries of nematocysts. Longitudinal muscle of tentacles ectodermal, Sphincter and retractors diffuse. Species : D. armata, Wass., 1908, p. 14. This is a genus not very clear in its exact relationships, but it does not seem to fuse readily with any other, and the papillose dise and tentacles are a distinct feature. Ixanactis, Haddon, 1898, p. 445. Actiniidae with the wall smooth below, with suckers above, and a definite crenulated parapet. Disc flat when fully expanded, but often thrown into lobes, and not fully retractile. Tentacles numerous, the aboral side of each being smooth, the oral side flattened, and with symmetrical lateral swellings, so that the whole looks not unlike a knotted cruciferous seed-pod in some conditions. Sphincter moderately developed, circumscribed. Species : I. simplex, H.and§&., 1893, p. 123. (See Haddon, 1898, p. 443.) A very distinct genus. Probably the form photographed by Saville-Kent as Condylactis, sp. It is a possibility that the genus is identical with Ragactis as represented by Andres’s figures of R. pulchra—at any rate they suggest it, and it would be an interesting point to follow up. GuyPHostyiuM, Roule, 1909, p. 14. Actiniidae. In the one form described the body is a long trumpet with short stout tentacles. The tentacles are thicker on one face than the other, and their longitudinal musculature is ectodermal and much stronger on the thick side than the other. Smooth body-wall and no sphincter, Species : G. calyx, Roule, 1909, p. 16. Roule (1909, p. 2) has set up a sub-family Glyphactininae of the Antheidae (Actiniidae), to rank equal with two other sub-families, the Bolocerinae and Actininae. His sub-family 278 T. A. STEPHENSON includes two genera, Glyphoperidium and Glypho- stylum. In the first place Glyphoperidium seems undoubtedly identical with Epiactis, and is here included in that genus, with its two species, G. vas and G. bursa. Moreover, the sub-family seems to have been erected as a result of laying too much stress upon some apparently trivial characters, especially connected with the actinopharynx. It is hard to find any justification for such a sub-family, and it is not adopted here. The other genus erected, Glypho- stylum, seems more worthy of distinction, and although its separateness is not very marked it is defined above, provi- sionally at any rate. ANTHEOMORPHE (Hertw., 1882, p. 30) seems barely if at all distinguishable from Gyrostoma. Comactis. C. flagellifera, Hertw., 1882, p. 32, might be almost anything. Dana’s is probably Anemonia sulcata. Ponystomipium (Hertw., 1882, p. 67) can hardly stand. The stomidia seem to be the remains of torn-off tentacles and the oesophageal openings probably ruptures (I have seen the specimen). Of what genus it is a battered representa- tive is another matter. InyantHopsts (Hertw., 1888, p. 13) has to lapse. I. longi- filis, Hertw., is probably a Condylactis, and I. elegans, Wass., is a Synhalcurias (see p. 260). Pax refers to I. longifilis in his 1910 paper, pp. 171, 173, &c., as probably C. passiflora. Myrractis (Haddon, 1888, p. 248) is not easily allocated. It may be a Stichodactyline like Radianthus, or it may stand among the Actiniidae near Condylactis, but there are not quite enough data to make a certainty of it. Gyractis (Boveri, 1893, p. 246) (see Haddon, 1898, p. 445) is very near Bunodactis, if not absolutely identical with it. The fact that it has no directives or siphonoglyphes cannot keep it apart, these things are too much matters of specific or individual idiosyneracy. The doubt about CLASSIFICATION OF ACTINIARIA 279 definitely fusing it up with Bunodactis is that the existence of regular vertical rows of verrucae seems uncertain, although both verrucae and acrorhagi are present. Regular vertical rows are extremely charac- teristic of Bunodactis, but otherwise Gyractis has the organization of that genus. Trauropsis polaris, Dan., and Kynrnprosactis elegans, Dan., are, according to Carlgren’s examination of the original specimens, identical with Stomphia. Mavpont- actis lofotensis, Dan., seems to be a name covering Tealia, Metridium, and Hormathia, ard is invalid. Family 5. AtictIDAk, sens. strict. Aliciidae, Duerden+Phyllactidae, Andres, as used by Haddon, 1898, pp. 483 and 435, both pro parte. Endomyaria with definite base and more or less delicate tissues, The column may be divided into a scapus with vesicles and a smooth capitt.- lum ; or the scapus may be smooth and the vesicles occur where it joins the capitulum, and somewhat higher up as well in some cases. The form may be very changeable. Tentacles simple, variable, may be long, their longitudinal musculature ectodermal. Ectodermal longitudinal muscle may be present in actinopharynx and capitulum, also spirocysts at least in the latter. Sphincter absent or feeble endodermal diffuse. Not more than one tentacle to each exo- and endocoel. Mesenteries NOT divided into macro- and microcnemes. Only six pairs of mesenteries perfect. Genera: Alicia, Phyllodiscus. Aticta, Johns., 1861. Cladactis, Panc., 1868, not Cladactis, Verr., 1868. Aliciidae with delicate column capable of elongation, and divided into scapus and capitulum. The scapus bears vesicles, varying in form and in detail, but at least some of which are compound and stalked. Capitu- lum naked, may have weak longitudinal muscle and spirocysts in its ectoderm; the muscle may also be present in the actinopharynx. The vesicles have numerous sting-cells. which may be very large. Margin may be tentaculate. Tentacles typically long and slender, retractile. The six pairs of perfect mesenteries may be sterile. Retractors not strong, diffuse. Sphincter absent or weak diffuse. NO. 262 x 280 T. A. STEPHENSON Species : A. mirabilis, Johns., 1861, p. 303. (See Duerden, 1897, A. costae, Panc., 1868, p. 30. (See Duerden, 1895, p. 213.) A. sansibarensis, Carlgr., 1900, p. 28. A. rhadina, H. and 8., 1893, p. 127, Haddon, 1898, p. 433. And probably others. Puy uopiscus, wiet., 1898, p. 407. (See Part II, Text-fig. 18.) Hoplophoria as used by Haddon (1898, p. 438) for H. cineta, not as used by Wilson for H. coralligens (1890, p. 379). Aliciidae in which the lower part of the body or scapus is smooth ; at its junction with the upper part or capitulum, which may be delicate and extensile, there is at least one ring of stalked vesicles ; there may be one ring of about six vesicles only, or one complete ring containing a good many more than that, and a few outside and above the ring ; or there may be several series of them, formed by one vesicle com- municating with each of the older endocoels, four or more with each of the younger endocoels and the exocoels. Form of vesicles variable as to detail, more or less ecmpound, Capitulum may have ectodermal longitudinal muscle, its margin tentaculate. The six pairs of perfect mesenteries may be sterile. Retractors:weak diffuse. Sphincter absent or weak diffuse. At their best the vesicles form a wide frill or ruff round the body (see Part II, Text-figs. 18 and 2 4), Species : P. semoni, Kwiet., 1898, p. 407. P. cincta, H. and &., 1893, p. 127; Haddon, 1898, p. 438. P. indicus, n. sp. IT am uniting under this genus (erected by Kwietniewski for P. semoni) three species. One is the Hoplophoria eineta of Haddon and Shackleton, which is a quite distinet form, but does not agree with the type of the genus Holo- phoria. That type, H. coralligens, Wilson, is taken by Duerden (see 1898, p. 456, and 1902) to be a Lebrunia. H. cincta does, however, fit in as a Phyllodiseus, possibly an immature one. The third species is a new one wnich [ have from the Maldive Islands (out of a collection kindly lent me by Professor Stanley Gardiner), and which, though perhaps not fully developed, is much further on than P. eineta, and forms a link between that and P. semoni, CLASSIFICATION OF ACTINIARIA 981 It seems not unlikely that Phyllodiscus is identical with Triactis, but it would be well to wait for the anatomy of T. producta before assuming that and changing the name. A species which might possibly come in here is the one described by Hargitt as Cradactis variabilis. Family 6. PHyLLActTIpAn, sens. strict. Phyllactidae, Andres + Aliciidae, Duerden + Dendromeliudae, MeM., as used by Haddon, 1898, pp. 485, 433, 440, all pro parte. Including Thaumactiniae, Fowler. Endomyaria with definite base. Body-wall variable; it may be wide and provided with vesicles below, narrower and naked above ; or there may be vesicles all over it, with or without acrorhagi at the margin; or the lower part of the body may be devoid of vesicles, and provided only with verrucae, while the vesicles are confined to the sub-marginal zone, really representing foliose acrorhagi, and sometimes forming a very definite collar or ruff ; or again, the sub-marginal region may bear only about six vesicles or ‘ pseudotentacles ’, which at their best form large branching bush-like structures. In spite of this variation vesicles are always present, and both they and any acrorhagi there may be can be simple or compound. Tentacles simple, provided with sphincters in one genus only; so that usually they are non-deciduous ; their longitudinal musculature ectodermal or mesogloeal. There may be ectodermal longitudinal muscle in body-wall and actinopharynx. Mesen- teries Not divided into macro- and microcnemes, more than six pairs, and usually twelve or more pairs perfect, with occasional exceptional individuals. Sphincter absent, diffuse, or circumscribed. Genera: Phyllactis, Cradactis, Phymactis, Cystiactis, Bunodeopsis, Thaumactis, Le- brunia. This family and its contaimed genera present a good deal of difficulty. [have attempted a revision of them, but it may need carrying a good deal further in the light of new knowledge. A number of genera have been described under Phyllactidae, Aliciidae, Bunodidae, Dendromelidae, and Thaumactidae, which need a good deal of sorting out. The principle upon which one must work, of having two families (Aliciidae and Phyllactidae) was introduced in Part II, p. 530, and the x 2 282 T. A. STEPHENSON Aliciidae, sens. strict., have been dealt with above. There remains the set of forms now to be called Phyllactidae, ~To begin with, one suspects that names have been needlessly multiplied, and the mass of forms seems to present seven good genera, with some synonyms. Of these seven, one can say that they exhibit the same general grade of structure as the Actiniidae, but with vesicles added ; but beyond that there are differences and one notes five sets of them. At least two of these sets are the logical outcome along slightly different lines of a further development of Actiniid forms, and may be looked upon as a natural family representing a stage further than the Acti- niudae. In one of these sets (Phyllactis and Cradactis) the acrorhagi of some Actiniid ancestor seem to have developed complications so as to form a sort of ruff, while the verrucae remained the same; in the other set (Cystiactis, Phy- mactis) the verrucae have developed into vesicles, and some- times there are acrorhagi as well. In connexion with the first set, it 18 interesting to note that one gets, now and then, an abnormal individual of Actinia equina in which some of the acrorhagi have become compound, in just such a way as one would expect a beginning to be made in the Phyllactis direction. It is when we come to the other genera that the chief diffi- culty arises. Thaumactis is a small, possibly a young form, of uncertain affinities. Bunodeopsis is very distinet and is now, thanks to Duerden, a well-studied genus; but it is possible to think of it on the one hand as an Aliciid (sens. strict.) which can develop more than six pairs of perfect mesenteries, or on the other as the outcome of an Actiniid which has developed along a line all its own—the ancestor being, even,a pre-Actiniid Boloceroides-like form. Lebrunia could well enough be derived from some Actinud or pre-Actiniid in a special way. Taking them as a whole, all these forms might be derived from forms like Actimidae or pre-Actiniidae, the suggestion of Aliciid origin only coming in strongly in the case of Bunodeopsis. Since we can never know their exact history, and since it seems reasonable to think . EE CLASSIFICATION OF ACTINIARIA 283 of them diverging along different lines from somewhere near the same place, it is probably best to include them all in one family, the Phyllactidae. It will show a good deal of range as to detail, but with the fundamentals in common. An extended discussion of other families which have been involved near Phyllactidae seems hardly necessary. It was pointed out in Part II (p. 580) that Dendromeliidae and Thaumactidae could hardly be upheld. Any inclusion of vesicled genera ike Bunodosoma in the Bunodidae seems to have been a mistake. Some genera here placed in Phyl- lactidae were referred to Aliciidae before, but the revised sense in which the families are here taken necessitates an alteration. Puyuxactis, M. Edw. and H., 1851. Oulactis, M. Edw. and H., 1851; Asteractis, Verr., 1868; Lophactis, Verr., 1868; ?Actinostella, Duch., 1850. Phyllactidae. Column may be capable of a good deal of elongation ; it has verrucae which usually occur in vertical rows and may attach foreign bodies to themselves. Above the verrucae and below the margin proper there is a very definite ruff, frill, or collar, which may be quite wide and conspicuous, and is formed of a number of radiating series of vesicles separated from each other by grooves, and which apparently represent complicated and extended acrorhagi; the whole ruff is separated from the tentacles by a fosse; the detail of the acrorhagi or ‘fronds’ varies in different cases. Sphincter usually circumscribed, more or less, not very strong, but it may be diffuse. Tentacular longitu- dinal muscle ectodermal. Retractors typically strong, diffuse to circum- scribed diffuse. Species : P. praetexta, Dana, 1849, p. 150. (See McMurrich, 1905 on D. and M. Actinians.) P. flosculifera, Les., 1817. (See McMurrich, 1889 (‘ Journ. Morph.’); 1889 (Bermudas).) (P. fasciculata, MeM., 1889 (Bermudas), p. 108.) P. conchilega, D. and M., 1860. (?P. expansa, Duerden, 1898, p. 455.) (See Pax, 1910, p. 194; MeMurrich, 1905; Duer- den, 1898, p. 455, and 1902.) (P. foliosa, Andres, 1883, p. 505.) P. bradleyi, Verr., 1868, p. 465; 1899. P. concinnata, Dana, 1846, p. 152. (See Pax, 1912, p. pv. 12.) 984 T. A. STEPHENSON P. radiata, D. and M., 1860. (See McMurrich, 1905.) P. californica, McM., 1893, p. 196. And probably others. [have done the best [ can with this genus, but the synonymy of its species is a matter for special study. I think MeMurrich has made it clear that Phyllactis is identical with Ou- lactis, Asteractis, and Lophactis—probably also with Actinostella, in which case the latter name would have priority ; but until more is known of the type, A. formosa, it seems better to keep to the well-known Phyllactis. P. striata, Wass., seems more like a Cradactis. Crapactis, MeM., 18938, p. 197. Saccactis, Lager, 1911, p. 220. Phyllactidae in which the column has verrucae, usually in vertical rows, to which foreign bodies may adhere. At the margin there are vesicles in a ring—probably modified and developed acrorhagi; they vary in form, the uppermost at least being somewhat lobed or branched or foliose, and there may be concentrations of nematocysts on them ; when these ‘fronds’ are at their best development they may form a wide frill round the animal. Sphincter diffuse or cireumseribed, usually well developed. Longitudinal musculature of tentacles ecto- dermal. Retractors usually strong, diffuse to circumscribed diffuse. Gonads may or may not appear from the first cycle onwards. Species : C. digitata, McM., 1893, p. 198. . plicatus, Hutton, 1878. (See Stuckey, 1909, p. 392.) . magna, Stuckey, 1909, p. 394. . memurrichi, Lager, 1911, p. 220. . australis, Lager, 1911, p, 223. . musculosa, Lager, 1911, p. 223. . excelsa, Wass., 1908, p. 23. ?C. striata, Wass., 1908, p. 22. De) Gye) Gree I have joined this genus and Saccactis because | cannot find any very serviceable distinction between them, and I think any variation there may be in the form of the vesicles and their continuity with the rows of verrucae probably finds a parallel in Bunodactis, as do also the variability of nematocysts in the fronds and the variations of the sphincter. Cradactis CLASSIFICATION OF ACTINIARIA 985 seems distinct from Phyllactis in that, apparently, the fronds or acrorhagi have not attained such clear distinction from the rest of the column as in that genus, where they make such a very definite zone or ruff. C. variabilis, Hargitt (1911, p. 51), seems of rather uncertain standing—it may come here, or, possibly, under Phyllodiscus. Puymactis, M. Edw., 1857. ivetia, Pax, 1912, p. D5; Bunodosoma, Verr., 1399; Hucladactis, Verr., 1899, p. 49. Phyllactidae in which the column is covered thickly with vesicles, which may be quite without arrangement, or may form more or less definite vertical rows, and sometimes the rows are of different sizes in a regular way according to mesentery cycles. The vesicles may be simple or more or less compound, and sometimes they fuse inseparably with each other. Acrorhagi, which may be compound, may be present (not always), being sometimes well developed and at others hardly distinguishable, within the same species. Above the acrorhagi a fosse. Sphincter weak or strong diffuse, circumscribed diffuse, or small to moderate circumscribed. Tentacular longitudinal muscle ectodermal. Retractors diffuse, weak or well developed, or stronger and circumscribed diffuse. The older mesenteries may be sterile, or most mesenteries may be fertile. There may be more than two siphonoglyphes, which need not correspond to directives—the latter may be absent. Species : P. clematis, Drayton in Dana, 1846, Syn. p. 6. (P. florida, Dana, 1849.) (See Carlgren, 1899, p. 17; McMurrich, 1904; Stephenson, 1918 A, p. 23.) P. granulifera, Les., 1817, p. 173. (Bunodes taeniatus, McM., 1889, p. 23.) (See Pax, 1910, p. 184; McMurrich, 1889 (‘ Journ. Morph.’), p. 23; Duerden, 1902, p. 348.) P. sphaerulata, Duerden, 1902, p. 350. Py kikenthali, Pax, 1910; p. 189: P. papillosa, Les., 1830, p. 78. (See Pax, 1912, p. D. 6.) P. grandis, Verr., 1868, p. 473; 1899, p. 49. IT have included Verrill’s Euecladactis grandis with some hesitation in this genus, followmg MeMurrich. It seems to have almost enough to merit distinction in its very definite cyclic rows of vesicles, rather comparable to the verrucae of Bunodactis gemmacea in arrangement—but there are 286 T, A. STEPHENSON at any rate tendencies in this direction in other species (cf. P. sphaerulata), and I leave it at that for the time being. I think my fusion of Bunodosoma with Rivetia and Phymactis can be supported in the same sort of way as Torrey’s fusion of Bunodactis with Antho- pleura, &c. It has been seen from time to time that sphincter-detail alone cannot always separate species; that acrorhagi are too variable (cf. P. granulifera in which they may be well developed or barely discernible) to be invari- able ground for separation ; and there do not in this case seem to be any valid distinctions to be based on the vesicles. The siphonoglyphe-variation and lack of directives in P. (Rivetia) papillosa has parallels elsewhere, and is hardly in itself of generic weight. Cystiactis, M. Hdw., 1857. (See Haddon and Duerden, 1896, p. 154.) Phlyctenactis, Stuckey, 1909, p. 396. Phyllactidae with the column covered by simple sessile or slightly pedunculate vesicles. No acrorhagi. Longitudinal musculature of tentacles mesogloea]l. Sphincter absent or diffuse. Primary mesen- teries may be sterile. Retractors diffuse. Species : C. tuberculosa, Q. and G., 1833, p. 159. (See Haddon and Duerden, 1896, p. 156; Lager, 1911.) C. retifera, Stuckey, 1909, p. 396. C. morrisoni, Stuckey, 1909, p. 396. ?7C. Koellikeri, Pax, 1910, p. 180. And other older species not yet well known. I have placed Stuckey’s Phlyctenactis under Cysti- actis, with which it seems to be identical, especially if the tentacular muscle is mesogloeal, as his figures lead one to believe. C. Koellikeri may belong here or under Phy- mactis. Cystiactis is distinguished from Phymactis by its simple vesicles, constant absence of acrorhagi, meso- gloeal tentacle-muscle, and on the whole weaker musculature. Bunopeopsis, Andres, 1880. (See Duerden, 1897 and 1902.) Phyllactidae with the body broad and flattish below, covered with CLASSIFICATION OF AOTINIARIA 287 vesicles ; narrower and devoid of vesicles above, thus having a smooth extensile capitulum. The capitulum has a tentaculate margin, and both it and the long tentacles are retractile. The latter have little stinging spots on them, and moreover each has a sphincter at its base, as in Bolocera, so that it is deciduous. The vesicles may be simple and sphaeroidal or compound, sessile or stalked, and have nematocysts in at least parts of their ectoderm. Tentacular longitudinal muscle ecto- dermal. There is ectodermal longitudinal muscle in body-wall and actinopharynx, and there are no siphonoglyphes; but the mesenterial filaments have ciliated tracts as usual, Sphincter absent or weak diffuse. Number of perfect mesenteries variable, about four to twenty pairs, the irregularity probably connected with fission and laceration as modes of reproduction. Retractors diffuse. Tissues delicate. Habitat, weeds, stones, &c. Species : B. strumosa, Andres, 1880, p. 315. (See Duerden, 1897 and 1902. ‘ Trans, Linn. Soc.’) B. antilliensis, Duerden, 1897, p. 7; 1902, ‘ Trans. Linn. Soe.’ B. globulifera, Verr., ‘Trans. Connect. Acad.’, x, p. 559. (See Duerden, 1902.) ?B. australis, Haddon, 1898, p. 435. B. australis may not really be a Bunodeopsis; its anatomy is unknown and it has only a single circle of vesicles near the base, and does not seem quite like the others. Viatrix globulifera, D. and M., is perhaps the same as B. globuli- fera. Tnaumactis, Howler, 1889, p. 143. Phyllactidae probably. Only described species a small form flattened like a dise, with the mouth in the middle of the upper side; perhaps free-swimming. The column has irregularly arranged simple or slightly compound vesicles. There are a few marginal tentacles, with ecto- dermal longitudinal musculature. Body-wall and actinopharynx have ectodermal longitudinal musculature. No siphonoglyphes. Weak diffuse sphincter. Mesenteries not numerous, with weak musculature, Species : T. medusioides, Fowler, 1889, p. 143. Lepruntia, D. and M., 1860. Probably Hoplophoria as used by Wilson, 1890, p. 379, for H. coralligens, not as used by Haddon, 1898, p. 488, for H. cincta. 288 1, A. STEPHENSON Phyllactidae with no marginal acrorhagi or fosse, but with six, or four to eight hollow outgrowths just below the margin, these at best forming large and complicated bunches or * pseudotentacles’. They vary as to detail, but are dichotomous in their early branchings ; there is usually some development of nematocysts on them, these latter being sometimes definitely concentrated into acrorhagi on the pseudoten- tacles, which may differ in colour from the rest of the pseudotentacles. No sphincter. Tentacles not retractile. Retractors diffuse. Mesenteries may be all fertile, save perhaps the directives. Species : L. danae, D. and M., 1860, p. 47. (See Pax, 1910, p. 209; MecMuwrich, 1889, ‘Journ. Morph.’, p. 31, &c.; Verrill, * Trans, Connect. Acad.’, x; Verrill, 1899; MeMurrich, 1905.) L. neglecta, D. and M., 1860, p. 48. (See same references as for L. danae.) ?L. coralligens, Wilson, 1890, p. 379. (See Haddon, 1898, p. 437, and Duerden, 1898, p. 456, and 1902.) Verrill considers L. danae and L. neglecta distinet species because of the acrorhagi on the fronds of danae; MeMurrich seems to think they run into each other. Duerden thinks that Hoplophoria coralligens is a Lebrunia. Family 7. MINYApDIDA#, sens. strict. Minyadidae, Andres, as used by Haddon, 1898, p. 463. Endomyaria (?) in which the base forms a float. It is hollowed out and has in-drawn edges with only a slight opening, the cavity being filled by a chitinous mass which is porous and exhibits a more or less definite structure. The body is smooth in the one form best known ; it has forty tentacles, one siphonoglyphe, ten pairs of perfect mesenteries with strong circumscribed to circumscribed-diffuse retractors, and ten pairs of imperfect mesenteries with more diffuse retractors. The endocoels are larger than the exocoels, this giving a curious appearance to a trans- verse section. Not more than one tentacle to each exo- and endocoel. Sphincter endodermal—well-developed circumscribed according to Carlgren, consisting only of a single fold according to Haddon. Genus: Stichophora. Sricuopuora, Brandt, 1835. S. torpedo, Bell, 1885, p. 114. (Minyas torpedo.) See above definition, and also Part II, p. 533, and Carlgren, 1894, p. 19, and Haddon, 1898, p. 465. CLASSIFICATION OF ACTINIARIA 289 Family 8. PHYMANTHIDAE, Phymanthidae as used by Carlgren, 1900, p. 66. Phymanthidae, Andres, as used by Duerden, 1900, p. 183. Including Thelaceridae, Mitchell, 1890. Endomyaria with definite but sometimes reduced and_half-physa- like base, which is more usually, however, well developed. Form of body variable. Cinclides may be present. Verrucae usually present. No sphincter or only a trace. Tentacles of two sorts, marginal and discal. Marginal tentacles in cycles in the usual way, rarely smooth, usually with greater or less development of paired lateral swellings or outgrowths, which may be simple or ramified, insignificant or conspicuous, Oral dise with short papilliform or not much developed tentacles as a rule —they are occasionally absent ; when present they may be connected with endocoels only, or with both endo- and exocoels. Mesenteries typically with well-developed retractors, which in the best cases are circumscribed. A good many mesenteries are perfect. Genus: Phymanthus. Puymantuus, M. Edw., 1857. Thelaceros, Mitchell, 1890. Phymanthidae. Base variable, from well developed to small or reduced and capable of being half like a physa. Form of body variable— may be trumpet shaped or almost Halcam pa-like, and so on. Cin- clides may be present near the base. Upper part of body with verrucae, which may occur in vertical rows; they may attach foreign bodies ; they may be insignificant. Margin crenulated or provided with acro- rhagi which may even be somewhat compound; rarely no verrucae or acrorhagi; there may be a fosse. Marginal tentacles (arrangement may be hexamerous or octamerous) smooth (rarely) or provided with feebly or strongly developed lateral, usually paired, swellings, which may be merely low knobs or may amount to short ramified branches ; they may meet across the oral face of the tentacle ; and grades between their presence and absence are found. Discal tentacles usually sessile out- growths of the disc; they may resemble the marginal tentacles in miniature, or may be merely papilliform, or even scarce, reduced, or absent (see Part II, Text-fig. 14, Hu). The whole disc may become some- what folded. The mesenteries are a good many of them perfect, and the stronger ones have usually strong retractors, sometimes diffuse but at their best circumscribed (see Part II, Text-fig. 4, ); the older ones or all of them fertile, save sometimes the directives. Little or no sphinc- ter. Radial musculature of disc and tentacles ectodermal or with 290) T, A. STEPHENSON a mesogloeal tendency. May be retractile or not. Actinopharynx and siphonoglyphes may have weak ectodermal muscle. Species : P. crucifer, Les., 1817, p. 174. (See Duerden, 1900, p. 139; Pax, 1910, p. 222; MeMurrich, ‘ Journ, Morph.’, 1889, &c.) P. sansibaricus, Carlgr., 1900, p. 67. P. strandesi, Carlgr., 1900, p. 68. P. loligo, Ehr., 1834, p. 41. (See Carlgren, 1900, p. 70.) P. muscosus, H. and §., 1893, p. 122. (See Haddon, 1898, p. 496, and Kwietniewski, 1898.) P. rhizophorae, Mitchell, 1890, p. 557. P. levis, Kwiet., 1898, p. 421. ?P. caeruleus, Q. and G., 1833, p. 157. (See Pax, 1912, p. 312.) And perhaps others. Phymanthus is an easily identified genus, and interesting as giving hints as to its evolution in the species which verge in structure in the direction of Halcampa. The genus Cram- bactis of Haeckel is invalid; Carlgren states that the queer inner tentacles were only extruded filaments (1900, p. 58). Family 9. HrreranrHipas, Carlgr, Heteranthidae, Carlgr., 1900, p. 72, 1900 (* Ofv. Vet.-Akad. Y'orh.’), p. 278. Rhodactidae, Andres, as used by Haddon, 1898, p. 476, pro parte. Endomyaria with definite base. The column in the only known form has verrucae and fosse, the distal margin with little warted lobes; the sphincter is not very strong, endodermal circumscribed’; the tentacles are distinctly marked off into marginal and discal, the marginal short conical, the discal wart-like, in rows; the mesenterial musculature is well developed. Genus: Heteranthus. Heterantuus, Klunz., 1877. H. verruculatus, Klunz., 1877, p. 84. See above definition for chief characteristics. I know nothing of this family save the details here given, which are taken from Carlgren, 1900, p. 72. = ——EE— ee CLASSIFICATION OF ACTINIARIA 291 Family 10. HomosticHaNntuipan, Carler. Homostichanthidae, Carler., 1900, p. 118. Discosomidae as used by Duerden, 1900, p. 154, pro parte. Endomyaria with definite base. The known form has smooth body but for possible acrorhagi. Sphincter not strong, circumscribed diffuse. Retractors diffuse. Tentacles all of one sort, simple, may be short and more or less papilliform, in radiating series on the exocoels as well as the endocoels. Numerous perfect mesenteries. Genus: Homostichanthus. HomosticHantuus, Duerden, 1900, p. 166. Homostichanthidae in which the body may be elongate, smooth ; distal part may be somewhat folded. Margin with elevations, possibly acro- rhagi, and slight fosse. Retractile. Tentacles short, smooth, slightly capitate, knob-like, their stems glandular and heads nematocystic, their longitudinal musculature ectodermal. Slight circumscribed-diffuse sphincter. Numerous perfect mesenteries and diffuse retractors, - Species : H. duerdeni, Carlgr., 1900, p. 117. (See Duerden, 1900, p. 167.) Family 11. AvtRrmrntanrpanr. Aurelianidae, Andres, as defined by Carlgren, 1900 (small paper on Stichodactylines), p. 279. Endomyaria with definite base, which may be large or small. Body may have more or less cuticle, or may have small vesicle-like verrucae below the margin. The tentacles are small vesicular outgrowths, often lobed, two or three or many communicating with each of the main exo- and endocoels. Sphincter strong circumscribed (see Part II, Text- fig. 13). Main mesenteries very strongly muscular, the retractors exhibiting at their best the extreme of circumscription and distinction from the mesenterial surface. All stronger or all mesenteries perfect and fertile. Only one siphonoglyphe. Radial musculature of dise and tentacles, such as it may be, ectodermal or mesogloeal. Genera: Aureliania, Actinoporus. AURELIANIA, Gosse, 1860, p. 282. Probably Leiotealia, Hertw., 1882, p. 37. Aurelianidae with a very wide base, so that the body slopes inwards more or less to the narrower disc, No verrucae, Body divided more or 292 T, A. STEPHENSON less definitely (the distinction not necessarily externally clear) into an extensive scapus (to which usually adheres a roughish brown cuticle, much or little of it) and a more delicate capitulum or puffy marginal region, the ectoderm of which may contain spirocysts. Distinct fosse. The tentacles are short knobs, sometimes with stems, arranged so that two communicate with each main exocoel, and two or three with each main endocoel (see Part Il, Text-fig. 14, a). Besides being in short radial rows they are so placed as to form cycles which alternate, though not in the genuine ‘ Actiniine’ way. They are some simple, some lobed, and in a living specimen one can distinguish different tentacle- forms for the different concentric rings. Sphincter fairly to very strong, circumscribed, with a heavy central axis of mesogloea (see Part II, Text-fig. 13, B). Radial musculature of disc and tentacles where present curious, chiefly mesogloeal. All stronger mesenteries perfect, fertile, with filaments and retractors ; the retractors very unusual, powerfully circumscribed (see Part II, Text-fig. 4, B), and attached to the mesentery by one edge only for part of their extent, and with an axis of mesogloea. There may be additional weak mesenteries beyond the main macro- cnemes. Species : A. augusta, Gosse, 1860, p. 283. (See Faurot, 1895.) A. heterocera, Thompson, 1853. A. regalis, Andres, 1883, p. 496. (See Carlgren. 1900, short paper on Stichodactylines, p. 279, &c.) 2A. nymphaea, Hertw., 1882, p. 38. I have personally studied A. augusta, both alive and anatomically, and I do not know why Andres assumed it to be the same as his A. regalis. It 1s not impossible that Capnea, Forbes, is the young of Aureliania; beyond this suggestion it cannot yet be allocated. ‘Then there is the question of Hertwig’s genus Leiotealia. It has generally been assumed (and I shared the idea formerly) that this is identical with Epiactis or [sotealia, but this overlooks certain details of its structure. I have recently been able to investigate two species of Aureliania, A. augusta and a possibly new one, and on re-reading Hertwig’s description in the light of this, it becomes evident that in all features one can be sure about the two genera really share essentials. An examination of the original Challenger specimen confirmed the idea, The one uncertain though necessary poimt is that it is el te te te el ee CLASSIFICATION OF ACTINIARIA 293 not known whether Leiotealia has more than one tentacle to each main endocoel and exocoel or not ; Hertwig apparently thought only one, but the specimen is small and so contracted (and Aureliania has cycles as well as radical rows) that it is hard to tell. The other things seem to poimt to its being an Aureliania, which in that case should be called A. nymphaea, Hertw.; whether it is A. nvmphaea, Drayt., is another matter. It has smooth body without verrucae or acrorhagi, small button-like tentacles, pinnate sphincter with stout mesogloeal axis, mesogloeal radial dise- muscle; only the mesenteries of cycles 1-3 have distinct retractors, and these are great circumscribed things attached only at one edge. “hese things are all found in Aureliania (not necessarily only three cycles of mesenteries with retractors of course), as are also the wide base and pyramidal form of nymphaea, and some of them are very characteristic features. The sphincter is less developed in nymphaea than in the others. In view of the general evidence it seems probable that the Stichodactyline tentacle-plan may be assumed. Actinoporus, Duch., 1850. (See Duerden, 1900, p. 174; Carlgren, 1900, short paper on Stichodactylines, p. 283.) Aurelianidae with definite but not specially wide base—it may even be somewhat reduced. The body may be long. There may be rather vesicle-like verrucae below the sphincter, of which the main ones some- times form a sort of collar. Deep fosse. Disc not extensive, but notched into little permanent lobes or lappets at its margin, which correspond in number to the endocoels and exocoels. Tentacles short vesicular knobs, may be lobed, many communicating with each exocoel and endocoel, the tentaculate areas thus formed separated from each other by radial grooves. Sphincter strong circumscribed (see Part II, Text- fig. 13, 4). Mesenteries all perfect and all or mostly fertile, with very strong circumscribed retractors which may be broadly or narrowly attached to the mesenteries, partly according to region. Disc and tentacle muscle very weak, ectodermal if present. Species : A. elegans, Duch., 1850, p. 10. (See Duerden, 1900, p. 175.) A. elongatus, Carlgr., 1900 (small paper on Stichodactylines), p. 283. 294 T, A. STEPHENSON Family 12. AcTINODENDRIDAE. Actinodendridae, Haddon, 1898, p. 488; Carlgren, 1900, p. 96. Acremodactylidae, Dendrianthidae, Kwiet., 1897-8. Endomyaria with definite base, not always well marked off from the column, and it may be small. Smooth wall. No special margin. No sphincter or a slight diffuse one. Dise produced into permanent arm- like lobes, which are arranged in cycles like large tentacles, each bearing numerous tentacles or branches on it (see Part IT, Text-fig. 14,K); these tentacles may be arranged all round the lobes or be more or less absent from parts of them, and may be themselves simple or branched, and in the latter case giving the whole arm a dendritic effect (see Part IT, Text-fig. 19). Mesenteries all (or twelve pairs only perfect) perfect and fertile, except sometimes the directives, with strong retractors. Radial musculature of disc and tentacles ectodermal. There may be concentrations of nematocysts in the tentacle-lips, in little thickenings. The discal arms correspond one to each endocoel in the inner cycles, one to each exocoel in the outer, the size varying according to cycle. Genera: Actinodendron, Actinostephanus, Megalactis. ACTINODENDRON, Blainy., 1830, p. 287. Acremodactyla, Kwiet., 1898. Actinodendridae in which the tentacles are arranged all round the arms, and are themselves branched, and may have nematocystic thickenings at their tips. Very strong diffuse retractors. (See Part I, Text-fig. 19, for appearance of A. plumosum.) Species : A. plumosum, Haddon, 1898, p. 490. (See Saville Kent, 1893.) A. glomeratum, Haddon, 1898, p. 492. (See Saville Kent, 1893 and 1897.) A. hansingorum, Carlgr., 1900, p. 98. A. ambonensis, Kwiet., 1898, p. 401. (See Carlgren, 1900, p. 96.) And probably others. AcTINOsTEPHANUS, Kwiet., 1898, p. 405. Actinodendridae in which the tentacles are irregularly arranged on the arms and are simple. Strong retractors. Species : A. haeckeli, Kwiet., 1898, p. 403, CLASSIFICATION OF ACTINIARIA 995 Mrearactis, Ehr., 1834. Actinodendridae with the arms longer than in Actinodendron, the oral faces of the arms freer from tentacles, the ultimate branches of the tentacles simple and pointed, not bifid. (See Part IT, Text-fig. 14, K.)- Species : M. griffithsii, Saville Kent, 1893, pp. 35, 147. (See Haddon, 1898, p. 493.) And probably at least one other, I have followed Haddon and Saville Kent in keeping Megalactis separate from Actinodendron, asI think it should be, although not yet well known. Family 18. HALASSIANTHIDAE. Thalassianthidae, MceM., as used by Haddon, 1898, p. 482, and Carlgren, 1900, p. 86. Endomyaria with definite base. Body with or without verrucae above. Disc circular or waved or puckered. Not more than one den- dritic tentacle to each exocoel—exocoelic tentacles all dendritic. Several simple or dendritic tentacles and several (sometimes many) modified tentacles or nematospheres on many of the endocoels, these sometimes gathered up on to a definite permanent elevation or even finger-like lobe of the disc (the nematospheres being aboral), which may cover an endocoel and the two adjacent exocoels. Sphincter of variable develop- ment, from more diffuse to more circumscribed, not very strong. Mesen- terial musculature well developed, but not unusually strong, retractors more or less diffuse. Numerous perfect mesenteries. Genera: Thalassianthus, Cryptodendron, Actineria. THALASSIANTHUS, Leuck., 1828. Heterodactyla, Khr., 1834. Thalassianthidae with verrucae above, which may be in rows, or with none, and they may vary in distinctness. Margin may be distinct or notched. Disc may be folded or puckered up, or not. Marginal tentacles dendritic, without nematospheres, not more than one per exocoel. The endocoelic tentacles are also dendritic, but many of them arranged on elevations or permanent lobes of the disc, and occupying the more oral part of the elevation, which typically possesses aborally a bunch of grape- like stinging batteries or nematospheres (see Part II, Text-fig. 14, 2). Siphonoglyphes two or several, directives present in the first case, NO. 262 ve 296 T. A, STEPHENSON absent in the second. Sphincter weak to moderate, more diffuse to more circumscribed, Radial muscle of dise ectodermal. Retractors diffuse or more like circumscribed diffuse, numerous mesenteries perfect. gonads chiefly on stronger imperfects, or on the older mesenteries save sometimes the directives. Species : T. aster, Leuck. in Riippel, 1828, p. 5. (See Carlgren, 1900, p. 87.) T. kraepelini, Carlgr., 1900, p. 91. T. senckenbergianus, Kwiet., 1897, p. 337. T. hemprichi, Ehr., 1834, p, 42. (See Carlgren, 1900, p. 94; Haddon, 1898, p. 485.) T. hypnoides, Saville Kent, 1893, p. 148. (See Haddon, 1898, p. 486.) I have jomed Thalassianthus and Heterodactyla because I cannot find any really important differences between them. The definition given covers both. The presence of several siphonoglyphes im some species, and no directives, of two siphonoglyphes and two pairs of directives in others, seems no valid ground of separation. CRYPTODENDRON, Klunz., 1877. Thalassianthidae with or without suckers on upper part of body, and with somewhat crenulated margin. Wide irregularly-folded disc. Three sets of tentacles: (a) aset of short exocoelic marginal dendrites ; (b) radial rows of short, simple, and dendritic tentacles on the inner part of the disc; and (c) an intermediate zone of nematospheres. Sphincter weak to moderate, circumscribed. Well-developed diffuse retractors. The nematospheres especially have apical batteries of sting-cells and glandular stems. Radial musculature of disc and tentacles ectodermal. The nematospheres are in sessile packets. and they and the discal dendrites are endocoelic. Species : C, adhaesivum, Klunz., 1877, p. 86. (See Haddon, 1898, p. 483, and Kwietniewski, 1896.) ActTINERIA, Blainy., 1830. Thalassianthidae with vertical rows of verrucae in upper part of column, parapet notched a little. Wide folded disc, bare in the middle, with small permanent lobes at the edge. Exocoels with dendritic tentacles, endocoelic lobes with dendrites (which run inwards on the disc) on the oral side and a mass of nematospheres aborally. Sphincter CLASSIFICATION OF ACTINIARIA 297 not very strong, sessile circumscribed. Numerous perfect mesenteries. Older mesenteries fertile, but probably not the directives. Species : A. dendrophora, H. and S., 1893, p. 123; Haddon, 1898, p. 487. And probably also A. villosa, Q. and G., 1833, p. 156. Family 14. SrorcHactipar. Stoichactidae, Carler., 1900, p. 72; 1900 (‘ Ofv. Vet.