14 iy . . 4 an . . A j 7 Vi eo = | ; ot “s . re : . . hs evan ty . 3H » 3 ‘ a os ea TOES: é ih) Ses ; Ra ) : +E AN hh ae Te LO ELON, ' ' , , pues wy at Aner * ae ' , ake b isl. jerry aee $8 per 0K “ss eer : ie pret egies.) ae Fear pen apie Hin eee Cao cae 1 lweetted tp de ry DEY, bie, ae a i ees se et ae ee a shoe & ae? he ir) iui toate ee Asie one i ee i aN AE eae ik bo 4 My diale “6 Roe FOR THE PEOPLE FOR EDVCATION FOR SCIENCE LIBRARY OF THE AMERICAN MUSEUM OF NATURAL HISTORY QUARTERLY JOURNAL OF MICROSCOPICAL SCIENCE. EDITED BY Size RAY LANKESYER, K.C.B., M.A., D.Sc., LL.D., F.RS., HONORARY FELLOW OF EXETER COLLEGE, OXFORD 5 MEMBER OF THE INSTITUTE OF FRANCE (associf ETRANGER DE L’ACADEMIE DES SCIENCES) 3 CORRESPONDENT OF THK IMPKRIAT ACADEMY OF SCIENCES OF ST. PETEKSBUKG, AND OF THE ACADKMY OF SCIENCES OF PHILADELPHIA, AND OF THE ROYAL ACADEMY OF SCIENCES OF TURIN; FORKIGN MEMBER OF THE ROYAL SOCIETY OF SCIENCES OF GOTTINGEN, AND OF THE ROYAL BOHEMIAN SOCIETY OF SCIENCES, AND OF THE ACADEMY OF THE WLINCEI OF ROME, AND OF THR AMERICAN ACADEMY OF ARTS AND SCIENCES OF BOSTON } ASSOCIATE OF THE ROYAL ACADEMY OF BELGIUM: HONORARY MEMERR OF THE NEW YORK ACADEMY OF SCIKNCES, AND OF THER CAMBRIDGE PHILOSOPHICAL SOCIETY, AND OP THE ROYAL PHYSICAL SOCIETY OF EDIN- BURGH, AND OF THE KIOLOGICAL SOCIETY OF PAKIS, AND OF THE CALIFORNIA ACADEMY OF SCIENCES OF SAN OF THK ROYAL ZOOLOGICAL AND MALACOLOGICAL SOCIETY OF BELGIUM; CORRESPONDING MEMBER OF THE SENKENBERG ACADEMY OF FRANKFURT-A-M} FOREIGN ASSOCIATE OF THK NATIONAL ACADEMY OF SCIENCES, U-S., AND MEMBER OF THE AMERICAN PHILOSOPHICAL SOCIETY 3 HONORARY FELLOW OF THE ROYAT SOCIETY OF EDINBURGH; LATE DIRECTOR OF THE NATURAL HISTORY DEPARTMENTS OF THE BRITISH MUSKUM, LATE PRESIDENT OF THE BRITISH ASSOCIATION FOR THE ADVANCEMENT OF SCIENCE; LATE FULLERIAN PROFESSOR OP PHYSIOLOGY IN THE ROYAL INSTITUTION OF GREAT BRITAIN; LATR LINACRE PROFESSOR OF COMPARATIVE ANATOMY AND FELLOW OF MERTON COLLEGE, OXFORD) EMERITUS PROFESSOR OF ZOOLOGY AND COMPARATIVE ANATOMY IN UNIVERSITY COLLEGE, UNIVERSITY OF LONDON. WITH THE CO-OPERATION OF ADAM SEDGWICK, M.A., F.R.S., FELLOW OF TRINITY COLLEGE, CAMBRIDGE. AND PROFESSOR OF ZOOLOGY IN THE IMPERIAL COLLEGE OF SCFENCE AND TECHNOLOGY, LONDON ; SYDNEY J. HICKSON, M.A., F.R. BEYER PROFESSOR OF ZOOLOGY IN THE UNIVERSITY OF MANCHESTER, E. A. MINCHIN, M.A., PROFESSOR OF PROTOZOOLOGY IN THE UNIVERSITY OF LONDON, AND GILBERT C. BOURNE, M.A., D.Sc., F.R.S., OXFORD. FRANCISCO, AND LINACRE PROFESSOR OF COMPARATIVE ANATOMY, AND FKLLOW OF MERTON COLLEGE, VOLUME 55.—New SERIEs. With Vithographic Plates und Text-Figures AL AE in AS LONDON: J. & A. CHURCHILL, 7, GREAT MARLBOROUGH STREET. 1910: if eee AAU ut Men ts TVA, GU Aen ie Tcthinc A i 7 KUHEW UGE, BAAR | ; | YAOT ETE SANTA io CON Lewy tS f CONTENTS OF No. 217, N.S., APRIL, 1910. MEMOIRS : On the Anatomy and Systematic Position of Incisura (Scissu- rella) Lytteltonensis. By GitBertr C. Bourne, Fellow’ of Merton College, Oxford, and Linacre Professor of Comparative Anatomy. (With Plates 1-5) The Eye of Pecten. By W. J. Daxin, M.Sc., ipamoristeatae aad Assistant Lecturer in Zoology, University of Belfast. (With Plates 6 and 7, and 2 Text-figures) : Observations on Certain Blood-Parasites of Biahiee occurring at Rovigno. By Prof. E. A. Mincurn, M.A., and H. M. Woopcock, D.Sc. (With Plates 8-10) : : On Ganymedes anaspidis (nov. gen., nov. ere a Gregarine from the digestive tract of Anaspides tasmaniz (Thomp- son). By Junian S. Huxusy. (With Plate 11, and 5 Text- figures) : 2 : The Fetal Membianes of tiie Vetteitatas By A. A. W. Husrecur CONTENTS OF No. 218, N.S., JUNE, 1910. MEMOIRS : The Structure and Life-History of Crithidia melophagia (Flu), an Endo-parasite of the Sheep-Ked, Melophagus ovinus. By ANNIE Porter, B.Sc.Lond., Zoological Research Laboratory, University College, London. (With Plates 12 and 13, and 15 Text-figures) : : Studies in the Experimental epee of Bee. By GEOFFREY Smiru, Fellow of New College, Oxford. (With Plate 14) Some Observations on a Flagellate of the Genus Cercomonas. By C. M. Wenyon, M.B., B.S., B.Sc., Protozoologist to the London School of Tropical Medicine. (With 19 Text-figures) ‘ _ Some Observations on a New Gregarine (Metamera schubergi nov. gen., nov. spec.). By H. Lynpuurst Dvuxs, B.A., B.C. Cantab. (With Plates 15 and 16) . : ‘ On the Anatomy of Histriobdella Homans By CRESSWELL