-Akad. mork.”))'p. 278. Discosomidae, Klunz., as used by Haddon, 1898, p. 469, pro parte. Endomyaria with definite base. Column usually but not always verrucose above. Size sonietimes very large. Tentacles simple, but for scattered bifid or multifid tentacles which sometimes occur sporadically among the others ; they may be fairly long and quite ordinary, or may be short or wart-like, or even short columns with spherical heads, They are all of one sort in the same animal, and there is not more than one to each exocoel ; the endocoels may in rare cases have only one tentacle each, but usually at least some of them have more—the stronger ones may have radial rows (see Part II, Text-fig. 14, F), or all the endocoels may have either one or several rows. Sphincter strong or not very strong, more or less diffuse to more or less circumscribed. Mesenterial musculature well developed, retractors weaker or stronger but not unusually strong, diffuse. Numerous perfect mesenteries. Gonads may occur on all mesenteries—usually the older ones are fertile save some- times the directives, but not always. Tentacular longitudinal muscle ectodermal. Genera: Stoichactis, Radianthus, Antheopsis. This family is itself very clearly marked off from others, but within it, it is difficult to satisfactorily separate off genera. The difficulty is increased because some descriptions of the forms do not give enough data. At best, it seems that there are only three sound genera to be distinguished, three stages im the evolution of very similar creatures ; they form a series really, and I do not feel perfectly confident that they do not all form one large genus. At any rate more than three it is unwise to insist on; some pairs of names have been given to similar forms, and some of these must now become synonyms. Y 2 298 T. A. STEPHENSON In Antheopsis (=Stichodactis) the condition is some- times purely ‘ Actiniine —not more than one tentacle to each exo- and endocoel; but the more normal state is for there to be more than one tentacle ; or a row, on the older endocoels but not the younger. In Radianthus (=Helianthopsis) comes the stage where all the endocoels have radial rows, but there is only about one row on each. In Stoichactis (=Disco- somoides) the last stage is attained, and there are not only radial groups on all the endocoels, but usually more than one row abreast in each group; and the tentacles have often specialized in small size. As far as sphincters are concerned, [ think comparison with other families will show that their exact form cannot be used here as a generic distinction. In the lists of species given below it should be remembered that a form here and there may be allocated to the wrong genus because of insufficient data about it ; but some re-arrangement has been made intentionally to get the three grades clearly separated off. The readjustments mainly mean a transference to Antheopsis of some forms originally described under Radianthus, Stichodactis, and Helianthopsis, and a consequent depletion of the true genus Radianthus. This has been necessary to get together all the forms with radial rows of tentacles on the older endocoels only. It is not much of a distinction, but if the two genera are to be kept apart at allit must be insisted on. That the sporadic occurrence of a few cleft tentacles in some species should be used as a generic character would be a mistake. Srorcuactis, Haddon, 1898, p. 472. Discosomoides, Haddon, 1898, p. 470. Stoichactidae. Some species attain enormous size, up to about two feet across, and often crustacea or fish are commensal with them. The body is usually wider above than below, and above with suckers which may be in vertical rows—these may, however, be rudimentary or absent, present or not evenin the same species. Margin barely or slightly or dis- tinctly marked, may be notched. Disc simple or little or much folded. Tentacles not very long at best, usually short or very short; digitiform CLASSIFICATION OF ACTINIARIA 299 or subulate; or wart-like ; or wider at the tip than at the base; or forming short stems with spherical heads. Only one tentacle per exocoel, a good many on each endocoel, usually more than one row abreast per endocoel ; the rows may be very irregular. Sphincter weaker or stronger, circumscribed diffuse to well circumscribed. A cleft tentacle sometimes occurs among the others. Siphonoglyphes and directives variable in number. Species : S. kenti, H. and6S., 1893, p. 119. (See Saville Kent, 1893, p. 144 ; Haddon, 1898, p. 473.) 8S. haddoni, Saville Kent, 1893, pp. 32, 144. (See Haddon, 1898, p. 474.) S. helianthus, Ellis, 1767, p. 436. (S. anemone, Ellis, 1767, p. 436.) (See Duerden, 1900, p. 162; Pax, 1910, p. 227; McMur- rich, ‘ Journ. Morph.’, 1889.) . ambonensis, Kwiet., 1898, p. 410. . tuberculata, Kwiet., 1898, p. 412. . giganteum, Forsk., 1775, p. 100. (See Carlgren, 1900, p. 77.) . tapetum, Ehr., 1834, p. 32. (See Carlgren, 1900, p. 74.) . laevis, Lager, 1911, p. 240. . intermedia, Lager, 1911, p. 238. . australis, Lager, 1911, p. 241. . fuegiensis, Dana, 1846. (See McMurrich, 1893, p. 200.) RAN RRARDRNRM Rapiantuus, Kwiet., 1897, p. 331. Helianthopsis, Kwiet., 1898, p. 417, pro parte. Stoichactidae with or without suckers on upper part of body. Margin fairly well marked. Tentacles shorter or longer, but not mere papillae. More than one tentacle communicates with every endocoel (not more than one per exocoel), but only about one row on each. Here and there may be cleft tentacles. The disc may be lobed. Sphincter more or less diffuse to weak or medium circumscribed. Species : R. lobatus, Kwiet., 1898, p. 414. R. mabrucki, Carlgr., 1900, p. 82. ?R. parvitentaculatus, Q. and G., 1833, p. 165. (See Pax, 1912, p. 314.) ANTHEOPSIS, Simon, 1892. Stichodactis, Kwiet., 1898, p. 415. Radianthus, Kwiet., 1897, p. 331, pro parte. Helianthopsis, Kwiet., 1898, p. 417, pro parte. Stoichactidae with suckers in the upper part of the body or not; if 300 T. A. STEPHENSON they are there foreign bodies may adhere to them ; margin distinct, may be crenulated. Dise circular or more or less lobed. Tentacles shorter or longer, may attain good length; at any rate not mere papillae. Not more than one tentacle per exocoel. As to the endocoels (see Part I, Text-fig. 14, F), there are never radial rows on all of them; usually there are radial rows on the older ones or some of them, but these vary in length—the larger ones may contain a good many tentacles or only a few; the rows are more or less single, and sometimes they are quite absent so that the form is not * Stichodactyline’ as to tentacles at all, having only one per endocoel. Cleft tentacles may occur here and there among the others. Sphincter weak or moderate, diffuse, cireum- scribed diffuse, or circumscribed. Number of siphonoglyphes and directives variable. Species : A. koseirensis, Klunz., 1877, p. 77. (See Simon, 1892, and Carlgren, 1900, p. 85.) A. ritteri, Kwiet., 1898, p. 417. (See Carlgren, 1900, p. 81.) A. kuekenthali, Kwiet., 1897, p. 332. A. papillosa, Kwiet., 1898, p. 415. A. macrodactylus, H. and 8., 1893, p. 120; Haddon, 1898, p. 471. A. malu, H. and&S., 1893, p. 120; Haddon, 1898, p. 472. A. carlgreni, Lager, 1911, p. 243. A. concinnata, Lager, 191], p. 244. A. glandulosa, Lager, 1911, p. 246. A. kwietniewskii, Lager, 1911, p. 247. Sub-order MADREPORARIA. I do not wish to suggest, even vaguely, to which of the skeleton-forming corals the genera defined below are related. The ground for placing them under Madreporaria will be found in Part. II, p. 510. ‘To save repetitions, a general statement covering Corallimorphidae and Discosomidae is given first, but it is not meant as the definition of a sub-tribe, although it would serve that purpose if such a sub-tribe were needed. Madreporaria which secrete no definite skeleton. They may live quite a solitary life, or may live together in numbers. They frequently repro- duce by fission, and compound individuals with several] mouths may be found, or individuals connected by a coenosare. There is a definite base. The body is smooth, and variable in form and consistency. The rarer CLASSIFICATION OF ACTINIARIA 301 tentacles are arranged so that more than one communicates with some at least of the endocoels, and sometimes more than one with exocoels also; they may be simple, knobbed, or branched, and so on, and there may be more than one sort in the same species ; they may be reduced and wart-like (Part II, Text-fig. 3), or even reduced to nothing externally visible. There are typically no siphonoglyphes—these are recorded in some cases but their existence probably needs confirming. The mesen- terial filaments have no ciliated tracts. Sphincters are absent or weak diffuse. Sting-cells of a size characteristic more of Madreporaria than of Actiniaria are usually present somewhere in the body (see Part I, Text-fig. 6). There are usually a good many perfect mesenteries, as a rule twelve or more pairs, and there is no distinction of them into macro- and microcnemes. The longitudinal mesenterial musculature consists typically of a feeble layer, not forming the sort of sheet or retractor characteristic of Actiniaria (see Part II, Text-figs. 4 and 5). Basilar muscles are absent. Ectodermal muscle present at least some- times in the body-wall, sometimes probably absent. Directives usually present, varying in number. The large sting-cells may occur in tentacles, actinopharynx, mesenteries, body-wall. Family 1. CorALLIMORPHIDAE. Corallimorphidae, Hertw., 1882, p. 21; Carlgr., 1900, p. 19. Size larger or smaller; habit solitary or gregarious, individuals may be connected by ceenosarc. Ectodermal muscle in body-wall present at least in some cases. Tentacles simple, knobbed at the tips. Not more than one tentacle per exocoel, more than one on at least the older endocoels, Genera: Corallimorphus, Isocorallion, Corynactis. CoraLLimorpHus, Moseley, 1877, p. 299. Corallimorphidae with weak musculature throughout. Body-wall ectoderm has weak longitudinal musculature. No sphincter. Body-wall and oral disc may be very thick and cartilaginous, and animal may attain fairly large size. Tentacles simple, and all knobbed at the tip (see Part II, Text-fig. 14, a), divided into two sorts, marginal and discal. There is never more than one tentacle of each sort arising from one and the same endocoel. The exocoelic tentacles are the smallest of the marginal series, taken on the whole, and the discal tentacles correspond to the endocoels of the inner marginal tentacles. There may be a good deal of irregularity. 302 T, A. STEPHENSON Species : C. rigidus, Moseley, 1877, p. 301. (See Hertwig, 1882, p. 23, and 1888, pp. 9, 10; Stephenson, 1920 B, p. 178.) C, profundus, Moseley, 1877, p. 300. (See Hertwig, 1882, p. 28, and 1888, pp. 9, 10; Stephenson, 1920 B, p. 178.) C. obtectus, Hertw., 1888, p. 9. (See Stephenson, 1920 B, p. 178.) C. ingens, Gravier, 1918, p. 23. The above definition of the genus is practically that given in my short note on the genus Corallimorphus (‘ Proce. R. I. Acad.’, 1920, B. 9). I began it there with the words ‘Stichodactyline Actiniaria ’, this being provisional, as I had not then worked out my idea of its beg a skeleton-less coral fully enough for publication. I have listed the four species here for reference purposes, but as before suggested, I am inclined to think they are all one, and the more so since dealing with another specimen from an Antarctic collection and looking at the Challenger specimens. C. ingens is probably the same as the others. If the four listed are to be separate, my Insh form would make a fifth. IsocoraLLion, Carler., 1900, p. 19. Chalmersia, Del. and Hér., 1901, p. 536. Corynactis as used by Hertwig for Corynactis, sp., 1888, p. 10. Corallimorphidae differing from Corallimorphus in having the ectodermal muscle in the body-wall stronger, and with normally two disc-tentacles on each of the oldest radii of the disc. Species : I. hertwigi, Carlgr., 1900, p. 19. (See Hertwig, 1888, p. 10, Corynactis, sp.) I feel doubtful of the distinctness of this genus from Coralli- morphus, but hardly enough is yet known of it to justify their fusion. Corynactis, Allm., 1846. (See Duerden, 1898, p. 635, &e.) Corallimorphidae of small size, often gregarious in habit, sometimes forming large sheets of individuals ; often clusters or pairs of individuals are found attached to each other by a basa] coenosare ; fission is a usual CLASSIFICATION OF ACTINIARIA 303 method of increase. The individuals are very variable in form, often trumpet shaped in expansion, and more or less retractile. The tentacles are knobbed, the outer larger than the inner, and the exocoelic tentacles largest of all. Some or all of the endocoels have more than one tentacle connected with them. Tentacle-heads usually with large sting-cells and little or no muscle, shafts with ectodermal longitudinal muscle. Perhaps very weak ectodermal muscle in the body-wall. Sphincter absent or weak diffuse. Species : C. viridis, Allm., 1846, p. 417. (See Gosse, 1860, p. 289, and Rees, 1915, p. 543.) C. globulifera, Ehr., 1834, p. 39. (See Carlgren, 1900, p. 20, and Haddon, 1898, p. 467.) (?C. hoplites, H. and &., 1893, p. 118.) C. myrcia, D. and M., 1866, p. 124. (See Duerden, 1900, p. 181.) C. carnea, Studer, 1879, p. 542. (See MceMurrich, 1904, and Kwietniewski, 1896.) C. australis, H. and Duerden, 1896, p. 151. C. haddoni, Farquhar, 1898, p. 532. (See Stuckey, 1909, p. 390.) C. mollis, Farquhar, 1898, p. 534. (See Stuckey, 1909, p. 390.) C. gracilis, Farquhar, 1898, p. 534. (See Stuckey, 1909, p. 390.) C. albida, Stuckey, 1909, p. 390. And perhaps others. Possibly haddoni, mollis, gracilis, and albida are all one species. Family 2. Discosomrpag, sens. strict. Discosomidae as used by various authors, pro parte. Used here in the sense taken by Carlgren, 1900, p. 58. Including Phialactidae, Howler, 1889. Rhodactidae, Andres, as used by Haddon, 1898, p. 476, pro parte. Size variable. Living singly or in patches. With one or more mouths. Sphincter absent or weak diffuse. Tentacles simple or dendritic (see Part II, Text-fig. 14, B, c) or somewhat capitate or curious and urn-like, or reduced to warts (see Part II, Text-fig. 3), or to little or nothing, so that they do not show above the surface of the disc at all; more than one sort may occur in the same species, and more than one may com- municate with endocoels and exocoels or with endocoels only, there being often radial rows. Presence of ectodermal muscle in body-wall doubtful. 304 T. A. STEPHENSON Genera: Discosoma, Paradiscosoma, Ricordea, Orinia, Rhodactis, Actinotryx. Discosoma, Leuck., 1828. Discosomidae with tentacles all of one sort, not branched, not knobbed, may be swollen towards the tips; short, usually wart-like, sometimes reduced or even vanished, so that only traces of them remain as endo- dermal evaginations in the mesogloea of the disc. Margin of body straight or more or less notched or irregular. Tentacles in radial rows on at least some endocoels, sometimes on exocoels too. Sphincter absent or weak diffuse. Species : D. nummiforme, Leuck., 1828, p. 3. (See Simon, 1892, and Carlgren, 1900, p. 62.) D. Yuma, Carlgr., 1900, p. 63. D. Unguja, Carlgr., 1900, p. 64. And probably others. I do not feel clear that all the genera that follow are really distinct from Discosoma, but am listing them in full. Taking the family as a whole, the two clearest genera are Discosoma and Actinotryx. Beyond this there is less certainty. Rhodactis is probably distinct but is little known. Ricordea and Paradiscosoma seem doubtfully distinct from Discosoma. Even Orinia might be only a curious state of Discosoma, but is more likely to be distinct than the others: even in Paradiscosoma one sometimes sees the tentacles collapse on themselves so that they form little double-walled cups, and it would not take much to make this into Orinia; and MeMurrich says some of the more peripheral of them are tuberculiform and not crateri- form. If there is a naked zone between the marginal and diseal sets, however, that will clinch the distinction. There are other genera and species which have been referred at one time and another to this family, before it was properly under- stood, but these have been cast out as time went on, and are in this paper referred to their new positions, e.g. Stoichactidae. Parapiscosoma, Carlgr., 1900, p. 60. (n. nom. for Isaura.) Discosomidae with margin of disc thrown into small lobes. Otherwise like Discosoma. (See Part Ll, Text-figs. 3, 6, B, 5.) CLASSIFICATION OF ACTINIARIA 305 Species : P. neglecta, D. and M., 1860, p. 51. (Isaura neglecta, D. and M.) (See Carlgren, 1900, p. 60, and Pax, 1910, p. 214.) A vertical section of a species of Paradiscosoma is given in Part lI, Text-fig. 3. Ricorpga, D, and M., 1860. (See Duerden, 1898, p. 635, &c.) Discosomidae which often live aggregated together in patches. The majority of individuals have more than one mouth, there may be up to seven or so, the disc being consequently sinuous in outline. Some- times individuals are found connected by a basal membrane. No sphincter, though the animal is retractile. Tentacles short and may be somewhat capitate or rounded at the tip, in radial rows on at least some endocoels, Stems of tentacles may be glandular, their tips nemato- cystic. Species : R. florida, D. and M., 1860, p. 42. (See Duerden, 1900, p. 156; Pax, 1910, p. 219; MeMurrich, 1889, ‘ Journ. Morph.’) OrintA, D. and M., 1860. Discosomidae with tentacular, simple structures in the periphery of the disc. Inner part of the disc provided with characteristic large urn- like outgrowths. Between the simple tentacles and the urns a tentacle- free area. (See Carlgren, 1900, p. 60.) Species : O. torpida, D. and M., 1860. (See Carlgren, 1900, p. 60, and MeMurrich, 1905.) Ruopactis, M. Edw. and H., 1851. ?Phialactis, Fowler, 1889. Discosomidae with tentacles of two sorts, simple ones round the mouth and the edge of the disc, branched ones in the middle, which may arise from pits in the disc; the two sorts not gathered up into sharply-separated zones, and no naked area between marginals and discals. Tentacles may be somewhat: capitate in certain states. The animals may live massed together in patches. Species : R. rhodostoma, Ehr., 1834. R. howesii, Saville Kent, 1893, p. 150. (See Haddon, 1898, p. 478.) ?R. neglecta, Fowler, 1889, p. 148.. (See Carlgren, 1900, p. 59- 61, &e.) And perhaps others. 306 T, A. STEPHENSON Actrnorryx, D. and M., 1860, p. 3821. (See Duerden, 1898, p. 635, &c.) Discosomidae which may occur in scores together, crowded so as to form a carpet, and some individuals have two or more mouths, More or less retractile. There are simple or nearly simple tentacles or tenta- culiform outgrowths connected with the margin; within these is a well- marked clear zone, then the main part of the disc has dendrites, some at least in radial rows. Sphincter absent or weak diffuse. (For details of an Actinotryx see Part II, Text-figs. 14, B and oc, 4, p, and 6, A.) Species : A. sancti-thomae, D. and M., 1860, p. 45. (See Duerden, 1900, p. 148; MeMurrich, 1889, ‘ Journ. Morph.’) A. bryoides, H. and§&., 1893, p. 121 ; Haddon, 1898, p. 479. And probably others. 2. APPENDIX. There are some anemones recently described by Professor Gravier, whose papers I did not know about, unfortunately, when Part I of this paper was written, and which should be mentioned now. I am at the same time giving a few further details which seem worthy of note about some of Verrill’s genera which can hardly be finally allocated yet, but are interest- ing as showing the direction which some future work should take to clear them up. I regret that by a mischance I over- looked the genus Euphellia of Pax before, and that also is included here, together with a few other points. (i) Professor Gravier’s forms. Professor Gravier has established five new genera and some new species, as follows : 1. Nectactis (1918, p. 18). N.singularis, 1918) page This has the form of a disc thicker in the middle than at the edge, where the capitate tentacles are, the lower surface of it representing the column and having a little pit-like base in its middle. Smooth wall and no sphincter. A good many mesenteries with indiscernible muscles. It is very difficult to even suggest a position for this form in classifica- tion. Gravier suggests Minyadidae, but it would not do for that family as understood here. If there were disc-tentacles one might suggest Corallimorphidae, and possibly that would be best even without them— but more details are needed. CLASSIFICATION OF ACTINIARIA 307 2, Vhoracactis (1918, p. 12). T. topsenti, 1918, p. 12. A small form living on the surface of a sponge. It is disc shaped, in- crusted, the foreign bodies even getting embedded in the mesogloea. Sphincter mesogloeal, seemingly double. No acontia or cinclides. Weak mesenterial musculature. Gravier believes that the gonads develop from the endoderm of the body-wall. There is not much guide, but the form may be a tiny Paractid or even, possibly, a Zoanthid ? 3. Telmatactis (1916, p. 286). T. valle-flori, 1916, p. 236. This seems to me to be probably identical with Phellia, in which case the species becomes Phellia valle-flori. 4, Sicyopus (1918, p. 21). S. commensalis, 1918, p. 21. This lives on a Holothurian, in a hollow of its skin near the mouth. It has the form of a thick disc, strong mesogloeal sphincter, no acontia or cinclides, diffuse retractors, all mesenteries fertile. It seems like a small Paractid of uncertain affinities. seGiractrs (1918, p. 7). G. crassa, 1918) p. 7. Here the base envelops Acanella. There are no verrucae, the column wall is thick. Good mesogloeal sphincter. Apparently twenty pairs of perfect mesenteries, probably diffuse retractors. If there are no acontia or cinclides this seems eligible for one of the Paractid genera, and probably does not merit generic distinction. I have not suggested very definitely about the above forms,. but they are not all very fully studied as yet, and the time has not come to decide for or against them; but they will probably fit into known families. In addition Professor Gravier has described new species in old genera as follows: 1. Paractis flava (1918, p. 4). Either a Paractis in the strict sense, or belonging to a neighbouring genus. 2, Paractis vestita (1918, p. 5) may have some sort of invest- ment on the column, and seems to have only six pairs of perfect mesen- teries, no acontia and cinclides, mesogloeal sphincter; in which case it is no Paractis, but an Actinoscyphid near Paranthus, perhaps eligible for that genus. 3. Actinernus verrilli (1918, p. 6) is not an Actinernus (=Porponia). since it has a mesogloeal sphincter and is apparently 308 T. A. STEPHENSON not endocoelactous. Nor is it an Actinoscyphia since it has numerous perfect mesenteries. It must therefore belong to Catadio- mene or Polysiphonia, as it has basal swellings to the tentacles ; and from the description I gather that it is more likely to be Poly- siphonia than the other, but further details are needed for decision. 4, Sagartia sociabilis (1918, p. 10). No cinclides. Seemingly six pairs of perfect mesenteries with weak musculature. If it has acontia it must be a Sagartiomorphe—certainly not a Sagartia. 5. 8. sobolescens (1918, p. 11) is perhaps a Sagartiomorphe also. 6. Chitonanthus incubans (1918, p. 11) is very exceptional as a Chondractinian in having the three oldest cycles of mesenteries fertile. Since Chitonanthus is only a synonym of Hormathia, the right name for the species is Hormathia incubans, 7. Chitonanthus indutus (1918, p. 12) should, similarly, be Hormathia induta. 8. Chitonanthus abyssorum (1918, p. 13) seems to be either Hormathia abyssorum oran Actinauge. 9. Hormathia elongata (1918, p. 14) seems correctly named. 10. Hormathia? musculosa (1918, p. 15) has apparently no acontia, so cannot be a Chondractiniid. It has numerous perfect mesenteries and a mesogloeal sphincter, which bring it to Paractidae ; its circumscribed retractors and some of its externals suggest Hormo- soma or Tealidium or Pseudoparactis, but this is uncertain, and it may need a new genus. 1l. Stephanactis impedita (1918, p. 16) becomes Stephan- auge impedita, since Verrill has shown that the name Stephan- actis was pre-occupied. 12. Stephanactis inornata (1918, p. 17) becomes Stephan- auge inornata. 13. Corallimorphus ingens (1918, p. 23); see this paper, p. 302. 14. Anemonia insessa (1918, p. 3) is more likely a Gyrostoma, (ii) Details from Verrill. 1. Verrill (1899) has explamed that the name Stephan- actis is pre-occupied (1868), and renamed Hertwig’s genus Stephanauge. ‘There are now recorded, as forms with mesogloeal sphincter, six pairs of perfect mesenteries (not macrocnemes), no acoutia, and a very few (up to about eight) cinclides, Stephanauge impedita, Grav., 8S. inornata, Grav., 8S. abyssicola, Hertw. (=Actinauge nexilis, Verr.), S. tuberculata, Hertw., &e. In Part I of this CLASSIFICATION OF ACTINIARIA 309 paper I mentioned (p. 487) this genus without being very definite about it. I do not think a final decision can be made even now, but if these cinelidal non-acontiated forms are established they will probably need a family Stephanaugidae, one of the further combinations foreshadowed in Part I. From the fewness of their cinclides, and from their general characters one imagines the cinclides to be vestiges not to last much longer, and probably the forms are descendants of Metridiid ancestors which have lost the acontia before all the cinclides ; but it is not even certain yet that there are not really rudimen- tary acontia, easily overlooked, present, in which case the forms are actually queer Metridiidae on the way to forming Chon- dractiniid or Actinoscyphiud or other stages. §. tuber- culata, at least, has basal mesogloeal swellings to some of the tentacles. If the others have not they need separation, and the whole genus and its relations need careful revision, The related (?) Amphianthus seems to be an Actino- scyphid, so far as it is at present known. 2. Synanthus, Verr., is probably Paranthus. 3. Ammophilactis, Verr., 1899, p. 213. Body may be long, with small base, divided into smooth scapus with a collar in which is the mesogloeal sphincter, and capitulum with suckers which can attach grains of sand. Tentacles in more than two cycles in the adult. Numerous perfect mesenteries. Strong apparently diffuse retractors. Older mesenteries fertile. A. rapiformis, Les., 1817, p. 171. (See Verrill, 1899, p. 213.) This seems clearly a Paractid, differmg from Pseudo- paractis inits single sphincter and suckers. 4. Archactis, Verr., seems very near or identical with Antholoba. 5. Raphactis, Verr., 1899, p. 144. Definite base, broadly expanded or stem-clasping. Column with a capitulum which may be more or less ridged, and a scapus which is often ridged at the top, where the mesogloeal sphincter lies, and may also be tuberculate. Twelve or more pairs of perfect (and atleast mostly) fertile mesenteries, others may be fertile too. Diffuse retractors. Tentacles in more than two cycles in the adult. R. nitida, Verr., 1899, p. 144. R. caribaea, Verr., 1899, p. 205. 310 T. A. STEPHENSON This may be the same as Pseudoparactis, in which case it takes priority. But it seems distinguished by its single sphincter, and distinct from Ammophilactis in its lack of suckers. 6. Verrill says Stomphia may have fertile perfect mesen- teries, perfect mesenteries 16-24 pairs in large specimens. 7. Antiparactis, Verr., is probably a synonym of Paranthus. (ii) Other details. 1. Euphellia, Pax, 1908, p. 475. Diadumenidae with definite base. Wall may be wrinkled. No papillae or suckers. Distinctly divided into scapus and capitulum, the scapus with an easily-shed investment. No acrorhagi or fosse. Long strong mesogloeal sphincter. Six pairs of macrocnemes. Acontia not specially strong. There are cinclides in longitudinal rows. E. cinclidifera, Pax, 1908, p. 475. The definition of Diadumenidae will need slight alteration of detail to admit this form. It seems to be, if it really has cinclides, a link between Diadumenidae and Phelliidae, a Diadumenid on the way to becoming a Phellia. 92. Pax describes a Paraphellia polyptycha (1908, p. 493), which may be a Paraphellia or possibly a Sagar- tiomorphe. 3. Andvakia is of quite uncertain standing and more needs to be known of it. 4, Allantactis seems to be the same as Sagartio- morphe, and if this is so the name has priority. 5. Octineon, Moseley, M.S. (See Fowler, ‘ Quart. Journ. Mier. Sci.’, vol. 35, 1894, p. 461.) (=Ammodiseus, Carp., 1871, p. 159.) The body has the form of a thin disc a little raised in the middle, and encrusted with sand and other things which may get into the mesogloea. Sphincter seemingly mesogloeal. Probably twelve tentacles. Twelve larger primary and perfect mesenteries, but only the eight Edwardsia mesenteries provided with true retractors. Very few of the mesenteries beyond the twelve primaries perfect, and these are thin, with no gonad or filament and little muscle. Of the two couples of primaries over and above the Edwardsia eight, one couple has a modified kind of muscula- [3) CLASSIFICATION OF ACTINIARIA 311 ture and no filaments, and the other couple has no gonad, filament, or well-developed muscle. The eight Edwardsia mesenteries have huge circumscribed retractors of curious form, which seem to be tending to shift off the mesenteries ; they also have gonads and filaments, O. lindahli, Carp. 1871, p. 159. (See Fowler, 1894, p. 461.) This genus seems to be eligible for Marsupiferidae. As far as I can understand the account of it, I take it that it has a mesogloeal sphincter, and the rest fits in fairly well. It is a queer form with a reduced number of macrocnemes ; ef. Decaphellia and some Halecampas. 3. LITERATURE. This list is additional to that given in Part I, and the numbering carries on from the end of the former list. References in Parts II and III are to both lists taken as one whole. 101. Agassiz, L.—‘‘ Lettre 4 M. Alexandre de Humboldt sur le développe- ment de la Rhodactinia Davisii’’, ‘Comptes rendus de l’ Académie des Sci.’, 25, 1847, p. 677. 102. Agassiz, Alexander A., and Elizabeth C.—‘ Seaside Studies in Natural History ’, Boston, 1865. 103. Boveri, Th.—‘‘ Das Genus Gyractis, eine radial-symmetrische Actinienform ”’, ‘ Zool. Jahrb.’, Abth. Syst., vol. 7, p. 241, 1893. 104. Carlgren, O.—“ Protanthea simplex, n. gen., n. sp., eine eigen- tiimliche Actinie, Vorl. Mitteilung’’, ‘Ofv. Kongl. Vet.-Akad. Forh.’, 1891, no. 2, p. 81. 105. “Uber das Vorkommen von Brutriiumen bei Aktinien”’, ibid., 1893, no. 4, p. 231. 106. —— “ Zur Kenntnis der Minyaden ”’, ibid., 1894, no. 1, p. 19. 107. —— “Zur Mesenterienentwicklung der Aktinien”, ibid., 1897, no. 3, p. 159. 108. —— ‘“‘ Uber abschniirbare Tentakel bei den Actiniarien”’, ‘ Zool. Anzeiger ’, Bd. xxii, no. 578, 1899, p. 39. 109. —— “ Tafelerklarung der Actiniarien und Zoantharien’’, ‘Symbolae physicae seu Icones adhuc ineditae etc., von Hemprich u. Ehren- berg, fol., Berolini, 1899. G. Reimer. Supplement, Zool. Phytozoa ’. 110. “Zur Kenntnis der stichodactylinen Actiniarien ’’, ‘ Ofv. Kongl. Vet.-Akad. Forh.’, 1900, no. 2, p. 277. iUER “Zur Mesenterienmuskulatur der Actiniarien’’, ‘Zool. An- zeiger ’, Bd. xxvili, no, 14/15, 1905, p. 510. 112. Clubb, J. A.—‘‘The mesenteries and oesophageal grooves of NO. 262 Z 312 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 128.) —— 125. 126. 2. —— 128. 129. 130. 131. 132. T. A. STEPHENSON Actinia equina”’, ‘Trans. Liverpool Biol. Soc.’, vol. xii, p. 300. Clubb, J. A.—‘‘ Actiniae’’, ‘ Report on the ‘Southern Cross” Antarctic Expedition ’, London, 1902. Cuvier, G.— Tableau élémentaire de histoire naturelle des animaux’, Paris, 1797-8. Dana, J. D.—‘‘ Zoophytes”’, Philadelphia, 1846. ‘U.S. Explor. Exped. during the years 1838-42’. (Atlas published 1849, and Synopsis 1859), Duchassaing, P., and Michelotti, J.—‘‘ Mémoire sur les Coralliaires des Antilles’, ‘Mem. Real. Accad. Torino’, 2 a, vol. xix, 1860. Duchassaing de Fontbressin and Michelotti.— Supplément anx Coralliaires des Antilles”, ibid., vol. 23, p. 97 (° Actinideae’, p. 118), 1866. Duerden, J. E.—‘‘ On the genus Alicia (Cladactis), with an anatomical description of A. costae, Panc.”, © Annals and Magazine of Nat. Hist.’, 6th series, 1895, p. 213. —— “The Actinian Family Aliciidae ’’, ibid., (6), vol. 20, 1897. —— ‘On the Relations of certain Stichodactylinae to the Madreporaria”’, ‘ Journ. Linn. Soc.’ Zool., vol. xxvi, 1898, p. 635. —— ‘On the Actinian Bunodeopsis globulifera, Verr.”’, ‘Trans. Linn. Soe.’, (2), vol. viii, 1902, p. 297. —— ‘Monograph on West, Indian Madreporaria”’, “Mem. Acad, Wash.’, vol. viii, 7th memoir, 1902. —— “Report on the Actinians of Porto Rico’’, * Bull. U.S. Fish Comm., 1900’, vol. 2, 1902. “West Indian Sponge-incrusting Actinians”’, ‘ Bull. Amer. Mus. Nat. Hist.’, vol. 19, 1903. Ehrenberg, C. G.—‘ Die Corallenthiere des rothen Meeres physio- logisch untersucht und systematisch verzeichnet ’, Berlin, 1834. Fowler, G. H.—‘‘ Two new types of Actiniaria”’, ‘Quart. Journ. Micr. Sci.’, 2nd series, 1889. “Octineon Lindahli (W. B. Carpenter): an undescribed Anthozoan of novel structure ’’, ibid., N.S., vol. 35, 1894, p. 461. Gemmill, J. F.—‘‘ The Development of the Sea-Anemones Metri- dium dianthus (Ellis) and Adamsia palliata (Bohad.)”, ‘Phil. Trans. Royal Soc.’, Ser. B, vol. 209, 1920, p. 351. —— ‘The Development of the Mesenteries in the Actinian Urti- cina crassicornis’”’, ‘ Proc. Zool. Soc.’, 1920, p. 453. Gravier, C.—‘‘ Sur un type nouveau d’Actinie de He San-Thomé”’, * Bull. Mus. d’Hist. Nat.’, Paris, vol. 22, 1916, p. 234. —— ‘ Bull. Inst. Océanogr. Monaco’, no, 344, 1918. —— Ibid., no. 346, 1918. ———————————— 133. 134. 135. 136 137. 138. 139. 140 141. 142 143. 144, 145, 146, 147. 148. 149. 150. 151. CLASSIFICATION OF ACTINIARIA ole Hargitt, C. W.—‘ Biol. Bull. Woods Holl, Mass.’, vol. 14, 1908, p. 965. ——‘‘Cradactis variabilis; an apparently new Tortugan Actinian”’, ‘Wash. Carnegie Inst. Pub.’, 132 (Papers Tortugas Lab. 3), 1911. Hornell, J.—‘‘ The Structure of Anemones ”’, Study IX in “ Micro- scopical Studies in Marine Zoology’’, ‘ Journ. Marine Zool. and Microscopy ’, vol. i, no. 3, 1894. Klunzinger, C. B.—‘ Die Corallthiere des rothen Meeres. 1t¢* Theil : Die Alcyonarien u. Malacodermen ’, Berlin, 1877. Lager, E.—“‘ Actiniaria”’, ‘Fauna Siidwest-Australiens’, hrsg. v. W. Michaelson u. R. Hartmeyer, Bd. 3, Lfg. 8, Jena, 1911. Lesson, R. P.—‘ Zoologie. Voyage autour du Monde sur la Corvette de 8.M. “‘la Coquille ’’ pendant les années 1822-5, &c.’, Paris, 1828 et seq. Lesueur, C, A.—‘‘ Observations on several species of Actinia’’, * Journ. Acad. Nat. Sci. Philad.’, vol. 1, 1817. MacMurrich, J. P.—‘*‘ A revision of the Duchassaing and Michelotti Actiniarian types in the Museum of Nat. Hist., Turin’’, ‘ Boll. Mus. Zool. Anat. comp. Torino’, vol. 20, 1905. —— ‘The Actiniaria of Passamaquoddy Bay, &c.’’, ‘ Ottawa Proc. and Trans. Royal Soc. Can.’, 3rd series, vol. 4, 1910. Mitchell, P. Chalmers.—‘‘Thelaceros Rhizophorae, n.g,, n. sp., an Antinian from Celebes’’, ‘ Quart. Journ. Micr. Sci.’, N.S., vol. 30, 1890, p. 551. Moseley, H. N.—‘‘ On new forms of Actiniaria dredged in the deep sea; with a description of certain pelagic surface-swimming species ’’, ‘ Trans. Linn. Soc.’, 2nd series, vol. 1, Zool., 1879, p. 295. Miller, O. F.---‘ Zoologiae Danicae Prodromus, seu Animalium Daniae et Norwegiae ’, Havniae, 1776. Pax, F.—‘“ Anthozoa. Die Actinienfauna Westafrikas’’ (in L. Schultze, ‘Zool. u. anthrop. Ergebnisse e. Forschungsreise in Siidafrika ’, Bd. i, Lfg. 2), Jena, ‘ Denkschr. med. Ges.’, vol. 13, 1908, p. 463. —— “ Actinia kraemeri, die essbare Seeanemone der Samoainseln ”’, “Zool. Anz.’ Leipzig, vol. 44, 1914. Roule, L.—‘ Hexactinides’. Expéd. antarct. francaise 1903-5, commandeée par le Dr. Jean Charcot. Paris, 1909. Saville-Kent, W.—‘ The Great Barrier Reef of Australia’, London, 1893. —— ‘The Naturalist in Australia’, London, 1897. Simon, J. A.—‘ Ein Beitrag zur Anatomie und Systematik der Hexactinien’. Inaugural-Dissertation, Miinchen, 1892. Stephenson, T. A.—‘ On the Classification of Actiniaria, Part 1: ih 7 314 T. A. STEPHENSON Forms with Acontia and Forms with a Mesogloeal Sphincter ’”’, “Quart. Journ. Micr. Sci.’, vol. 64, part iv, 1920, p. 425. 152. Stephenson, T. A.—“The Genus Corallimorphus”, ‘ Proe. Royal. Irish Acad.’, vol. xxxv, Section B, no. 9, 1920, p. 178. 153. Stuckey, F. G. A.—‘‘ Notes on a New Zealand Actinian, Bunodes aureoradiata’”’, ‘Trans. New Zealand Inst.’, vol. xli, 1909, p. 367. 154. Torrey, H. B.—‘ The Californian shore anemone Bunodactis xanthogrammica”, Berkeley, ‘Univ. Calif. Pub.’, Zool, vol. 3, 1906, p. 47. 155. Verrill, A. E.—‘ Trans. Connect. Acad.’, x (2). 156. —— Ibid., xi (1). 157. Wilsmore, L. J.—‘ Journ. Linn. Soc.’, vol. 32, 1911, p. 39. 158, Wilson, H. V.—‘‘ On a new Actinian, Hoplophoria coralli- gens”, ‘Studies from the Biol. Lab. Johns Hopkins Univ.’, vol. iv, 1890. APPENDIX To List or LITERATURE. The following list consists of references to literature which it has been neither possible nor necessary for me to consult personally, but which is mentioned in the foregoing paper in connexion with some of the original descriptions of certain species listed after the generic definitions, In cases where the original description has not been consulted, one or more later and usually fuller descriptions of the species have been seen. These references are simply included for convenience in case they should be needed, and are taken from Andres, 1883, and other works. I cannot, of course, guarantee that the descriptions referred to will be found in the works quoted, but I have done my best to select the right references. The lettering is continued on from the end of the similar list given in Part I. (Zb) Allman, G.—‘ Description of a new genus of Helianthoid Zoophytes (Corynactis) ”, ‘ Ann. Mag. Nat. Hist.’, la, vol. xvii, 1846, p. 417. (Zc) Andres, A.—‘‘ Prodromus neopolitanae, &c.”’, ‘ Mittheil. Zool. Stat. Neapel ’, vol. ii, 1880, p. 305. (Zd) Bell, F. J.—‘* Description of a new Species of Minyad (Minyas torpedo) from North-west Australia’, ‘Journ. Linn. Soe. Zool.’, vol. 19, 1885, p. 114. (Ze) Blainville, H. M.—‘ Manuel d’Actinologie et de Zoophytologie’, Paris, 1834-7. (Zf) Brandt, J. F.—‘ Prodromus Descriptionis animalium ab H. Mertensio in orbis terrarum circumnavigatione observatorum’, Fas. 1, Petropoli, 1835, p. 17. CLASSIFICATION OF ACTINIARIA 315 (Zg) Carpenter, W. B., and Jeffreys, G.—‘* Report on Deep Sea Researches, &c.”’, ‘ Proc. Royal Soc.’, 1871, p. 159. (Zh) Cocks, W. P.—‘ Rep. Cornw. Polytech. Soc.’, 1849, p. 94. (Zi) Contarini, N.—‘ Trattato de Attinie ed osservazione sopra alcune di esse ecc.’, Venezia, 1844. (Zj) Danielssen and Koren.—“ Nye Actinier. Siphonactinia et Actinopsis ” ‘Fauna litt. Norvegiae ’, Bergen, 1846-56, Part IJ, 1856, p. 87. (Zk) Duchassaing, P.—* Animaux radiaires des Antilles’, Paris, 1850. (Zl) Ehrenberg, C. G.—‘ Abhandl. Kgl. Akad. Wissensch.’, Berlin, 1832. (Zm) Ellis, J., and Solander, D.—‘* The natural history of many curious and uncommon zoophytes, &c.’, London, 1786. (Zn) Gosse, P. H—‘On Aegeon Alfordi, a new British Sea- Anemone ”’, ‘ Ann. Mag. Nat. Hist.’, 3a, vol. xvi, 1865. (Zo) Hutton, F. W.—‘‘ The Sea Anemones of New Zealand”, ‘ Trans. New Zealand Inst.’, vol. xi, 1878. (Zp) Johnson, J. Y.—‘* Notes on the Sea-Anemones of Madeira ”’, ‘ Proc. Zool. Soc. Lond.’, 1861, p. 298. (Zq) Johnston, G.—* History of the British Zoophytes ’, London, 1847. (Zr) Jourdan, E.—‘* Recherches zoologiques et histologiques sur les zoanthaires du Golfe de Marseille’’, ‘ Ann. d. Se. Nat.’, 6a, vol x, 1880, p. 1. (Zs) Leuckart.—See Riippel. (Zt) Linnaeus, C.—‘ Systema naturae sive regna tria, &c.’, edit. xii, Holmiae, 1766-8. (Zu) Milne-Edwards and Haime.—‘‘ Monographie des Polypiers fossiles ... précédée d’un tableau général de la classification des Polypes”’, * Arch. du Mus.’, vol. v, 1851, p. 1. (Zv) Panceri, P.—‘‘ Nuovo genere di polipi Actiniari (Cladactis) ”’, ‘ Rend. d. R. Accad. di Sc. f. e mat. Napoli’, 1868, p. 30. (Zw) Riippel, W. P. H.—‘ Atlas zu der Reise im nordlichen Africa’, Frankfurt, 1826-31 (Wirbellose Thiere, 1828). (Zx) Pennant.—‘ A British Zoology ’, London, 1766. (Zy) Pfeffer—‘* Zur Fauna von Siid-Georgien”, ‘Jahrb. Hamburg. Anstalt ’, VI. Jahrg., 2. Halfte, 1888, 1889. (Zz) Sars, M.—‘ Beskritelser og Jagttagelser over nogle moerkelige eller nye i Havet ved den Bergenske Kyst levende Dyr af Polypernes, Acalephernes .. . Classer’, Bergen, 1835. ‘* Beretning om en i sommeren 1849 foretagen zoologisk Reise i Lofoten og Finmarken”’, ‘ Nyt. Mag. Nat.’, ecc. vi, 2°, no. 10, 1851, p. 122. (Zzb) Stimpson, W.—‘‘ On some Marine Invertebrata inhabiting the shores of South Carolina ’’, ‘ Proc. Soc. Nat. Hist. Boston ’, vol. v, 1854-6, p. 110. (Zza) 316 T. A. STEPHENSON (Zzc) Studer, T.--‘ Anthozoa polyactinia, welche waihrend der Reise S.M. Corvette ‘Gazelle’ um die Erde gesammelt wurden ”’, “ Monats- berichte der Akademie der Wissensch. in Berlin’, 1878, p. 524. (Zzd) Thompson, W.—“ Description of a new species of Corynactis, &c.”, ‘Proc. Zool. Soc.’, vol. xxi, 1853, p. 107. (Zze) Verrill, A. E.—‘ Revision of the Polypi of Eastern Coast of U.S.”, “Mem. Soc. Nat. Hist.’, vol. i, Boston, p. 1. Read 1862, published 1864, (Zzf) ——‘ Notice of recent additions to the marine invertebrata of the North-eastern Coast of America, &c.”’, ‘ Proc. U.S. National Mus.’ Washington, 1879-80. 4. InpEX To GENERA DEALT WITH IN Parts II anp III. Numbers printed in heavy type refer to the pages on which the generic definitions are to be found. Words printed in italics are synonyms or genera of uncertain position. Acremodactyla, U1, 294 Actinauge, ITI, 308 Actineria, IT, 548 ; ILI, 295, 296-7 Actinernus, II, 523, 545; III, 259. 307-8 Actinia, IT, 501, 524, 527, 531, 546: 5) III, 266—7, 282, 293 Actinioides, II, 524, 526-8; ILI, 270- 71 Actinodendron, Il, 547, 562; III, 294-5 Actinoporus, IL, 534, 547; M1, 291, 293 Actinoscyphia, II, 560 ; IIT, 259, 308 Actinostella, I11, 283-4 Actinostephanus, I], 547; LIL, 294 Actinotryx, II, 511-12, 515, 535, 544 ; IIT, 304, 306 Aegeon, III, 270 Aegir, III, 251, 253 Aiptasia, IT, 504 Alicia, II, 530-31, 546 ; ILI, 279-80 Allantactis, ILI, 310 Ammodiscus, III, 310 Ammophilactis, I11, 309-10 Amphianthus, III, 309 Andresia, II, 518, 521, 545; IL], 255. 