Surarer, M.A., Trinity College, Cambridge. (With Plates 17— 20, and 5 Text-figures) : On the Artificial Culture of Marine Plankton @npunisrad: By E. J. Auten, D.Se., Director of Laboratories and Secretary of the Marine Biological Association, and E. W. Netson, Assistant Naturalist PAGE 49 113 241 1V CONTEN''S. CONTENTS OF No. 219, N.S., SEPTEMBER, 1910. MEMOIRS: Notes on the Free-living Nematodes. I. The Hermaphrodite Species. By F. A. Ports, M.A., Fellow of Trinity Hall, Cam- bridge, and Demonstrator of Comparative Anatomy in the University. (With 11 Text-figures) : : : J Observations on Trypanoplasma Congeri. Part I. The Division of the Active Form. By C. H. Martin, B.A., Demonstrator of Zoology, University of Glasgow. (With Plate 21, and 1 Text- figure) . ; : ; The Development ‘of Agila Se tert By A. M. Carr SaunpERs and Maraaret Pootr. (With Plate 22,and 20 Text- figures) y The Relation pate: een Light and eee for acetate in Creni- labrus and Hippolyte. By F. W. Gamsuz, F.R.S., Mason Professor of Zoology, University of Birmingham. (With Plate 23) Is the Trophoblast of Hypoblastic Origin as Assheton will have it? By A. A. W. Huprscutr. (With 7 Text-figures) . The Origin and Formation of Fibrous Tissue produced as a Reaction to Injury in Pecten Maximus, as a Type of the Lamellibranchiata. By G. H. Drew, B.A., Beit Memorial Research Fellow, and W. Dr Moraan, F.Z.S. (With Plate 24) CONTENTS OF No. 220, N.S., NOVEMBER, 1910. MEMOIRS : The Division of the Collar-Cells of Clathrina coriacea (Montagu); a Contribution to the Theory of the Centrosome and Blepharoplast. By Murizt Roperrson, M.A., and E. A. Mincuin, M.A. (With Plates 25 and 26) : Studies on Avian Hemoprotozoa. I.—On Certain Parasites of the Chaffinch (Fringilla celebs) and the Redpoll (Linota rufescens). By H. M. Woovcock, D.Se.(Lond.), Assistant to the University Professor of Protozoology. (With Plates 27-31) Studies on Ceylon Hematozoa. No. II.—Notes on the Life-Cycle of Hemogregarina nicoriz, Cast. and Willey. By Murren Rosertson, M.A. (With Plates 32-41 and 1 Text-figure) On the Origin and Migration of the Stinging-Cells in Craspedote meduse. By CuHarueEs L. BoutencErR, M.A.Camb., Lecturer on Zoology in the University of Birmingham. (With Plates 42 and 43 and 5 Text-figures) The Researches of Bouvier and Bowiage on Maikationss in Crustal of the Family Atyide. By W. T. Cauman, D.Sc. (With.4 Text-figures) : : : Tire, INDEX, AND CONTENTS. PAGE 433 485 497 611 641 785 INCISURA (SCISSURELLA) LYTTELTONENSIS. 1 On the Anatomy and Systematic Position of Incisura (Scissurella) lytteltonensis. By Gilbert C. Bourne, Fellow of Merton College, Oxford, and Linacre Professor of Comparative Anatomy. With Plates 1—5. Wuen Mr. Geoffrey W. Smith was in Tasmania in 1907-08 T asked him to collect for me any rare or remarkable speci- mens of gastropod molluscs and preserve them in a form suitable for anatomical and histological examination. Among other forms Mr. Smith obtained for me, through the kind offices of Mr. C. Hedley, of the Australian Museum, Sidney, a number of specimens of the little gastropod which is the subject of the present memoir. They were preserved in Perenyi’s fluid, which of course dissolved the shells, but except for the difficulty of staining always resulting from a prolonged immersion in this reagent, the histological condition of the specimens leaves little to be desired. Scissurella lytteltonensis was described in 1893 by H. A. Smith (16), who noted certain differences between the shell of this and other species of the genus Scissurella, but evidently did not consider them of generic importance. In 1904 C. Hedley (8) recalled attention to these differences, and founded the new genus Incisura for the reception of the species which, he maintained, is marked off from all other Scissurellide as also from all Pleurotomariide by the brevity of the slit in the shell, by the absence of raised rims or keels on either side of the sht, by the subterminal apex, voL. 55, PART 1.—NEW SERIES. 