264-5 Andvakia III, 310 Anemonia, IT, 501, 524, 527-8, 531, 546; III, 266, 267-8, 270, 278, 308 Anthea, I11, 267 Antheomorphe, U1, 278 Antheopsis II, 535-6, 548; LIL], 265, 297-8, 299, 300 Antholoba, ILI, 309 Anthopleura, Il, 526-8 ; 111, 270-71, 286 Antiparactis, III, 310 Archactis, III, 309 Artemidactis, II, 515, 557 Asteractis, III, 283-4 Aulactinia, III, 270-72 Aureliania, II, 512, 534-5, 547; I11, 274, 291-3 Bicidium, U1, 257 Bolocera, II, 506-8, 526-7, 531, 546 ; III, 266, 272-3, 275-7, 287 Boloceroides, II, 505-8, 526, 531, 540, 545, 564; III, 250, 262, 263 Boloceropsis, II, 531, 546; ILI, 266, 276 Bunodactis, I], 500-V1, 515, 526-9, 531, 546; III, 266, 269, 270-74, 276, 278-9, 284-6 Bunodella, III, 270 EEE —— CLASSIFICATION Bunodeopsis, II, 506-7, 526, 530-31, 546 ; ITI, 281-2, 286-7 Bunodes, II, 526 ; III, 270-71, 285 Bunodosoma, II, 283, 285-6 Capnea, II, 292 Carlgrenia, II, 523, 545, 550, 560; 257, 258 Caryophyllia, IT, 511 Catadiomene, II, 554, 560; III, 308 Cereactis, III, 268 Chalmersia, ITI, 302 Charisea, III, 262 Chitonactis, II, 530 Chitonanthus, II, 530; III, 308 Chondractinia, II, 530 Cladactis, III, 279 Comactis, IIT, 267, 278 Condylactis, Il, 524, 527-8, 531, 546 ; IIT, 266, 268, 269-70, 277-8 Condylanthus, II, 524, 531, 545; IL], 262 Corallimorphus, Il, 515, III, 301-2, 308 Corynactis, IT, 543; LI, 301, 302-3 Cradactis, II, 5380-31, 546; ITI, 281-2. 284-5 Crambactis, III, 290 Oribrina, III, 270-71 Cryptodendron, II, 502, 512, 534, 548 ; III, 295, 296 Cystiactis, II, 530-31, 546; 281-2, 286 Doo,- O40 5 IU, Dactylanthus, II, 508-9, 543; III, 249, 250 Decaphellia, IT, 520; IIT, 311 Diplactis, III, 266 Discosoma, IT, 511, 544; III, 304 Discosomoides, III, 298 Dofleinia, IT, 531, 546; ITLL, 266, 277 Edwardsia, II, 550, 557; III, 248 Eloactis, II, 518, 520-21, 544, 560 ; III, 254, 255 Endocoelactis, III, 257 Koactinia, IT, 532, 563, 565 Eosagartia, II, 532, 557, 563, 565 Epiactis, II, 512, 526, 529, 531, 546 ; III, 266, 274-5, 278, 292 OF ACTINIARIA olt Lpigonactis, II, 274 Hucladactis, III, 285 Huphellia, IIT, 306, 310 Hvactis, III, 270 Fenja, IIT, 251, 253 Gliactis, III, 307 Glyphoperidium, TI, 274, 278 Glyphostylum, IT, 531, 546 ; III, 266, 277-8 Gonactinia, IT, 504—5, 543, 550, 554-6; ITT, 247, 248 Gyractis, III, 278-9 Gyrostoma, II, 524, 531, 546; IIT, 263-4, 266, 267-8, 276, 278, 308 Halcampa, IT, 499, 504, 515, 518-20, 544, 550, 554-6; IIT, 251, 252-3, 257, 290, 311 Halcampactis, II, 517; III, 265 Halcampella, TIT, 251, 253 Halcampoides, II, 518-20, 532, 544 ; TIT, 251-2, 253 Halcampomorphe, LI, 253 Haleurias, II, 523, 545, 560; III, 257-8 Halianthella, VII, 251 Halianthus, 111, 251 Haloclava, II, 518, 520-21, 544; IIT, 253-4, 255-6 Harenactis, II, 501, 518, 520-21, 544, 560; III, 254, 256 Helianthopsis, III, 298-9 Heteranthus, II, 548; III, 290 Heterodactyla, III, 295-6 Homostichanthus, IT, 547; III, 294 Hoplophoria, III, 280, 287-8 Hormathia, II, 580; III, 266, 279, 308 Hormosoma, III, 308 Ilyanthopsis, III, 260, 268, 270, 278 Ilyanthus, JI, 518, 520-21, 544; IIT, 254, 255, 264 Isactinernus, IT, 545 ; III, 259, 260 Tsactinia, III, 267 Tsaura, III, 304—5 Isocorallion, II, 543 ; III, 301, 302 318 1", A. STEPHENSON Isotealia, II, 526, 531, 546; III, 266, 275, 292 Txalactis, IL, 531, 546; III, 266, 277 Kylindrosactis, ITI, 279 Lebrunia, II, 530-31, 546; ILI, 280— 82, 287, 288 Leiotealia, III, 274, 291-3 Leipsiceras, II, 531, 546; IIT, 266, 276 Liponema, III, 275-6 Lophactis, 111, 283-4 Macrodactyla, II, 524, 531, 545; IIT, 263, 265, 272 Madoniactis, III, 272, 279 Megalactis, Il, 535, 547; ILI, 294, 295 Metridium, III, 279 Minyas, Il, 501; II, 288 Myonanthus, II, 506-7, 524, 531, 545 ; III, 263, 265 Myriactis, II1, 278 Nautactis, Il, 533 Nectactis, III, 306 Nevadne, II, 531, 545; III, 263, 264 Octineon, III, 310, 311 Ophiodiscus, IT, 530 Oractis, II, 504, 505 ; III, 248, 249 Orinia, LI, 511, 544; III, 304, 305 Oulactis, III, 283-4 Paractis, III, 307 Paradiscosoma, II, 509, 511, 514-15, 544, 557; ILI, 304-5 Paranemonia, III, 267-18 Parantheopsis, IT, 531, 546 ; III, 266, 269, 270 Paranthus, III, 307, 309-10 Paraphellia, III, 310 Parazoanthus, IT, 550 Peachia, IT, 501, 517-18, 520, 544, 552, 660, 574; ILI, 251, 254, 256-7 Pentactinia, II, 518-20, 544; III, 251, 254 Phellia, IIT, 307, 310 Phialactis, IIL, 305 Philomedusa, TU, 257 Phlyctenactis, I11, 286 Phyllactis, II, 530-31, 546; III, 281-2, 283, 284-5 Phyllodiscus, II, 502, 530-31, 534, 546, 561; III, 279, 280-81, 285 Phymactis, IL, 525, 5380-31, 546, 554 ; ITI, 281-2, 285-6 Phymanthus, II, 501, 512, 585, 547 ; III, 289-90 Polyopis, III, 254 Polysiphonia, IIT, 259, 308 Polystomidium, III, 278 ’ Porponia, III, 259, 307 Protanthea, II, 504-5, 543, 555; ITI, 247, 248 Pseudoparactis, III, 308-10 Pseudophellia, IT, 531, 546; ILI, 266, 275 Ptychodactis, Il, 508, 543; ITIL, 249 Radianthus, II, 548 ; III, 278, 297-8, 299 Ragactis, III, 277 Raphactis, TIL, 309-10 Rhodactinia, III, 272-3 Rhodactis, I, 511, 544 ; ILI, 304, 305 Ricordea, II, 511, 544; III, 304, 305 Rivetia, III, 285-6 Saccactis, III, 284 Sagartia, ITI, 271, 308 Sagartiomorphe, ILI, 308, 310 Scytophorus, II, 518-20, 544; II, 251, 253, 254 Sicyopus, III, 307 Siphonactinia, III, 256 Stephanactis, III, 308-9 Stephanauge, III, 308-9 Stichodactis, III, 298-9 Stichophora, IT, 533, 546 ; III, 288 Stoichactis, II, 548 ; ILI, 297, 298-9 Stomphia, III, 279, 310 Synactinernus, IT, 545; III, 259 Synanthus, III, 309 Synhalcurias, II, 545; III, 259, 260, 278 ll CLASSIFICATION Tealia, II, 501, 518, 526, 528-9, 531, 546, 554; III, 266, 272-4, 276, 279 Tealidium, ITI, 308 Tealiopsis, III, 279 Telmatactis, III, 307 Thalassianthus, Il, 535, 548, 560 ; TIT, 295-6 Thaumactis, IJ, 530-31, 546; 281-2, 287 III, OF ACTINIARIA 319 Thelaceros, III, 289 Thoracactis, III, 307 Triactis, III, 281 Urticina, III, 271, 273 Viatrix, III, 287 A LG Ta SA ae #0, Jy il, AA On the Post-Embryonic Development of certain Chalcids, Hyperparasites of Aphides, with Remarks on the Bionomics of Hymenopterous Parasites in General. By Maud D. Haviland, Research Fellow of Newnham College. With 7 Text-figures, INTRODUCTION. In the summers of 1919 and 1920, certain hyperparasitic Chalcidoidea were reared from material collected in the field for the study of two hyperparasites of aphides, the Procto- trypid, Lygocerus (5), and the Cynipid, Charips (6). The following is an account of the post-embryonic develop- ment of two common forms, which were obtained in consider- able numbers from the cocoons of the Braconid, Aphidius ervi, Hal., a parasite of Macrosiphum urticae, Kalt., an aphid that infests the stinging nettle. I would here express my sincere thanks to Professor Stanley Gardiner, who gave me facilities to carry out the work in the Zoological Laboratory, Cambridge ; and to Mr. J. Waterston of the British Museum (Natural History), who kindly deter- mined the species of Chaleidoidea submitted to him. BIONOMICGAL AND SYSTEMATIC PosITION. The two species now considered belong to the sub-family Sphegigasterinae of the family Pteromalidae, which is, accord- ing to Ashmead, the largest group of the Chalcidoidea, and the most difficult to classify. Asaphes vulgaris, WIk., belongs to the tribe Asaphini, the majority of which are said by Ashmead to be parasitic on Aphidiidae and Coccidae (1). 322 MAUD D. HAVILAND Pachycrepis clavata, Wlk., is included in the allied tribe Pachyneurini, which Ashmead says are regarded as chiefly parasites of the same Rhynchota, but he adds that these insects have other hymenopterous parasites, through which the Pachyneurini are probably hyperparasitic. In addition to Asaphes and Pachycrepis, two females of a species of Pachyneuron were reared. ‘The eggs and early larval stages of the two former species are indistinguishable. The egg of the Pachyneuron is charac- teristic, but its development was not observed. Various Chalcidae have been recorded as reared from aphides, and it is possible that some of them may yet prove to be primary parasites; but the forms described here are hyperparasites of the plant-lice through the larvae of Aphi- dius, and allied genera of Braconidae, which develop internally in aphides. The Chalcidae do not oviposit until the aphid is dead and the Aphidius has woven its cocoon, and is ready | to transform inside the empty skin of its late host. Their true relation to the aphid was shown as long ago as 1834 by Nees ab Esenbeck for Asaphes or a similar form, and his observations have been confirmed by Walker and Buckton, and subsequently by other writers. These hyperparasites do not appear to be specific for different Aphidiidae or aphides. In 1919 IT reared Asaphes vulgaris from an Aphidius in Rhopalosiphum sonchi, Kalt., and also from Aphidius salicis, Hal., a parasite of Aphis saliceti, Kalt. This Braconid and aphid are less than half the size of A. ervi and M. urticae, but the Chalcid seems to adapt itself to either form, and thus probably has considerable latitude in the choice of a host. PAIRING. All observed ovipositions of Asaphes and Pachy- crepis took place after pairing. Only two examples of Pachyneuron were obtaimed, and both were females. One laid a single egg parthenogenetically and died soon after- wards. The other lived for some days but did not oviposit. DEVELOPMENT OF CHALCIDS 323 OVIPOSITION. The female Chalcid selects a cocoon containing an Aphi- dius larva on the point of metamorphosis, but sometimes a newly-transformed pupa may be chosen. ‘The hyper- parasite shows considerable excitement in her search, and runs round the cocoon, tapping it eagerly with her antennae. Finally she mounts upon it, facing the head of the aphid, and, boring through the integument with its silk lining, she deposits a single egg upon the upper surface of the body of the Aphidius larva, as it lies curved head to tail within the cocoon. The whole operation lasts from one to three minutes. Only one egg is inserted at each oviposition, and when more are found they are the result of different attacks. The number of eggs laid by each female seems to be between thirty and forty, but it is difficult to be precise on this point as the insects will live for some days in captivity, and the eggs in the ovarian tubes do not all mature at the same time. Tur Hae. The eggs of Asaphes and Pachycrepis are indistin- suishable from one another. They are white, elliptical bodies TExt-ric. 1. TEXT-FIG, 2. Egg of Asaphes vulgaris, Egg of Pachyneuron sp. < 100. x 100. with a smooth chorion, having dimensions, -29 x-12 mm. (Text-fig. 1). The single example of the egg of Pachyneuron was long, oval, and slightly curved. On the concave side, the 324 MAUD D. HAVILAND chorion is smooth, but the rest of the surface is covered with minute scales or papillae. Dimensions, -31 x-10 mm. (Text- fig. 2). This egg is very similar to that of Pachyneuron cifuensis, Ashm., figured by Howard and Fiske (7). Tuer First Instar. Dimensions -45 mm. «-23 mm. The egg hatches about sixty hours after oviposition. The larva in the first instar much resembles in general form that of the Lygocerus previously described (Text-fig. 3). It is TEXT-FIG. 3. TEXxtT-FIG, 4, The larva of the first instar. » 300. Mandibles of the newly-hatched larva. x 600, white, semi-transparent, and consists of thirteen segments in addition to the head, which is furnished with two tactile papillae. The mouth is small and oval, and the mandibles are somewhat more curved than those of the larva of Lygocerus (Text- fig. 4). The tracheal system consists of a pair of longitudinal trunks, united by an anterior commissure between the first and second segments, and a posterior commissure in the eleventh segment. At this stage there are four pairs of functional spiracles, namely between the first and second segments, and on segments 4-6 inclusive. These segments are supplied with dorsal and ventral lateral branches, and the developing spiracular trunks of segments 3 and 7-9 are visible. The larva makes an incision in the skin of the host, and as the DEVELOPMENT OF CHALCIDS 325 body-fluids of the latter fill the midgut the hyperparasite is tinged pale yellow. INTERMEDIATE STAGES. The exact number of eedyses of these Chalcids was not determined. There is no marked change of form during development, but the body becomes more globose, and the head less conspicuous. The cephalic papillae do not disappear as in Lygocerus, but persist until metamorphosis. The spiracles on segments 7 and 8 become functional, and those on segment 5 open shortly afterwards. The ninth pair (on segment 10) open as development proceeds, but the tenth pair are closed until shortly before metamorphosis. The host dies a day or two after the Chalcid larva has begun to feed, and decomposes rapidly. These hyperparasites penetrate more deeply into the decaying tissues than do the larvae of Lygocerus at the same stage. The larvae are also more fragile and transparent, and are easily crushed or ruptured when handled. Tor FunLL-GRowN LARVA. Dimensions, 1:26 mm x -60 mm. The larva when fully fed is creamy white and opaque, slightly curved, and with a smooth glabrous cuticle. The body tapers somewhat to the anus, and the segmentation is well marked. The head bears a pair of conspicuous papillae, and a pair of similar, though smaller, appendages are found on the first segment. In addition, each segment from the first to the fifth or sixth is furnished with one or two pairs of minute spines (Text-fig. 5). The labrum and labium both bear palps, as do also the maxillae. The mandibles are simple, and strongly chitinized, though less massive than in Lygocerus (text-fig. 6). The ramifications of the tracheal system are more elaborate than in the preceding stages, and the tenth spiracle (on seg- ment 11) becomes functional. The internal structure is of the type usual among hymeno- 326 MAUD D. HAVILAND pterous larvae. The narrow oesophagus opens by a valve into the vast mesenteron filled with food, which is churned to and fro by muscular contractions. The mesenteron is closed posteriorly and does not communicate with the proctodaeum until immediately before metamorphosis. A pair of short Malpighian tubules enter the hindgut at its anterior end. The salivary glands extend backwards to the ninth segment, and lie on either side of the gut ventrally as a pair of long straight tubes. Behind the head their ducts unite to form the common salivary duct, which opens on the floor of the mouth. The ventral nerve-cord appears as a broad unconstricted band extending backwards into the tenth segment. The rest of the internal structure calls for no particular comment. TEXxtT-FIG. 5. TEXtT-FIG. 6. Head of the full-grown larva, x 75. In a cocoon opened carefully when the hyperparasite was almost full grown, it was possible to watch the transformation into the pupa, and by this means if was determined that the mature larvae of the two forms examined were identical in appearance. Attempts to follow the earlier development in the same way always failed, because exposure to the air caused the decaying tissues of the Aphidius to dry up and thus brought about the death of the hyperparasite. The larval development of the Chalcidoidea has been more studied than that of other Hymenoptera parasitica, but so much diversity exists within the family, owing to secondary modifica- tions induced by various hosts and habits, that a comparative account can throw little light on their affinities. The forms here described agree very closely with that of Torymus propinquis, an ectoparasite of certain Cecidomyiidae, DEVELOPMENT OF CHALCIDS oor studied by Seurat (10). The general form and the number and order of opening of the spiracles are the same in both cases. Certain parasites of Coccidae, described by Imms (8) show a reduction in the number of spiracles from behind forwards ; but in one, Aphyeus melanostomatus, rudimentary trunks of the tenth pair appear during development, though they never become functional. Ectoparasitic Chalcidae, such as Asaphes and Torymus, have probably retained certain primitive features, such as the full number of spiracles, which have been lost in the more specialized and frequently hypermetamorphic forms, found among the endoparasitic members of the super-family. PUPATION AND EMERGENCE. When the remains of the Aphidius have been completely devoured, the gut of the hyperparasite opens, the meconium is voided, and the Chalcid pupates within the cocoon pre- viously woven by the Aphidius inside the skin of the aphid. The pupal stage lasts from fourteen to sixteen days, for Asaphes and Pachycrepis; but in a single observed instance of Pachyneuron the period of pupation was only ten days. When ready to emerge, the imago gnaws a hole in the cocoon and creeps out. The adults lived in confinement for from four to seven days, and fed on the sap oozing from cut leaves, and on honey-dew which had fallen from the aphides. At least two generations may occur in the year, but the exact, number was not ascertained : it is probably dependent on the number of hosts obtainable. There is no evidence to show how these Chalcids pass the winter. REMARKS ON THE Bionomics oF HyMENOPTEROUS PARASITES IN GENERAL. The relations of any animal to its enemies, predatory or parasitic, form what may be termed a bionomical complex ; although the limits of such a complex are often difficult to determine, especially when the enemy has a wide choice of alternative food or host species. NO, 262 Aa 328 MAUD D. HAVILAND Aphides, with their parasites and hyperparasites, form a bionomical complex of considerable intricacy ; but its limits are well defined, and it is thus convenient for the study of the bionomics of parasitism. ‘The Aphididae, which are a large and distinct sub-family of Braconidae, are all obligative parasites of Aphides, and have no alternative hosts ; and the hyperparasites, which belong to the three super-families of Cymipoidea, Chalcidoidea, and Proctotrypoidea, are exclusively confined to the Aphididae, with the exception of certain Cynipids (Charipinae) and Chalcids, which possess allied forms parasitic upon Coccidae. The bionomics of some members of this complex are com- paratively simple. Thus, the species of Charips (Cynipidae) described elsewhere (6) are invariably parasites of Aphidius, and thus hyperparasites of the aphid, and, so far as is known, never prey upon another hymenopteron. Thestatus of such Proe- totrypids as Lygocerus (5), and Chalcids such as Asaphes and Pachycrepis, is more difficult to determine, because although usually parasites of Aphidius, and _ therefore standing in the same relation to the aphid as Charips, they may on occasion be parasitic on each other. The interrelations of these forms are shown in the accompanying diagram (Text- fig. 7). An Aphidius cocoon is sometimes found to contain two hyperparasites of either, or both these species, the result of two successive ovipositions. Fiske (8) has called this phase of parasitism ‘superparasitism’; but as the word means neither more nor less than hyperparasitism, a term already employed in cases where the parasite is itself attacked by a parasite, I would suggest replacing this etymological hybrid by ‘epiparasitism ’. In such a case, in the aphid complex, only one imago emerges from the cocoon. Either one parasite is sufficiently advanced to devour the host before its rival can compete with it ; or else, if both parasites are of the same age, there is insufficient food to nourish both up to meta- morphosis, and they starve to death. One seems never to make a direct attack on the other. But in certain instances a Chalcid hyperparasite larva, eae habs DEVELOPMENT OF CHALCIDS 329 generally one three-quarters grown, may be found with the egg or larva of a Proctotrypid, or of another Chalcid, on its body. It may be that the second hyperparasite deliberately oviposits upon the larva of the first, if the Aphidius host TEXT-FIG. 7. \ Aphid ; Aphidius - Lygocerus Asaphes Pachycrepis Diagram to illustrate the bionomical complex of an aphid, its parasites, and hyperparasites. Endoparasitism is indicated by a double margin to the host. has already succumbed to the attack, and originally I thought that this was the case; but further observation led me to modify this conclusion. Thus instances of this kind are rare compared with those of simple epiparasitism and attempts to induce the Chaleid or Proctotrypid to oviposit on the full- srown larya of another hyperparasite that had already devoured A a 2 330 MAUD D. HAVILAND the host, always failed. The more probable explanation seems to be that the intention of the second hyperparasite is to ovi- posit upon the Aphidius, but if by chance her ovipositor comes in contact with the larva of the first, she is unable to distinguish between it and the proper host, and places her egg upon it. Certain observations support this view. For instance, young larvae were never found thus parasitized, possibly because they escaped discovery owing to their small size ; and the mature larva of Lygocerus was never found to be infected. There is very marked increase in the size in this species between the early and late stages, and the latter is of peculiar form with a dorsal conical appendage to the last segment. The full-grown larva and the pupa are capable of active movement, and jerk the abdomen violently when irritated. It is possible that this action warns off the ovipositor of another hyperparasite. I have observed only three instances where Lygocerus was parasitized, and then always by its own species. In two cases, larvae were observed on newly-trans- formed pupae, and here, contrary to the usual rule, the egg must have been placed on the larva when nearly full grown. In the third case, an egg was found upon a younger larva, whose power of movement was not yet developed. The Chaleid larvae, which are sluggish at all stages, are more frequently attacked in their later instars by Ly gocerus and by other Chaleids. The incidence of mortality from epiparasitism is high in the Cynipidae, since they invariably perish within the host when the latter is attacked by an ectoparasite. Exceptionally, a full-grown larva of Charips may be found epiparasitized by a Chalcid or by Lygocerus, and in such eases it is probable that the oviposition of the second hyperparasite coincided with the emergence of the Cynipid from the host, and before it had demolished the remains of the latter.” 1 It should be pointed out that other forms not dealt with here are involved in this bionomical complex. Thus Silvestri (“‘ Contribuzioni alla conoscenza biologica degli Imenotteri Parassiti’’, ‘.Boll. Lab. Scuola Agric. Portici’, vol. iii, 1909) has described the development of a Chalcid, DEVELOPMENT OF CHALCIDS 381 It is clear that this phase of parasitism differs somewhat from ordinary epiparasitism. It has been called * accidental superparasitism ’ by W. D. Pierce, quoted by Fiske (8), but might better be termed ‘ metaparasitism’. Epiparasitism then may be defined as successive infestations of a single host by two or more species, or by several individuals of the same species, of parasite. Metaparasitism is a development of epiparasitism, and may be defined as the direct attack of one epiparasite upon another. Objection may be taken that the distinction is too fine to warrant the coining of a new word in a science already burdened with technical names; but of late years the practice of introducing parasites to control insect pests, in countries or continents where the latter have become troublesome, has been much extended; and, before importing a parasite into a new area, it is of the first importance to ascertain to what extent it is potentially metaparasitic upon other species. Thus, suppose that two forms of primary parasites A and 5 are imported into a certain locality. There will be a shght reduction of their total efficiency, in proportion to the incidence of epiparasitism between them ; but as long as plenty of hosts are available, the loss due to this will be small, and in any case little harm will result, as a pest destroyer will be reared ultimately. But supposing that B is potentially metaparasitic, while A is not, then in course of time, B, since it will always be successful in contest with A, will reduce the latter species, or even supplant it altogether. The mischief will be even greater from an economic standpoint, if B should prove to be less efficient than A in destruction of the host pest. In fact, this is what has actually taken place in Hawan, Eucyrtus aphidivorus, Mayr., which like Charips is an endo- parasite of Aphidius; but as its other bionomical relations are not known, it has not been included in this discussion, and the same applies to other Chalcidae, recorded as reared from Aphides, but many, if not all of which, are probably hyperparasites. However, as Arrow (‘ Entomolo- gist’s Monthly Magazine ’, vol. lvii, September 1921) observed Aphelunis chaonia, WIk., ovipositing in aphides, this form may prove to be a primary parasite. 382 MAUD D. HAVILAND according to the recent investigations of Pemberton and Willard (9). Among the parasites introduced to control the Mediterranean fruit-fly (Ceratitis capitata, Wied.) were two species, Opius humilis, Silvestri, and Diachasma tryoni, Cameron. It has now been shown that epiparasitism is common between Opius and Diachasma, and that im such a case Diachasma is nearly always victorious. Thus Diachasma is gradually suppressmg Opius in Hawaii ; and, as the authors point out, this result is the more deplorable in that Opius is not only equally efficient as a parasite, but is actually more prolific than its rival, and if left to itself would destroy a larger number of fly larvae. The situation has been further complicated by the introduction of a Chaleid, Tetrastichus giffordianus. This form is very prolific ; but, as it is almost always epiparasitic, it is ineffective as a con- trol of the pest, and generally causes the death of the Opius or Diachasma larva when it comes into competition with them. Fiske and Thompson (4) have shown that the larvae of certain Saturniidae are parasitized by the hymenopterons, Ophion, Theronia, and Spilocryptes. All three are primary parasites, but epiparasitism is frequent, and when it occurs, Theronia and Spilocryptes_ respectively overcome Ophion. In competition between Spilocryptes and Theronia, the first generally is the conqueror; but Theronia, it appears, dies of starvation from destruction of its food-supply rather than by direct attack. Timberlake investigated the bionomics of Coccophagus lecanii, Fitch (14), a parasite of Coccus hesperidum, which is more frequently reared as a hyperparasite from another primary parasite, Microterys. According to this observer, Coccophagus is thelyotokous when a primary parasite, producing generations of females only; but when it is reared as a hyperparasite, the resulting imagos are all males—a state of things so far unparalleled. Howard and Fiske (7) in their report on the measures taken to control the. gipsy and brown-tail moths in the United States, record many interesting observations on the bionomics DEVELOPMENT OF CHALCIDS 333 of native and imported parasites. Thus the Chaleid Schedius Kuvanae, How., is primarily an egg parasite of the gipsy moth, but it will also oviposit in Anastatus bifasciatus, Forst., another egg parasite. In this complex, two other species, Tyndarichus novae, How.,and Pachyneuron gifuensis, Ashmead, are hyperparasitic upon Anastatus, but epiparasitism is frequent, and they have been reared not only from Schedius, but also from one another. Monodontomerus aereus, Walk., and Pteromalus egregius, Forst., are also primary parasites of the gipsy moth and brown-tail moth respectively; but both forms are also hyperparasitic through certain Tachnidae, and, addition, the latter form is sometimes reared from other hymenopterons, such as Mesochorus and Apanteles. Smith (12) has shown that Perilampus hyalinus, Say., although strictly speaking an obligative hyperparasite of certain lepidopterous larvae, through their hymenopterous and dipterous parasites, may, when epiparasitism occurs, become metaparasitic. Thus in one instance a cocoon of the Ichneumonid, Limnerium validum, was first parasi- tized by Perilampus, and subsequently by the Pteromalid Dibrachys boucheanus. The latter devoured the Limnerium host, but was shortly afterwards itself destroyed by Perilampus. The following table gives the synonyms used by previous writers on the bionomics of the Hymenoptera parasitica for the terms suggested here. Bera dita iene parasitisin. Parasitism. Superparasitism. Secondary parasitism. Secondary hyperparasitism. Hpiparasitism ( Accidental superparasitism. Metaparasitism Tertiary hyperparasitism. le eae eee eee parasitism. Hyperparasitism a Hyperparasitism. 334 MAUD D. HAVILAND These terms may be illustrated with examples from the aphid complex as follows : Parasitism aphid+ Aphidius Epiparasitism aphid+ A phidius+ rthbes thy s and Metaparasitism aphid+Aphidius+ Asaphes+Lygocerus ( Asaphes or Hyperparasitism aphid+Aphidius+ / Lygocerus or ( Charips The possibility of * hyper-hyperparasites ’ has been suggested by some writers, but although obligative hyperparasitism of the second degree may occur, [ am not aware that it has been definitely proved. The records that seem to point to it are probably due to epiparasitism among hyperparasites. Apart from their economic importance, cases such as those described are of much biological interest, as throwing light on the origin of parasitism in the Hymenoptera parasitica. Thus the epiparasitism of Lygocerus and Asaphes may exceptionally become metaparasitism, if, by chance, one species oviposits directly upon the larva of the other: and a stage further has been reached in Coccophagus and Theronia which are as often hyperparasites as parasites. Viske says of the latter (8) that it 1s so frequently a ‘ super- parasite’ that it is im danger of becoming a hyperparasite. From such forms as these it is not a great step to the obligative hyperparasitism of, for example, Charips. Kpiparasitism is brought about by a high proportion of parasites to the host population. Fiske (8) has made an ingenious calculation, showing that as the incidence of para- sitism rises, the chances of epiparasitism rise likewise. Thus, given a hundred hosts, by the time that the parasite has laid ten eggs, there is an even chance that one will have been placed in a host already infected, and so on, until with fifty eggs the odds are even that no less than ten ovipositions will have been duplicated in this way. But although hyperparasitism may have arisen from epiparasitism, through metaparasitism, primary parasitism cannot be accounted for thus. eee ee DEVELOPMENT OF CHALCIDS 335 Wheeler (15) has put forward a theory of the origin of parasitism in the Aculeata. He supposes that parasitism arose within the species, when certain individuals acquired the habit of laying their eggs in the brood cells of their neighbours, instead of working for themselves ; and he supports his sugges- tion by the significant fact that the existing parasites are frequently generically allied to the host species. But this theory can hardly be extended to the Parasitica, even if we regard them as a heterogeneous group, derived from different ancestral stocks, and classified together in virtue of characters acquired independently by members of different families in adaptation to parasitic life. The existing Parasitica are a vast class, of imfinite variety of size, structure, and habit ; and with the exception of most of the Cynipids and a few Chaleids, which are gall-formers on plants, all are parasitic upon insects, frequently upon families distantly related to them. To suppose that the parasitic habit arose spontaneously in a common ancestor, and was perpetuated by natural selec- tion, involves the assumption of a considerable initial mutation. If, as among the bees and wasps, we found that phylogenetic relationship between host and parasite was the usual rule, we might suppose that parasitism arose within the species in the Parasitica, as Wheeler suggests for the Aculeata ; but there is as much to be said against as for this view, since the modern Parasitica include, not only their own allies, but almost every stage of almost every family of insects among their hosts. Nevertheless, parasitism must have had a beginning, and the suggestion may be put forward that the parasitic habit arose among these Hymenoptera from the inquiline habit. In other words, the proto-Parasitica were phytophagous, and oviposited on plants. A further stage was reached, when, for the better protection of the eggs, they resorted to the shelter of galls and other deformities produced by members of their own tribe, and by other insects. Here they became established as commensals or inquilines, and from the inquiline habit to the parasitic habit is possibly not a great step. The Chalcid, 336 MAUD D. HAVILAND Torymus propinqguis, previously alluded to, is now an ectoparasitic of a gall-forming Cecidomyid of the nettle, but if this view is correct, its ancestors inhabited this, or a similar gall, as inquilines, and later acquired the habit of devouring the maker of the growth that harboured them. The intra-specific origin of parasitism in bees may find a parallel among inquilines, for it is quite conceivable that certain individuals may have adopted the habit of ovipositing in a ready-formed gall, and thus became inquilines to their own species. Cameron (2) remarks that among the Cynipidae, the known inquilines are species of Synergus, Ceroptres, or Sapholytus, which are all forms nearly related to the true gall-formers. The view that parasitism is derived from inquilinism would account for the diversity of the hosts of the Parasitica. Galls, and similar plant deformities, are caused by insects of other groups, such as many Hemiptera, Diptera, and Lepidoptera. ‘The ancestors of the Parasitica may have used these as well as the galls produced by members of their own family, and later become parasitic upon the insects which formed them. It will be very desirable in future to investigate fully the bionomics of the forms reared from, for example, Cynipid galls. If any, generally found to be inquiline, are proved on occasion to devour the maker of the gall, it will support the suggestion that the Parasitica are descended from inquiline ancestors. SUMMARY. 1. Asaphes vulgaris, Wlk., Pachycrepis clavata, Wlk., and Pachyneuron, sp., are hyperparasites of aphides through the larvae of certain Braconidae (A phidius). 2. Oviposition took place after mating for Asaphes and Pachycrepis, and parthenogenetically for Pachyneuron. 3. The eggs are deposited upon the body of the host when the latter is fully fed and about to undergo metamorphosis within the skin of the aphide. 4. The larvae feed ectoparasitically upon the host, which soon becomes a decomposing mass. ——E———————————— ae DEVELOPMENT OF CHALCIDS Bol 5. The newly-hatched larvae are maggot-shaped forms, with four pairs of open spiracles and two cephalic papillae. 6. In the later stages small tubercles are developed on the prothorax and succeeding segments, and there are ten pairs of functional spiracles. 7. The total period of development is a little over three weeks, and at least two broods may occur in the summer. 8. The bionomics of aphides and of their parasites and hyper- parasites are discussed. 9. The term ‘ epiparasitism ’ is proposed instead of * super- parasitism ’ which has been used by other writers, and it 1s suggested that it should be restricted to cases where two or more species, or two or more individuals of the same species, independently attack the same host. 10. The term ‘ metaparasitism ’ is suggested for cases where one parasite or hyperparasite in epiparasitism, becomes secondarily hyperparasitic upon the other. 11. Instances are given of the occurrence of epiparasitism and metaparasitism among other hymenopterous parasites. 12. The origin of parasitism in the Hymenoptera parasitica is discussed, and it is suggested that it arose from an earlier inquiline mode of life. BIBLIOGRAPHY. —_ . Ashmead, W. H. (1904).—‘‘ The Classification of the Chalcid Flies ”’, “Mem. Carnegie Museum ’, vol. 1, mem. 4. 2. Cameron, Peter (1893)—‘ British Phytophagous Hymenoptera ”’, “Ray Society Publications ’, vol. 4. . Fiske, W. F. (1910)—\‘‘ Superparasitism: an important factor in the natural control of insects’, ‘ Journ. Econ. Entom. Concord ’, vol, 3. 4, Fiske, W. F. (1910) and W. R. Thompson (1909)—\*‘* Notes on the Parasites of the Saturniidae ”’, ibid., vol. 2. 5. Haviland, Maud D, (1920)—‘‘ On the Bionomics and Development of Lygocerus testaceimanus, Kieff., and lLygocerus cameroni, Kieff., parasites of Aphidius’”’, * Quart. Journ. Micr, Sci.’, vol. 65, pt. i. ls (1921)—* Preliminary Note on a Cynipid hyperparasite of Aphides ”’, ‘ Proc. Camb. Phil. Soc.’, vol. xx, no. 2. [ee] a 598 fis 12. 14, 15 MAUD D. HAVILAND Howard, L. O., and Fiske, W. F. (1911)—‘‘ The Importation in the United States of the Parasites of the Gipsy Moth and Brown-tail Moth ”, ‘ U.S. Dep. Agric. Entom. Bureau’, Bull. 91. Imms, A. D. (1918)—‘‘ On Chalcid parasites of Lecanium capreae’’, ‘ Quart. Journ. Micr. Sci.’, vol. 63, no. 251. . Pemberton, C. E., and Willard, H. F. (1918).—‘‘ Inter-relations of Fruit-fly Parasites in Hawaii”’, ‘ Journ. Agric. Research, Washing- ton’, vol. xii, no. 5. . Seurat, L. G. (1899)—‘‘ Contributions 4 l'étude des Hyménoptéres entomophages ”’, ‘ Ann. Sci. Nat.’, 8™¢ série, t. 10. . Sharp, David (1899).—‘ Camb. Nat. Hist.’, “‘ Insects ’’, pt. 1. Smith, Harry 8S. (1912)—‘‘The Chalcidoid genus Perilampus, and its relations to the problem of parasite introduction”, * U.S. Dep. Agric. Bur. Entom.’, Tech. Series, no. 19, pt. iv. . Smits van Burgst, C. A. L. (1921)—‘‘ Hyperparasitism noticed in Primary Parasites of the Pine Caterpillar, Pinolis flammea”, ‘ Tijdschr. Plantenziekten Wageningen’, vol. xxvii, no. 4, Sum- mary in English, R.A.E., series A, vol. ix, pt. 7, 1921. Timberlake, P. H. (1913)—‘‘ Preliminary Report on the Parasites of Coccus hesperidum in California ’’, “ Journ. Econ, Entom, Concord ’, no. 3. Wheeler, William Moreton (1919)—*“‘ The Parasitic Aculeata: a study in Evolution ’’, ‘ Proc. Amer. Phil, Soc.’, vol. lvii. Animal Chlorophyll: its Relation to Haemo- globin and to other Animal Pigments. (Contribution from the Bermuda Biological Station for Research, No. 132). By John F. Fulton, Jr., Magdalen College, Oxford, CoNTENTS. PART I. THE PIGMENTS OF ANIMALS HAVING NO BLOOD-VASCULAR SYSTEM. PAGE 1. INTRODUCTION ; é ; ‘ : ‘ . 5 atk) 2. PROTOZOA AND eee ‘ ; : 3 5 : . odAl 3. COELENTERATA : é ; : é : . 344 (a) Condylactis < ypentileine : . : : : . 345 (b) Actinia bermudensis . : 5 ° ; . 347 4, PLATYHELMINTHES ‘ ; ; : : j ; made 5. ECHINODERMATA . : : : é ‘ ‘ a) 2D) (va) Tripneustes ele mertee . é : : . 85d (b) Other Echinodermata. j : , ; é ODS PART II. THE PIGMENTS OF ANIMALS WHICH HAVE A BLOOD-VASCULAR SYSTEM. 1. INTRODUCTION ; : - ; , F : ; . 360 2. NEMERTINA ; : F : : : ‘ ; —. stay 3. Momuusca . ; ; ‘ : : : ool (a) The iacinonenehe ‘ : ; : F : . 361 (b) The Cephalopods . 4 : é ‘ : : . 862 (c) Other Mollusca ; : , : : : ¢ . 363 4, ANNELIDA . F ; : ; : : . é . 364 5, ARTHROPODA ‘ : ; 3 : : ‘ : . 368 (a) Crustacea. : : , : , : : - o0S (b) Insecta é : : : : ‘ ; : . 374 6. TUNICATA . 5 : i j : F ‘ on (a) Ascidia ie ‘ ; é F : : 2 : “Sd 16 (b) Other tunicates ; ; 5 : - ; . 9379 (c) Discussion. : ‘ A . o19 dle 12 ca REO ees OF veneers é ; ; . 3880 8. CONCLUSIONS . F 4 : ; 5 ‘ 3 5 | Ber 9, BIBLIOGRAPHY 3 : ? ; : ‘ , ‘ . 385 340 JOHN F. FULTON, JR. PART I. THE PIGMENTS OF ANIMALS HAVING NO BLOOD-VASCULAR SYSTEM. 1. INTRODUCTION. In the study of marine invertebrates one of the most impres- sive things encountered is the great richness and variety of colour; it is not surprising, therefore, that the question of animal coloration has long engaged great attention. Investi- gated at first superficially by those who sought an explanation of the so-called phenomenon of ‘ protective coloration ’, the problem attracted, during the latter part of the nineteenth century, the attention of several English physiologists, and it is to the investigators of this group—Lankester, Sorby, Mae- Munn, Mosley, Griffiths, Poulton, and Halliburton are the more important names—that we are indebted for very real contribu- tions to our knowledge of animal pigments ; especially for the introduction of the microspectroscope into this field of biological research. Since 1900 the work on pigmentation has been to a very large extent spasmodic, and while certain valuable additions have been made (in particular, those of Gamble and Keeble) there are still many problems which invite further investigation. The present paper aims to show that the pigment which is responsible for the colour of certain representative inverte- brates comes from the blood-stream, and that in many cases the pigment cells of the blood arise (while in circulation) from unpigmented corpuscles. This view concerning the origin of pigment occurred to the writer after noting that a pigment of the blood appeared to be identical with the body pigment in three representative phyla: (1) in another paper the writer (Fulton, 1921 b) has shown that the pigmented corpuscles in the blood of a Bermuda tunicate, Ascidia atra, arise in the blood from colourless cells, and that the blue pigment cell, so common in the blood-stream, is identical with the blue eells of the tunic—the cells which give to this ascidian its intense blue colour ; (2) Crozier (1916 b) has demonstrated that the blue ANIMAL CHLOROPHYLL 341 pigment granules which colour the nudibranch Chromodoris zebra are also to be found in great abundance in the pigment cells of the blood—the identity of the two pigments having been determined by the spectroscope; (3) lastly, while ex- amining the body-fluid of one of the common sea-urchins, Tripneustes esculentus, it was found that the large red amoeboid cells (so well deseribed by Geddes, 1880) gave strong indication that their pigment is the same as that which colours their spines and tube-feet. Thus, in a tunicate, molluse, and an echinoderm there seemed to be very good evidence that the body-pigment is found also in the blood. The experimental work was carried on at the Bermuda Biological Station for Research during the summer of 1920, and I wish to express my warmest thanks to Dr. E. L. Mark, who gave me the facilities of the laboratory, and who made possible my trip to Bermuda. 