1 2 . GILBERT C. BOURNE. by the absence of spiral sculpture, and by the remarkable solidity of the shell. He further asserted that his new genus cannot, because of the above-mentioned differences, be in- cluded among the Scissurellide, and suggested that it is a member of the Fissurellide in which development has been arrested, so that the larval characters of the shell have persisted in adult life. Hedley was evidently unacquainted with Pelseneer’s (12) memoir, containing an account of the anatomy of this very species and of Scissurella costata, which, brief as it is, leaves no doubt that the New Zealand and the Mediterranean species are members of the same family, but at the same time discloses so many anatomical as well as conchological differences that they may well be placed in different genera. After some consideration I am of the opinion that Hedley’s genus should stand, because the New Zealand species, in addition to the conchological characters enumerated above, differs from the Mediterranean species in the following particulars: (1) In the shape of the radular teeth. (2) In the shape of the foot, which is long and narrow in 8. costata and S. crispata, but short and broad in Incisura lytteltonensis. (3) In the absence of cirrhi below the epipodial tentacles in Incisura. (4) The greater development of the right columellar muscle, and the more symmetrical disposition of the mantle in Incisura. In its general anatomical features Incisura bears much the same relation to Scissurella as Septaria bears to Paranerita among the Neritide. The systematic position of the Scissu- rellidee will more conveniently be discussed at the end of this paper. Scissurella is placed by most authors among the Pleuro- tomariide, though a few recognise the Scissurellide as a separate but closely allied family. A full description of its anatomy is therefore much to be desired, but the accounts that have hitherto been published are insufficient. Vayssiére (18) has given a short and, as far as it goes, a good account of the external features of S. costata var. levigata, and has figured and described the radula and jaws of this species. INCISURA (SCISSURELLA) LYTTHLTONENSIS. 3 Pelseneer (12), in his well-known memoir on the morphology of primitive mollusca, gives seven figures of sections of S. costata and two of Incisura lytteltonensis in addi- tion to three figures of the external features of the latter species. ‘The description he gives in the text is concise, and furnishes a good general idea of the anatomy of the family ; but he does not give sufficient detail to enable one to make a critical examination of its systematic position. Hence, having sufficient material at my disposal, I have thought it worth while to make a thorough study of the anatomy of Incisura lytteltonensis. Incisura, as Mr. Hedley states in a letter accompanying the specimens, is found on the seaweed Cystophora in rock-pools in Lyttleton Harbour, where it is associated with Rissoina, Cantharides, and Gibbula. It may be inferred from its shape and structure that it is semi-sessile in habit, but it is not attached to one spot lke a limpet. On the contrary, it is fairly active, and one of the specimens was observed to crawl for a distance of nearly half an inch in the space of a quarter of an hour. When alive it is of a pink colour, and this tmge is sometimes preserved in the shell. The length of the animal, when contracted in spirit, is about 1 mm. External features.—These have been correctly if somewhat diagrammatically figured by Pelseneer. A three-quarter ventral view of the animal is given in fig. 1. Attention may be called to the following points: The visceral spire is attenuated and much reduced, its coiled apex containing only some lobes of the liver and, in some specimens, a portion of the gonad. ‘The last whorl is greatly expanded laterally, compressed dorso-ventrally, and contains all the important organs of the body. The snout is moderately long, termina- ting in a trumpet-shaped expansion, on the ventral side of which is the mouth. The mantle is large, and in contracted specimens completely covers the head and the greater part of the snout. The mantle slit, corresponding to the labral incision in the shell, is short, and situated nearly opposite the right eye; its margins are furnished with short digitiform 4, GILBERT C. BOURNE. processes bearing projecting sense-papille, such as have been described by Vayssiére in Scissurella costata. The cephalic tentacles are moderately long, reaching in their contracted state as far forward as the end of the snout. They are fringed with a large number of small, conical sense- papille, which, in Incisura, are not scattered all over the surface of the tentacles as figured by Vayssiére for S. costata, but are arranged in two multiple rows on the inner and outer margins of each tentacle (fig. 27), somewhat like the pinnules on the tentacles of an Alcyonarian polyp. The structure of these sense-papillee will be described further on. The eyes, which are closed and provided with a cornea and lens, are situated on prominences at the outer sides of the bases of the tentacles. Just below and behind the tentacle of each side is a short sub-ocular tentacle which does not bear sense-papillz like the cephalic tentacles, but is richly ciliated and glandular in structure. In the single male specimen of which I have cut sections, the sub-ocular tentacle of the right side is somewhat enlarged, spatulate in form, and more abundantly provided with gland-cells than in the females. In all the females I have examined the sub-ocular tentacle of both sides is digitiform. The foot, as is shown in fig. 1, is rather short and triangular in shape, the apex of the triangle being posterior. In shape and in the size of the broad, creeping sole it differs consider- ably from the narrower elongated foot of S. costata and S. crispata. The epipodium begins as a low ridge in about the middle third of the foot, and increases in size posteriorly. As described by previous authors it bears three moderately long epipodial tentacles on each side of the body towards the posterior end of its course. These tentacles bear lateral rows of sense-papillee exactly like those of the cephalic tentacles, but there are no ventral cirrhi in connection with them as in 8. costata. The epipodial folds meet posteriorly above the posterior end of the foot, and just dorsal to and in front of their union is a muscular opercular lobe bearing the operculum. The last-named structure is small, horny, and INCISURA (SCISSURELLA) LYTTELTONENSIS. 3) multispiral, as in other Scissurellide. It must be regarded as vestigial since, as is the case in Pleurotomaria, it cannot be of any use in closing the aperture of the shell. There are two columellar or shell-muscles (fig. 2) symmetrically disposed right and left of the middle of the body, the right muscle being slightly larger and extending rather further back than the left. As it is almost impossible to make dissections of an animal scarcely exceeding 1 mm. in length, the following account of the anatomy of Incisura is mainly founded on reconstructions from sections, but I succeeded in making some satisfactory whole preparations of the ctenidia, and have checked the results of my reconstructions as far as possible by the study of whole specimens cleared in various ways.. Fig. 2 is a camera drawing of a specimen stained in picro-carmine and mounted in oil of cloves; it shows as much of the general anatomy as can be made out by this method. Figs. 3, 4, and 5 are reconstructions from sections showing respectively the anatomical relations of the alimentary tract, the kidneys and pericardium, and the nervous system. Figs. 6 to 12 are camera drawings of some of the sections from which the re- constructions were made. Organs of the pallial complex.—lIncisura is typically zygo- branchiate, and the position and general characters of the ctenidia, hypobranchial glands, left kidney, and pericardium have been correctly described by Pelseneer. The ctenidia.—Both right and left ctenidia take their origin from the roof of the mantle-cavity, close to the anterior end of the columellar muscle of their respective sides of the body. The left ctenidium lies almost transversely across the neck of the animal, its anterior extremity reaching nearly as far as the base of the right tentacle (fig. 2), and it is closely com- pressed between the body-wall and the roof of the mantle. The right ctenidium, on the other hand, lies for the most part in front of the right columellar muscle, and the bulk of it hangs vertically downwards in the space enclosed between the mantle and the outer side of the foot (fig. 7). Pelseneer 6 GILBERT C. BOURNE. has described the right ctenidium as mono-pectinate, but, as may be seen in fig. 7, it is really bi-pectinate; the external lamelle, however, are few in number, and in some specimens are so feebly developed that they might easily be overlooked. It is at first rather difficult to make out the details of the structure of the ctenidia and to institute an exact comparison between them and those of closely allied Aspidobranchs, but a careful study of