2. PRotTozoA AND PORIFERA. With any effort to trace animal pigment back to the blood- system there arises at once a very serious difficulty. Pigmenta- tion as such appears phylogenetically long before the rise of the blood-system. How, then, is it possible to assume that all coloration comes from the blood ? Among the Protozoa the only class—with rare exceptions— which contains chromatophores is the Mastigophora (Minchin, 1912, p. 13). Here, however, the pigment is probably in every case chlorophyll, or a substance closely allied to chlorophyll. A typical species of chlorophyllogenous protozoa is Archerina boltoni, which was described by Lankester (1885). One notable exception to the assumption that all protozoan pig- ments are closely related to chlorophyll is found in Stentor cwruleus, which possesses a blue pigment called by Lan- kester (1873) ‘stentorin’. Spectroscopically the absorption bands of this pigment, quite unlike chlorophyll, resemble those of the blue algal pigment, phycocyanin, which, according to Phillips (1911, p. 596), when present even in minute amounts, greatly alters the spectrum of chlorophyll. Consequently 342 JOHN F. FULTON, JR. stentorin is not to be considered an important exception to the rule that all protozoan pigments are chlorophyllogenous in nature. But how does the chlorophyll of protozoans originate ? Is the animal itself capable of manufacturing chlorophyll, or is it the result of outside infection? Geddes (1882) and Lankester (1882 a, and 1882) maintained strenuously that Hydra viridis and Spongilla fluviatilis were capable of synthesizing their own chlorophyll :* that the green deposits found in these animals are chloroplasts belonging to the animal and consequently are not of plant origi. In support of his contention Lankester asserted the absence of nucleus and cellulose wall in the green corpuscles. Though no histological evidence has been adduced to show the presence of a nucleus in these bodies, it seems fair to conclude, since Beyerinck (1890, note 1, p. 784) has succeeded in obtaining cultures of algae from the green corpuscles of Hydra viridis,” that the chlorophyll of Hydra is algal in nature and due to an infection from the outside. The algae probably represent a phase in the life-history of Chlorella viridis. A similar condition undoubtedly holds for most of the green protozoa: Famintzin (1889, 1891), Dantee (1892), and Dan- geard (1900) all report having obtained colonies of algae from the macerated tissues of Stentor, Paramoecium, and Fron- tonia ; Schewiakoff (1891) found that if colourless Frontonia are fed upon macerated green specimens they become infected with green algae which subsequently divide within the cell. Similar results have been reported for Paramoecium (Dantee). Certain contrary evidence is also on record for the Protozoa— the puzzling cases of Vorticella campanula (Engle- mann, 1883), and of Pelomyxa viridis (Bourne, 1891) 1 For a more complete discussion of this question see Gamble and Keeble (1903), Keeble and Gamble (1907), and Fulton (1921 @). 2 Entz (1881 and 1883) reports similar results, but in his experiments little precaution was taken against infection and consequently the results are to be accepted with caution. Recently Goetch (1921) has succeeded in infecting colorless Hydra with Chlorella, and in so doing has corroborated in a very substantial way the views of the earlier investigators cited above, ANIMAL CHLOROPHYLL 343 are the most significant examples—which favours Lankester’s hypothesis of the intrinsic nature of certain of the animal chlorophylls. Nevertheless the balance of evidence seems to favour the algal theory to account for chlorophyll in animals ; and it is probable, moreover, that further vestigation with a more refined technique will show conclusively that Lankester was wrong. Particularly does this seem probable in the light of the classic researches of Gamble and Keeble (which will be dis- cussed later) on Convoluta, the green cells of which were shown to be intruding algae. In short, the Protozoa do not present any real difficulty to assuming that body-pigments arise in the blood. The sponges possess nothing in the nature of a blood-system : nutrition and respiration being accomplished by water currents within the body. What, then, is the nature of their coloration ? Is it an animal pigment, which has arisen independently of the blood-system ; or is it, as in the Protozoa, a chlorophyllous substance ? Though it has not been demonstrated that every species of sponge contains chlorophyll, the spectroscopic investigations of Sorby (1875 a), Krukenberg (1884), and MaeMunn (1888) have established the presence of a chloro- phyllous pigment in eighteen species of sponge, it being most common in the genera Halichondria and Halina. The more highly-coloured sponges possess pigments the absorption bands of which (as indicated by the figures of MacMunn and Kruken- berg) resemble in many respects those of certain of the pigments from blue and red algae, the pigment spectra of these sponges is, in addition, quite different from the spectra of chlorophyll. Inasmuch as it has been shown that the presence of small — quantities of such algal pigments as phycocyanin, phycophaein, and phycoerythrin (Phillips, 1911; also Willstitter and Stoll, 1918) greatly alter the spectrum of chlorophyll, the fact that the pigment of certain coloured sponges fails to show the bands characteristic of the plant pigment does not necessarily indicate its absence. As in the Protozoa, one is also confronted in the Porifera with a question concerning the nature of the pigment itsel{—a point which is as yet unsettled. Lankester (1882) and his school were vigorous in upholding the animal origin NO. 262 Bb 344 JOHN F. FULTON, JR. of sponge chlorophyll, while Brandt (1881 a, 1881 b, 1882, and 1883) supported the view that chlorophyll in the Porifera and other animals results from a symbiotic association with green algae.! Zooxanthellae—symbiotic holophytic flagellates—have been reported for several sponges, also recently by Kirkpatrick (1912) for Merlia normani.? Cotte (1904) likewise gives an account of an interesting association of this sort. There is no doubt but that Lankester had every reason to question the evidence of Brandt—which in the light of later investigations was most decidedly inconclusive—and he ‘ has done valuable service by his championship of the opposed view, that of the intrinsic nature of the corpuscles under discussion. For his view compels those who hold the “ algal ”’ theory to investigate each case separately and to vindicate their view by the synthesis of the green animal’ (Keeble and Gamble, 1907, p. 171). Now, however, there is little question but that true chlorophyll in animals owes its existence in every case to plants. It seems evident, therefore, that the pigmenta- tion of most sponges has resulted from an association, symbiotic or otherwise, with plant cells ; and that, as with the Protozoa, the Porifera present no serious obstacle to the assumption that animal colour arises in the blood. Consequently a diseus- sion of the phylogenetic aspect of pigmentation must of necessity commence with the coelenterates. 3. COELENTERATA. Since the coelenterates are organisms having but two cell layers, ectoderm and entoderm, it is at once obvious that they 1 Since the present writing the work of Van Trigt (1918) has been brought to the writer’s attention. He has shown that in Spongilla the green cells very clearly are invading organisms, and has made an extensive series of experiments with cultures of the green cells derived from the macerated sponge tissue. He has also given conclusive proof of an oxygen-carbon dioxide exchange between the algal cells and the sponge tissue. His evidence further corroborates the view just expressed concerning the symbiotic nature of the green cells in sponges. 2 Winter (1907) has shown that Zooxanthellae are symbiotic in the foraminifer Peneroplis. ANIMAL CHLOROPHYLL 345 can possess no blood-system in the sense in which it is used for the higher animals. Nevertheless, as Griffiths (1892, pp. 128 and 184) has emphasized, the nutritive or ‘ chylaqueous ’ fluid is analogous to the blood of the higher forms in that it carries nourishment, supplies oxygen, carries off the waste products of metabolism, and in many cases, as Kollmann (1908) and others have since shown, is a corpusculate fluid. The impor- tance of this analogy between the chylaqueous fluid and the blood of higher animals is uncertain; however, a question immediately presents itself concerning the relation of this fluid to the pigmentation of the coelenterates. It is therefore desirable to consider first the nature of coelenterate pigments. The animals on which the greater share of my work has been done are two species of actinians common in the Bermuda Islands: Condylactis passiflora Duch. and Mich.," and Actinia bermudensis Verill. (a) Condylactis passiflora. Condylactis occurs in great abundance in all parts of the Bermudas, and is usually found firmly attached to the under side of rocks and in crevices just below the level of low tide. If the gastrovascular (chylaqueous) fluid of Condylactis be withdrawn at any point on the body with a hypodermic syringe and examined, two types of cell are usually to be observed: one a yellow cell with several large granules, and the other, an unpigmented element. The pigment cells of this body-fluid might easily be confused with wandering pigment cells of the body-wall. In reality, however (as Rand, 1909, has noted) these cells are Zooxanthellae, and it is this fact which in part explains a very striking phenomenon pre- sented by a fresh smear of the gastrovascular fluid: viz. the very marked oscillation of the mdividual cells. These yellow cells (Zooxanthellae) gyrate usually in a counter-clockwise direction on a single axis, while the colourless cells as a rule 1 For an excellent description of this species MeMurrich’s (1889) paper should be consulted. Bb2 346 JOHN F. FULTON, JR. vibrate much less regularly. The latter, inasmuch as they are motile, have in all probability been torn from the ciliated lining of the gastrovascular cavity ; with acetic acid or neutral red their cilia may very easily be demonstrated. The coloration of Condylactis, it should be emphasized, is due largely, but not entirely, to this yellow flagellate. The tissue of the tentacles is itself colourless (Rand, 1909 ; Parker, 1917), as is shown when a tentacle is transected. ‘The brownish- yellow colour, which is a constant feature of the uninjured tentacle, is therefore due to the presence of the flagellate organ- ism in the internal fluid. This may readily be shown by examining the liquid contents of the tentacle. When an animal is withdrawn from the aquarium with the tentacles in expanded condition the internal pressure on the gastrovascular fluid causes minute streams of water to issue from the terminal pore of each tentacle. If some of the fluid so exuded be caught in a watch-glass it occasionally contains the * symbiotic’ organisms ; under normal conditions of exudation, however, they probably do not escape when the tentacle contracts. Not only are the tentacles coloured by the presence of Zooxanthellae, but the column itself owes much of its colour to this organism. However, the column also possesses large collections of red pigment granules, some patches being as much as 2mm. in diameter. These are more highly con- centrated in the lower parts of the column than in the upper, which gives to the basal region an intense red colour, while the upper parts tend toward the brownish yellow of the tentacles. In two Siphonophora, Velella spirans and Por- pita umbella, Kuskop (1921) has found Zooxanthellae in abundance; they reside chiefly in the ‘hepatic canals’, so called, and in the gonophores. Their occurrence in the latter organs strongly suggests that the association of the Zooxan- thellae with these coelenterates is continuous from one genera- tion to another. Before discussing the significance of these observations the condition of the pigment in Actinia ber- mudensis will also be described, ANIMAL CHLOROPHYLL 347 (b) Actinia bermudensis. Actinia bermudensis is a deep red anemone, which is found on the rocks just about the level of low tide. At high tide, when the water splashes over them, their tentacles open up for feeding ; when out of contact with the water they draw their tentacles into the interior of the column and have the appearance of a deep-red gelatinous mass hanging limply from the rocks. ‘The specimens used in the present study were obtained from the caves on the north side of Long Island (Bermuda) where they occur in considerable numbers. This Species is distinguished by a very remarkable power of resisting unfavourable surroundings; as an example, it will remain alive sealed in a 100 ce. of sea-water for from six to seven days (Fulton, 1921 a). A. bermudensis is coloured uniformly by red pigment granules, which are spread through the entire ectoderm. The granules are not of a definite size, however, and the out- lines of the cells which contain them are never clear in the living tissue and can be discerned only with great difficulty in tissue which has been fixed. In shade, the pigment is precisely the same as the red pigment patches of Condylactis. Con- sequently a series of experiments was performed with a view to determining whether or not the two pigments are identical. Small pieces of tissue from each species were teased out and placed side by side upon a slide. ‘I'heir action in the presence of an acid was first tested. In both cases when either 10/N hydrochloric acid or 10/N valeric acid? were added a decided increase in the depth of colour took place. When treated with alkahs (NH,OH and NaOH, 10/N and 50/N) no change could be observed. Neither of the pigments could be dissolved with any of the following solvents : ether, chloroform, methy], ethyl or amyl alcohol, petroleum ether, xylol, or pyridine. In acetone, however, the pigment of A. bermudensis 1 Tt has been shown by Crozier (1915, 1916 a, and 1916c) that, of twenty-two of the more common acids, valeric is the most penetrating to tissues. 348 JOHN F. FULTON, JR. proved readily soluble; the condylactid pigment was also dissolved by acetone, but not so readily. The difference in rate of dissolution is probably to be accounted for by the sreater thickness of the Condylactis tissue. Therefore, since the two pigments are found in species of the same order, since they are of the same colour, since in the presence of acid they are uniformly deepened in shade, and since they are each to be extracted by only one (the same one) of nine solvents, it seems reasonable to conclude that the two pigments are identical. Among the earlier authors the observations of Moseley (1873) upon actinian pigments are perhaps the most significant. He described a red colouring matter (from Actinia mesem- bryanthemum and from Bunodes crassicornis) which he called ‘ actiniochrome’; MacMunn found this pig- ment insoluble in many of the ordinary solvents for animal pigments.’ All of the solvents which he employed also gave negative results when applied to A. bermudensis and C. passiflora. This suggests that the pigment of the Bermuda actinians is probably identical with Moseley’s actinio- chrome, but only a spectroscopic examination can determine this with certainty. The two species worked on by Moseley were subsequently investigated by MacMunn (1885 c) with the result the Moseley’s actiniochrome was identified, but, in addition, a haematin-yieldmg pigment was isolated and given the same ‘ actiniohaematin ’.* This pigment, which, it should be emphasized, is closely related to haemoglobin, performs a respiratory function, being capable of existing in a state both of oxidation and reduction. On finding these two pigments together m one animal MacMunn drew the conclusion that one is a respiratory substance (actiniohaematin), and that the other (actiniochrome) is purely for ornament. The most notable contribution of MacMunn, however, was his observa- tion concerning the relation of Zooxanthellae (‘ yellow cells ’) 1 MacMunn (1885 c¢, p. 643), alcohol, ether, chloroform, and carbon bisulphide. 2 This pigment, on treatment with a metallic hydroxide and sodium sulphide, gives haemochromogen which on oxidation gives haematin. s ee ANIMAL CHLOROPHYLL 349 to the respiratory pigment. He found that in actinians which were not infected with Zooxanthellae a respiratory pigment is present, but that in forms which are ‘ packed with “ yellow cells ’’’ the pigment had ceased to perform a respiratory func- tion. The most striking case which he observed was that of Bunodes balli, in which occurs a facultative association between itself and the Zooxanthellae : the larger variety of that species has many Zooxanthellae and, as a result, is almost without trace of any of the pigments common among other forms; the smaller variety of the same species is uninfected, and as a result possesses a respiratory pigment. The results of MaeMunn’s work are briefly as follows: (1) that a respiratory pigment is present in most actinians ; (2) the pigment is not a carrier of oxygen, but serves simply to store oxygen in the tissue which is subsequently to use it; (8) in those actinians in which yellow cells are present the chlorophyllous pigment of these organisms seems to replace the respiratory pigment ; (4) besides this respiratory substance there are other pigments (such as Moseley’s actiniochrome) which serve for decoration." In a recent paper on Actinia equina and Anemonia suleata Ehnhirst and Sharpe (1920) record several observa- tions which are not in agreement with those of MacMunn. They find that the non-haematin pigment, instead of being purely ornamental, produces oxygen, possibly by photo- synthesis. However, they report the presence of Zooxan- thellae and fail also to find any haematin derivative, which accords with the results of MacMunn. In addition, these authors hold that the intensity of colour in A. equina varies with exposure to light, and the pigment, therefore, functions as a light sereen. Certain observations made in the course of the present study support this latter conclusion ; when A. bermudensis is kept in the dark (three days) the animal loses its deep red tinge and acquires a brownish-red shade. Conversely, if an individual is exposed to direct sunlight, its colour changes to a brilliant carmine. 1 Quoted also by Griffiths (1892) ; an excellent description of MacMunn’s work will be found in Griffiths’ book, especially in Chapter VIII. 350 JOHN F. FULTON, JR. Having considered the physical and chemical aspects of the actinian pigments, what deductions can be drawn as to their origin ? It is clear, inasmuch as there is no internal circulatory system between the ectoderm and entoderm, that the pigment must be manufactured from substances of the outside world which come actually into contact with the imdividual cells. Likewise it is evident that these substances are absorbed from within, carried, that is, in the chylaqueous fluid, since the ectoderm of an anemone serves for protection rather than for absorption. It is highly improbable that the pigment comes directly as food (as Crozier (1917) holds for a species of polyclad and Poulton (1893) for certain insects); if that were true, it should exist in solution in the gastrovascular fluid; but this seems definitely not to be the case (MacMunn). The more reasonable hypothesis, it seems to me, is that the cells con- taining the colour themselves synthesize the pigment from certain food substances. ‘This means that in the absence of a blood-system each actinian cell has to elaborate its own pigment. Little is known concerning the nature of the food from which the cells manufacture pigment. Various speculations, however, have been made on this point, particularly in the case of the insects. In this class of animals, as Poulton (1893) has shown, the pigment of the body appears to be a modified chlorophyll. That such a condition should obtain among the actimians seems at first impossible. Many actinians live in an obligate associa- tion with Zooxanthellae, an association in which the anemone is probably parasitic upon chlorophyllous cells (Fulton, 1921 a); that is, im times of starvation they turn upon the cells from which they possibly receive nourishment (by photo- synthesis) and engulf them. Also in many actinians the * yellow cells’ are lodged directly in the tissue of the ectoderm and entoderm (Hertwig, 1883). It is a matter of common knowledge, too, that actinians feed upon pelagic forms which contain chlorophyll. From these facts it is evident that actinians use chlorophyll as food. As has already (p. 348) ll Ee ANIMAL CHLOROPHYLL 351 been stated MacMunn (1885 c) demonstrated in many actinians the presence of a haematin-yieldmg pigment, which was designated as actiniohaematin. Recent investigations have shown that haemoglobin and chlorophyll are similar chemically, each having as a base a substance known as a porphyrin. This derivative is composed of four pyrrol groups in complex linkage. The exact similarity which exists between haemo- globin and chlorophyll may be shown by the following : * Haemoglobin Chlorophy!! a (by K,CO,) (by alkali) | | \ . 4 Haematin (+ proteid) phyllins (+ phytol) (by acid) (by acid) | | | | it: Leer a (Haemato) porphyrin (+ Fe) (phyllo) porphyrin (+ Mg) (by soda-lime) (by soda-lime) \ \ \ v *‘aetioporphyrin ° Thus haemoglobin, by the loss of a proteid (globin) and its iron, forms a porphyrin; in the same way chlorophyll a, by the loss of phytol and its magnesium, forms several porphyrins, one of which (phyllo) is spectroscopically and chemically very closely related to haematoporphyrin; both phyllo- porphyrin and haematoporphyrin by the action of soda-lime give the same substance, aetioporphyrin. From this it seems evident that there is no a priori reason for assuming that the tissues of actinians could not convert chlorophyll into such a substance as actiniohaematin, which is closely related to haemoglobin. The obvious objection is that exchange of metals (iron for magnesium) would make such a transformation impossible. But one should recall that it is a far more simple 1 The best discussion in English of the chemistry of haematoporphyrin and chlorophyll is that of Plimmer (1915); Bayliss (1918) is good, but the standard work is that of Willstatter and Stoll (1913). 852 JOHN F, FULTON, JR. process to drop off an atom of magnesium and add one of iron than it is to build up an enormously complex molecule such as the porphyrins present. If such a view 1s capable of experimental proof it will have an important bearmg upon the phylogenetic origin of animal pigments ; it will give fair indication that many animal pig- ments were derived originally from plant chlorophyll as the result of some symbiotic association (perhaps for the purpose of facilitating respiration) of an animal with a chlorophyll- bearing organism—a condition probably similar to that found to-day among the sponges and certain of the Protozoa. On the basis of this theory it is interesting to speculate concerning the origin of haemoglobin in the higher animals. Is it not possible, for instance, that our blood-pigment is derived from the chlorophyllous substances which are taken in’as food, a condition not unlike that which the writer believes to exist in the coelenterates? The recent feeding experiments of Burgi and his co-workers (1919) indicate that such is the ease, for they give strong indication that the animal body is depen- dent upon chlorophyll for the bwlding of haemoglobin ; of three sets of anaemic rabbits, one was fed alone upon a chloro- phyll diet, the other upon iron pills, and the last group upon a mixture of iron pills and greenstuffs. ‘The anaemic condition of the first two groups was very slow to improve, whereas the animals in the last group within a short time lost all symptoms of anaemia and the haemoglobin content of their blood came back to normal. This means that chlorophyll with its four pyrrol groups is quite as necessary for the manu- facture of haemoglobin as elemental iron. This conclusion is further substantiated by Grigoriew (1919), who has repeated Burgi’s feeding experiments with positive results. If this be true, methods can very well be devised to control the formation of haemoglobin in disease. What, then, must be the conclusion as to the origin of pigment in the coelenterates ? In the first place the chyl- aqueous fluid, which in function at least is the analogue of the blood of higher animals, carries to the tissue the components ANIMAL CHLOROPHYLL 353 from which it elaborates its pigment; the components, in addition, are probably highly-organized substances. ‘T'he gastrovascular fluid derives these pigment-making substances from, the chlorophylls which enter as food. The pigment, therefore, in addition to bearing a close relationship to haemo- globin, is probably itself derived from chlorophyll. This applies to MacMunn’s actiniohaematin ; and it will also be recalled that Elmhirst and Sharpe (1920) have shown that the non-haematin pigment of A. equina releases oxygen as a result of photosynthesis, which likewise suggests .an intimate relation to chlorophyll. 4, PLATYHELMINTHES: Huxley (1877, p. 57), writing of the digestive cavity in the Coelenterata, remarked that the ‘ fluid which it contains represents blood’. Concerning the next higher group he states : ‘Tn the Turbellaria, Trematoda, and Cestoidea, the lacunae of the mesoderm and the interstitial fluid of its tissues are the only representatives of a blood-vascular system.’ The observa- tion of Huxley is interesting, but it must be recalled that the mesodermal lacunae represent merely the morphological homologue of the blood-system ;* the functional precursor of the vascular system, as in the coelenterates, is to be found in the gastrovascular cavity. Pigmentation is common among the flat worms; many of the marine polyclads, in particular, are distinguished by a brilliant coloration. The only investigations concerning the pigment of the animals belonging to this class with which the writer is acquainted are those of Gamble and Keeble, and Crozier. The latter author (Crozier, 1917) has shown that the polyclads 1 Though the supply of pigment-forming substances is undoubtedly given by the gastrovascular system, it is interesting to note that among the Rhadocoele Turbellaria the parenchyma (in which the lacunae are formed) is the seat of the body-pigment (Parker and Haswell, vol. i, p. 265). This seems to be the first instance in which the function of provid- ing pigment has been taken over by the morphological fundament of the future blood-system. 354 JOHN F. FULTON, JR. commensal with the orange colonies of Ecteinascidia turbinata and the purple colonies of Rhodozona picta are themselves orange and purple, respectively, and of a shade very similar to that of the animal with which they are com- mensal. On starvation (i.e. when removed from the colony of tunicates) these polyclads lose their colour, but when allowed to feed again with the tunicate colonies they regain their colour in a very short time. This, Crozier believes, is an example of a pigment which is formed directly from food, and it accounts for the colour being the same as that of the animal with which the polyclad is commensal. The writer has made certain other observations which in part support Crozier’s conclusion. The colouring matter in the tunic of Eetein- ascidia turbinata is made up of stellate orange chromato- phores. Now, if the body-cavity in one of the polyclads recently taken from a colony of Ecteinascidia be observed under the microscope, not infrequently small pieces of orange pigment can be observed, many of which show clearly that they are portions of the chromatophores from the tunic. Owing to the great frailty of the polyclads, the fate of these small pieces of pigment could not be followed completely ; as a result it was impossible to settle definitely whether the pigment was ingested bodily by the entodermal lining as Congo red is ingested by the young of Convoluta roscoffensis (Gamble and Keeble, 1903), or first went into solution. If the former assumption were true, it might be possible to raise a race of colourless imdividuals of this species (as Poulton, 1893, has done for certain of his insect larvae) by preventing their association with Ecteinascidia. The writer, however, is in- clined toward the belief that the fragmentary pieces of chroma- tophore go into solution in the water-vascular system and are subsequently taken up by the cells which need them. The latter explanation, if correct, would accord with the fact that the pigment does not extend promiscuously over the body, but is found in definite and regular designs. Among some Turbellarians (Convoluta and Vortex) the green or yellow colour is occasioned by the presence of symbiotic algae. ANIMAL CHLOROPHYLL 355 5. EcHINODERMATA. It is generally accepted that the first phylogenetic appear- ance of a vascular system is to be found among the echinoderms. The animals of this phylum are provided with distinct organs of circulation consisting of two radiating canal systems (haemal and perihaemal), the most important of which arises from a ring surrounding the oral end of the digestive tube. Although these vessels always contain a corpusculate fluid, it is certain that they are incapable of either peristalsis or of any other contractile manifestation. Also the sinus which accompanies the madreporic canal, while usually looked upon as a rudimentary heart, certamly performs no pumping func- tion. Consequently serious doubt has arisen as to the correct- ness of looking upon the haemal and perihaemal systems of the echinoderms as true blood-vascular systems. The corpusculate fluid of the haemal and perihaemal vessels is likewise found throughout the entire peritoneal cavity. ‘The corpuscles are nucleated cells, which exhibit amoeboid movements ; and the fluid so obviously represents the blood of higher animals, that I know not why the preposterous name of “ chylaqueous fluid” should have been invented for that which is in no sense “ chyle”’, though, like the other fluids of the living body, it contains a good deal of water’ (Huxley, p. 480). (a) The Sea-urchin Tripneustes esculentus Leske. The red pigment cells are the most noticeable constituent of the body-fluid of this animal. With them are found non- pigmented cells, vibratile cells which are supposed to facilitate circulation, and, less frequently, yellow cells which are not unlike Zooxanthellae.t. The red cells are closely packed with small granules; when protruding its pseudopodia, the cell first sends forth a thin, transparent lamella of hyalin ectoplasm, 1 The most recent work on the body-fluid cells of sea-urchins is that of Kollmann (1908). He recognizes five types of cell. 356 JOHN F. FULTON, JR. and into this the round red granules subsequently flow. The conformation of the pseudopodia resembles closely that described by Goodrich (1919) for the coelomic corpuscles of Asterias glacialis. Thus (as first shown by Geddes, 1880) the cell is truly amoeboid, being able both to protrude and to withdraw its pseudopodia. It seemed a curious fact that the cells within the body-fluid of the sea-urchin should be of identically the same colour as the pigment granules which give to the animal its characteristic coloration. Consequently an effort was made to determine whether or not the two pigments are identical. The same technique was employed as was made use of in settling the identity of the pigments from A. bermudensis and C. passiflora (p. 347), viz. that of testing their action in the presence of certain solvents, and the results seemed to indicate clearly that the two pigments are one and the same. In examining the external pigmentation, small pieces of the tube-feet were employed, since the behaviour of their coloured eranules can be watched much more closely than can those of the spines. For testing the perivisceral fluid, fresh smears were used to which the reagents were added with a capillary pipette while under observation. Of the alcohols, amyl was the only one which dissolved the pigments, they being very readily soluble, however, in this reagent. Neither of the pigments were extracted by the lipochrome solvents, ether, chloroform, petro- leum ether, or xylol. The colouring-matters are readily dissolved by 10/N solutions of the acids, dissolving with particu- larly great rapidity in valeric acid.’ In the presence of alkalis both pigments were darkened, but not extracted. After these experiments had been made, it came to the author’s attention that Geddes, in a personal communication to Gamgee (1880, p. 134), stated his belief that the pigment of the red amoeboid cells was identical (in Echinus) with that of the epidermal spmes. This view resulted from a very thorough study of the body-fluid of sea-urchins (Geddes, 1880). Con- 1 This again corroborates Crozier’s (1916 a) conclusion that valeric is the most penetrating of all acids, dA. ww wee? + Bi i iii ii ee is DD irs bit tt DS ce en a ti A lth he ie 2 oe) rem — ae ee a ANIMAL CHLOROPHYLL 373 association with chlorophyll. It was first isolated from carrots (sometimes spelled ‘ carrotin ’), from which it received its name. Willstatter und Meig (1907) have shown that carotin is a crystalline unsaturated hydrocarbon (Cy) H;,) melting at 174°. It is of wide occurrence not only in the vegetable, but also in. the animal kingdom, being found in the blood sera of most birds and mammals (Schunck, 1903; Palmer, 1915, 1916; and Hymans van den Bergh und Miller, 1920), and in mammalian milk ;+ the yellow pigment of butter fat likewise is composed largely of carotin (Steenboch, Sell, and Buell, 1921). It is also found in mammalian ovaries ; however, the ‘lutem’ of the corporea lutea themselves is isomeric with xanthophyll, the oxide of carotin. The colored substance of egg-yolk is made up largely of xanthophyll, though it too contains some carotin (Palmer and Kempster, 1919 a, b, and c) ; carotin also is sometimes present in human urine, appearing there after carotin has been ingested (Hess and Meyers, 1919), and it has been isolated from gall-stones (Plimmer, 1915, p. 582). But of much greater interest is the fact that carotin is present abundantly in mammalian nerve-cells (Dolley and Guthrie, 1919 a and b), and in the fovea centralis of the human eye. Also the yellow pigment cells (xantholeucophores) in the epidermis of Fundulus contain carotin, but what concerns us more immediately is its occurrence in the chromatophores of crustaceans. Blanchard (1890) made it evident that the epidermal pig- ment of the copepod Diaptomus lacillifer contains a large percentage of carotin. Keeble and Gamble (1902, 1904, and 1905) have corroborated this result, finding that the chromatophores of many other crustaceans also possess the pigment. They noted, too, that mobile fat globules are usually to be found in the branching pigment cells. It has been shown by Kohl (1902) that carotin was capable of photosynthesizing fats. With this fact before them, Keeble and Gamble made 1 Palmer and Kckles (1914) have shown that the amount of carotin and xanthophyll in milk (human and cow) is in exact proportion to the amount of green food consumed. 374 JOHN F. FULTON, JR. an effort to find the origin of the fat globules in the crustacean chromatophores. The investigation proved that the carotin present in these cells produces the fat globules by photo- synthesis. The work was carried out on Hippolyte varians, and the experiments were briefly as follows :+ when Hippolyte is starved in the dark practically all of the oil disappears from the chromatophores ; when starved in sun- light, however, the globules continue to exist as before ; when a ‘dark-starved’ Hippolyte is exposed to sunlight (without feeding) the fat returns. Thus the crustaceans present a remark- able phenomenon: the plant pigment which has been eaten by the animal, is stored in the liver ; later it is carried by the blood-system and deposited in the epidermal chromatophores, where it functions exactly as in the plants from which it was derived! The only other instance of such a phenomenon known to the writer is that (already referred to on p. 373, note) described by Palmer and Eckles (1913), who have shown that carotin and xanthophyll pass in the blood from the intestine to the mammary glands. It is extremely interesting, also, to note the recent observation of Findlay (1920), that carotin and xanthophyll are found in the mammalian adrenals, and are largely responsible for the colour of the glands. (b) The Insecta. In many ways the insects offer a more favourable condition for the study of animal pigments than any other group of animals. They are small, many are briliantly coloured, and in addition the physiological processes of insects are on the whole less complicated than in other forms; moreover the more important pigments of insects are concentrated in the wings, which are thin and therefore well adapted for the purposes of observation. The greater share of the work on the pigmentation of insects ! Tt is not within the scope of this paper to attempt to summarize Keeble and Gamble’s results on the mechanism of colour change in this animal. They will be found in the papers cited above and in Gamble (1910) ; a more recent study of colour variation in crustaceans is that of Potts (1915). Pan Me ANIMAL CHLOROPHYLL 375 has been carried on by the English investigator Poulton,’ who found that the pigment of a large number of insects (in all stages of development) is a modified chlorophyll derived from the plant on which the animal feeds. The chlorophyllous substances are eaten, absorbed into the blood-system, and deposited in the regions of the body exhibiting pigmentation. The most valuable results in Poulton’s work came from a spectroscopic examination of insect blood. His (1884) work on the pigment of Sphinx ligustri is interesting. The blood from the pupae of this form was examined and its spectrum recorded ; then an extract was made of the calceolaria leaves upon which the larva feeds. When the spectrum of the extract was superimposed upon that of the blood, the bands were found to correspond in a very striking way. In Poulton’s own words (p. 290) : ‘Considering the chemical change which must have taken place in the chlorophyll during digestion, rendering possible the passage of the walls of the digestive tract, and considering its chemical union with the proteid constituent of the blood, the resemblances of the spectra are very striking ; in fact, the two spectra are far nearer each other than the ordinary spectrum _ of chlorophyll in alcoholic solution is to the unaltered chloro- phyll of leaves ’. It was held by Poulton that the power of utilizing chloro- phyll in building up pigments is an adaptation on the part of the insect which enables it to assume the colour of the leaves on which it feeds. As evidence for this he brought forward the fact that if larvae of Trypoena pronuba are fed respectively on green, brown, and white cabbage leaves, green, brown, and white larvae result according to the colour of the leaf on which they were fed. This result has since been corro- borated by Levart et Conte (1902), who worked on Attacus orizaba and Bombyx mori. Peterson (1918) has found that the chlorophyll which passes into the intestine of certain red caterpillars is modified into 1 Poulton has written a large number of papers on this subject. The more important ones will be found in the bibliography under Poulton, 1884, 1889, 1892, and 1893. NO. 262 pd 376 JOHN F. FULTON, JR. a red substance (vanessa red), which is later absorbed and transported to the epithelium, where it is deposited and becomes the pigment of the wings and of the other body-parts. The unabsorbed portion of the red pigment is voided. ‘The investigations of Gortner (1911 and 1912 6) on insect melanins are also of interest, since he has shown that they are sometimes formed from chlorophyllous substances (oxidizable chromogens) acted upon by the plant ferment, tyrosinase. More ‘recently Schmidt (1919) has conducted similar investigations on insect melanins, and his results accord with those of Gortner. It appears to the writer that the most significant part of all Poulton’s work on insects is the demonstration that chlorophyll resists the digestive enzymes, and passes practically unchanged into the blood-system. Evidence has been cited to prove that this also is the case in the Mollusca and Crustacea, but the evidence is not conclusive (except in the case of carotin). In insects, however, there is undisputable proof, and it is of particular importance, since it shows that the theory that haemoglobin as a derivative of chlorophyll must not be ruled out by the fact that chlorophyll is incapable of passing through the digestive tract. In general, then, the epidermal pigments . of Crustacea are derived from food and are carried to their destination by the blood-system. 