sections and whole preparations shows that they are constructed on the familiar pattern. Hach ctenidium consists of an axis, the posterior part of which is fused to the roof of the mantle-cavity and extends back in the angle of that side of the mantle-cavity to which it belongs, lying just above the columellar muscle. The anterior end of the axis is free, and the large osphradial ganglion, as is always the case in Aspidobranchia, is situated at the point where the axis becomes free from the mantle. ‘This point, in Incisura, corresponds with the anterior end of the columellar muscle. In the case of the left ctenidium that part of the axis which is fused to the mantle bears no filaments, but, as will be des- cribed further on, this statement does not hold good for the right ctenidium. Taking the left ctenidium for the purpose of description: its free apex projects into the mantle- cavity in front of the columellar muscle as a thin, tri- angular lamina, which, as already explained, is bent over to the right, and also is twisted about its own axis from right to left in such wise that the morphologically outer row of filaments become posterior in position, the morphologically inner row anterior. ‘The efferent branchial vessel runs, as is always the case, along the dorsal, here the posterior margin, and the afferent vessel along the ventral, here the anterior margin of the axis. ‘lhe inner and now anterior filaments borne on the free portion of the axis are short and not more than four or five in number, and are folded backwards over the upper (morphologically ventral) side of the axis, appa- rently as a result of the latter being twisted from right to left in a narrow space. ‘lhe morphologically dorsal edges of the anterior filaments are consequently maintained in a dorsal INCISURA (SCISSURELLA) LYTTELTONENSIS. 7 position. But in the case of the posterior filaments, which are eight in number and much longer than the anterior fila- ments, the twisting of the axis has brought the ventral surfaces into a dorsal position. Fig. 16 represents a section through the anterior and fig. 17 a section through the posterior row of filaments. Hach is more or less quadrangular in out- line, its lateral walls formed of long columnar cells bearing long and fine cilia, which in contracted specimens appear to interlock like the cilia of the ciliated discs of filibranch Lamellibranchia. I do not think, however, that their function is to hold the filaments together, but simply to create currents over the surfaces of the filaments. Their inter- locking is simply due to their becoming matted together in consequence of the contraction of the gillin spirit. On the ventral surface of each filament is a band of very short cilia. The dorsal edge of the filament bears no cilia externally, but, as shown in the figures, is produced to form a peculiar bolster- shaped swelling, which, as far as I am aware, has no analogue in the gills of any other mollusc. ‘his dorsal glandular ridge, as I will call it, takes its origin from near the free distal end of the filament, and extending along the dorsal face of the latter is closely fused to it for the greater part of its length, but on approaching the proximal end of the fila- ment the glandular ridge becomes free and ends in a small rounded projection. ‘lhe ridge is traversed throughout its length by a small ciliated canal, which makes no communica- tion with the blood channel of the filament, but opens into the mantle-cavity in the angle between the free proximal extremity of the ridge and the filament. ‘This communication with the mantle-cavity, as seen in section, is shown in the central filament in fig. 17. In the filament on the right hand in the same figure the section passes through the middle of the glandular ridge, and the ciliated canal is seen to be closed in on all sides and to be situated near the ventral, i.e. the filamentary side of the ridge. ‘The same features are shown in the ridge attached to the right-hand filament in fig. 18, but in the case of the left-hand filament in this figure the 8 GILBERT C. BOURNE. section passes through the more distal part of the ridge, and the ciliated canal is seen to be smaller and situated near the dorsal side of the ridge. A little further on it ends blindly. As the figures show, the ridge is made up of a sheath or cortex of elongated, fusiform cells, which pass nearly trans- versely round the periphery of the ridge, and a medulla of large, closely packed ovoid or fusiform cells having large nuclei and granular cell contents. The cells abutting on the lumen of the ciliated canal are usually larger and more granular than those more peripherally situated, and their histological characters leave little doubt that they are secretory. It is noticeable that there are very few if any glandular cells interspersed among the columnar ciliated cells of the filament, and the glandular ridge appears to have taken over the secretory functions, and to replace the secretory cells scattered over the surface of gill-filaments of other Mollusca. The extreme specialisation exhibited by the formation of a closed canal into which the secretory cells discharge their products is certainly a remarkable feature in Incisura. The central blood-channels of the filaments, as may clearly be seen in the figures, are elongate-oval in shape, and their walls are strengthened, for about half their extent, by flattened, chitinous, skeletal bars, which, as in other molluses, may be traced to the proximal end of each filament, where they diverge from one another, and curve round to run up in the walls of the blood-spaces of the adjacent filaments (fig. 19). As M. F. Woodward (19) has shown that in Pleurotomaria these skeletal bars run along the dorsal edges of the gill-filaments, whereas in Nucula they run along the ventral edges, it is of some interest to determine the position of these bars in Incisura, which is usually reckoned as belonging with the genera Scissurella and Schismope to the Pleurotomariide. It is clear from an inspection of fig. 17, representing a transverse section through the posterior gill- filaments of the left ctenidium, which, as explained above, are turned upside down, that the skeletal bars lie on the dorsal sides of the filaments, and the same thing can be INCISURA (SCISSURELLA) LYTTELTONENSIS. 9 seen still more clearly by inspection of fig. 7, in which the relations of the gill-filaments to the axisare obvious. In the anterior gill-filaments of the left ctenidium the skeletal bars appear to be ventral in position, but this is because these filaments are reflected backwards and their natural surfaces are reversed. Incisura, then, agrees with Pleurotomaria, and also with Trochus (fide Fleure and Gettings) and Fissurella, and differs from Nucula. But it must be observed that Woodward went further than the facts warranted when he asserted that the position of the gill-bars indicated a more remote affinity between Pleurotomaria and the primitive Lamellibranchia than is generally supposed. As a matter of fact the skeletal bars differ considerably in position in some not remotely related mollusca. In Solenomya, for instance, they are shifted to a more dorsal position than in Nucula, and in the Filibranchia they are actually dorsal. The fact is, as Woodward himself pointed out, these skeletal bars have a physiological rather than a morphological significance, and are always developed in close relation to the tracts of cells bearing specially long or functionally important cilia. Hence, in Filibranchia we find them related to the ciliated discs, which are near the dorsal edges of the filaments. In so small an object as Incisura it is very difficult to make sure of the presence or absence of a septum dividing the blood-channel into an afferent and an efferent moiety, but I am tolerably certain that such a septum exists, as shown in fig. 18. But it is not always placed transversely, but may be oblique or even nearly longitudinal. The attached portion of the axis of the right ctenidium extends far back in the extreme right-hand corner of the mantle cavity, lying close above the columellar muscle of that side, and gives off some three or four short filaments before reaching the level of the osphradial ganglion. At this spot there is a break in the continuity of the filaments, none being formed in the immediate proximity of the ganglion, but in front of it the ctenidial axis becomes free, and drops vertically down in front of the columellar muscle to hang in 10 GILBER! C. BOURNE. the space between the foot and mantle, as shown in fig. 7. The basal portion of the axis is also enlarged at this point, and gill-filaments are given off from both sides, both of the free apex and of the broad basal portion. These filaments