6. Tue TuNICcATA. The strongest evidence in support of the view that the epidermal pigments are deposited by the blood-system is to be derived from a study of the tunicates, where it is possible to predict that there will be found in the blood-stream pigment cells of a colour corresponding to that of the tunic. Moreover one can show that the pigmented corpuscles arise from colour- less cells while they are in the circulation. The tunicate to be first considered is Ascidia atra. (a) Ascidia atra. This species is distinguished from other tunicates, and in fact from almost all other animals, by the possession of an ANIMAL CHLOROPHYLL oVere exceptionally large variety (ten distinct types) of blood-cells. In the vascular fluid of this animal there are three kinds of highly-pigmented cells : green, orange, and blue ; in addition there are four kinds of non-motile white corpuscles, and three other types which are distinctly amoeboid (Fulton, 1921 b). The writer has shown that all of the pigmented cells in A. atra arise directly in the blood-stream from unpigmented corpuscles. In the case of the green * chromocyte ’ the metamorphosis from the colourless cell may be stimulated artificially and the complete process watched under the microscope. When an acid, preferably an organic acid of N/10 to N/20 strength, is added to a fresh smear of blood, all the non-motile colourless cells of one variety may be observed to take on a light shade of green, which gradually deepens; at the same time the cell fragments into large green lumps and finally assumes the characteristic form of the green pigment cell. Therefore, it may be inferred that in nature the green pigment cell arises from an unpigmented corpuscle as a result of an increase in acidity. The orange and the blue cells also arise from unpigmented corpuscles, but in a slightly different manner. Various methods were employed in an effort artificially to stimulate the change from the colourless to the orange and to the blue cells. No response was secured from acids or bases, but it was observed in smears of blood taken from an animal which previously had been weakened by the loss of blood, that there occurred many intermediate stages between the unpigmented corpuscles and the orange cell or the blue cell. If one of the intermediate stages be carefully watched, under very favourable conditions, there is some indication that it gradually increases its depth cf colour. With this evidence | the conclusion is unavoidable that all of the pigmented cells arise in the blood-stream directly from colourless ones. The deep purple-blue colour of the tunic of A. atra is 1 The details of the experiments and a more complete statement of the evidence for this conclusion will be found in the paper by Fulton (1921 6) on the blood of Ascidia atra. Didi 378 JOHN F. FULTON, JR. caused by the presence of large blue pigment cells (Hecht, 1918 a; Crozier, 1916 d) containing spherical granules, which migrate from one part of the cell to another. The blue cor- puscles of the blood-stream are like the pigment cells of the test in every detail of their structure. It remains, therefore, to establish the identity of these cells. In the first place, the presence in both cells of a very prominent vacuole is strong evidence in favour of their being identical. In testing the cells with various reagents another interesting resemblance was noted. It is known that calcium chloride in minute amounts has the power of greatly accelerating the activity of phagocytes (Hamburger, 1910 and 1916); it also causes a decided increase in the activity of the blue pigment cells both of the blood-stream and of the test. After a M/10 solution of CaCl, has been added to a fresh smear of blood the blue cells, which in their quiescent state are nearly spherical, immediately send forth pseudopodia, and at the same time the blue granules within the corpuscles commence to move from one end of the cell to the other. In the test, however, the pigment cells, inasmuch as they are fixed within the substance of the test, cannot move their processes ; there is, nevertheless, following the addition of CaCl, to a section of the test, a decided activity on the part of the blue granules,’ an activity which is similar to that displayed by the corresponding granules of the blood-cells. The conclusion, therefore, seems to be warranted that the two cells are identical, and consequently that the blue cells in the blood give rise to the pigment cells of the test.” From these two kinds of evidence it is a reasonable conclusion that in Ascidia atra the coloration of the animal is even- tually traceable to the colourless cells of the blood ; for, as has been shown, the unpigmented cells give rise, while in circula- 1 A description of a phenomenon of this kind is also to be found in Pizon’s (1898, 1901) papers on the pigment granules of tunicates. 2 Hecht (1918 a) states that when A. atra regenerates a portion of its test, there are a great many of the blue blood-cells in the area of the regenerating tissue. ANIMAL CHLOROPHYLL 379 tion, to the blue corpuscles ; these finally become lodged in the tunic, and in that way give rise to the surface pigmentation of the animal. (b) Other Tunicates. There are many other ascidians in the Bermuda waters which are highly coloured. In a cave on the west side of Agar’s Island’ five specimens of the brilliant red tunicate, Microcosmus miniatus, Verrill, were found. An examination of the blood revealed that its most prominent constituent was an amoeboid cell, containing many brilliant carmine-coloured granules, which was very similar to the pigment cell that colours the test. The animal from which the blood had been extracted was examined on the day follow- ing, and, asin A. atra, there were many intermediate stages between the colourless cells and the pigmented corpuscles. These observations confirmed in a substantial way those made upon A. atra. Other species of ascidians have been examined and in every case the colour of the pigment cells in the test was duplicated by the coloured cells of the blood. The colonial form, Ee tein - ascidia turbinata, Herdman, which is brilliant orange in colour, has as its only coloured cell in the blood-stream a corpuscle possessing orange granules. (c) Discussion. Concerning the origin of the pigments in the blood-stream of ascidians, no definite statement can be made. It has been observed that the pigment cells arise while in circulation from unpigmented corpuscles. Just what is the process involved in that colour change it is difficult to explain. Griffiths (1897) has described a colourless respiratory proteid (y-achroglobin) in the blood of ascidians, and the chromogen of Phallusia has been stated by Henze (1911, 1912) to be a proteid in com- bination with the element vanadium. It is possible, therefore, that the change from the colourless cell to the ‘ chromocyte ’ 1 Where the laboratory is situated, 380 JOHN F. FULTON, JR. is occasioned by the chemical union of the proteid with the metal. This seems extremely unlikely, however, since vanadium does not exist in the blood in its elemental state—the differences in colour of the corpuscles being due to vanadium in different states of oxidation—and also because the vanadium probably has the réle of catalyst in the respiratory phenomena of ascidians (Fulton, 1921 c). A more likely explanation is that in the colourless antecedents of the pigment cells there exists some colourless vanadium compound which, either on oxida- tion or reduction, is converted into one of the coloured oxides of that metal. 7. Discuss1IoN—PIGMENTATION OF VERTEBRATES. The results obtained in a study of the pigmentation of invertebrates cannot be entirely without application to the problem of coloration in vertebrates. Particularly does this seem true in view of the question regarding the origin of melanin. In recent years the origin of pigments in vertebrates has been much discussed. One school holds that the epidermis is capable of elaborating its own pigment ; * another maintains that the pigment is carried into the integument by wandering leucocytes ;? still another holds that melanin is derived from the haemoglobin of the blood.* For a very complete account of the historical development and the present status of the biological theories regarding the origin of melanin, the reader is referred to Dawson’s (1920) paper on the integument of Necturus. But, in addition to the purely biological discussion of the question, there are certain chemical investigations on melanin which demand attention. The chemical analyses which have been made upon melanin tend on the whole to support the view that the pigment is 1 Hooker (1914) and Eycleshymer (1906) are the more important advocates of this view. 2 See the papers of Reinke (1906), Négre (1906), and Borre) (1913). 3 Rabl (1894) maintained that the leucocytes phagocytized red blood- cells and converted their haemoglobin into melanin, This view is also supported by certain chemical investigations, ANIMAL CHLOROPHYLL 381 a derivative of haemoglobin. In the first place the same elements are present in melanin and in haemoglobin (Ham- marsten and Hedin, 1915, p. 84). Aside from nitrogen and sulphur the most noticeable element present is iron. The earlier observers (Scherer, 1841; Berdez und Nencki, 1886) failed, to detect iron, but, as Halliburton (1898, vol. i, p. 121) points out, their failure was due to the fact that they extracted the pigment with hydrochloric acid and thus removed the iron. Morner (1887), and Brandl und Pfeiffer (1890) found that melanin contained a large amount of iron, and believe as a result that melanin is a derivative of the blood-pigments. Schmiedeberg (1897) obtained similar resuits for the sarco- melanin from a sarcomatous liver, finding that it contained 2°7 per cent. iron. The more recent work on the subject likewise corroborates the observation that iron is present in melanin (Gortner, 1912a@; von Firth und Jerusalem, 1907; Piettre, 1911 a). There is, therefore, both biological and chemical evidence in favour of the view that melanin is derived from the blood- pigments. Moreover, as the present paper has attempted to show, the great majority of invertebrate pigments are not only derived from the pigments of blood-systems but the invertebrate blood-pigments are themselves derived from food. Unless a profound change has occurred in the physio- logical processes of the vertebrates as compared with those of the invertebrates—which is not probable—it appears reason- able to the writer to admit that some at least of the vertebrate pigments likewise owe their origin to the pigments of the blood. Urochrome.—the recent feeding experiments of Roaf (1921) have given strong evidence that the output of urochrome from the urine of guinea-pigs and of man is roughly propor- tional to the amount of chlorophyll taken in as food, and Roaf suggests, in view of the chemical similarity between the two pigments (the pyrrol reaction), that urochrome is derived from chlorophyll. It is evident that this observation throws quite a new light upon the debated question of urimary pigments, and it gives an added instance of the dependence of animals 382 JOHN F. FULTON, JR. upon the pigments of plants. The further relation of chloro- phyll to the bile and urinary pigments is a subject which will well repay further investigation. In conclusion it may be said with reasonable certainty that many animals and probably man do normally use the four pyrrol groups of the chlorophyll molecule to synthesize haemoglobin and allied pigments; however, though most evidence points in this direction, no one has actually demonstrated that the animal body is itself incapable of synthesizing haemoglobin in the absence of chlorophyll. 8. CONCLUSIONS. Part; What deductions may be made concerning the animals which have thus far been considered? In the first place it must be recalled that in no case has the writer dealt with forms which have a true blood-vascular system. The most important function of the blood-system im the higher animals is that of carrying nutriment to the tissues. In the lower invertebrates this function is accomplished either by the direct contact of the tissue with the surrounding sea- water, or by circulatory fluids—less highly specialized than blood—which move within the body. These fluids, therefore, are the ones which represent the functional antecedents of the blood-vascular system; and it is to these that one should look in seeking the origin of many of the body-pigments. From the foregoing pages the following conclusions seem reasonable : 1. The pigmented protozoans owe their colour, probably in every case, to an algal pigment which has resulted from an out- side infection. 2. Though the evidence in the Porifera is not conclusive,1 1 The evidence recently adduced by Van Trigt (1918) removes any reasonable doubt concerning the chlorophyllous nature of the pigments of a large number of sponges, ANIMAL CHLOROPHYLL 3838 it is probable that their pigment is chlorophyll or a substance closely allied (chlorophyll has been demonstrated in more than twenty species of sponge), and there is evidence which indicates that it, too, is obtained from external sources. 3. The constituents from which actinians manufacture their pigment are carried to the tissues by the gastrovascular fluid ; it is likely that the constituents themselves are derived from the chlorophyllous substances which enter as food; thus they are in a highly-organized state when they reach the tissues, making the synthesis of the pigment less difficult. Certain of the acti- nian pigments—aside from being derivatives of chlorophyll— are Closely related to haemoglobin (actiniohaematin). 4. In certain flatworms which are not coloured by algal symbionts, the pigment is derived from food and is carried to the tissues by the gastrovascular system. 5. The pigment of the red cells of Tripneustes escu- lentus, so numerous in the perivisceral fluid, is identical with the pigment of the epidermis, and since it has been shown by Geddes and others that the pigmented cells arise while in circulation from yellow cells, direct evidence is thereby afforded that the body pigment arises in the nutritive fluid. 6. The red pigment, echinochrome, though probably not respiratory (McClendon), nevertheless bears a close chemical relationship to haemoglobin (Griffiths). 7. Since there is every probability (nm T. esculentus) that the yellow cells from which the reds arise are chloro- phylloid corpuscles, it seems clear that chlorophyll is capable not only of giving rise to an animal pigment but to a pigment which is closely akin to haemoglobin (echinochrome breaks down into haemochromogen, a reduction product of haemo- globin). Birgi’s feeding experiments show that chlorophyll facilitates the formation of haemoglobin in anaemic rabbits. 8. The theory that haemoglobin is derived from chlorophyll is further strengthened by the fact that in many echinoderms there is present, simultaneously with haemoglobin and chloro- phyll, a substance, haematoporphyrin, which is an intermediate product chemically between chlorophyll and haemoglobin, 384 JOHN F, FULTON, JR. 9. In many other echinoderms Asteroidea, Ophiuroidea, Echinoidea, and Holothuroidea, there are found pigmented body-fluid cells which likewise give rise (in part at least) to the external pigmentation by becoming deposited in the epidermis. Part If. In the preliminary observations of the present work, it was found that in nearly every’ invertebrate imvestigated there occurs in the blood a chromogen—either completely formed, or in the process of formation—which is similar to, and not infrequently identical with, the pigment of the epidermis. Though this in itself is an interesting fact, it is at once apparent that behind the phenomenon there les something of much sreater significance : Why should the body-pigments occur in the blood-system, and whence do they come? The pigments of the invertebrates, so far as they have been investigated, appear to be derived very largely from food, bemg absorbed into the blood-stream and carried by that tissue to the epidermal regions, where they are deposited. The more specific conclusions are as follows : 1. In the absence of a blood-system, as in the Echinoderms and lower forms, the nutritive fluids supply the epidermis with its pigments (Part I). 2. The coloration of several nemertean worms is due to the presence of haemoglobin in the epidermis and in the blood. 3. In certain nudibranchs (Chromodoris zebra) and cephalopods, pigments are found in the blood-stream which are identical with the epidermal pigments. There is strong evidence that the pigments of the blood-stream of the Mol- lusca owe their origin to the chlorophyllous substances taken in as food. 4. The lipochromes of annelids are derived from food substances, being absorbed into the blood-stream and trans- ported to the epidermis. Annelid haemoglobin is found both in the blood and in the epidermis. 5. The enterochlorophyll found in the liver of many crusta- ceans and molluses is of vegetable origin. There is evidence ANIMAL CHLOROPHYLL 385 that it is the base from which the animal synthesizes many other of its pigments, including haematin. 6. The red lipochrome of the blood and the chromatophores of crustaceans are derived from carotin, a pigment associated with chlorophyll (found also in the liver). Carotin and the red lipochrome of crustaceans are chemically identical (Verne). 7. Many pigments of insects are modified chlorophylls derived directly from the chlorophyll of the food (Poulton). 8. The pigments of the tunicates are found first in the blood- system. ‘The pigmented cells arise, while in the circulation, from unpigmented corpuscles, and certain of the pigment cells which arise in this way are subsequently deposited in the test. 9. Strong evidence exists that the respiratory pigment haemo- globin is derived both phylogenetically and physiologically from chlorophyll. I cannot close this paper without mentioning my great indebtedness for advice and inspiration to Professor Benjamin Moore whose recent untimely death will be most keenly felt in America as well as in England. He more than any one has helped to clarify the perplexing question of the relation which chlorophyll bears to the protoplasmic system of plants and animals. 9. BIBLIOGRAPHY. The following list of references is for both Parts I and II of the paper. Alsberg, C. L., and W. M. Clark (1914).—“ The solubility of oxygen in the serum of Limulus polyphemus and in solutions of pure Limulus haemocyanin’”, ‘ Journ. Biol. Chem.’, vol. 19, pp. 503-10. Bayliss, W. M. (1918).—‘ Principles of General Physiology’, 2nd ed. London: Longmans, Green & Co., xxiv + 858 pp. Berdez, J., und M. Nencki (1886).—‘* Uber die Farbstoffe der melanotischen Sarkome ’’, ‘ Arch. f. exp. Path. u. Pharm.’, Bd. 20, pp. 346-61. Bert, M. P. (1867).—‘“‘ Sur la physiologie de la Seiche (Sepia offici- nalis)”’, “C. R. Acad. Sci. Paris’, tom. 65, pp. 300-3. Bertrand, G. (1897).—“ Sur Vintervention du manganése dans les oxyda- tions provoquées par la laccase ’’, ibid., Paris, tom. 124, pp. 1032-5. —— (1912).—“* Sur le role capital du manganése dans la production des conidies de |’ Aspergillus niger”, ‘ Bull. Soc. Chim. France ’, tom, 11, pp. 495-9. Also ‘C, R, Acad. Sci. Paris’, tom. 154, pp. 381-3. 386 JOHN F. FULTON, JR. Beyerinck, M. W. (1890).—‘‘ Culturversuche mit Zoochlorellen, Lichenen- goniden, und anderen niederen Algen”’, © Bot. Zeitung’, Jahrg. 48, pp. 725-54, 757-68, 781-5, Taf. 7. Biedermann, W. (1914).—‘‘ Farbe und Zeichnung der Insekten”’, ‘ Hand. Vergl. Physiol. (Winterstein) ’, Bd. 3, Halfte 1, Teil 2, pp. 1656-1994, Blanchard, R. (1890).—‘‘ Sur une matiére colorante des Diaptomus, analogue a la carotine des végétaux’”’, “C. R. Acad. Sci. Paris’, tom. 110, pp. 292-4. Borrel, A. (1913).—‘ Cellules pigmentaires et associations cellulaires ”’, ‘C. R. Soe. Biol. Paris,’ tom. 74, pp. 1215-18. Bourne, A. G. (1891).—‘“‘ On Pelomyxa viridis, sp. n., and on the vesicular nature of protoplasm ”’, ‘ Quart. Journ. Micr. Sci.’, vol. 32, pp. 357-74, pl. 28. Brandl, J., und L. P. Pfeiffer (1890).—‘ Beitrag zur Kenntniss der Farbstoffe melanotischer Sarkome nebst Bemerkungen iiber einige Eigenschaften der sogenannten melanogenen Substanz im Horn ”’, ‘ Zeits. f. Biol.’, Bd. 26,pp. 348-76. Brandt, K. (1881 a).—‘‘ Untersuchungen an Radiolarien”’, * Monatsb. Akad. Wiss. Berlin’, Jahrg. 1881, pp. 388-404, 2 Taf. —— (1881 b).—‘* Ueber das Zusammenleben von Thieren und Algen”’, ‘Verhandl. Physiol. Gesell. Berlin’, Jahrg. 1881-2, in * Arch. f. Anat. u. Physiol.’, 1881, ‘ Physiol. Abth.’, pp. 570-74, Text-figs. 1-3. Also in ‘ Sitzungsb. Gesell. naturf. Freunde Berlin’, Jahrg. 1881, pp. 140-6. —— (1882).—‘‘ Uber die morphologische und physiologische Bedeutung des Chlorophylls bei Thieren ”, ‘ Arch. Anat. u. Physiol.’, Jahrg. 1882, Physiol. Abth., pp. 125-51, Taf. 1. —— (1883).—‘‘ Uber die morphologische und physiologische Bedeutung des Chlorophylls bei Thieren”’, ‘II. Mitt. Zool. Sta, Neapel’, Bd. 4, pp. 191-302, Taf. 19, 20. Biirgi, E., C. F. v. Traczewski, 8S. Bass, A. Braunstein, und §. Fridkiss (1919).—‘ Uber die biologischen und pharmakologischen Eigenschaften des Chlorophylls ”’, * Biochem. Zeits.’, Bd. 98, pp. 256-83. Cotte, J. (1904).—‘‘ Contribution a létude de la nutrition chez les Spon- giaires ’, ‘ Bull. Sci. France et Belg.’, tom. 38, pp. 420-574. Cowdry, E. V. (1911).—** The colour changes of Octopus vulgaris, Lmk.”, ‘ Univ. Toronto Studies’, Biol. Ser., no. 10, pp. 1-53. Crozier, W. J. (1914).—** Note on the pigment of a Bermuda Nudibranch Chromodoris zebra, Heilprin”’, ‘Journ. Physiol.’, vol. 47, pp. 491-2. —— (1915).—“‘ On cell penetration by acids’’, ‘Science’, N.S., vol. 42, pp. 735-6. —— (1916 a).—‘‘ Cell penetration by acids”, ‘Journ. Biol. Chem.’, vol. 24, pp. 255-79. —~— (1916 b).—*‘ Cell penetration by acids. II. Further observations of the blue pigment of Chromodoris zebra ”’, ibid., vol. 26, pp. 217-23, <== ANIMAL CHLOROPHYLL 387 Crozier, W. J. (1916 c).—* Cell penetration by acids. ILI. Data on some additional acids ”’, ibid., pp. 225-30. —— (1916 d).—‘ Some indicators from animal tissues ”’, ibid., vol. 24, pp. 443-5. (1917).—** On the pigmentation of a Polydad”’, * Proc. Amer, Acad., Arts and Sci.’, vol. 52, pp. 725-30, 1 pl. Cuénot, L. (1890).—‘‘ Le sang et la glande lymphatique des Aplysies ”’» *C. R. Acad. Sci. Paris’, tom. 110, pp. 724-5. —— (1891).—“ Etudes sur le sang et les glandes lymphatiques dans la série animale’’, “ Arch. Zool. Exp.’, sér. 2, tom. 9, pp. 13-90, 365-70, pls. 1-4, 15-23. —— (1901).—“ Etudes physiologiques sur les Astéries”’, ibid., sér. 3, tom. 9, pp. 233-59. Dangeard, P. A. (1900).—‘* Les Zoochlorelles du Paramoecium bursaria’”’, ‘ Le Botaniste ’, sér. 7 (fasc. 3, 4), pp. 161-91. —— (1902).—“* Recherches sur les Eugléniens ”’, ibid., tom. 8, pp. 77-357, Dantec, F. le (1892).—“‘ Recherches sur la symbiose des Algues et des Protozoaires ’’, ‘ Ann. Instit. Pasteur’, tom. 6, pp. 190-8. Dastre, A. (1899).—‘* La chlorophylle du foie chez les Mollusques ”’, ‘ Journ. Physiol. et Path. gén.’, tom. 1, pp. 111-20. —— et N. Floresco (1898).—*‘ Pigments du foie en général. II.—Pigments hépatiques chez les Invertébrés ”’, * Arch. de Physiol.’, vol. 10, pp. 288- 303. —— (1899 a).—‘ Recherches sur les matiéres colorantes du foie’, Paris: G. Steinheil, 86 pp. (1899 b).—*‘ Contribution a l’étude des chlorophylles animales. Chlorophylle du foie des Invertébrés ”’, “ C. R. Acad, Sci. Paris ’, tom, 128, pp. 398-400. Dawson, A. B. (1920).—“ The integument of Necturus maculosus’”, * Journ. Morph.’, vol. 34, pp. 487-577, 6 pls. ‘Dhere, C., et G. Vegezzi (1916).—*‘ Sur la composition de l’hépatochloro- phylle ”’, ‘C. R. Acad. Sci. Paris ’, tom. 163, pp. 399-401. Dolley, D. H., and F. V. Guthrie (1919 a).—“‘Origin of nerve cell pigments ”’, “Science ’, N.S., vol. 50, pp. 190-2. (1919 b).—“‘ The pigmentation of nerve-cells. The lipochrome a plant carotinoid pigment ”’, ‘ Journ. Med. Res.’, vol. 40, pp. 295-309. Elmhirst, R., and J. 8. Sharpe (1920)—‘“‘ On the colours of two sea anemones, Actinia equina and Anemonia sulcata. I. En- vironmental (Elnhirst). II. Chemical (Sharpe) ”’, ‘ Bio-Chem. Journ.’, vol. 14, pp. 48-57, 2 pls. Engelmann, T. W. (1883).—** Ueber thierisches Chlorophyll ”’, ‘ Pfliiger’s Arch. ges. Physiol.’, Bd. 32, pp. 80-96. Entz, G. (1881).—‘‘ Ueber die Natur der * Chlorophyllk6rperchen ’ niederer Tiere’, ‘ Biol. Centralb.’, Bd. 1, pp. 646-50. 388 JOHN F. FULTON, JR. Entz, G, (1883).—* Das Consortialverhiltniss von Algen und Tieren ”’, ibid., Bd. 2, pp. 451-64. Eycleshymer, A. C. (1906).—** The development of the chromatophores in Necturus ’’, ‘ Amer. Journ. Anat.’, vol. 5, pp. 309-13. Famintzin, A. (1889).—“ Beitrag zur Symbiose von Algen und Thieren ”’, ‘Mém. Acad. Sci. St. Pétersbourg ’, sér. 7, tom. 36, no. 15, 36 pp., 2 Taf. —— (1891).—‘‘ Beitrag zur Symbiose von Algen und Thieren”’, ibid., tom. 38, no. 4, 16 pp., 1 Taf. Fauvel, P. (1899).—‘‘ Sur le pigment des Arénicoles ”’, ‘C. R, Acad. Sei. Paris, tom. 129, pp. 1273-5. Findlay, G. M. (1920).—‘* Pigments of the adrenals’, “ Journ, Path, and Bact.’, vol. 23, pp. 482-9. Focillon, A. (1851).—‘‘ Sur les couleurs du test des Crustacés”’, *C. R. Acad. Sci. Paris ’, tom. 33, pp. 384-5. Foettinger, A. (1880).—‘ Sur l’existence de ’hémoglobine chez les Echino- dermes ”’, ‘ Arch. de Biol.’, tom. 1, pp. 405-13, pl. 17. Franz, V. (1910).—‘‘ Zur Struktur der Chromatophoren bei Crustaceen ”’, * Biol. Centralbl.’, Bd. 30, pp. 424-30. Fredericq, L. (1878).—** Recherches sur la physiologie du poulpe commun (Octopus vulgaris) ’’, ‘ Arch. de Zool. exp.’, sér. 1, tom. 7, pp. 535-83. Fuchs, R. F. (1914).—‘‘ Der Farbenwechsel und die chromatische Haut- funktion der Thiere’’, ‘ Handb. Vergl. Physiol. (Winterstein) ’*, Bd. 3, Hiilfte 1, Teil 2, pp. 1189-1656. Fulton, J. F., Jr. (1921 a).—‘‘ Concerning the vitality of Actinia bermudensis: A study in symbiosis’, ‘Journ. Exp. Zodl.’, vol. 33, pp. 353-64. —— (1921 6)—‘‘ The blood of Ascidia atra, Lesueur: dealing especially with pigmentation and phagocytosis”, *‘ Acta Zoologica ’, vol. 1, no. 4, pp. 381-431, 2 pls. —— (1921 c).—‘‘ The catalytic properties of the respiratory metals ”’, “Science ’, N.S., vol. 53, pp. 444-6. ; Firth, O. von, und E. Jerusalem (1907).—** Zur Kenntniss der melanoti- schen Pigmente und der fermentativen Melaninbildung ’’, ‘ Beitr. Chem. Physiol. u. Path. (Hofmeister) *, Bd. 10, pp. 131-73. Gamble, F. W. (1910).—* The relation between light and pigment forma- tion in Crenilabrus and Hippolyte ’’, ‘Quart. Journ. Mier. Sci.’, vol. 55, pp. 541-88, 1 pl. and F. Keeble (1900)—‘‘ Hippolyte varians: a study of colour-change ”’, ibid., vol. 43, pp. 589-698, 5 pls. (1903).—“* The bionomics of Convoluta roscoffensis, with special reference to its green cells”, ibid., vol. 47, pp. 363-431, pls. 30, 31. (Abstract, ‘ Proc. Roy. Soc. London’, vol. 72, pp. 93-8). Gamgee, A. (1880).—‘ Text-book of the physiological chemistry of the animal body’. London: Macmillan, vol. 1, 487 pp. ANIMAL CHLOROPHYLL 389 Geddes, P. (1880).—‘* Observations sur le fluide périviscéral des oursins ”’ * Arch. Zool. exp.’, sér. 1, tom. 8, pp. 483-96, pls. 37-8. —— (1882).—‘‘ On the nature and functions of the ‘ Yellow Cells’ of radiolarians and coelenterates ’’, ‘ Proc. Roy. Soc., Edinburgh ’, vol. 11, pp. 377-96. Goetch, W. (1921).—‘‘ Grime Hydra fusca L.”, ‘Zool. Anz.’, Bd. 53, pp. 173-6. Goodrich, E. 8. (1919).—** The pseudopodia of the leukocytes of inverte- brates’, ‘ Quart. Journ. Micr. Sci.’, vol. 64, pp. 19-26, pl. 1, 2. Gortner, R. A. (1911).—* The origin of the brown pigment in the integu- ments of the larva of Tenebrio molitor’”’, ‘Journ. Biol. Chem.’, vol, 7, pp. 365-70. —— (1912 a).—“‘ Studies on Melanin. I. Methods of isolation. The effect of alkali on melanin ”’, ibid., vol. 8, pp. 341-63. -—— (1912 b).—‘‘ Studies on Melanin. IV. The origin of the pigment and the colour pattern in the elytra of the Colorado potato beetle (Leptinotarsa decemlineata, Say)”, ‘Amer. Nat.’, vol. 45, pp. 743-56. Griffiths, A. B. (1892).—‘ Physiology of the Invertebrata’. London : Reeve & Co., xvi+ 477 pp. —— (1897).—‘ Respiratory Proteids’. London: Reeve & Co., viii+ 126 pp. Grigoriew, R. (1919).—‘‘ Uber die blutbildenden Eigenschaften des Chlorophylls ”’, ‘ Biochem. Zeits.’, Bd. 98, pp. 284-93. Halliburton, W. D. (1885).—‘‘ The blood of decapod crustacea ’’, ‘ Journ. Physiol.’, vol. 6, pp. 300-35, 1 pl. —— (1898).—‘ Text-book of Physiology ’, ed. by E. A. Schafer. London: MacMillan, 2 vols., 1036+ 1365 pp. Hamburger, H. J. (1910).—“‘ The influence of small amounts of calcium on the motion of phagocytes”, ‘ Proc. Konink. Akad. Amsterdam ’, pp. 66-79. —— (1916).—“‘ Researches on Phagocytosis ”’, * British Medical Journ.’, 1916, vol. i, pp. 37-41. Hammarsten, O., and 8. G. Hedin (1915).—‘* A text-book of Physiological Chemistry ’, 7th English ed. Trans. J. A. Mandel. New York: Wiley & Sons, 1026 pp. Harvey, HK. N. (1915).—‘‘ The Permeability of cells for acids’’, ‘ Publ. Carnegie Instn. Wash.’, no. 212, pp. 143-56. Hecht, S. (1918 a).—“ The physiology of Ascidia atra Legueur. I. General Physiology ”’, ‘ Journ. Exp. Zoél.’, vol. 25, pp. 229-59. (1918 b).—‘‘ The physiology of Ascidia atra Lesueur. III. The blood system ”’, ‘ Amer. Journ. Physiol.’, vol. 40, pp. 157-87. Heilprin, A. (1889).—‘ The Bermuda Islands’, Philadelphia: Acad. of Nat. Sci., Phil., 231 pp., 17 pls. 390 JOHN F. FULTON, JR. Henze, M. (1911).—‘* Untersuchungen iiber das Blut der Ascidien, “I. Die Vanadiumverbindung der Blutkérperchen”’. ‘ Zeits. physiol. Chem.’, Bd. 72, pp. 494-501. —— (1912).—‘* Untersuchungen iiber das Blut der Ascidien”’, LL. * Zeits. physiol. Chem.’, Bd. 79, pp. 215-28. Hérouard, E. (1889).—‘‘ Recherches sur les Holothuries des cétes de France ”’, ‘ Arch. Zool. exp.’, sér. 2, tom. 7, pp. 535-704, pls. 25-32. Hertwig, O. (1883).—‘‘ Die Symbiose oder das Genossenschaftsleben im Thierreich ’’, ‘ Vers. deutsch. Nat.-Aerzte Freiburg’. (Also separate.) Jena: Gustav Fischer, 50 pp., 1 Taf. Hess, A. F., and V. C. Meyers (1919).—** Carotinemia: A new clinical picture ”’, ‘ Journ. Amer. Med. Assoc.’, vol. 73, pp. 1743-45. Hooker, D, (1914).—*‘ The development of stellate pigment cells in plasma cultures of frog epidermis ”’, ‘ Anat. Rec.’, vol. 8, pp. 103-4. Hopkins, F. C. (1894).—“*‘ Pigments of the Pieridai. A contribution to the study of excretory substances which function in ornament”, * Proc. Roy. Soc. London ’, vol. 57, pp. 5-7. Hubrecht, A. A. W. (1874).—“* Untersuchungen iiber Nemertinen aus dem jolf von Neapel’”’, *‘ Niederl. Arch. f. Zool.’, Bd. 2, pp. 99-135. ——and A. E. Shipley (1911).—*‘ Nemertina”, ‘ Ency. Brit.’, vol. 19, pp. 363-8. Huxley, T. H. (1877).—*‘ A manual of the Anatomy of invertebrated Animals’. London: Appleton & Co., 596 pp. Hymans van den Bergh, und P. Miller (1920).—‘‘ Das lipochrome Pigment in Blutserum und Organen, Xanthosis, Hyperlipochromimie ”, ‘ Bio- chem. Zeits.’, Bd. 108, pp. 279-303. Johnson, M. E. (1913).—*‘ Pigment formation in amphibian larvae”’, ‘Univ. Calif. Publ., Zool.’, vol. 11, pp. 53-83. Keeble, F. (1908).—‘‘ The Yellow-brown cells of Convoluta para- doxa’’, ‘ Quart. Journ. Micr. Sci.’, vol. 52, pp. 431-79, pls. 26-8. —— (1910).—‘ Plant-Animals, a study in symbiosis’. Cambridge : Univ. Press, viii+ 163 pp., 1 pl. ——and F. W. Gamble (1900).—* Colour-physiology of Hippolyte varians’’, ‘ Proc. Roy. Soc. London’, vol. 65, pp. 461-468. —— —— (1902).—** The colour-physiology of higher crustacea’’, ibid., vol. 71, pp. 69-71 (1904).—** On the presence of mobile fat in the chromatophores of the crustacea (Hippolyte varians)”, ‘Zool. Anz.’, Bd. 27, pp. 262-4. (1905).—** The colour-physiology of higher crustacea, Part III”, ‘Phil. Trans. Roy. Soc. London’, vol. 198, pp. 1-16. (1907).— The origin and nature of the green cells of Con- voluta roscoffensis’”’, ‘Quart. Journ. Micr. Sci.’, vol. 51, pp. 167- 219, pls. 13, 14. ANIMAL CHLOROPHYLL 391 Kirkpatrick, R. (1912).—*‘ Note on Merlia normani and the ‘ Monti- culiporas’”’, ‘ Proc. Roy. Soc. London ’, vol. 85 B, pp. 562-3. Kohl, F. G. (1902).—‘ Carotin und seine physiologische Bedeutung in der Pflanze ’, Leipzig. Kollmann, M. (1908).—‘‘ Recherches sur les leucocytes et le tissu lymphoide des invertébrés”’, ‘ Ann. Sci. Nat., Zool.’, sér. 9, tom. 8, pp. 1-240, pls. 1-2. Krukenberg, C. F. W. (1884).—“ Grundziige einer vergleichenden Physio- logie der Farbstoffe und der Farben”, in ‘ Vergl. physiol. Vortrige ’ (Abth. III), pp. 83-184. Kuskop, M. (1921).—“ Uber die Symbiose von Siphonophoren und Zooxan- thellen ”, ‘ Zool. Anz.’, Bd. 52, pp. 257-66. Lankester, E. R. (1869).—“‘ Abstract of a report on the spectroscopic examination of certain animal substances”’, ‘Journ. Anat. Physiol.’, vol. 4, pp. 119-29. ; —— (1872).—‘ A contribution to the knowledge of haemoglobin ”’, * Proc. Roy. Soc., London’, vol. 21, pp. 71-81, 1 pl. —— (1873).—“ The colouring matter of Stentor caeruleus’”’, ‘ Quart. Journ. Micr. Sci.’, vol. 13, pp. 139-42. —— (1882 a).—“ Chlorophyll corpuscles of Hydra’, ‘ Nature’, vol. 27, pp. 87-8. —— (1882 b).—* On the chlorophyll corpuscles and amyloid deposits of Spongilla and Hydra ’’, ‘ Quart. Journ. Micr. Sci.’, vol. 22, pp. 229-54, pl. 20. —— (1885).—“‘ Archerina boltoni, nov. gen. et sp., a chlorophyllo- genous protozoon, allied to Vampyrella, Cienk.”’, ibid., vol. 25, pp. 61-72, plot: —— (1886).—“‘ On green oysters ”’, ibid., vol. 26, pp. 71-94, pl. 7. Leitch, I. (1916).—‘‘ The function of haemoglobin in invertebrates with special reference to Planorbis and Cheronomus larvae.”, ‘Journ. Physiol.’, vol. 50, pp. 370-4. Levart, D., et A. Conte (1902).—* Sur l’origine de la coloration naturelle des soies des Lépidopteéres ”’, ‘ C. R. Acad. Sci. Paris ’, tom. 135, pp. 700-2. MacMunn, C. A. (1883 a).—“‘ Studies in animal chromatology”’, ‘ Proc. Birm. Phil. Soc.’, vol. 3, pp. 351-407, 3 charts. (1883 b).—“‘ On the presence of haematoporphyrin in the integu- ment of certain invertebrates’, ‘ Journ. Physiol.’, vol. 7, pp. 240-52. pl. 8. (1883 c).—‘‘ Observations on the colouring matters of the so-called bile of invertebrates, on those of the bile of vertebrates, and on some unusual urine pigments ”’, “ Proc. Roy. Soc. London’, vol. 35, pp. 370- 403. (Abstract, pp. 132-4.) (1885 a).—“‘ On the chromatology of the blood of some inverte- brates’, ‘ Quart. Journ. Micr. Sci.’, vol. 25, pp. 469-90, pls. 33, 34. NO, 262 Ee 392 JOHN F. FULTON, JR. MacMunn, C. A, (1885 b).—“ Further observations on enterochlorophyll and allied pigments ”’, ‘ Proc. Roy. Soc. London’, vol. 38, pp. 319-22. (1885 c).—‘*‘ Observations of the chromatology of Actiniae”’, ‘ Phil. Trans. Roy. Soc., London’, vol. 176, pp. 641-63. (Abstract, * Proc. Roy. Soc.’, vol. 38, pp. 85-7.) —— (1888).—‘‘ On the Chromatology of some British Sponges ’’, ‘ Journ. Physiol.’, vol. 9, pp. 1-25, pl. 1. —— (1889).—‘“* Contributions to animal chromatology”’, ‘ Quart. Journ. Micr. Sci.’, vol. 30, pp. 51-96, pl. 6. —— (1899 a).—“‘ On the gastric gland of Mollusca and decapod Crustacea, its structure and function”, ‘ Proc. Roy. Soc. London’, vol. 64, pp. 436-9. (Abstract.) —— (1899 b).—“‘ The pigments of Aplysia punctata”, ‘Journ. Physiol.’, vol. 24, pp. 1-10, pls. 1, 2. Mandoul, H. (1903).—‘‘ Recherches sur les colorations tégumentaires ”’, ‘Ann. Sci. Nat., Zool.’, sér. 8, tom. 18, pp. 225-463, 2 pls. McClendon, J. F. (1912).—Echinochrome, a red substance in Sea-Urchins ”’, ‘Journ. Biol. Chem.’, vol. 11, pp. 435-41. McMurrich, J. P. (1889).—** The actiniaria of the Bahama Islands, W.I.’, * Journ. Morph.’, vol. 3, pp. 1-80, pls. 1-4. Mérejkowski, C. de (1881).—‘*‘ Sur la tétronérythrine dans le régne animal et sur son role physiologique”’, ‘C. R. Acad. Sci. Paris’, tom. 93, pp. 1029-32. Milne-Edwards, H. (1838).—‘‘ Recherches pour servir 4 Vhistoire de la circulation chez les Annélides’’, ‘ Ann. Sci. Nat., Zool.’, sér. 2, tom. 10, pp. 193-221, 4 pls. Minchin, E. A. (1912).—‘ An introduction to the study of the Protozoa with special reference to the parasitic forms.’ London: ed. Arnold, 520 pp. Moglia, H. 8. (1910).—‘ Sul significato funzionale del pigmento nei gangli nervosi dei Molluschi Gasteropodi. Ricerche sperimentali”’, ‘ Archivio zoologico’, vol. 4, pp. 317-34, tav. 7-8. (Through ‘Zentralbl. f. Biochem. u. Biophys.’, vol. 10, p. 929.) Morner, K. A. H. (1887).—‘‘ Zur Kenntniss von den Farbstoffen der melanotischen Geschwiilste”’, ‘Zeits. f. physiol. Chem.’, Bd. 11, pp. 66-141. Moseley, H. N. (1873).—** On Actiniochrome, a colouring matter of acti- nias, which gives an absorption spectrum ’”’, ‘Quart. Journ. Micr. Sci.’, vol. 13, pp. 143-4. Moseley, M. A. (1877).—‘‘ On the colouring matters of various animals, and especially of the deep-sea forms dredged by H.M.S. Challenger’, ibid., vol. 17, pp. 1-23, 2 pls. Négre, L. (1906).—‘‘ Morphologie des pigmentophores de la peau des verté- brés et leurs rapports avec les cellules épidermiques ”’, ‘C. R. Soc. Biol. Paris’, tom. 60, pp. 26-8. ANIMAL CHLOROPHYLL 393 Nencki, M. (1897).—‘ Sur les rapports biologiques entre la matiére colorante des feuilles et celles du sang’’, ‘ Arch. Sci. biol. St.-Péters- bourg’, tom. 5, pp. 254-60. Newbigin, M. I. (1897).—** The pigments of the decapod crustacea ”’, * Journ. Physiol.’, vol. 21, pp. 237-57. Paladino, R. (1908).—“‘ Uber das spektroskopische und chemische Ver- halten des Pigmentsekretes von Aplysia punctata’”’, ‘ Zeit. Chem. Physiol. u. Path. (Hofmeister) ’, Bd. 11, pp. 65-70. —— (1910).—“ Zur Kenntniss der Leberpigmente der wirbellosen See- tiere’, ‘ Biochem. Zeits.’, Bd. 28, pp. 56-9. Palmer, L. 8. (1915).—** Xanthophyll, the principal natural yellow pig- ment of the egg yolk, body fat, and blood serum of the hen. The physiological relation of the pigment to the xanthophyll of plants ”’, ‘Journ. Biol. Chem.’, vol. 23, pp. 261-79. —— (1916).—“The physiological relation of plant carotinoids to the carotinoids of the cow, horse, goat, pig and hen”’, ibid., vol. 27, pp. 27-32. and C. H. KEckles (1914 a).—** Carotin—the principal natural yellow pigment of milk : its relation to plant carotin and the carotin of the body fat, corpus luteum, and the blood serum. I. The chemical and physio- logical relations of the pigments of milk fat to the carotin and xantho- phylls of green plants ’’, ibid., vol. 17, pp. 191-210. —— —— (1914 b).—“ Carotin. V. The pigments of human milk fat”’, ibid., pp. 245-9. —— and H. J. Kempster (1919 a).—* Relation of carotinoids to growth, fecundity, and reproduction in fowls’’, ‘ Journ. Biol. Chem.’, vol. 39, pp. 299-312, pl. 2. (1919 6).—* The physiological relation between fecundity and natural yellow pigmentation of certain breeds of fowls ’’, ibid., pp. 313- 29, pls. 3, 4. (1919 c).—‘** The influence of specific feeds and certain pigments on the colour of the egg-yolk and body-fat of fowls ’’, ibid., pp. 331-7. Parker, G. H. (1917).—‘* The movements of the tentacles in Actinians ’’, ‘Journ. Exp. Zodl.’, vol. 32, pp. 95-110. Parker, T. J.. and W. A. Haswell (1910).—‘ A Text-book of Zoology ’. London: Macmillan & Co., 2nd ed., 2 vols., 839+ 738 pp. Paul, J. H., and J. 8. Sharpe (1919).—* Relationship of lecithin to the growth cycle in crustacea’”’, ‘ Biochem. Journ.’, vol. 13, pp. 487-90. Peterson, H. (1913).—‘‘ Chlorophyll in caterpillars”’, ‘Nature’, vol. 91, p. 643; ‘Pharm. Journ.’, vol. 91, pp. 373-7. Phillips, R. W. (1911).—“‘ Algae ’’, ‘ Ency. Brit.’, vol. 1, pp. 585-98. Piettre, M. (1911 a).—‘*‘ Sur les pigments mélaniques d’origine animale ”, *C. R. Acad. Sci. Paris’, tom. 153, pp. 782-5. —— (1911 6).—‘ Sur les mélaines ”’, ibid., pp. 1037-40. 394 JOHN F. FULTON, JR. Pizon, A. (1898).—“‘ Sur la coloration des Tuniciers et la mobilité de leurs granules pigmentaires”’, ibid., tom, 129, pp. 395-8. (1901).—‘*‘ Origine du pigment chez les Tuniciers. Transmission du pigment maternel a l’embryon”’, ibid., tom, 132, pp. 170-2, Plimmer, R. H. A. (1915).—‘ Practical organic and Bio-Chemistry ’. London: Longmans, Green & Co., xii+635 pp., 1 pl. Potts, F. A. (1915).—‘‘ The fauna associated with the crinoids of tropical coral reef: with especial reference to its colour variations ’’, * Publ. Carnegie Instn., Wash.’, no. 212, pp. 71-96, 7 figs., 1 pl. Poulton, E. B. (1884).—‘‘ The essential nature of the colouring of phytophagous larvae (and their pupae) with an account of some experiments upon the relation between the colour of such larvae and that of their food-plants’’, ‘ Proc. Roy. Soc. London’, vol. 38, pp. 269-315. —— (1889).—‘ The colours of animals’. New York: Appleton (Internat. Sci. Ser.), xii+ 360 pp. —— (1892).—“‘ Further experiments upon the colour-relation between certain lepidopterous larvae, pupae, cocoons and imagines and their surroundings’, ‘Trans. Entom. Soc. London’, 1892, pp. 293-487, pls. xiv, xv. —— (1893).—“* Experimental proof that the colours of certain Lepi- dopterous larvae are largely due to modified plant pigments, derived from food ”’, ‘ Proc. Roy. Soc. London ’, vol. 54, pp. 41-2 (abstract) ; pp. 417- 30, pls. 3, 4. Prouho, H. (1887).—‘‘ Recherches sur Dorocidaris papillata et quelques autres Echinides de la Méditerranée”, ‘ Arch. Zool. exp.’, sér. 2, tom. 5, pp. 213-80, pls. 14-24. Przibram, H. (1913).—“ Griine tierische Farbstoffe ”’, ‘ Arch. ges. Physiol. (Pfliiger) ’, Bd. 153, pp. 385-400. Rabl, H. (1894).—“‘ Ueber die Herkunft des Pigmentes in der Haut der Larven der urodelen Amphibien ”’, ‘ Anat. Anz.’, Bd. 10, pp. 12-17. Rabuteau et F. Papillon (1873).—“‘ Observations sur quelques liquides de Vorganisme des Poissons, des Crustacés et des Céphalopodes ”’, “C. R. Acad. Sci, Paris’, tom. 77, pp. 135-8. Racovitza, E. G. (1895).—“ Sur le rdle des amibocytes chez les Annélides polychétes ’’, ibid., tom. 120, pp. 464-7. Rand, H. W. (1909).—‘‘ Wound reparation and polarity in tentacles of Actinians ’’, ‘ Journ. Exp. Zodl.’, vol. 7, pp. 189-238, pls. 1-2. Regnard, P., et B. Blanchard (1883).—‘‘ Note sur la présence de ’hémo- globine dans le sang des crustacés branchiopodes ’, “C. R. Soe. de Biol. Paris’, tom. 35, pp. 197-200. Reichert, E. T., and A. P. Brown (1909).—* Differentiation and specificity of corresponding proteins and other vital substances in relation to biological classification and organic evolution. The Crystallography of ANIMAL CHLOROPHYLL 395 Haemoglobins ”’, ‘ Publ. Carnegie Instn., Wash.’, no. 116, xix+ 388 pp., 100 pls. Reinke, F. (1906).—‘* Ueber die Beziehungen der Wanderzellen zu den Zellbriicken, Zellliicken und Trophospongien”’, ‘ Anat. Anz.’, Bd. 28, pp. 369-78. Roaf, H. E. (1921).—** Urochrome as a derivative of chlorophyll ”’, ‘ Bio- chem. Journ.’, vol. 15, pp. 687-8. Robertson, T. B. (1920).—‘ Physical chemistry of the proteins’. New York: Longmans, Green & Co., xv+ 483 pp. Rouget, C. (1859).—‘‘ Note sur l’existence de globules du sang colorés ”’, * Journ. de Physiol.’, tom. 2, pp. 660-70. Saint-Hilaire, C. (1896).—‘‘ Ueber die Wanderzellen im Darme der Echinoiden’’, (Ausz.) ‘Trav. Soc. imp. nat. St.-Pétersbourg’, vol. 27, livr. 1, no. 2, pp. 84-6. Scherer, J. (1841).—‘‘ Chemisch-physiologische Untersuchungen ”’, ‘ Anna- len d. Chemie u. Pharmacie’, Bd. 40, pp. 1-64. Schewiakoff, W. (1891).—‘ Bemerkung zu der Arbeit von Professor Famintzin iiber Zoochlorellen ’’, * Biol. Centralb.’, Bd. 11, pp. 475-6. Schmidt, W. J. (1919).—“‘ Ueber Chromatophoren bei Insekten”’, ‘ Arch. Mikr. Anat.’, Bd. 93, pp. 118-36, Taf. 5. Schmiedeberg, O. (1897).—‘* Ueber die Elementarformeln einiger Eiweiss- korper und iiber die Zusammensetzung und die Natur der Melanine ”’, ‘ Arch. f. exp. Path. u. Pharm.’, Bd. 39, pp. 1-84. Schunck, C. A. (1903)—‘“ The xanthophyll group of yellow colouring matters ”’, ‘ Proc. Roy. Soc. London ’, vol. 72, pp. 165—76, pls. 6, 7. Smith, G. W. (1913).—‘ Studies in the experimental analysis of sex. Part 10.