are not simply digitiform like those of the left ctenidium, but are plate-like, with the glandular ridge running along their dorsal margins, as shown in fig. 7. As the skeletal bars and glandular ridges are on the inner side of the filaments of the inner row, the free axis must have been rotated through 45° to bring the dorsal surface inwards. The plate-like filaments springing from the expanded base of the free part of the axis spread out on, and are attached to, the adjacent parts of the mantle; the filaments, or as they more appropriately might be called, the ‘‘ gill-lamelle” of the inner row extending dorsally along the inner surface of the mantle, while those of the outer row, two or three in number, pass round the front edge of the columellar muscle and run back for some distance below it as ridges projecting inwards from the dependent margin of the mantle (fig. 8, m. br.) The blood supply to the ctenidia will be described in connection with the heart. The rectum runs diagonally from left to right in the roof of the mantle-cavity, and the anus opens opposite the slit in the mantle edge. In much contracted specimens, such as that from which fig. 2 was drawn, the anus is situated some dis- tance from the slit, but in other less contracted specimens it is close to it. The hypobranchial glands lie in the roof of the mantle on either side of the rectum, between it and the ctenidia. Both consist of a more or less extensive modified glandular patch of the internal epithelium of the mantle. The glaud-cells are very large relatively to the size of the animal, and’are of two kinds: large ovoid cells filled with large granules which stain deeply in hematoxylin and green in picro-indigo-car- mine; these are therefore mucigenous cells. ‘he other gland-cells are- of nearly the same size and shape, but have clear or minutely granular contents. The left hypobranchial INCISURA (SCISSURELLA) LYTTELTONENSIS. 11 gland is much the smaller of the two (figs. 7 and 8) ; posteriorly it is a narrow strip of glandular epithelium lying between the terminal part of the rectum; anteriorly in front of the anus it becomes broader and extends about as far forward as the level of the mantle slit, but stops far short of the anterior border of the mantle. In this pre-anal region the right and left hypobranchial glands are very closely approximated in the middle line. The right hypobranchial gland has approxi- mately the same anterior extension as the left, but runs back- wards on the right side of the rectum nearly to the posterior end of the mantle-cavity. Comparing the arrangement with that described by Woodward for Pleurotomaria, it is obvious that the pre-anal portions of the two glands of Incisura correspond to the large anterior hypobranchial gland, “ par- tially divided by a median longitudinal furrow into two halves,” of Pleurotomaria, and the posterior portions of the two glands of Incisura correspond to the two “ additional mucous glands” lying on either side of the rectum of Pleuro- tomaria. But whereas in the latter genus the left additional gland is conspicuously the larger, in Incisura it is the right posterior portion of the gland which preponderates in size, the left gland being small, no doubt because of the relatively large size of the left kidney, for the hypobranchial gland does not extend beneath this organ. The pericardium, as in all Rhipidoglossa except the Helici- nide, is traversed by the rectum. It is relatively of large size, and can always be distinguished in whole specimens as a clear space surrounding the first bend of the rectum on the left side of the body behind the columellar muscle. At this point it lies close to the surface of the body, and its outer wall is very thin and transparent (fig. 11). The exact limits of its extension to the right are very hard to make out, because the left kidney projects into it from above, and its cavity is largely blocked by the auricles. Its extent, as far as I am able to determine it by reconstruction from sections, is represented by the thick black line in fig. 4. The large transverse extension of the pericardial space, as compared Le GILBERT C. BOURNE. with its narrow limits in Pleurotomaria, Haliotis, or Trochus, is correlated with the tendency towards a secondary bilateral symmetry, the development of two columellar muscles, and the position of the ctenidia wide apart from one another on the right and left sides of the body. The necessary result is an increased breadth of the body, and the blood returning to the heart by the efferent branchial vessels has to traverse a considerable distance before reaching the ventricle. In other words, the auricles are considerably elongated, and the peri- cardium has to be extended to receive them. Very similar relations are seen in Fissurella. The heart and circulatory system.