—The effect of Sacculina on the storage of fat and glycogen, and on the formation of pigment by its host’’, ‘ Quart. Journ. Micr. Sci.’, vol. 59, pp. 267-95. Sorby, H. C. (1875 a).—‘‘ On the chromatological relations of Spon- gilla fluviatilis”, ibid., vol. 15, pp. 47-52. (1875 b)—“‘ On the colouring matter of Bonellia viridis”, ibid., pp. 166-72. (1876).—“ On the evolution of haemoglobin ”’, ibid., vol. 16, pp. 76-85. Steenboch, H., M. T. Sell, and M. V. Buell (1921).—“ Fat-soluble vitamine. VII. The fat-soluble vitamine and yellow pigmentation in animal fats, with some observations on its stability to saponification”, ‘Journ. Biol. Chem.’, vol. 47, pp. 89-101. Van Trigt (1918).—‘‘ A contribution to the physiology of the freshwater sponges (Spongilidae) ’’, ‘Proc. Kon. Akad. Wetensch. Amsterdam ’, vol. 20, no. 7, 15 pp. (Also separate.) Leiden: F, J. Baill (1919), 260 pp. Verne, J. (1920).—“‘ Sur la nature du pigment rouge des Crustacés ’’, *C. R. Soc. Biol. Paris’, tom. 83, pp. 963-4, 988-90. 396 JOHN F. FULTON, JR. Vies, F. (1919).—‘‘ Remarques sur les propriétés spectrales de quelques hémoglobines d’Annélides”, ‘C. R. Acad. Sci. Paris’, tom. 169, pp. 303-5. Willstatter, R., und W. Meig (1907).—*‘ Untersuchungen tiber Chlorophyll. IV. Ueber die gelben Begleiter des Chlorophyll”, ‘Ann. d. Chem.’, Bd. 355, pp. 1-28. Willstitter, R., und A. Stoll (1913).—‘ Untersuchungen iiber Chlorophyll’. Berlin: Springer, 424 pp. Winter, F. W. (1907).—‘‘ Zur Kenntnis der Thalamophoren. I. Unter- suchung iiber Peneroplis pertusus (Forskal)”, ‘ Arch. f. Prot.’, Bd. 10, pp. 1-113, Taf. 1, 2. Winterstein, H. (1909).—‘* Zur Kenntnis der Blutgase wirbelloser Seetiere”’, ‘ Biochem. Zeits.’, Bd. 19, pp. 384-424. Wurm (1871).—‘ Tetronerythrin, ein neuer organischer Farbstofi”’, * Zeits. f, wiss. Zool.’, Bd. 21, pp. 535-7. Further observations on Chromosomes and Sex- determination in Abraxas grossulariata. By the late Professor L. Doncaster, F.R.S., University of Liverpool. Tue following paper was left by the late Professor Doncaster in an incompleted condition, and, as I was one of his assistants during his last year’s work, it has been entrusted to me to prepare for publication. The paper as it now stands is exactly as Professor Doncaster left it, except that I have added the account of the staining methods used to test the nature of the elimination plate. I was familiar with his staining methods because I was myself testing for chromatin in some entirely different work when Professor Doncaster was testing the elimination plate, and he kindly passed on all his stains to me as he used them, showed me his preparations, and discussed the whole matter with me. In his rough notes I find a full account of all the stains used, and carefully labelled figures showing the results obtained from the different staining methods: there is, therefore, no uncer- tainty about the facts which I have added. A summary of the paper included amongst Professor Doneaster’s rough notes shows that he intended to add three other sections on ‘ Conjugation, &c., of polar nuclei’, * Binu- cleate eggs’, and ‘Gynandromorphs’. These sections were unfortunately not written even in note form, and therefore cannot possibly be produced; but the paper as it stands is of such obvious interest that its publication even in this very incomplete form is more than justified. Rutn C. BaMBER (Mrs. Bisbee). NO, 263 F f 398 L. DONCASTER FuRTHER OBSERVATIONS ON CHROMOSOMES AND SEx- DETERMINATION IN Abraxas grossulariata. In a previous paper (‘ Journ. of Genetics’, iv, 1914, p. 1) I deseribed the inheritance of a tendency to produce families consisting chiefly or entirely of females in Abraxas grossu- lariata, and attempted to correlate it with the behaviour of the chromosomes. It was found that females of the strain in which unisexual families occurred have fifty-five chromosomes as the somatic number, while all males and most other females — have fifty-six. In the maturation of the fifty-five-chromosome strain, twenty-eight chromosomes travel to one pole of the first polar spindle and twenty-seven to the other. Since all spermatozoa were found to have twenty-eight, if seemed evident that eggs with twenty-seven must be female-determin- ing, since the union of an egg having twenty-seven with a sperm having twenty-eight would give fifty-five, the number found in females of the strain in question, while eggs with twenty-eight meeting sperms with twenty-eight would give the fifty-six found in the male. Evidence was also given that in families in which great excess of females was produced, a majority of the eggs matured in such a way that twenty-eight chromosomes were extruded in the first polar nucleus, and twenty-seven remained in the egg-nucleus, and it was therefore inferred that the condition in some families, in which only females were produced, was caused by the invariable extrusion of the twenty-eighth chromosome in the polar body, leaving all eggs with only twenty-seven, and therefore female-producing. This hypothesis was supported by the observations of Morgan on Phylloxera, in which one chromosome is always extruded in the polar body of male-producing eggs, although it is already determined in some other way that these eggs will become males. When the paper referred to was published, I had been able to obtain no completely conclusive evidence that in families consisting wholly of females all the eggs had only twenty-seven chromosomes in the egg-nucleus, and I spent the next two seasons in collecting material which it was hoped would give CHROMOSOMES AND SEX-DETERMINATION 399 an unequivocal result. The method adopted was to pair females belonging to all-female families, allow them to lay eggs as far as possible under observation, and to preserve the first 50 or 100 eggs at an age (about two hours) when the maturation-divisions would be in progress. The moths were then allowed to continue laying, the eggs counted, and reared either to the imago or to larvae in which the sex could easily be determined by dissection. Some of these families produced both sexes, others either females only or females in great excess. ‘The preserved eggs of families which proved all-female were then sectioned, and counts made of the chromosomes in the polar division-spindles. By the summer of 1915 I had already enough material to show that the hypothesis put forward in the 1914 paper was almost certainly incorrect, and since work in connexion with the war prevented the immediate continuation of the investiga- tion I published a preliminary note in a letter to ‘ Nature’ (June 10, 1915) in which I wrote as follows: ‘I have now examined the eggs of several such families [i.e. all-female families], and find, contrary to expectation, that the equatorial plate of the mner spindle contains twenty-eight chromosomes about as frequently as twenty-seven. The new material con- firms the observation that twenty-seven occur in one spindle and twenty-eight in the other, but it seems to make it certain that the presence of twenty-eight chromosomes in the inner spindle does not necessarily cause the production of a male— at least in the strain which produces all-female families. A possible explanation of the anomaly is that in all-female families a chromosome is eliminated at a later stage, but at present I have no direct evidence for this.’ From that time to the summer of 1919 the work was inter- rupted, but enough material had been collected to provide the required observations, and examination of the sections confirms the account shortly given in the letter quoted. There are two questions at issue: (1) whether the all-female families are so because all the fertilized eggs are truly female, or whether they arise through non-viability of male zygotes; (2) if all Fi2 400 L. DONCASTER zygotes in such families are female, whether the egg-nuclei before fertilization contain always twenty-seven chromosomes, or sometimes twenty-eight. EVIDENCE THAT ALL-FEMALE FAMILIES ARE NOT CAUSED BY NoN-VIABILITY OF MALE ZyGores. In the earlier papers a number of families were recorded in which considerably over half the eggs were reared either to imagines or to larvae in which the sex was definitely deter- minable, but there still seemed some slight chance that all- female families might arise through death of male zygotes _at an early age. This, however, seems to be definitely excluded by the results of later experiments, as is shown in Table I, which gives a list of the all-female, or almost exclusively female families in which at least two-thirds of the eggs were reared to larvae or adults of ascertainable sex. TaBLe I. Family. Number of Eggs. Eggs Hatched. Larvae or Imagines. ) o Poor, ¢. : 97 97 66 1 1912.8 (2) . : 40 31 28 — 1912.29. : 110 110 77 = 1912.29 8B . : 83 82 66 — 1913.30. i aT ii 54 2 1914.9 ‘ , 37 not recorded 28 = 1914.16. : 47 - 37 — 1914.18. : 63 f 58 — 1914.7 ‘ : 14 " 9 l 1914.28. 62 5 45 — 1916.7 ; : 61 46 42 -- 1916.9 ; é 27 22 22 — 1916.10. ; 62 56 39 4 In view of the fact that in most cases almost all the eggs hatch, and that as soon as the larvae are old enough to be dissected the sex is already clearly distinguishable, these results make it practically indubitable that the all-female families do not arise in consequence of the death of male eggs or larvae. But the matter can be tested in another way,! which makes this conclusion doubly sure. ‘ [t seems as though a paragraph has been omitted here, but there is no CHROMOSOMES AND SEX-DETERMINATION 401 The evidence just given seems to prove beyond the possibility of reasonable doubt that all zygotes in the all-female families are female, and that these families do not arise by the death of male zygotes. ‘The problem then presents itself whether all eggs of these families before fertilization contain twenty- seven instead of twenty-eight chromosomes. In the letter to ‘ Nature ’ referred to I announced that I found evidence that this was not so, and further work has confirmed this conclusion. In 1914 I preserved eggs from four pairings, of which the egas subsequently laid yielded only females. The data with regard to these families are as follows, excluding the eggs preserved for microscopic examination : No. of No. of Eggs Males Females Family. Eggs. Hatched. Reared. feared. 14.9 37 not recorded 0 28 14.22 74 nearly all 0 28 14.28 62 62 0 45 14.29 38 38 1? 21 It will be noticed that in families 14.9 and 14.28 over two- thirds of the eggs kept for rearing were reared to imagines (or in 14.28, thirty-six imagines and nine pupae). The eggs of these same families preserved for microscopic examination gave the following chromosome counts in the equatorial plates of the second maturation division. 14.9. In the inner spindle 27, in the outer 28—four cases recorded as ‘ good’. In the inner spindle 27, in the outer 28—two cases recorded as ‘ probable ’. trace of it in the manuscript unless it be the following, which I find on a page of note-paper along with the manuscript: ‘Summaries to 1916 show that all-female families are not due to mortality, due to “* lethal” or other causes, of male. -Apart from such cases as 14.16, and 14.18 (37 and 58 females from 47 and 63 eggs),* the fact that in all-female families in which over 50 per cent. of the eggs are reared to imagines there are twice as many females per cent. of eggs (64:6 per cent.) as compared with per- centage of females in bisexual families (32-3 per cent.) proves this.’ * See table given on previous page, 1914.16 and 1914.18. 402 L. DONCASTER ‘In the inner spindle 27, outer not countable—one case recorded as ‘ good ’. In the inner spindle 28, in the outer 27—three cases recorded as ‘ good’. In the inner spindle 28, in the outer 27—four cases recorded as * probable ’. In the inner spindle 28, outer not countable—one case recorded as ‘ good’. In the outer spindle 27, inner not countable—one case recorded as ‘ good ’. Total, seven cases with the inner spindle containing 27, five of these being ‘ good’ cases in which there is no reasonable doubt as to the number, and nine cases, in which the inner spindle has 28, four of these being ‘ good ’ cases. 14.22. The counts were less satisfactory; they gave three cases in which the inner spindle had 27 or the outer 28, and five in which the inner had 28 or the outer 27, but in only one could both inner and outer be counted with con- fidence in the same egg; in this egg the inner spindle had 28 and the outer 27. 14.28. In the inner spindle 27, or the outer 28—four cases (three in which both plates could be counted with fair certainty). In the inner spindle 28, or the outer 27—four cases (one countable in both plates). 14.29. In the inner spindle 27, or the outer 28—three cases, in two of which both inner and outer plates were countable with fair certainty. In the inner spindle 28, in the outer 27—one case (fairly good). Although the number of ‘ good ’ counts in which the chromo- somes could be counted with confidence in both inner and outer spindles is not large, some of them, especially in family 14.9, are so certain that no doubt can remain that in many eggs of all-female families the inner spindle contains twenty-eight chromosomes, and adding up all counts in the four families we get seventeen cases in which the inner spindle had twenty- CHROMOSOMES AND SEX-DETERMINATION 403 seven (or the outer twenty-eight) and nineteen with the converse arrangement. In the bisexual family 14.35, in which 15 0707 and 19 99 were reared from forty-six eggs, four eggs were found in which the inner spindle had twenty-seven, the outer twenty-eight chromosomes, and four with the converse arrange- ment (all ‘ good’ counts including both spindles of each egg), so it does not appear that the all-female families have twenty- seven in the inner spindle with any greater frequency than in bisexual families of the same stock. It seems evident from the facts given above that the deter- mination of sex in the fifty-five-chromosome strain of Abraxas grossulariata does not depend on the passage of the odd chromosome to one or other pole of the first polar division. At the same time, since females of this strain have fifty-five chromosomes in their diploid nuclei and males have fifty-six, a chromosome must be eliminated at some stage from those eggs in which twenty-eight travel to the inner pole of the first polar spindle. Attempts to find a chromosome which does not divide in the second maturation division have not been success- ful, and it seems clear that the elimination does not occur at that stage. Only two possibilities remain: either a chromo- some is eliminated at some division after fertilization—pre- sumably the first segmentation division, or the odd chromo- some must degenerate so that the twenty-eight chromosomes present in about half the eggs at the mner pole of the first polar spindle are reduced to twenty-seven by the degeneration of one of them. Neither possibility seems hkely on general srounds, but there are some facts which make the hypothesis of the degeneration of a chromosome less entirely improbable than would appear at first sight. These will be discussed in the next section. With regard to the hypothesis of the elimina- tion of a chromosome in the first segmentation division, I can only say that I have not succeeded in obtaining figures in which the chromosomes in this division can be accurately counted ; in the few segmentation divisions present in my material the chromosomes tend to become aggregated into small groups, apparently of two or three, so that counts give numbers not 404 L. DONCASTER much greater than the haploid complement (twenty-eight). Probably mitotic figures embedded deeply in the yolk are fixed less rapidly than the maturation mitoses near the surface of the egg, with the result that observations on the number and behaviour of the chromosomes in the segmentation divisions become untrustworthy. ‘CHROMATIN ELIMINATION’ IN THE MATURATION DIVIsIons OF THE Kaa. In my 1914 paper? I mentioned that ‘ during the first polar division, a mass of granules which stain deeply with iron haematoxylin is left in the equatorial plate as the chromosomes travel to the poles ’ (fig. 14 of that paper). No further investiga- tion was made at the time on the nature or mode of origin of these granules, but in a paper published almost simultaneously ” Seiler describes them in considerable detail in the eggs of the moths Phragmatobia fuliginosa, Orggia antiqua, Lymantria monacha, and L. dispar. He gives evidence that these granules are separated from the chromo- somes in the early anaphase of the first polar division, and maintains that in favourable cases it is possible to see that each chromosome, as it divides, leaves behind on the equator of the spindle a chromatin mass which for a time at least preserves its identity, so that im sections of a mitosis in anaphase cut at right angles to the axis of the spindle it is possible to see three plates each containing the same number of chromatin bodies similarly arranged—the two anaphase groups of chromo- somes and between them an ‘elimination plate ’ consisting of chromosome-like bodies having the same number and arrange- ment as the chromosomes in the true chromosome plates. Careful search among my preparations—both old ones and new sections made for the purpose—has not revealed the existence of plates with such definite, chromosome-like granules in Abraxas, but in other respects my sections, when stained 1 Doncaster, L., ‘ Journ. of Genet.’, 1914, p. 1. 2 Seiler, J., ‘ Archiv fiir Zellforschung.’, xiii. Band, 2. Heft (p. 159). Leipzig und Berlin, 1914. 7 CHROMOSOMES AND SEX-DETERMINATION 405 with iron haematoxylin, give very nearly the same series of figures as are represented by Seiler. I have a few cases in very early anaphase (just after metaphase) in which each chromosome seenis to be leaving behind, as its halves diverge on the spindle, a mass of staining substance (ef. Seiler’s figs. 19- 22), and in later anaphase there is always an equatorial plate of staining granules lying across the middle of the spindle. Not infrequently these granules are elongated, so as to appear like short threads, and some or all of them seem to lie on or in the spindle fibres. Towards the end of the anaphase they generally form a plate of fine-stamed dots, of varying size, and always more numerous than the chromosomes, as if they had become broken up and scattered. During the second division they sometimes become aggregated into a sort of network (cf. Seiler’s fig. 35), or they may apparently have become more finely divided and comparatively incon- spicuous. Like Seiler, I find great variation among different polar mitoses in respect of the amount of this eliminated substance. In some spindles there is a dense equatorial mass, staining with iron haematoxylin almost as deeply as the chromosomes around the poles. In others the granules are much less con- Spicuous, in others again so few and small as hardly to be noticeable. The amount of staining matter in the ‘ elimina- tion plate ’ varies in different eggs of the same female, and even in eggs mounted on the same slide, though on the whole it is more abundant in the eggs of some females than in those of others. It is important to notice, however, that the apparent amount varies with depth of staiming, and when sections of several eggs are mounted together on a slide it may happen that some spindles are fully washed out, so that the chromo- somes alone remain clearly stained, while a spindle ina neigh- bouring egg may retain so much stain as to be useless for the study of chromosomes. This variability probably arises from differences of fixation due to variation in the penetrability of the egg-shells to the fixative, and therefore it is not impossible that the variation in the apparent amount of eliminated sub- 406 L. DONCASTER stance may be in part at least due to the technique of fixing and staining. Seiler has no doubt that the stamed matter in the elimina- tion plate is chromatin, and reviews the literature of maturation divisions of insect eggs, and also such examples of chromatin elimination as those seen in the segmentation of Ascaris and Miastor, in order to discuss the significance of the pro- cess. He does not, however, discuss at all fully the question whether the substance eliminated is in fact chromatin, or if it is, whether it is of the same nature as the chromatin of the anaphase chromosomes. His account of his staining methods is meagre— Gefirbt wurde vorwiegend mit Heidenhains Hisenhimatoxylin und Kontrollfairbungen wurden mit Kern- farbstoffen vorgenommen. Als Plasmafarbstoff verwendete ich $.-Fuchsin’’. Unless the elimination process is in reality an artefact, which seems very unlikely in view of the almost invariable presence of staining granules in the equator of the spindle and the definite manner in which they appear to be left behind by the diverging chromosomes, it appears to be of considerable importance to determine the true nature of the eliminated substance, for if it be chromatin, it seems not impossible that the process may supply the clue to the anomaly presented by the presence of twenty-eight chromosomes in eggs which nevertheless yield females. If the chromosomes do in fact leave behind on the polar spindle a considerable part of their substance, it is at least conceivable that the sex- chromosome, in the eggs of all-female broods, eliminates so much that it becomes functionless as regards sex-determination, and that, having lost so large a part of its substance, it ceases to function and disappears, so that im the oogonia only fifty- five instead of fifty-six can be counted. With the object of determining whether the elimination plate does or does not consist of chromatin I stained eggs with a number of com- binations of stains, the more important of which were as follows :? [All the sections used had been previously stained with iron haematoxylin and were decolourized with acid alcohol.) ! Professor Doncaster’s manuscript ends here. CHROMOSOMES AND SEX-DETERMINATION 407 I. Ehrlich’s Triacid Stain. Sections were immersed for eighteen hours in the stain, blotted, and passed through absolute alcohol and xylol into balsam. The chromosomes and the elimination plate were stained purple, the surrounding protoplasm brown. Il. Safranin and Lightgreen. Sections were placed in safranin for from twelve to twenty- four hours, followed by lightgreen for one to two minutes. The chromosomes were stained a bright red, the elimination plate a lighter red in some cases and in others green with a distinct admixture of red. ‘The spindle-fibres were green. Both the above methods gave very clear results which strongly suggest that chromatin was present in the elimination plate. III. Mann’s Methyl Blue Kosin. Sections were stained for a few minutes only. This method gave very erratic results. In most cases the spindle-fibres were blue, but whereas in some sections the chromosomes and elimination plate were also blue, in others the chromosomes were purple and the elimination plate red ; and in others again the chromosomes were red with a bluish tint here and there, and the elimination plate purple. In spite of the varying results obtained with this stain it is clear that in any given section there is a very close corre- spondence between the chromosomes and the elimination plate. IV. Ehrlich’s Haematoxylin. Sections were stained for eighteen hours with Ehrlich’s haematoxylin, differentiated for from one to two minutes in acid alcohol, washed in 70 per cent. alcohol and counterstained with eosin in 90 per cent. alcohol for about one minute. The spindle-fibres always stained pink, the chromosomes were usually black, and the elimination plate pink with grey dots ; sometimes, however, the chromosomes were a bright pink and the elimination plate a paler pink, and at other times the chromosomes and the elimination plate were purple. Here again, as with Mann’s methyl blue eosin, the chromo- 408 L. DONCASTER somes and the elimination plate in any given section corre- spond very closely in their stainimg properties, although in different sections very different results were obtained from the same combination of stains. V. Borax Carmine and Picro-indigo-carmine. Sections were stained in borax carmine for forty-eight hours, followed by picro-indigo-carmine for ten minutes. ‘he chromosomes were found to be dark red, and the elimination plate and spindle-fibres yellowish. Although this method did not appear to give support to the view that chromatin is present in the elimination plate, it does not disprove that hypothesis, for in some sections even the chromo- somes themselves were barely stained with the carmine, so that it is not surprising to find the elimination plate unstained even though it may contain chromatin, for this eliminated chromatin would almost certainly be undergoing disintegration. In sections stained by other methods as given above, it was not unusual! to find that the elimination plate was unstained, even though it had previously stained deeply with iron haema- toxylin. These staming experiments, although not conclusive, give a considerable weight of evidence in support of the hypothesis that there is a certain amount of chromatin left behind, on the equator of the spindle, by the chromosomes when they move apart at anaphase. If this be true ‘ it is at least conceiv- able that the sex-chromosome, in the eggs of all-female broods, eliminates so much that it becomes functionless as regards sex-determination ’,' and that here may lie the explanation of the production of all-female families from eggs some of which contain twenty-seven and others twenty-eight chromosomes. From conversation with Professor Doncaster, as well as from his own argument in this paper, I know that this was the conclusion at which he himself had arrived. R.G. 1 p. 406. The Infra-cerebral Organs of Peripatus. By William J. Dakin, D.Se., F.Z.S., F.L.S., Derby Professor of Zoology, University of Liverpool. With 4 Text-figures. ATTacHED to the ventral surface of the supra-oesophageal ganglion of Peripatus and hanging therefrom are two small vesicles. They were discovered as far back as 1853 before Peripatus was regarded as an Arthropod, and Grube, their discoverer, considered them to be auditory organs (8). In 1883, Balfour, in his classical description of the anatomy of Peripatus capensis (2), described the structure of these vesicles, and after a statement detailing their shape and position added that each consisted mainly of ganglion cells. He continued with the followmg words: ‘In its interior is a cavity with a distinct bounding membrane... . . At its free end is placed a highly refractive, somewhat oval body, probably forming what Grube describes as a dark spot, half embedded in its substance, and kept in place by the sheath of nervous matter surrounding it. It is difficult to offer any interpretation of the nature of this body. It is removed considerably from the surface of the animal, and is not, therefore, so far as I can see, adapted to.serve as an organ of hearing.’ Three years after the appearance of Balfour’s paper, Kennel (11) followed the development of the infra-cerebral vesicles and discovered that they were apparently homologous with certain sroups of cells situated between the legs—and known as» the ‘ ventral organs ’ (see T'ext-fig. 1). This was confirmed by Sedgwick (15) in 1888, and the 410 WILLIAM J. DAKIN discovery has very considerably modified the views of the function and meaning of the infra-cerebral vesicles. If any definite theory of the function of these structures can be said to be generally accepted, it is that they represent the ectoderm from which the nervous system arose in the embryo, and in a recent paper by Duboseq the suggestion is made that the infra-cerebral vesicles remain, even in the adult stage, structures for the renovation or increase mm size of the supra-oesophageal ganglia. Cells are supposed to be eut off from the vesicle cells and to migrate into the ganglia, there to become either new nerve-cells or supporting cells. Trxt-Friac. 1. Peripatoides occidentalis: section of so-called ventral organ, VO; ©, cuticle; 5, ectoderm ; M, muscles of body-wall ; s, strand connecting ventral organ with lateral nerve-cord. The present note has been written because in several of our best preparations from the head of Peripatoides occidentalis, Dendy, of Western Australia (5), the histology of the organs in question is not the same as that illustrated by Duboseq (6). And a little more may be said in explanation of the presence of these curiously definite structures. According to Duboseq (who examined Opisthopatus cinetipes, Purcell) one can distinguish in these organs two distinct regions, (a) the vesicle, (b) the ganglion intermediare. ‘The latter is the part which former writers have called the stalk or peduncle of the vesicle. The term ‘ ganglion intermediare ’ is unsuitable, especially since it appears that Duboseq himself INFRA-CEREBRAL ORGANS OF PERIPATUS 411 is not sure whether the cells of this part are nerve-cells or merely supporting cells. There is nothing that might be termed the ganglion inter- mediare in Peripatoides occidentalis (Text-figs. 2 and 3). The vesicle is described by Duboseq as containing nothing within the cavity but serous fluid, there being no refringent oval body of the kind referred to by Balfour. (Unfortunately Balfour’s figure gives no idea of the histology of the infra- cerebral vesicles.) Now we have found occasionally that TEXT-FIG. 2, pe Hf, MLL Wis }, buon hehe 3 est Ae iy) nas oD a 8 SEs ee SS Boe ON Infra-cerebral vesicle, with enclosure, from adult Peripatoides occidentalis. 530. «a, ganglion; v, infra-cerebral vesicle. bodies do occur within the cavity, reminding one of Balfour’s description, and the Text-fig. 2 is from the best preparation of this character. It is part of a transverse section through the head. ‘The structure is referred to below. The infra-cerebral organs appear in dissections to hang from the supra-oesophageal ganglion by short stalks. In sections, however, they appear more closely attached. The difference in appearance is due to the transparency of the suspending membrane which is the structureless, almost non-staining, sheath of the ganglion. In Peripatoides occidentalis there is generally a region that one might term the peduncle, within which are a few scattered nuclei and a small number 412 WILLIAM J. DAKIN of delicate fibre-like strands. They might be nerve-fibres or, on the other hand, merely processes of non-nervous cells. In young and small Peripatoides (Text-fig. 3) the conditions are somewhat different, however, and the walls of the vesicles are not so distinctly separated from the cerebral ganglia (see T'ext-figs. 2 and 3). In the adult the infra-cerebral vesicle is covered by the neurilemma or sheath of the supra-oesophageal ganglion, and TEXT-FIG. 3. Infra-cerebral organ from young Peripatoides occidentalis. x 400. this layer almost cuts it off from the latter. According to Saint-Rémy (14) and Duboseq (6) the sheath is pierced by numerous pores, through which bipolar cells are to be seen migrating into the brain. This is hardly the case in the adult Peripatoides, as the figure shows. There are only a few fibres passing from the vesicle into the supra-oesophageal ganglion and but a few nuclei occur here and there. The cells of the vesicle itself are not of the same depth throughout. Ventrally the walls are thin whilst laterally they are thick, and the cells are slender, so that the nuclei lie at different levels. The nuclei resemble closely those of the ganglion cells of the brain mass. INFRA-CEREBRAL ORGANS OF PERIPATUS 413 The cavity of the vesicle is most usually empty, but some- times contents are present, and the most conspicuous example of this kind has been figured (Text-fig. 2). In this specimen there is a non-granular mass surrounded by a number of concentric lamellae—almost like a decalcified concretion. The occasional presence of an enclosed body is very interesting. TEXxtT-FIc. 4. Diagram illustrating development of infra-cerebral vesicles in Peripatus occidentalis, At first it was thought to be due to fixation, but there is no reason to believe that such is the case. Evidently the vesicle cells may sometimes secrete into the central cavity. Thus a feature recorded by Balfour has again been found and in another species. The few writers who have mentioned the infra-cerebral vesicles since the date of Balfour’s paper seem to have doubted its occurrence. No cilia are found within the vesicle nor is there any very definite liming membrane. NO. 263 Gg 414 WILLIAM J. DAKIN The development of these infra-cerebral vesicles is now well known. They arise as invaginations of the ectoderm (Text- fig. 4) which is concerned in the formation of the supra- oesophageal ganglia, and at first they are open to the exterior (to the buccal cavity or near it), so that at this stage the supra- oesophageal ganglia possess cavities which are open below. The cavities become closed and then, whilst increase in size of the ganglia takes place, they remain almost of the original size (a slight decrease takes place if anything), and the surrounding cells, which are indistinguishable from the ganglion cells, become separated and pinched off from the brain mass, until finally two small distinct vesicles lie appended as we have seen. The diagrams illustrate how this takes place. One other feature of considerable importance remains to be stated. In the adult one may occasionally find cells in the infra-cerebral body undergoing mitotic division. This was first described by Duboseq. We find, however, that the number is much reduced as the animal becomes larger, and they are only occasionally found in the full-sized specimens. In the small Peripatus, not long born, they are more numerous (see Text- fig. 8). It may be stated here that the same feature is to be met with in the so-called ventral organs (Text-fig. 1). This fact has not been recorded before and it completes the resemblance between the ventral organs and the infra-cerebral vesicles. There can be little doubt of their homology. DISCUSSION. The infra-cerebral vesicles of Peripatus were once con- sidered to be sense organs concerned in hearing. Probably this was by analogy with the little vesicles often found close to the supra-oesophageal ganglia in the Polychaets and certain other Invertebrata and once termed Otocysts. They are now usually regarded as Statocysts or organs of orientation. Duboseq concludes, however, that in Peripatus they are not sense organs, nor glandular structures, but that in the adult as well as in the embryo they are organs for the production of either nerve-cells or neurogloea cells (supporting cells). INFRA-CEREBRAL ORGANS OF PERIPATUS 415 The vesicles known as statocysts or otocysts in the Inverte- brates are still of questionable function in many cases. ‘This is particularly so in the case of certain Nemertines (Metanemer- tines) (8), where the walls of the vesicles are surrounded by the ganglion cells of the brain mass and no cilia are present. In fact they are not unlike the infra-cerebral vesicles of Peripatus. On the other hand, in Molluses such as Pterotrachea, where each statocyst contains a statolith supported on bunches of cilia, the circumstances are altogether different, as experiment has shown. Amongst Polychaet worms statocysts are known in Sabellidae, some ‘l'erebellidae, Arenicola, Aricidae, and some Alciopidae. In some cases cilia are found within the vesicles, and statoliths are present (either secreted, or consisting of sand grains from the exterior). In the Arenicola species, however, the state of development of the ‘ statocysts ’ varies within very wide limits and it is difficult to express any opinion about the function of these organs. They appear to develop from invaginations of the ectoderm, but there is not the close connexion with the development of the cerebral ganglia which is so characteristic of Peripatus. Are these vesicles homologous? It is interesting to look at the condition of things amongst the Tracheata. In none of the Tracheata do the organs of orientation take the form of statocysts associated with the supra-oesophageal ganglion. But in the development of the supra-oesophageal ganglion of the Myriapoda it is certainly very striking that pit- like depressions occur on the ventral surface which afterwards become closed vesicles and later disappear (10). The same thing is true of the Insecta and Arachnida (1 and 12). In the Crustacea there is, so far as I am aware, no evidence of pit-lke depressions of this kind during development. Stato- cysts are found, but these are not at all homologous with the organs we are considering and occur in very different situations. Curiously enough, there is a striking exception. Thus, according to Claus (4), two otocysts are found connected with the cerebral ganglion in certain Amphipoda—the Platyscelidae. The same author mentions two vesicles as of similar function in the Gg2 416 WILLIAM J. DAKIN brain of a Copepod—Eucalanus attenuatus (Dana), but Esterly (7) considers these to be optical in function and apparently their structure is quite different from that of the other brain vesicles we have dealt with. We would suggest, therefore, that the infra-cerebral vesicles of Peripatus are homologous with the cephalic pits of other Tracheate embryos. In these cases the pits become closed off, the walls become parts of the cerebral ganglion, and the cavities disappear altogether. In Peripatus, on the other hand, the vesicles remain, but they are gradually constricted off from the rest of the supra-oesophageal ganglion. The adult condition in Peripatus is, then, an embryonic stage in the Myriapoda. In the adult it is probable that the infra-cerebral vesicles serve no special function—they are not really ‘ organs’ at all—they may still be regarded as parts of the supra-oesopha- geal ganglion. Possibly their wall contains a few ganglion cells from which fibres pass into the deeper parts of the brain mass. The occasional presence of bodies within the cavity is interesting, but this suggests nothing more than the ectodermal origin of the cells of which the vesicle is composed, and the tendency for the secretion of a chitinous cuticle. In the earlier stages before growth is complete this portion of the supra-oesophageal ganglion retains some of its former power of growth and continues to give rise to cells by division (as observed by Duboseq), but it is not a special organ for this purpose and loses its function in the adult. Whether this character, apparently common to the Tracheates, is homologous with the statocysts found occasionally in the worms is another matter—quite impossible as yet to decide. It has been affirmed, however, that in certain Annelids (Lopado- rhynchus) the supra-oesophageal ganglia develop in connexion with ciliated pits, which degenerate somewhat afterwards (18). This is very suggestive. 14. 15. INFRA-CEREBRAL ORGANS OF PERIPATUS ALT BIBLIOGRAPHY. . Balfour.—‘‘ Notes on the Development of the Araneina’’, ‘ Quart. Journ. Micr. Sci.’, 1880. ‘** Anatomy and Development of Peripatus capensis”’, ibid., 1883. . Burger—‘‘ Nemertini”’, ‘ Bronn’s Tierreich’, Bd. 4, Suppl., 1897— 1907. . Claus——‘ Die Platysceliden’, Vienna, 1887. . Dakin—‘‘ Onychophora of Western Australia’’, ‘ Proc. Zool. Soc.’, 1920. Duboseq.—‘‘ Notes sur Opisthopatus cinctipes”’, ‘ Archiv. de Zool. Exper.’, vol. 59, 1920. . Esterly—‘‘ Light Recipient Organs of Eucalanus ’’, ‘ Bull. Mus. Comp. Zool. Harvard ’, vol. liii, 1908. . Grube.—‘‘ Unters. itb. d. Bau von Peripatus edwardsii”’’, ‘ Arch. f. Anat. u. Phys.’, 1853. Heathcote.—‘‘ Post-embryonic Development of Julus_ terrestris’’, * Phil. Trans. Roy. Soc. Lond.’, 1888. . Heymons.—*‘ Entwick. d. Scolopendra ”’, * Zoologica’, Heft 33, 1901. . Kennel—‘‘ Entwick. von P. edwardsii’’, ‘ Arch. Zool. Inst. Univ. Wiirzburg ’, Bd. 7 and 8, 1885 and 1886. Kishinouye——‘‘ On the Development of Araneina”’, ‘ Journ. Coll. Sci. Imp. Univ. Tokio ’, vol. 4, 1891. Kleinenberg —‘*‘ Entsteh. des Annelids a. d. Larve von Lopado- rhynchus ”’, ‘ Zeit. wiss. Zool.’, Bd. 44, 1886. Saint-Réemy.—* Contribution a l’étude du cerveau chez les Arthropodes Trachéates ’’, ‘ Archiv. de Zool. Exper.’ (2), tom. 5, Suppl. Sedgwick—‘‘ Development of the Cape Species of Peripatus’’, ‘ Quart. Journ. Micr. Sci.’, vol. xxvii, 1888. va imi ta A ay if ; “a . 1} i= i 4 ’ s » ' A si\ Weal livia. we iinfaagphes ae ’ ' ‘ i"? weet win ut Ss A a ignes 7 : ia » Ta? {# we -s 2 | = ; ie Ove ee . i ‘ : 8 A Critical Study of the Facts of Artificial Fertilization and Normal Fertilization. By J. Gray, M.A., Fellow of King’s College, Cambridge. With 1 Text-figure. Facts concerning the process of fertilization and of artificial parthenogenesis have steadily accumulated during the past thirty years, and although numerous suggestions have been put forward as partial explanations of this imposing mass of experi- mental evidence, yet there are only two theories which claim to give an adequate picture of even a majority of the known facts. These theories we owe to J. Loeb (24) and to F. R. Lillie (20). According to Loeb, the activation of an unfertilized egg is effected by the introduction into the egg of two substances, (i) a specific cytolysin, which brings about the destruction of the surface layer of the egg, and (ii) a substance which limits or controls the destructive influence of the cytolysin. On the other hand, Lillie holds that the union of the egg and sperma- tozoon is only possible in the presence of a specific substance or fertilizin which is secreted by the unfertilized egg; if all three elements are present fertilization and normal develop- ment take place. Loeb’s theory is based upon the facts of artificial partheno- genesis ; Lillie’s theory is based upon the behaviour of the normal gametes. It is not surprising to find that each theory encounters its chief difficulties when confronted with the facts which constitute the main argument of its rival. Both theories are essentially chemical, although the door is left open, at rare intervals, to the intervention of physical factors. In 1915 (8) I suggested that although the theories urged by R. 8. Lillie (21) 420 J. GRAY and by McClendon (25) were inadequate, yet the facts appeared to indicate that the activation of the egg, by a spermatozoon, or by artificial parthenogenetic agents, was essentially a physical rather than a chemical process. It now seems possible to put forward a more comprehensive scheme. The activation of a resting cell, by contact with another cell in a state of activity, is not limited to the reproductive cells. All contractile cells exhibit the same phenomenon ; if localized fibres at the surface of a large muscle are stimulated, the whole of the muscle is rapidly thrown into a state of activity ; the ciliated combs of Pleurobrachia illustrate the same fact (Gray, 10) ; also cells in contact with each other usually divide at the same moment. ‘There can be no doubt that such co-ordination of activity is due to the responsive cells them- selves, and is not due to any nervous or controlling influence. Thus spermatozoa in contact with each other rapidly acquire a synchronous rhythm; similar examples are readily found in the case of ciliary or muscular elements. There can be but little doubt that the influence of one cell upon the activities of its neighbours has a very profound bearing on the behaviour of the animal as a whole. There is, however, no reason to regard such co-ordination as essentially vital, since a ready parallel is found in inorganic systems. Ostwald (27) found that when a strip of chromium was placed in hydrochloric acid the hydrogen was evolved at regular periods ; each period of activation was followed by a period of inactivity. This periodic condition of activity and inactivity was quite regular for each strip of metal: different strips of metal were, however, characterized by periods of different length. If several such strips are placed in a bulk of hydro- chlorie acid, the periodicity of each strip exhibits itself; if, however, the strips are in contact with each other then all the strips exhibit the same uniform periodicity. The activation of a passive strip of iron by contact with an active piece of the same metal has been discussed by R. 8. Lillie (28), and has a close bearing on the present problem. The activation of a passive cell or metal by contact with an STUDY OF FERTILIZATION 491 active unit is invariably accompanied by an electrical disturb- ance; and there seems good evidence for the belief that the electrical change is the essential condition of activity. When an inactive unit comes into contact with an active unit, an electro-motive force is established between the two ; the active unit is electro-negative to the inactive unit, and if activation of the latter occurs, the state of negativity is not restricted to the region of contact but spreads from it all over the originally inactive surface. Such facts are, of course, well known in the case of fibres in the same muscle, but Kithne (18) showed that the action current of one muscle could stimulate another muscle if the two were in close electrical contact. Now the E.M.F. set up between two cells in contact depends on, and is an expres- sion of, the difference in the activity of the two units; the sreater the difference in activity the greater is the H.M.F’. set up on contact. In the opinion of the writer an application of the above principles to the problem of fertilization is not without value. In the unfertilized egg metabolic activity is reduced to a — minimum, and unless fertilization takes place the cell dies with- out any recovery from its inert condition. The spermatozoon, on the other hand, is radically different: it is exceedingly active and metabolism proceeds at a rapid rate (Cohn, 8). When the two cells come into contact it seems legitimate to conclude that an E.M.F. will be set up between the two, and if the conditions be right it is to be expected that some form of activity will be induced in the inert egg. If the process of activation be analogous to that of other cells, then the egg will be activated whenever the E.M.F. set up by contact with a spermatozoon reaches a certain minimum value in a minimum time. It is, therefore, not surprising to find that a certain minimum of activity on the part of a spermatozoon is necessary for fertilization. Mobility and proximity of the egg are not sufficient—a fact difficult to explain on any chemical concep- tion; there must be a definite and rather high degree of activity on the part of the sperm, and this degree of activity differs between individual spermatozoa. Glaser (7) observed 422, J. GRAY that the eggs of Arbacia punctulata can be activated by means of highly active minute Infusoria. It is, therefore, the degree of activity of the sperm which determines one condi- tion of fertilization and not its structure or chemical constitu- tion. Whereas the normal activity of a spermatozoon is usually adequate for the activation of eggs of the same species, yet it requires to be increased to an abnormal degree to fertilize the eggs of another species. Now the activity of spermatozoa can readily be controlled by the hydrogen-ion concentration of the medium (Gray, 9), and correspondingly it is found that the addition of hydroxyl ions removes the block which normally exists between the eggs of Strongylocentrotus and the sperm of Sphaerechinus. ‘These and similar facts are difficult to explain on the basis that fertilization depends on the existence of specitic chemical substances in the egg or sperm (see Loeb, p. 204). They appear to the writer to be less formid- able when subjected to physical arguments. According to the present physical hypothesis, if it be mechanically possible for the sperm of one species to gain contact with the egg of another, then activation will take place if the E.M.F. set up between the two cells reaches a certain critical value within a minimum time. Consider two species, A and B. When normal fertilization of egg A is effected by sperm A, let the E.M.F. be E, and let it be developed in unit time ; similarly when egg B is fertilized by sperm B let the E.M.F. at contact be Ea. Let Ei >H2; then when egg B comes into contact with sperm A, the E.M.F. will probably be more than enough to activate the egg; when egg A comes into contact with sperm B, no activation will occur unless the activity of the sperm is artificially increased, since otherwise the requisite E.M.F. will not be reached. Such irreciprocal hybridizations have already been mentioned and are by no means uncommon [see Vernon (28), Doncaster (5)}. Further, the conditions which affect the ease with which hybridization can occur are such as support the view that some such physical factors are involved, e. g. seasonal variation of gametes, degree of maturity, staleness or freshness of gametes, STUDY OF FERTILIZATION 428 hydroxyl ions, dilution of sea-water, &c. It is exceedingly difficult to apply the chemical conceptions of Loeb or of Lillie to such facts. It is obvious that the self-sterility of the gametes of Ciona can also be analysed by a similar physical argument to hybridization. Under normal conditions only one spermatozoon enters an egg. In view of the very large number of spermatozoa which may be in the immediate vicinity of the egg-surface at the moment of fertilization, it is almost inconceivable that any chemical change could be set up, and carried to a conclusion between the time that two successive spermatozoa touch the egg. Neither Lillie nor Loeb has offered what would seem to be a reasonable explanation of monospermic fertilization. Once more, the facts appear to be amenable to physical treat- ment. Assuming that the rate at which an electrical change can travel round the egg is approximately that at which it travels along a piece of smooth muscle then within 0-00001 sec. after the effective spermatozoon has made its contact with the egg, no other spermatozoon will have any effect : if, however, the eggs are treated in such a way as to reduce the rate of pro- pagation of an electrical disturbance, e.g. by incomplete anaesthesia (cf. nerve or muscle), then the wave will not pass ‘completely over the egg before other spermatozoa can effect contact with unaffected portions of the egg-surface, and poly- spermy will result. Hertwig (17) showed that unfertilized eggs treated with chloral hydrate and other anaesthetics were markedly polyspermic. The first visible sign that fertilization has occurred, is at the surface of the egg. In the case of the echinoderm or annelid egg, fertilization is attended by the formation of a ‘ fertiliza- tion membrane’. It must be remembered, however, that the essential change at the egg-surface is completed long before any visible change is possible. What is the nature of the fertilization membrane? In the case of the egg of Nereis it seems certain (Lillie, 20) that this membrane is the vitelline membrane of the unfertilized egg, which is pushed away from the egg-surface by the disintegration and hydration of the 49,4 J. GRAY ege-surface immediately under the vitelline membrane. In the case of the echinoderm egg it seems probable that essen- tially the same change takes place. Many years ago Loeb (25) showed that the fertilization membranes collapsed when placed in sea-water containing albumen, and that on transference to normal sea-water the membrane regained its normal spherical shape. He concluded TEXtT-FIG. 1. A Diagram illustrating the origin of the fertilization membrane of Echinus miliaris. Lk ee hi q . + ay a? ORTY i 4 > ° is ftv 7 1a) ne T.J.Evans del. 102. a> SS : S&P WALR? YT) tig RI eis } - Huth lith. et imp x in a a far, 7 ; < oa P34 ¥ am 4 ‘ be =" * Pe The Segmentation of the Head in Squalus acanthias. By G. Rylands de Beer, B.A., B.Se., Christopher Welch Scholar, Demy of Magdalen College, Demonstrator in the Department of Zoology and Comparative Anatomy, University Museum, Oxford. With 13 Text-figures. THERE are two views with regard to the segmentation of the head. One has arisen out of Balfour’s (1) pioneer work, the other is due to Van Wijhe (18). They both agree on many points, and the difference between them lies in the interpreta- tion of the numerical relations of the different segmented structures, as pointed out by Goodrich (8). In order that they may be compared, a short account of these views will now be given. Balfour, in his classical ‘ Development of Elasmobranchs ’, expressed the opinion that the six visceral clefts are related to six consecutive somites situated dorsally and posteriorly to each respective cleft, the clefts being intersomitic. In front of the spiracle he recognized two somites (premandibular and mandibular), so that in all, from the anterior extremity to behind the last gill-slit, there are eight somites, of which the six posterior are simply and harmoniously related to six visceral arches and clefts. The recognition of the nature of the cranial nerves is due to the work of Marshall and Van Wijhe. Five nerves are regarded as dorsal roots, viz. ramus ophthalmicus profundus, trigeminal, facial, glossopharyngeal, and vagus, the latter being really compound and probably representing four segments. ‘There are then eight dorsal nerve elements, and if each is related to 458 G. RYLANDS DE BEER one of Balfour’s eight somites the following relations will result, as shown in Table I (ventral roots also included). This is the theory which has grown out of Balfour’s work, and the somites bearing the relations described above may be called Balfour’s somites. Among the supporters of this view are Ziegler (19), Koltzoff (10), Goodrich (8). TaBLeE I. Seg- Dorsal Ventral | Visceral | Visceral ment. Nerve. Nerve. Somite. Arch. | Cleft. >) Ole: Oculomotor Premandi- rol = bular 2 | Trigeminal Patheticus Mandibular | Mandibular Saieacie 3 | Facial Abducens Hyoid Hyoid Gull slit 1 4 | Glossopharyn- — 4th 3rd p geal Gill-slit 2 5 | Vagus 1 Hypoglossus | 5th 4th Wok 6 | Vagus 2 Hypoglossus 6th 5th por: . 7 | Vagus 3 Hypoglossus | 7th 6th Gill-slit 5 8 | Vagus 4 Hypoglossus | 8th 7th 9 Ist Spinal — 9th — Tasie II. Seg- Dorsal Ventral | _ Visceral | Visceral ment. Nerve. Nerve. Somite. | Arch. | Cleft. 1 |. ROP. Oculomotor | Premandi- | —. ae bular 2 | Trigeminal Patheticus Mandibular Mandibular | 7,4 3 | Facial Abducens | Hyoid Hyoid Seahedie 4 Facial = | 4th Hyoid Cull elit 1 5 Glossopharyn- — | 5th 3rd geal Wal: 6 | Vagus 1 — 6th 4th yee eo : 7 | Vagus 2 Hypoglossus | 7th 5th Gill-slit 4 8 | Vagus 3 _Hypoglossus | 8th 6th Gill-slit 5 9 | Vagus 4 Hypoglossus 9th 7th 10 | Ist Spinal | — 10th a But the majority of authors have followed Van Wijhe (18) in the interpretation of the relations. He regards the facial nerve as really of double nature, the existing nerve being related to the fourth somite while the nerve of the third somite has either disappeared or become merged with that of the ? fourth. The relations of these somites (Van Wijhe’s somites) a "> HEAD OF SQUALUS 459 are shown in Table Il. Braus (2), Hoffmann (9), Sewertzoff (17), and Neal (11) adopt this view. The difference between the two interpretations is centred in the region of the spiracle and hyoid arch. On the first view all the segmented elements are harmoniously and con- secutively related without gaps or discrepancies: on Van Wijhe’s there is one somite too many in the region of the spiracle. For since the mandibular or second somite corre- sponds to the first (mandibular) visceral arch, if the second (hyoid) arch corresponds to the fourth somite, as Van Wijhe supposes, then the third somite has no arch or cleft. Van Wijhe suggests that these have been lost. The question has been gone into thoroughly in the ease of Seyllium canicula by Professor Goodrich (8), and it was at his suggestion that I undertook to investigate Squalus acanthias(Acanthias vulgaris) in order to see whether the conditions were similar in this related form. The first part of this paper deals with the question of the correspondence in the region of the hyoid arch, and which somite forms the first permanent myotome. This is followed by a brief description of the occipital region, for the purpose of comparing the extent of the cranial region in Squalus and Seyllium. The work was done in the Department of Comparative Anatomy at Oxford. To Professor Goodrich, for advice and encouragement, I wish to offer my grateful thanks. I also had the privilege of consulting Professor Neal in person and to him, for valuable assistance and material, I express my deep gratitude. In a 45 mm. embryo (Text-fig. 1) reconstructed from longitudinal vertical sections all the somites of the head can be discerned. Two visceral clefts are present—spiracle and gill-slit 1 situated beneath the third and fourth somites respectively. In the next stage (Text-fig. 2,5 mm.) the second gill-slit has appeared beneath somite 5. Of the dorsal nerves trigeminal is related to the second somite, facial is situated between imm ut Text-fig. 1—Reconstruction of the head of anembryo of Squalus acanthias 4-5 mm. long seen from the left side. Text-fig. 2—Embryo 5 mm. long. First appearance of muscle-fibres in somite 5. The relation of somites to gill-slits and dorsal nerves is plainly seen. Text-fig. 3—Embryo 6 mm. long. The fifth somite is indistinct, the sixth is covered by the vagus except for the posterior dorsal corner, which begins to assume the hook-shaped appearance. Text-fig. 4—A single section through an embryo 5 mm. long, showing the single nature of the somite under the auditory capsule. EXPLANATION OF LETTERING. Ab, Abducens nerve. A.H.C., Anterior head cavity of Platt. Ap., Auditory Placode. A.S., Auditory Sac. F, Facial nerve. G, Gut. Gl, Glossopharyngeal nerve. G.S. 1-5, Gill-slits 1 to 5. H, Heart. N. 1-7, Neuromeres 1 to 7. Oc, Oculomotor nerve. S. 1-10, Somites 1-10. Sp, Spiracle. Sp. Gn. 2, Second Spinal Ganglion. Zr, Trigeminal nerve. V, Vagus nerve. V. 7-10, Ventral nerve-roots of segments 7 to 10. V.C.L., Vena Capitis Lateralis. HEAD OF SQUALUS 461 somites 8 and 4, glossopharyngeal between 4 and 5. The rudiment of the first branch of the vagus overlies somite 5, which is the most anterior of the post-otic somites to develop muscle-fibres. Text-fig. 8 shows clearly the relation of the vagus to the somites. It covers the posterior border of the fifth somite and most of the sixth. By tracing up through later stages I have arrived at the conclusion that it is the sixth somite which forms the first permanent myotome. For the greater part median to the vagus, its posterior dorsal corner is prolonged into a - hook-shaped process which, lapping round the posterior edge of the vagus, extends forwards laterally to it. The hook-shaped process can be seen in an incipient condition in this (Text-fig. 3) and in subsequent stages; likewise that the sixth somite corresponds to the third gill-slit, which is of course related to the second branch of the vagus. There is also serial correspon- dence between spiracle, gill-slits 1 and 2, and the third, fourth, and fifth somites respectively. The establishment of this corre- spondence is important, for some authors (Dohrn, Froriep) have described a varyingly large number of somites under the auditory capsule. J am convinced that there is only one somite between the facial and the glossopharyngeal in Squalus. Text- fig. 4 is a drawing of a single section, and the region beneath the auditory capsule from two sections of another embryo is shown in Text-fig. 5 and under higher magnification. The pecu- liar nature of the posterior corner of the sixth somite also is shown. Hach cleft lies between two visceral arches. ‘The first or mandibular arch contains a prolongation of the second or mandibular somite. Therefore, since clefts and somites corre- spond, the next posterior visceral arch (hyoid or second) must correspond to the next somite (hyoid or third somite). This is corroborated by the fact that these two consecutive arches, first and second, are related to two consecutive dorsal nerves, trigeminal and facial. Similarly the third visceral arch corre- sponds to the fourth somite and the glossopharyngeal nerve. This interpretation implies that the dorsal roots are related to NO. 263 K k 462 G. RYLANDS DE BEER the somites lying anterior to them, and it will be shown that this is the only view which avoids weighty assumptions and discrepancies. As development proceeds the interpretation becomes more difficult, and for two reasons : 1. The fourth and fifth somites lose their distinctness and the fourth breaks up unrecognizably into mesenchyme. This is possibly due to the pressure of the auditory sac, which appears TExtT-FIG. 5. The somite beneath the auditory capsule. between the facial and glossopharyngeal nerves, overlying the fourth somite. As the sac extends backwards the fifth somite also begins to break up, though some remnants of its muscle- fibres persist. 2. The somites appear in later stages to be situated more posteriorly with regard to the gill-slits. This is due partly to the development of the latter, which push them backwards, and partly to the fact that owing to the slight curvature which the head undergoes, the line of mesodermic somites finds itself situated on the outer side of the circumference of this curvature. Since the centre region of the head mesoderm (somites 4 and 5) is broken down into mesenchyme, the more anterior somites 3 ee TEXT-FIG. 6. 5mm. 7mm. 8 mum 10mm. Series of diagrammatic reconstructions of embryos of the lengths of 4-5, 5, 6, 7, 8, and 10 mm. drawn to the same scale. The rudi- ments of the facial nerve are joined by a chain line, the corre- sponding somites of the different embryos by dotted lines. K k 2 464 G. RYLANDS DE BEER 2, and 1 (which will be drawn off into the service of the eyeball), acquire a more anterior position. Similarly the more posterior somites, 6, 7, &c¢., move relatively backwards. By measuring somites in the region of the fifth and sixth, it can be seen that they are stretched and occupy more space along the long axis of the embryo than the remainder. Trxt-Fias. 7, 8. Sp Text-fig. 7.—Embryo 8mm. long. Ventral roots are present from the seventh somite backwards. Text-fig. 8.—Embryo 10mm. long. The tenth segment is the most anterior to develop a fully-formed mixed nerve. The vagus is represented as truncated to reveal the ventral portion of the sixth somite, which lies median to it. Text-fig. 6 shows a number of embryos in successive stages of development, all drawn to the same seale. ‘The corresponding somites of each embryo areinterconnected by dotted lines. It will beseen that the somite which laps round the vagusand forms the first permanent myotome in the 10 mm, and all later stages HEAD OF SQUALUS 465 cannot be any other than the sixth, provided that no great migration on the part of the somites has taken place. Several authors regard the seventh somite as the one which gives rise to the first permanent myotome, and this suggests a migration forwards, as described by Braus (2). But, as stated above, any relative movement which the somites undergo is backwards and not forwards, and is entirely passive. Text-figs. 7 and 8 are reconstructions of embryos in the 8 mm. and 10 mm. stages respectively. The sixth somite is very obvious, with its anterior margin drawn out and indistinct, anterior to which there are remnants of muscle-fibres of somite 5, and the same is true of the 20 mm. stage (Text-fig. 9). Up to a stage between 8mm. and 10mm. a ventral root can be traced to the myotome of the sixth somite. In later stages, however, J have been unable to find it. This is in agreement with Neal (11) and Hoffmann (9), both of whom state that it disappears. Presumably the sixth somite is innervated by a branch from the next posterior ventral root and somite (seventh), or the two nerves may combine, but I have not succeeded in determining this point. The fate of the ventral roots of the post-otic myotomes is of importance in determining the posterior limit of the cranium. Cartilage begins to appear in the L stage (Sewertzoff, 17 ; Gaupp, 5). Text-fig. 10 shows a reconstructed embryo about 50 mm. long. Three ventral roots are present, emerging through foramina in the cranial cartilage. The anterior one is very thin and belongs to the seventh somite, the remaining two are stout nerves. The foramina through which they pass become confluent with the vagus foramen, the vagus lying immediately lateral of their point of exit from the neural tube. These two nerves I regard as belonging to somites 8 and 9. These results are in agreement with those of Hoffmann (9). Fir- bringer (4) did not study very young specimens of Squalus, so that it is probable that his y and z are the same as Hoff- mann’s ¢ and d and my eighth and ninth. The next ventral root, the tenth, comes out of a deep notch in the posterior wall of the cranium, but behind the occipital Text-Fia. 9. 20 mm, embryo. Text-Fric. LO. (mm. elo 50mm, stage. View of the occipital region. Three ventral roots are included in the skull, v 7-v 9, and the next, v 10, joins a dorsal root to form the first mixed nerve. HEAD OF SQUALUS 467 arch so that it is not included in the cranium. Its fibres join those of the spinal ganglion to form the first mixed nerve, branches of which I have traced to the tenth somite. Since the eighth somite is the last of the vagus segments the ninth is morphologically the first spinal or post-vagal (in the case of Squalus included in the skull), and the tenth, which in Squalus forms the first mixed root, is really the second spinal or post-vagal. Rudimentary dorsal ganglia are present belonging to the seventh, eighth, and ninth somites (Text-fig. 9). In Squalus, therefore, there are nine segments included in the skull. Hoffmann (9), Sewertzoff (17), and others state that there are ten, but since they adopt Van Wijhe’s somites, and like him intercalate a somite be- tween those related to the fifth and seventh nerves, the number of their somites from the third backwards are the same as those of Balfour, plus one. Hence their results and mine are really in accordance since we both regard the same segment as being the last one included in the skull, though the numbers attri- buted to it are different. As compared with Scyllium Squalus has two more seg- ments included in the skull. But it is interesting to note that in both forms it is the tenth segment (second spinal) which gives rise to the first mixed nerve. It is another proof that homology does not depend on numerical correspondence (Goodrich, 6). DISCUSSION. The acceptance of Van Wijhe’s scheme of segmentation renders it necessary that two somites, the third and fourth, should be associated with a single cleft and visceral arch : the spiracle and hyoid arch. There is in this region one dorsal nerve, the facial, and this Van Wijhe regards as double and representing elements belonging to the third and fourth somites. The hyoid arch he assigns to the fourth somite, and in order to account for the third he assumes that a visceral cleft and arch have been lost. We shall return to this assumption later. Further, he regards the somites from the fourth to the eighth as related to the visceral arch and dorsal root lying in front of them. Now the 468 G. RYLANDS DE BEER mandibular somite corresponds undoubtedly to a visceral arch (the first) and a dorsal nerve (trigeminal) lying behind it. Similarly the ramus ophthalmicus profundus is situated posteriorly to the first somite. Again, in the trunk region the ventral root of a somite joins the dorsal root posterior to that somite to form a mixed nerve. ‘Therefore Van Wijhe’s scheme involves two discrepancies, viz. that in the regions between the mandibular and hyoid arches and between the posterior somite of the head and the first of the trunk there has been a reversal of the relations between somites and dorsal nerves. The first of these discrepancies concerns the trigeminal and facial nerves. The trigeminal is situated behind the mandibular somite, whereas the facial les in front of the fourth. ‘lo be consistent one would have to attribute the trigeminal to the somite posterior to it (third) and the ramus ophthalmicus profundus to the second, but this would leave the first somite without a corresponding dorsal nerve. Similarly, in the region between the trunk and the head, the last branch of the vagus would le anterior to its somite (Van Wijhe’s ninth), whereas the first spimal ganglion is situated posterior to its somite. Not only would the nerves from the facial to the vagus lie anterior to their somites, but they would also lie anterior to their corresponding ventral roots. In the trunk the ventral root is always more anterior than its corresponding dorsal root (Goodrich, 8). These relations of Van Wijhe’s somites are diagrammatically represented in Text-fig. 11. We see then that this scheme has to contend with serious difficulties, all of which are the outcome of regarding the third somite as having lost its visceral cleft and arch. Let us now examine this assumption. In the first place the missing gill-slit is not indicated by any of the structures which it must have involved and of which it is reasonable to expect that some vestige would remain. There is no trace of arch, cleft, afferent or efferent blood-vessels or nerve. This in itself is significant in view of the fact that the anterior visceral arches are con- served with constant regularity all through vertebrate phylo- geny. And even when the clefts disappear they leave traces of HEAD OF SQUALUS 469 their former existence in the form of blood-vessels, nerves, skeletal elements or modified structures. ‘Then authors are not agreed as to the exact position of this missing cleft. Whereas Van Wijhe considers the hyoid arch as double, Hoffmann and Platt (14 and 15) regard the mandibular arch as representing two elements fused. With regard to the ventral roots there is no question about Text-Fic. 11. Diagrammatic representation of (a) Van Wijhe’s somites, (b) Bal- four’s somites. the oculomotor being the premandibular somite’s nerve, and the patheticus, in spite of its curious course, doubtless belongs to the second and mandibular segment. The fourth and fifth somites since they disintegrate have no ventral roots as such (though the fifth is present in Scyllium). To the sixth somite a ventral root can be seen up till about the 10 mm. stage. The abducens has usually been regarded as the nerve of the third somite and therefore as the ventral root corresponding to the facial. It certainly innervates the external rectus muscle, but Neal (12) states that in Squalus this muscle is of composite origin, consisting of elements derived from the mandibular as well as the hyoid somite. 470 G. RYLANDS DE BEER The abducens is held to arise from the neural tube by many roots ; according to Neal four, corresponding to Van Wijhe’s somites 8, 4, 5, and 6. I have been able to discern three roots which in a 23 mm. embryo arise not very far behind the facial Trxt-Fic. 12. \ AG = . IS =a ane mm Ventral view of an embryo 23 mm. long to show the origin of the abducens nerve. (Text-fig. 12). In earlier stages their origin appears to be slightly more posterior. It has been suggested that the hypoglossus and abducens roots form a continuous series, implying that the abducens is a compound nerve derived from elements belonging HEAD OF SQUALUS 471 to three or more segments, but even going by topographical - relations alone it is not unreasonable to regard the abducens as being the genuine third ventral root. At any rate I do not see that the condition of the abducens furthers the assumption that a gill-slit has been lost. If Neal’s contention is true the whole question of the eye muscles mnervation and segmenta- tion will require revision. Lastly, one more train of thought has been brought to bear on the supposedly lost organs, and that is the ques- tion of the relation of neuromeres to the other segmental structures. Neal (11) describes seven neuromeres, of which the first corresponds to the anterior head cavities with the olfactory as its dorsal nerve: this is, of course, assuming that the anterior head cavities have the value of a somite anterior to the pre- mandibular somites. ‘To the second neuromere correspond the ramus ophthalmicus profundus premandibular somite and oculomotor. The third or cerebellar neuromere later under- goes subdivision (which Neal regards as secondary) and to it belong trigeminal, patheticus, and mandibular somite. The next dorsal root, the facial, arises from the fifth neuromere, and this led to the idea that the nerve of the fourth neuromere (which has none) has disappeared and that it was this nerve which was related to the lost gill-slit. To the sixth and seventh neuromeres belong the glossopharyngeal and vagus, though their topographical correspondence has been lost. The glosso- pharyngeal appears to arise from the seventh neuromere, and the vagus behind it ; but this is explained as being due to the pres- sure of the auditory sac and relative shifting of the elements of the neural crest and neural tube. These relations are shown in the embryo (6 mm.) reconstructed in Text-fig, 18. If it be granted that neuromeres have a primary segmental value, then it may be said that there is one neuromere too many overlying the hyoid arch; but that such a segmental value exists remains to be proved. To start with it rests on the assumption that the anterior head cavities represent a somite. These are present only in Galeus and Squalus, but in Amia, 472 G. RYLANDS DE BEER Reighard, and Phelps (16) have described the sucker as arising from muscle anterior to the premandibular somite. Goodrich (7) has produced good evidence to show that the anterior head diverticula of Amphioxus are homologous with the premandi- bular somites of Craniates, and it is more reasonable to agree with Dohrn (8) that no segmental significance must be attached to Platt’s anterior head cavities. Then, supposing that the TEXxT-FiG. 13. Tr Na F Ns Ne NvGl v Se \ No. L Imm. =| Embryo 6 mm. long showing the relations of the neuromeres to the remaining segmented structures. subdivision of the third neuromere is not secondary but primary and retarded (which it might be, for the third neuro- mere is just about twice as long as the following ones), it would be necessary to postulate yet another gill-slit lost, to correspond to the extra neuromere. But perhaps the greatest objection to the segmental value of neuromeres lies in the fact that they are altogether absent in Amphioxus, scarcely developed in Petromyzon, irregular and asymmetrical in Bdellostoma, and that they are best developed in the higher craniates, birds, and mammals (Neal, 18). This strongly discountenances their palingenetic value and suggests that they aze neomorphs. And so I cannot believe that the evidence from neuromeres favours the assumption of a lost gill-slit. ————— HEAD OF SQUALUS 473 The simplest explanation then, that origimating from the work of Balfour, is the most suited to the facts: the third somite is related to the hyoid arch and the fourth to the third arch, &¢.; somites correspond to the dorsal roots lying imme- diately behind them, as in the pre-otic and trunk regions, and there are no discrepancies. In dealing with such a subject as segmentation, the essence of which is a simple and orderly repetition of parts, if a simple explanation can be deduced which does not produce incon- sistencies or go against facts, the onus probandi must lie with those who would reject such an orderly state of affairs. SUMMARY. 1. Balfour’s interpretation of the somites of the head is correct and free from the objections which accompany Van Wijhe’s. 2. No gill-slit or arch has been lost in the neighbourhood of the hyoid arch. 3. Nine segments are included in the head of Squalus, of which three are pre-otic (first, second, third) and six post-otic. Of these One (fourth) breaks down completely into mesenchyme ; One (fifth) forms muscle-fibres but later breaks down ; Four (sixth, seventh, eighth, ninth) produce permanent myotomes ; The tenth somite (first of the trunk, and second post-vagal) corresponds to the first mixed nerve. List oF THE LITERATURE CITED IN THIS PAPER. 1. Balfour, F. M—*‘ Monograph on the Development of Elasmobranch Fishes ’’, ‘ Journ. Anat. and Phys.’, 1876-7 and 1878, 2. Braus, H—*‘ Die metotischen Urwirbel ”, ‘ Morph. Jahrb.’, vol. xxvii, 1899. 3. Dohrn, A——‘‘ Die PrimandibularhGhle”’, ‘ Mitth. Zool. Sta. Neapel ’, 17, 1904. 4, Fiirbringer, M.—‘‘ Ueber die spino-occipitalen Nerven”’, ‘ Festschr. f. C. Gegenbaur *, Leipzig, 1897. 474 G. RYLANDS DE BEER 5. Gaupp, F.—“‘ Allg. Entwicklung des Kopfskelettes”, ‘ Hertwig’s Handbuch d. Entw. d. Wirbeltiere ’, vol. 6, 1906. 6. Goodrich, E, S—‘‘ Metameric Segmentation and Homology ”’, ‘ Quart. Journ. Micr. Sci.’, vol. 59, 1913. “* Proboscis pores in Vertebrates ’’, ibid., vol. 62, 1917. ‘* On the development of the segments of the head in Scyllium ”’, ibid., vol. 63, 1918. 9. Hoffmann, C. K.—“‘ Beitr. zur Entwicklung der Selachier ’’, ‘ Morph. Jahrb.’, vol. xxv, 1897. 10. Koltzoff—‘‘ Metamerie des Kopfes von Petromyzon Planeri’’, ‘ Anat. Anz.’, 1899. 11. Neal, H. V—‘‘ The segmentation of the Nervous System in Squalus acanthias ’’, ‘ Bull. Mus. Comp. Zool. Harv.’, vol. xxxi, 1898, 12. ——‘‘ The history of the Eye muscles ”’, ‘ Journ. Morph.’, vol. xxx, 1918. 13. ———‘‘ Neuromeres and Metameres ”’, ibid., vol. xxx, 1918. 14. Platt, J. B—‘‘ Contribution to the morphology of the Vertebrate Head ”’, ‘ Journ. Morph.’, vol. v, 1891. 15. “* Further contributions ”’, ‘ Anat. Anz.’, vol. vi, 1891. 16. Reighard and Phelps——‘‘ The development of the Adhesive Organ and Head Mesoblast of Amia’”’, ‘ Journ. Morph.’, vol. xxx, 1908. 17. Sewertzoff, A. N.—‘‘ Die Entwicklung des Selachierschidels”’, ‘Festschr. f. L. v. Kupffer’, Jena, 1899. 18. Wijhe, J. van—‘ Ueber die Mesodermsegmente und die Entwicklung der Nerven des Selachierkopfes ’, Amsterdam, 1882. 19. Ziegler, H. E—‘‘ Die phylog. Entstehung des Kopfes”’, ‘ Jena. Zeitschr, Naturw.’, vol. xliii, 1908. Some Notes on the Gametogenesis of Ornithorhynchus Paradoxus. By J. Bronté Gatenby, M.A., D.Phil. (Oxon.), D.Se. (Lond.), Professor of Zoology and Comparative Anatomy, Dublin University. —_ (From the Department of Embryology, University College, London.) With Plates 12, 13, 14, and 1 Text-figure. CoNTENTS. . LytTRopuctIon 2. Previous WorK on inddodannais OF Oakimmosne noses AND OF ECHIDNA . GENERAL NOTE ON THE erdouien OF THE Ree OF THE gata: PSIDA . GENERAL Meanie OF THE ewiiasion, OF Begs MEMBRANES IN SAUROPSIDA AND MAMMALIA . GENERAL ACCOUNT OF THE YOLK ORALeON IN eee AND AMPHIBIA . THE STRUCTURE OF THE Ovary OF Oroutyenvs : . Tar APPEARANCE OF THE IMMATURE OvaRY . . THE SIZE OF THE LARGEST AND SMALLEST OVARIAN Garena : . THe YounG Oocyte . . ON THE Earty ESTABLISHMENT OF A Hoveenry IN THE Oagvin . FoRMATION oF EGG-MEMBRANES . YOLK ForRMATION . FORMATION OF THE Tteaeiea . A Futy-rormMep Eae (diameter 4 mm. i . NovTE ON SPERMATOGENESIS . Discussion . BIBLIOGRAPHY 1. InTRODUCTION. PAGE 475 476 478 480 482 483 484 485 486 486 487 489 490 49] 492 493 494 In this paper I have described in as much detail as was possible the oogenesis of the duck-billed platypus of Australia. Owing to the unique position of the Prototheria, any new facts 476 J. BRONTE GATENBY with regard to their germ-cells is sure to be of value. I must take this opportunity of thanking Professor J. P. Hill, F.R.S., for allowing me to study his material of Ornithorhynchus, without which I could not have published these notes. There is no account extant of the detailed structure of the ovarian egg of Ornithorhynchus, of the yolk formation, of the maturation stages, or of the corpus luteum. Such accounts of the ovary as are published are scrappy and full of errors, this, however, being chiefly due to the scanty and poor material at the disposal of the various observers who have attacked these problems. The material at my disposal, while not having been prepared by the most modern technique, is well preserved by routine methods and allows of a fuller description of various problems than hitherto given. The material consisted of one ovary preserved in Flemming’s strong fluid, and of several ovaries preserved by a variety of picric and bichromate fixa- tives. Most of the new results were procured by examination of the Flemming-fixed ovary. This work was partly carried out in the Embryological Laboratory, University College, London, and was finished in the Zoological Laboratory, Dublin University. Apart from his kindness in lending me the material, I have to thank Professor J. P. Hill for assisting me by lending some of the literature on Ornithorhynchus and Echidna. 2. Previous WorK ON GAMETOGENESIS OF ORNITHO- RHYNCHUS AND OF ECHIDNA. Thirty-seven years ago E. B. Poulton, in his paper on ‘The Structures connected with the Ovarian Ovum of Mar- supialia and Monotremata ’, gave some account of the general appearance of the ovary and follicle of Ornithorhynchus and Echidna. Poulton’s material consisted only of ovaries removed from spirit specimens, and he was consequently much handi- capped. Nevertheless, he succeeded in establishing several facts of great importance. The ovary of Ornithorhynchus, according to Poulton, is flat or compressed, oval, and about 13 mm. long, 7 mm. wide, and 2mm. thick. The follicles are GAMETOGENESIS OF ORNITHORHYNCHUS ATT confined to the edge of the transverse section of the ovary, i.e. on the surface of the ovary; there does not seem to be any distinct arrangement of follicles, according to size, but the small ones always seem to be near the surface. Poulton noticed that there was evidence that the large follicles were constricted off, in the presence of a deep furrow encircling some of them. By this I believe he means that the egg (and follicle) is con- stricted from outside, and tends to hang somewhat freely on the surface of the ovary. | Poulton identified a follicular epithelium, which he considered to be of one layer, ‘the whole of the time the ovum remains in the follicle ’. This author also describes faithfully the zona pellucida, follicle, basement membrane, and tunica fibrosa, and establishes the fact that the ‘ova of Monotremes practically fill their follicles, and are of considerable size’. The nucleus Poulton considered to be central in the small ova. He recognizes in the older egg a peripheral stainable granular area, and, deeper down, a lighter granular area, beneath which lies the yolk. It is remarkable that Poulton should have been able to describe so many interesting facts from such poor material. Three years later, in 1887, Caldwell published a paper on ‘The Embryology of the Monotremata and Marsupialia ’, in which he pointed out that Poulton and Guldberg had wrongly stated that the follicular epithelium remains always a single layer of cells. Guldberg and Beddard both described the ovary of Echidna. They showed that it resembled in its oogenesis the condition already described by Poulton for Ornithorhynchus. Probably the finest collection of Monotreme material is that procured by Semon about 1893; this observer had at his disposal a large number of eggs in all stages. He gives no account of the oogenesis, and his description of the structure of the egg consists of thirty-five lines of general comment, without any detailed account of his material. Tt is therefore difficult to know how much Semon ynderstood of the structure of the eg. Certain appearances drawn in his figures of the ege NO. 263 L | A768" J. BRONTE GATENBY are undescribed in the text. In some cases it is impossible to know whether Semon’s figures of supposed nuclei are cells or nucleoli; this applies especially to his Tafel [X, figuring early stages of development. Writing of the full-grown egg, Semon says: ‘ Die Keimscheibe ruht auf einem Lager von feinkérnigem, weissem Dotter, und dieser entsendet nach innen eine strangformige Fortsetzung, einen “‘ Dotterstiel’’, der im Centrum sich flaschenformig zu einer Latebra aufbliht. Die Elemente des gelben Dotters sind kugelrund ; gegen den weissen Dotter zu, besonders in der Gegend der Keimscheibe, nimmt der Durchmesser der Kugeln des gelben Dotters continuirlich ab. An der Grenze erblickt man haufig die Kugeln des gelben Dotters in allen Stadien des Zerfalls zu kleineren und kleinsten Elementen. In gleichem Maasse wie das Blastoderm den Dotter umwachst, breitet sich an der Oberfliche des letzteren und ersteren eine Schicht von weissem Dotter aus.’ In his figures of sections of the eggs of both Echidna and Ornithorhynchus, Semon draws, within the more central part of the cross-section of the egg, from one to as many as three concentric rings lying in the yellow yolk, as is well known to occur in the hen’s egg, but.he does not describe these rings in his text. His description of the structure of the egg is very poor. It would probably be worth while for a capable cytologist to re-describe the sections of eggs of both Echidna and Ornitho- rhynchus procured by Semon’s party in Australia. 38. GENERAL NOTE ON THE STRUCTURE OF THE EGG OF THE SAUROPSIDA. In both the Aves and the Reptilia, the egg, as is well known, has a very complicated structure, and for the purpose of comparison with that of Ornithorhynchus I have given a diagram in Text-fig. 1. The germinal disk (Gp) is formed of pure protoplasm free of any but the smallest yolk-spheres; this protoplasmic disk contains a very granular, generally somewhat basophil, type of protoplasm, which can readily be distinguished from the GAMETOGENESIS OF ORNITHORHYNCHUS 479 clear cone of protoplasm (acp) which lies below. This clear cone of protoplasm is not granular, and passes insensibly into the disk above, on the one hand, and into the neck of the latebra (NL) below, on the other hand (Nucleus of Pander). The latebra (1) is formed of a clear substrate containing numbers TExtT-FIG. 1. | | of fine yolk-granules. Completely surrounding the egg, and forming a peripheral area, is a thin layer of clear protoplasm containmg very fine yolk-spheres (cu). All the internal substance of the egg, excepting that part occupied by the latebra (x), is filled with enormous numbers of large coarse yolk-spheres ; and within this substance can be found con- centric rings of clear material (cr) which are said to mark areas of growth of the yolk (see Riddle, 5). Hype 480 J. BRONTE GATENBY The peripheral clear area (cu), the cone of protoplasm (acp), and the latebra are generally described as containing white yolk-spheres, the rest of the egg mainly yellow yolk-spheres. The clear thin layers of concentric stratification (cr) have been said to contain white yolk-spheres, though this has not been settled satisfactorily. Riddle (5), however, believes that the concentric layer does contain white yolk, and is a growth- mark, Semon’s description of the yolk of the egg of Ornitho- rhynchus does not include any mention of these concentric layers of stratification within the egg, but in his figures he shows eggs of Echidna and Ornithorhynchus which contain one (Tafel VIII, fig. 23), two (fig. 25 and fig. 19), and three (fig. 20) layers, as depicted in Text-fig. 1 (cr) of this paper. It is possible that the egg of Ornithorhynchus might contain these concentric lines of growth, if such they be. The varying number of lines are probably significant of the different periods or epochs of the year during which the eggs grew most; that with two rings possibly grew in two sudden well-marked periods, and so on. This opinion is supported by Riddle’s work on feeding Sudan IIT to laying fowls (5). 4, GENERAL ACCOUNT OF THE FORMATION OF EGG- MEMBRANES IN SAUROPSIDA AND MAMMALIA. In a recent paper (8) Miss Alice Thing has studied the forma- tion of the zona pellucida in various turtle eggs. When the young oocyte of the turtle has reached a size two or three times that of the oogonium, it becomes surrounded by a flattened epithelium which persists as one layer throughout the course of development of the egg. With the gradual growth of the oocyte, the epithelial cells take on a definite prismatic shape and increase in height in the axis perpendicular to the surface of the egg. Occasional mitoses prove, that to accommodate the increasing volume of the egg, the epithelium extends itself by division of its constituent cells ; in very large eggs numerous mitoses occur. The epithelial cells forming the follicle are ede GAMETOGENESIS OF ORNITHORHYNCHUS 481 sharply marked off from one another by intercellular channels filled with intercellular substance. The latter undergoes early a change of constitution and becomes transformed at the level of the surface of the cells into the special cement known as the terminal bars. The zona pellucida is formed by two or three different elements. It takes its origin as a veil-like formation consisting of a mosaic of terminal bars and polygonal fields within which may be recognized the small, pale areas, future canals of the adult membrane separated by pale and dark filaments giving origin to the future fundamental substance of the adult membrane. The fundamental substance of the zona_pellu- cida is developed as a cuticular element, by the terminal bars or primary network, that is by a definite special inter- cellular cement possessing the property of extension over the free surface of the epithelial cells and forming connexions there with the delicate secondary network apparently produced directly by the superficial cytoplasm of the epithelial cells. With regard to the origin of the zona pellucida in mammals, I believe that there are three possible methods of development : the zona might develop from the follicular epithelium, it might develop from the egg-cytoplasm, or it might be developed under the influence of, and from, both egg-cytoplasm and follicular epithelium. The majority of present-day workers appear to believe that the zona of Mammalia develops from the follicular epithelium alone. This is the view of such well-known older observers as Flemming, Retzius, Fischer, Von Ebner, Bonnet, and Rubasch- kin. But Van Beneden, Sobotta, Waldeyer, and Kolliker all believe that the zona pellucida is secreted by the egg-cytoplasm. In support of this view are such observations as that of Van Beneden, who described in a bat the fact that while there may be two eggs in such close contact that at one place the follicle is interrupted, yet at this region the zona is properly developed. This is not of very rare occurrence in ovaries of placentals, and is certainly difficult to explain if one believes that the zona is of purely follicular origin. 482, J. BRONTE GATENBY 5. GENERAL ACCOUNT OF THE YOLK ForRMATION IN BiRDS AND AMPHIBIA. In both birds and amphibians the egg is richly provided with yolk, i.e. macrolecithal. The formation of yolk in the egg of amphibians does not seem to have been followed out with any detail or pains by a modern worker, and though I have made numerous preparations by the best methods, it has been difficult to determine the exact source of origin of the yolk- spheres (see Gatenby, 15, p. 189). In the amphibian oogenesis the mitochondria spread out mainly to form a deep cortical zone on the periphery of the egg. It is in this zone that the first sign of yolk-granules appears, but it is wellnigh impossible to give an opinion as to whether the yolk originates directly from the mitochondria, or whether the latter only elaborate materials which, precipitat- ing in the ground cytoplasm, come to form the separate spheres of substance we recognize as yolk. Van Durme, in his monumental work on the oogenesis of birds (10), has entered into the subject with care, and has produced a paper which may be accepted as an authentic account of the steps in the formation of yolk in birds. He recognizes in the oocyte just before the beginning of the exten- sive yolk formation: (1) an attraction sphere containing a centrosome, (2) a yolk-forming region or vitellogenous cloud, (3) a quantity of fatty yolk. The vitellogenous cloud is formed of mitochondria of various types, e. g. chondriomites, chondrio- somes, and it soon undergoes a process of dissociation. This dissociation of the ‘ couche vitellogéne ’ invokes the appearance throughout the egg-cytoplasm of a uniform layer of mito- chondria. This uniformity does not last long, for soon after- wards three distinct mitochondrial zones appear; a cortical dense zone, an inner deeper, and an internal still deeper zone. The first vestiges of yolk formation are the appearance of clear yolk-vesicles (vacuoles) in the neighbourhood of the cortical fatty layer, thus constituting a peripheral vacuolated area, which spreads gradually towards the centre of the yolk ; tern. GAMETOGENESIS OF ORNITHORHYNCHUS 488 a second vacuolated area appears around the nucleus, known as the perinuclear vacuolated zone. These two zones meet at the animal pole of the egg, above the nucleus, forming the vacuolated nuclear cap. Subsequently in this second phase of vitellogenesis the first true yolk-spheres put in an appearance firstly in the region of the exoplasm, then more deeply in the endoplasmic region. Van Durme unhesitatingly states that these yolk-spheres partly arise from the larger mitochondria, and partly from the contents of the clear yolk-vesicles (vacuoles). From this stage onwards the more deeply-lying mitochondria become fewer, the yolk-elements more numerous, but the cortical mitochondrial zone persists throughout all stages. 6. THe STRUCTURE OF THE OVARY OF ORNITHORHYNCHUS. On taking up a slide of sections of the ovary of Ornitho rhynchus and examining it with the naked eye, one is first of all struck by the enormous size of the riper eggs. These are much larger than the full-grown ovarian oocytes of the frog, and of course infinitely larger than those of a rabbit or dog. As in the ovary of a Sauropsidan, the eggs project out around the surface of the organ in a way familiar to any one who has examined the ovary of a fowl or turtle. Thus, while the eggs may be very large, the stroma and general extent of the whole ovary is relatively small. This will be best understood by refer- ence to Pl. 12, fig. 3: in this ovary there was at least one egg nearly if not quite ripe (0), which measured 4-36 mm. in diameter in its shortest way, by 4:52 mm. in its longest way. In the ovary drawn in Pl. 12, fig. 3, no corpora lutea were to be found, and when these occur they protrude from the surface of the ovary almost as much as the full-grown egg. In several of the ovaries I have examined there are two corpora lutea close together, and these form by far the most prominent structures in the ovaries in question. Examined under the low power of a microscope the most striking features of the Ornithorhynchus ovary are the innumer- 484 J. BRONTE GATENBY wble lacunae or spaces in the ground-work or stroma. The biggest of these spaces are drawn in PI, 12, fig. 3, being cross- hatched (ca), but to gain a better understanding of this pecu- liarity one must examine fig. 4 of Pl. 13. Here the extraordinary structure of the ovary is demonstrated, a well-marked germinal epithelium is recognizable (x), and beneath it are a row of oocytes in various stages ; on the right of Pl. 18, fig. 4, the oocytes are found to lie in a more solid cortical area of the ovary, which is marked off at this region quite sharply by the wide and numerous lacunae, with their trabeculae in between (TR). These cavities do not contain blood, or lymph corpuscles, but seem to have been occupied by a non-corpuscular fluid, which leaves no trace of coagulum in the finished sections. As the young oocytes grow older they tend to become com- pletely surrounded by strands of much vacuolated tissue, as is indicated in the largest oocyte drawn in PI. 13, fig. 4. This feature is certainly one of the most remarkable in the ovary of Ornithorhynchus. It will therefore be clear that by the time an egg has reached the stage drawn in Pl, 13, fig. 4 (roughly one-eighth of its full size), it is already floating in a basket-like area formed by connective-tissue trabeculae and intervening lacunae filled with liquid. 7. Tue APPEARANCE OF THE IMMATURE OVARY OF THE PLATYPUS. In Pl. 12, fig. 2, is drawn an immature ovary measuring 3-250x1-0 mm. This shows remarkably well the almost amphibian character of the ovary at this stage. As was pointed out above with reference to the mature ovary, there is also to be seen in this immature specimen a cortical arrangement of oocytes ; around the ovary the eggs tend to lie in a thickened area, beneath which is a space occupying the centre of the organ. This cavity is only partly filled with loose strands of connective tissue. One is forced to look upon this peculiar structure of the imma- ture ovary of Ornithorhynchus as a very primitive feature. tise —@ GAMETOGENESIS OF ORNITHORHYNCHUS 485 In subsequent development the cavity becomes more and more filled with connective tissue, and this, together with the growth of the cortical walls of the ovary, caused the primitive type of arrangement to be disguised and partly obliterated ; but it should be pointed out that the lacunae figured on Pl. 18, fig. 4, are largely the remains of the early cavity within the gonad. 8. Tue Size oF THE LARGEST AND SMALLEST OVARIAN OocyTES OF THE PLATYPUS. In the adult ovary of Ornithorhynchus no oogonia are to be found; all these seem to have undergone their maturation prophases and to have become oocytes certainly long before the animal is full grown. Even in one very small immature ovary in Professor Hill’s possession there were no oogonia; this ovary measured only 8 mm. in depth (see Pl. 12, fig. 2), whereas the adult ovary is at least 12 mm. in depth. Possibly during an embryonic period all the oogonial divisions, as well as the prophases of the maturation division, have taken place, so that when the animal hatches there are already formed all the oocytes which it will possess and use during its life. This feature, with regard to the absence of true oogonia in the ovary, does not occur in forms like the frog, where numerous pockets of true oogonia exist in the ovary of the adult (vide Gatenby, 9). Were it not for these pockets of cells which continually proliferate new oocytes, the frog would be unable to lay three to five thousand eggs for so many seasons. In the case of Ornithorhynchus and other Mammalia, the number of offspring produced is so small as not to necessitate a con- tinuous new supply during each breeding season. Measurements have been taken of a number of the oocytes of the smallest dimensions I could find. The smallest was 0-07 mm., the average among the smaller being 0-08 mm. In the adult ovary the smallest oocytes measured from 0-08 to 0-09 mm. With regard to full-grown ovarian oocytes the largest I found was 4-5 mm.in diameter, not counting the theca (PI. 12, fig. 3); 4mm. seems an average diameter for the ovarian oocyte of 486 J. BRONTE GATENBY Ornithorhynchus. The one complete egg- and shell-membrane of which I examined sections was only from 4-5 to 5 mm. in diameter, though it was difficult owmg to the wrinkling to make an accurate measurement (PI. 12, fig. 1). 9. Tat YounGa OocytTE OF THE PLATYPUS. Some oocytes which had just undergone the prophases of the heterotypic division were discovered in the !'lemming-fixed material; two such oocytes are drawn on PI. 14, figs. 7 and 8. The nucleus is nearly always spherical, but occasionally irregular as shown in fig. 8; there is a well-marked nucleolus, Nu in fig. 7, of the fragmented type ; insome nuclei the nucleolus can be seen to be formed of two parts—a lightly-staining region, NuP in fig. 10, and a darkly-staining region, NusB. In fig. 11 the nucleolus consists of a very large darkly-staining sphere and a number of smaller pale elements; the chromatin is feebly staining and dispersed in all these nuclei. In nearly all the young oocytes observed a centrosphere is present, cs in figs. 7 and 8; the centrosphere at this stage les near the nucleus, often within a dent in the nuclear mem- brane, as in Pl. 14, fig. 8. In some cases contrioles or small granules within the centrosphere can be made out, as in fig. 8, cs. In the youngest oocytes the centrosphere may be surrounded by a cloud of granules which have been identified as mitochondria (M), fig. 7. In older oocytes the mitochondria, as happens in all verte- brate eggs, gradually pass away from the centrosphere, and become spread out into the cytoplasm (fig. 8); they tend to collect as matted granules and filaments, particularly in the region of the periphery of the egg, and become difficult to demonstrate at and after this period. 10. ON tHE Harty EstaBLISHMENT OF A POLARITY IN THE Puatypus OocyTeE. All the oocytes examined showed a distinct polarity, in that the nucleus had taken up a position to one side of the oocyte cytoplasm. I believe that this polarity has no relation- GAMETOGENESIS OF ORNITHORHYNCHUS 487 ship to the plane of the surface of the ovary, nuclei being found lying inwards, outwards, or sideways to an axis drawn directly down at right angles to the surface of the gonad. From the material examined it is impossible to understand completely the mode of origin of the polarity in the young oocytes, but from our knowledge of many vertebrate oogonia we are aware that when in this early stage the nucleus tends to lie to one side of the cell. The polarity of the Ornithorhynchus oocyte is therefore probably established during the oogonial stage, either as the accidental result of the position of the centro- somes and centrospheres of the daughter-cells during oogonial divisions, or as a subsequent and more expressly determined movement of the oogonial nucleus within the cytoplasm, at a stage just before the inception of the prophases of the heterotypic divisions. The former is most likely. This polarity of the oocytes persists throughout their entire growth, marking permanently the position of blastoderm and vegetative pole of the full-grown oocyte, and of the part of the egg in which the latebra will be formed. 11. ForRMATION oF HKGG-MEMBRANES. The egg-membranes on the ovarian oocyte of Ornitho- rhynchus are a theca (externa and interna), a follicle, and a zona pellucida. . : In all the youngest oocytes that have been observed the follicle is well formed ; it is shown in PI. 14, figs. 7 and 8, rou, and much enlarged in fig. 9. In the latter figure the follicle is seen to consist of one layer of flattened cells, overlying the substances of the oocyte (oc). In good preparations it is possible to recognize clearly a limiting or true cell-membrane around the egg-cytoplasm, om, in Pl. 14, fig. 9. Distinct cell- walls between the individual cell elements of the follicle were generally difficult to find, but are probably always present. In Pl. 14, fig. 12, the same region of an older oocyte is drawn. The follicle cells as such could not be identified in this prepara- tion, but the nuclei and general cell-substance have increased greatly in size. Just at this stage a new arrangement of the 488 J. BRONTE GATENBY individual elements of the follicle begins to take place; the nuclei dividing rapidly, soon become too large and too numerous to lie all in one row in the follicle, and gradually certain nuclei are displaced, as shown in PI. 14, fig. 12, and ultimately a two-layered follicle results (Pl. 14, fig. 11, rot). T'wo-layered the follicle remains all through its subsequent life. Now comes one of the events most difficult to understand and interpret—namely, the formation of the zona pellucida. Possibly, however, judging from the accounts of workers who have studied other material, Ornithorhynchus presents the problem in a less difficult form, though there are some points which are still far from clear to me. A glance at Pl. 14, fig. 9, gives one an impression of the condi- tion of the egg-membrane (om) at this early stage—the mem- brane is a true cell-wall, and nothing else at this period. Now in Pl. 14, fig. 12, the egg is considerably older, and two new structures have appeared: one is the substance marked pz, the other the fibrillae marked cr. The substance marked pz is the precursor of the zona pellucida, while the fibrillae, cp, grow to form the much larger structures shown in PI. 14, fig. 14, at cr. The fibrillae serve as connecting elements between the zona pellucida and the outer cell-membrane (om) of the oocyte cytoplasm. In none of the best slides I examined could I be sure that cell-walls existed at the stage drawn in PI. 14, fig. 12, just when the pre-zona substance is becoming clearly marked. The follicle nuclei appear to lie within a syncytium, but in my mind there exists no doubt that the pre-zona material is formed in or by the follicle cells. The substance might possibly be intercellular, as described by Miss Thing, but it is certainly derived from the follicle ; moreover, up to the last step in the development of the oocyte the follicle cells le in close relation- ship with the zona, as in Pl. 14, fig. 13, and when the egg is extruded the naked edges of the follicle cells are left, apparently supporting the view that the zona and the follicle were pre- viously most intimately related. This is all I can write with reference to the development of the zona. — O27 Invertase . . : : , : ; ; : 5. yar Lipase : : : : : : : : : . 539 Pepsin 3 : ; A ; ; : ; . . 540 Trypsin. : ; ; ; : ; 5 : «Bel Chymosin . : : : : : ; : , . 542 Emulsin ; 3 : : : 3 ‘ : 5 by Pe CONCLUSION : ‘ P : : P ; ; : =, OAS REFERENCES. : ‘ : : , ; : 5 | fatal EXPLANATION OF Thetis ; F 3 ; é : . 553 General meaning of Letters. , : : : : . 553 EXPLANATION OF PLATES . é p : ‘ F ‘ . 554 NO. 263 Nn 510 FE. N. PAVLOVSKY AND E. J. ZARIN Tris work was conducted by us with the same division of labour as the research on the intestine, its appendages and ferments in the scorpion, published previously. E. Pavlovsky undertook the zoological part of the work—the dissection of live bees, the preparation of the intestine, and the preparation of extracts from its parts. The chemical investigation of the ferments of these extracts was subsequently done by E. Zarin. ANATOMICAL Part. The intestine of the bee formed the subject of investigation for many scientists (for literature cf. Zander, Snodgrass), therefore its general anatomical relations may be considered to be sufficiently elucidated. We shall limit ourselves to the description of the general organization of the intestine and point out some peculiarities in its microscopical structure, whilst the literature on the question will be omitted. The intestine of the bee consists of the fore-, mid-, and hind- cuts (PI. 15, figs. 1-3). The fore-gut begins with the pharynx, which passes to the oesophagus dilating into the honey- stomach, crop, or ingluvies (Pl. 15, figs. 1-3; Pl. 16, fig. 4, 2). The latter passes by means of the cardial valve into the ventri- culus (mid-gut or stomach, -v). The hind-gut is divided into the anterior portion—the small intestine (it), and the posterior— the large intestine (-r) with the rectal glands (rq). The Malpighian vessels (mp) open on the border of the ventriculus and the small intestine. Fore-gut. The Fore-gut (pharynx, oesophagus, ingluvies) is lined within by a chitinous cuticle, to the exterior of which lies a layer of non-glandular epithelium (PI. 16, fig. 5, ep) resting on membrana basilaris. The latter is covered by a network of transversally striated muscle-fibres lymg in two layers— circular and longitudinal (Pl. 16, fig. 5, m4, ms). The valve of the ingluvies is represented by a capitulated ALIMENTARY CANAL IN THE BEE 511 eminence of the bottom of the ingluvies. The capitulum consists of four valves between which there is a cruciate slit (Pl. 16, fig. 5, pe). The valve is provided with three systems of muscular fibres—two longitudinal (PI. 16, fig. 5, m., ms) and one circular (Pl. 16, fig. 5, m,) between them. ‘The former serve to open the valve, the latter to close it. ‘The capitulum of the valve is set on a trunk connecting it with the stomach. From the circumference of the ventricular opening into the intestine hangs an intestiniform cardial valvule preventing the contents of the stomach from returning into the crop. All these data were already established by previous investigators. Mid-gut (Stomach). The stomach of the bee consists of a fairly thick cylindrical tube with numerous circular constrictions on it corresponding to which the epithelium of the stomach protrudes into its cavity in the form of folds. The epithelium consists of cylindrical cells which assume the shape of clubs on the ridges of the folds. At the bottom of the depressions between them are situated round groups of cells called cryptae. Exteriorly to the mem- brana basilaris is disposed the connective tissue in the form of small groups of cells. The muscular membrane of the stomach is formed by two layers of transversally striated muscle- fibres—interior circular, and exterior longitudinal. (a) The epithelium of the stomach consists of cells with an alveolar protoplasm of basophil character, (Pl. 16, fig. 7, ep ; figs. 9, 11, 18, ep). The oval nucleus with sparse chromatin granules, or a dense network of them, lies in the middle of the cell. In its protoplasm are produced oxyphil granules of secretion which are numerous in the superficial portion of the cells. In some sections the cells appear to be set on thin peduncles and to have truncated apices. ‘This picture, as well as the formation of evaginated swellings on the surface of the cells, is in most cases artificial (Pl. 16, fig. 12, bl). The superficial layer of protoplasm of the epithelium is transformed into a fairly broad band vertically striated and bearing the aspect of a brush of cilia (PI. 16, figs. 9,11, 12, wp). Nn2 512 HK. N. PAVLOVSKY AND E. J. ZARIN This hairy layer of protoplasm stains with iron haematoxylin in a grey colour, whereas the protoplasm remains black. The hairy band is covered above by a cuticle (not chitinous) ; together with the latter it is cast off into the cavity of the stomach in the form of a peritrophic layer (PI. 16, figs. 6, 7, p; figs. 9, 10, p). This casting off is repeated many times, on account of which in the stomach the membranes are disposed in concentrical layers, sometimes in very great numbers (Plo 16;figh 6)1p): The peritrophic membrane presents a structure known for a long time in the articulated animals. With regard to the bee Petersen has demonstrated that the said membrane of this insect contains a proteolytic ferment. The significance of the peritrophic membrane is interpreted in different ways. Some investigators believe it to serve for the defence of the tender stomach epithelium against mechanical injury by vegetable food, especially by the flower pollen in the bee. Such an interpretation cannot be extended to all Arthropods, since an analogous formation is also present in blood-sucking forms (Culex, Anopheles, according to Schaudinn), the liquid food of which cannot do any harm to the walls of the stomach. Probably those investigators are right who regard the peri- trophic membrane as a cuticle formed by the secretion of the stomach epithelium. Originating by transformation of the surface protoplasm of the cells, the membrane itself presents a hard secretion. In the depth of the hairy layer, as in a sponge (Pl. 16, fig. 9, p),is retained the liquid secretion of the stomach, on account of which the same quantity of ferment is capable of acting for a longer period upon the food contained in the cavity of the mid-gut. On account of the relative shortness of the intestine this mode of action of the ferments is of special significance, especially in herbivorous insects, since the food does not pass through the intestine so rapidly, being detained in the folds soaked with the digestive juices of the peritrophic membranes. Thus, in our opinion, they compensate the relatively small length of the intestine in insects. (b) At the bottom of the folds of the stomach epithelium ALIMENTARY CANAL IN THE BEE 518 are situated groups of smaller cells forming crypts which are weakly developed in the bee. The cells of the latter are disposed in two or three layers (Pl. 16, fig. 8; fig. 9, k). The deepest row situated on the basal membrane is represented by the smallest cells (in the section three or four of them are visible), covered above and laterally by larger cells bordering directly on the epithelium of the folds of the stomach (PI. 16, fig. 11, k). In general the protoplasm of the cells of the crypts are more basophil than the stomach epithelium (PI. 16, fig. 11, k,d). The nuclei of the eryptic cells are large and poor in chromatin. As described by Petersen we did not succeed in observing their karyokinesis. Nasonov (1898), however, observed the process of division of the nuclei in these cells. The cryptal cells present the sources from which the stomach epithelium is newly formed. Besides, their cells seem to produce a secretion themselves as well. The following facts confirm this supposition. ‘The most superficial cells of the cryptal are not adjacent to each other with their apices, so that there remains an ovoid lumen between them in the shape of a vacuole filled up with a drop of homogeneous secretion staiming pink with Giemsa’s stain (Pl. 16, fig. 7, k; fig. 11, ve). Besides, there is also observed an accumulation of secretion above the crypta which is revealed by displacement to the sides of the * hairs ’* of the superficial band of the stomach epithelium (PI. 16, fig. 9, k). In general the secretory processes in the stomach of the bee take the following course : (1) Separation of the peritrophic membrane (PI. 16, fig. 6, p ; fig. 9, p), (2) production of secretion by the surface of the glandular cells, (3) severance of the superficial portions of the . epithelial cells (Pl. 16, fig. 8, d), and (4) separation of a homo- geneous secretion by the cryptic cells (Pl. 16, fig. 11, ve). (c) The epithelium of the stomach lies on a basal membrane clothed exteriorly by a transversely striated muscular mem- brane, the muscle-fibres of which are very rich in sarcoplasm. The muscle-fibrils are disposed in bundles occupying the greater 514 E. N. PAVLOVSKY AND E. J. ZARIN part of the surface of the transverse section of the fibre (PI. 17, fig. 15, mf) at the point where the sarcoplasm and nuclei are scarce, and half the diameter of the fibre where the sarcoplasm and nuclei are strongly developed. ‘The nuclei always lie in the sarcoplasm nearer to the periphery of the muscle-columns (Pl. 16, fig. 14; Pl. 17, fig. 15, 16, sp), and not between the latter as in the analogous membrane of the small intestine (Pl. 17, fig. 18). Hind-gut. The hind-gut is divided into two parts—both in its anatomical and histological structure—the anterior—small intestine (Pl. 15, fig. 3; Pl. 16, fig. 4, a), and posterior—large intestine (Pi. 16, fig. 4, r). Small Intestine. The structure of the small intestine has already been estab- lished by previous investigators. We may add to these some details in the microscopical structure of its single-layered cylindrical epithelium. The cells of the latter are covered on their interior surface by a thick chitinous cuticle. The proto- plasm of the cells is divided into two portions, the exterior— sranular (Pl. 17, fig. 17, d), and interior—characterized by a rod-line striation (Pl. 17, fig. 17, ds). The fairly large rounded nucleus (n) lies nearer to the base of the cell. Interiorly to it in the layer of granular plasm are found large vacuoles with granules of secretion (vs). Both the protoplasm and secretion of the epithelium of the small intestine are oxyphil. The basilary membrane of the intestine (PI. 17, fig. 17, mb) is surrounded by circular muscle-fibres anastomazing with each other. They are thick and their nuclei are disposed along the axis of the fibres surrounded from all sides by bundles of myofibrils (Pl. 17, fig. 18, emf). In general the small intestine of the bee is characterized by the glandular character of its epithelium. The structure of the intestine described may serve as evidence either of its glandular function or of processes of absorption taking place ALIMENTARY CANAL IN THE BEE 515 in it, or, lastly, of its excretory rdle. We have hitherto only established one fact for certain—the complete absence of fer- ments in extracts from the small intestine of the bee. Large Intestine. The large intestine, similarly to the crop of the bee, presents a thin-walled sac which is capable of expanding to enormous dimensions, as seen by comparison of figs. 1 and 3 of Pl. 15. During the whole winter the bees do not evacuate their exere- ments, but continue taking food, on account of which their large intestine becomes overfilled with faeces and swells into a voluminous bladder. The scheme of structure of the large intestine is the same as in the crop. In the anterior third of its wall are situated six elongated cylindrical rectal glands (Pl. 17, fig. 19, rg), the microscopical structure of which was in general features correctly described by Snodgrass and Petersen. We have also succeeded in establishing certain interesting details elucidating the structure of these glands. From the part of the cavity of the rectum each gland is covered by a chitinous cuticle forming on the periphery of the organ a marginal fillet. Within the gland there is an axial cavity (Pl. 17, fig. 21, h) dividing it into two parts—an exterior thin wall (wa) and interior thick one (sn). The latter is formed by tall wedge- shaped cells polyhedral in transverse section. The exterior wall is formed by two layers of minute poly- gonal cells (Pl. 17, fig. 21, wa). At the point where both walls jom together there lies a syncytial layer of cells containing pigment inclusions (Pl. 17, fig. 21, s7). The exterior wall of the rectal gland (Pl. 17, fig. 20; fig. 22, wa) is perforated in some places by tracheae (tr) which pass into the cavity of the organ and penetrate with their branches into its inner wall. To these data, which are to be found in the literature, we may add the following : The ramifications of the tracheae passing to the thick inner wall of the rectal gland pass along the edge of the poly- 516 E. N. PAVLOVSKY AND E. J. ZARIN hedral ‘epithelial cells (Pl. 17, figs. 28, 25, tr). The layers of protoplasm of the latter adjacent to the tracheae consist of a substance staining deep black with Heideuhain’s iron haema- toxylin (Pl. 17, figs. 22, 25, z). These bordering layers differ from the alveolar-granular protoplasm of the cells in their dentate aspect; in some individuals they resemble coarse intercellular bridges; in others they are more weakly expressed; their striation, however, is always visible in a greater or less degree. It is possible to trace the course of the tracheae to four-fifths of the height of the cells. At this level the tracheae which have hitherto pursued a radial course give off lateral branches forming beneath the inner surface of the gland a network rich in anastomoses (PI. 17, fig. 24, tr). We did not observe anything like the opening of the tracheae directly into the cavity of the intestine in the rectum of the bee, as was described by Vallé in Diptera, The protoplasm of the large cells is in general granular in some individuals with a fairly distinctly expressed alveolar structure. ‘lhe protoplasm is oxyphil. To the chitinous cuticle is adjacent a layer of protoplasm staining less and bearing the aspect of vesicles lying close to each other. The nuclei of the cells described are of an irregular round shape. ‘They are disposed cither in the middle part, or basally, depend- ing upon the degree to which the protoplasm is filled up with granular inclusions. ‘lhe nuclei are poor in chromatin. The variation in the contents of the cells described probably is in Connexion with the seasons of the year. In the hibernating bee the large intestine of which had for several months been filled up with faeces, the protoplasm of the large cells of the rectal glands contains numerous globular inclusions and minute granules (Pl. 17, fig. 21, gr). Both are oxyphil, with the exception of some of the larger granules. In some of them are visible roundish portions not stained black with iron haematoxylin. All these formations occupy the middle two- thirds of the transverse section of the cell ; whilst in the basal quarter of it lies the displaced nucleus (Pl. 17, fig. 21). ALIMENTARY CANAL IN THE BEE 517 The protoplasm of the rectal glands of bees taken in ordinary condition, although granular, is devoid of the inclusions described above (Pl. 17, fig. 22, d). In bringing together these facts, we may speak of the absorp- tive role of the rectal glands, which appears to be correct a priori, onaccount of the long period during which the faeces remain in the rectum in bees hibernating in our latitudes. The microscopical structure of the tall epithelium of the gland points to a possibility of true glandular processes taking place in it. Below we shall discuss the conclusion according to which the rectal glands present the source of seasonal production of catalase, and the pomt of development of energetic oxidizing processes which is evinced by the intimate connexion between these organs and the tracheae. PHYSIOLOGICAL Part. There are few data in literature regarding the ferments found in the organism of the bee. The first works in this direction were conducted by Erlenmeyer and Planta in 1877. ‘The authors named dissected 152 worker-bees separating head, thorax, and abdomen, and infused them separately in glycerme. It was found that all the three extracts converted starch to dextrin and sugar, and saccharose to inverted sugar, the extracts from the head and abdomen being much more active than that from the thorax. The extracts from the head and abdomen algo contained a ferment dissolving fibrin of the blood, the latter extract being stronger than the former, whilst that from the thorax produced no effect. The methods applied by Erlenmeyer and Planta for the preparation of extracts is of no use at all, since the exterior division of the body of the bee into head, thorax, and abdomen does not correspond at all to the division of the intestine into its characteristic portions. In 1912 Petersen, whilst studying a question on the digestion in the bee, also made experiments on the determination of ferments in the digestive organs of the bee. In glycerine extracts from the stomachs of bees the author discovered the 518 E. N. PAVLOVSKY AND E. J. ZARIN following ferments: diastase, invertase, and a proteolytic ferment dissolving fibrin and splitting peptone. These are essentially all the data to be had in literature on the question discussed. Methods used in Preparing the Material for Chemical Investigation of the Ferments. The only fit material is presented by live bees, live for anatomical purposes. ‘They are chloroformed and dissected in physiological solution (0-75 per cent.) of common salt for the preparation of the intestine. The removed intestine is washed in a Petri dish with the physiological solution and separated into the following parts: crop, stomach, small and large intestines. Hach portion is further dissected in order to remove its contents, washed in a fresh portion of the same solution, and placed in a small evaporating glass with a small quantity of desiccated sand and several drops of glycerine. After the portions of the intestine of all the bees have been placed in glasses, they are GROUND with a glass mortar until a uniformly opaque emulsion is produced. Then to each glass is added the necessary quantity of glycerine or some other liquid, the whole is rapidly mixed and poured out into a jar with a hermetically closing glass stopper. Extracts were prepared with (a) glycerine, (b) distilled water, (c) a mixture of equal quantities of the liquids named, and (d) physiological solution of common salt. As an antiseptic a few (5-10) drops of toluol were added. ‘The jars were several times shaken thoroughly and left to stand in the dark- ness at the temperature of the room for different periods. The density of the extracts varied. We started from strong extracts, ninety bees per 80 c.c. of liquid, i.e. from each portion of the intestine taken from ninety bees an extract was prepared in 30 ¢.c. By degrees as the experiments progressed it proved to be more practicable to take weaker extracts, until finally we stopped at the proportion of 25 bees per 50 c.c. of liquid. ALIMENTARY CANAL IN THE BEE 519 Altogether thirty analyses were performed, a table relating to the periods of which is adduced below. No. of Date of Heaperi- | Preparation ment. | of Extract. 1 23 x 1916 2 71v 1917 3 22 v1 1917 4 ” 5 | 20 vir 1917 6 21 1x 1917 7 15 x1 1917 8 31 m1 1918 3) ” 10 141v 1918 ll 15 1v 1918 12 | 211v 1918 13 | 231v 1918 14 17 v1 1918 15 ” 16 29 17 2 vir 1918 18 3 vil 1918 19 » 20 6 vir 1918 21 oe) 22 9 23 24 | 20 vir 1918 25 » 26 ” 27 o> 28 p 29 =| 24x 1918 30 % Total No. of Bees from which Intestines were prepared. 90 workers 60 5 60 9 30 a 50 drones 40 workers 30 5 30 ; 30 D 15 °° 15 » 15 ; 15 : 15 »» 15 j 15 25 25 9° 15 drones 25 workers 25 ” 15 ” 15, 15 9 25 ” 15 “ 15 » 15 » 15s, 25 99 ZOE ss 795 bees TABLE I. Volume of fluid | Base on which Extract was mM C.C. prepared. 30 | Glycerine 25 oe 25 | Water 12-5 | Glycerine 25 7) 20 2° 30 | 15c.c. glyc.+ 15 ¢.c. water 30 2? ” 30 | Water 15 7-5 c.c. alyce.+ 7-5 ¢.c. water | 15 > 2? | 15 ” 29 15 » , 15 o is 15 » 15 » , 50 * 2 15 or) 2 50 - » 50 » 15__| Water 15 7°5 c.c. glye.+ 7-5 c.c. water 15 ‘| Glycerine | 50 =| 25 c.c. glyc.+ 25 c.c. water | 15 | Water 15 ___| Glycerine 15 | Physiol. sol. NaCl 15 | 7-5¢.c. glyc.+ 7-5 c.c. water 50 =| 25c¢.c. glyc.+ 25 c.c. water 50 =| Water Amount of Toluol in Drops. 10 10 10 10 ee ee | — As the primary aim of the present work we regarded the qualitative determination of the ferments in the different portions of the ventriculo-intestinal tract of the bee, omitting, meanwhile, the investigation of the ferments of their salivary glands. In the intestine were established catalase, inulase, lactase, invertase, lipase, pepsin, trypsin, chimosin, and emulsin. Of these ferments we have investigated in fuller detail the catalase and invertase. 520 EK. N. PAVLOVSKY AND HE. J. ZARIN Catalase. As is known, catalase is a ferment widely distributed in the animal and vegetable kingdom. Regarding its presence in the body of the bee no data are known in the literature. For the determination of catalase we employed a special apparatus constructed by one of us (Zarin). The process consisted in mixing 2 ¢.c. of corresponding extracts of ferments with 8 c.c. of water; to the filtered mixture were added 10 c.c. of freshly prepared 1 per cent. solution of hydrogen peroxide; the number of ¢.c. of oxygen evolved being marked after the expiration of twenty-four hours. It need not be mentioned that all the analyses were accom- panied by control experiments. The results obtained in the investigations are shown in Table IT. TaBLE II. Carauask IN THE INTESTINE OF THE Bes. Date of Concentvatian and Quantity of Oxygen evolved in c.c. No Baperi- Composition of | Small Large ment. Eetract. Crop. | Stomach. | J ntestine. | Intestine. 1 | 81v 1917 | 60 bees : 25 c.c. glyc. 0 Gili 0 2-0 2 | 23 vr11917 | 30 bees: 12-5 c.c. glye. 0 15 0 0 3 | 16x11917 | 30 bees : 15c.