—The ventricle is placed rather far forward on the rectum; no further for- ward than in Fissurella, but much further forward than in either Pleurotomaria or Haliotis. The walls of the ventricle are so thin and feebly muscular that they are difficult to recognise, even with the highest powers of the microscope. The auricles also have very thin walls but are more easily recognisable. ‘The left auricle is relatively very large (fig. 10), and its anterior border gives off a number of short and wide sinuses, which penetrate the folds of the wall of the left kidney and vascularise this organ. The right auricle is of smaller size. The course of the blood-vessels, as far as I was able to determine it, is of the usual diotocardiate type, and is diagrammatically represented in fig. 4, which is fully lettered and needs no further description. I was unable to trace the course of the aorta, but the blood, after passing to the foot and the various viscera, is evidently collected in a large sinus lying below the pedal ganglia, and is returned to the afferent branchial vessels by sinuses running over the dorsal side of the great mass of muscle-fibres which diverge on each side of the foot to form the columellar muscles. The kidneys.—The left kidney (figs. 8, 9, and 13) is of comparatively large size, but its structure and histological characters leave no doubt that it corresponds physiologically to the papillary sac of the Pleurotomariidz, Haliotide, and Turbonidz, for it is unquestionably phagocytic and not depu- INCISURA (SCISSURELLA) LYTTELTONENSIS. 13 ratory. It is a triangular sac lying close alongside of the rectum and projecting largely into the pericardium. It opens into the mantle-cavity by a simple slit-shaped aperture with somewhat tumid lips (fig. 9). The majority of the specimens of which I cut sections were females, and in all of them the cavity of the sac was large and but slightly broken up by ridges or papillae projecting into it. In all the specimens the epithelium lining the cavity of the sac and covering the papillae had the characters shown in fig. 14. The cells are large and pale, with pale nuclei, and most of them are stuffed with rod-shaped masses which stain very deeply with iron hematoxylin. Whatever may be the nature of these rods, which, as shown in the figure, have rhomboid outlines and are apparently crystalline, they have clearly been taken up by the amceboid cells of the left kidney from the adjoining blood- spaces, for these latter are also filled with similar rods, which, however, are smaller, more transparent, and stain less deeply in hematoxylin. The left kidney differs considerably in appear- ance according to its functional activity. In some specimens no rod-shaped bodies can be detected in the cells, and the walls of the kidney sac then appear pale and thin. In other specimens, again, no rod-shaped bodies can be seen in the blood-sinuses, but the kidney-cells are stuffed so full of them that their outlines are no longer distinguishable. In other specimens, again, the rod-shaped bodies are abundant in the blood-sinuses and more or fewer are present in the kidney- cells. A portion of the epithelium of a specimen in the last condition is represented in fig. 14. The fact that the histo- logical character of the left kidney or papillary sac in Haliotis and Trochus is different from that of the right kidney was established by Rémy Perrier in his careful studies on the kidneys of prosobranch Gastropoda, and Pelseneer (11) after- wards showed that the amcebocytes of the papillary sac take up solid particles, such as carmine or Indian ink, injected into the blood-sinuses, whereas the secretory cells of the right kidney eliminate sulphindigotate of soda injected in solution into the blood. Both kidneys of Patella are depuratory, that 14, GILBERT C. BOURNE. is to say, they take up sulphindigotate of carmine from the blood, but there is still some doubt as to the very rudimentary left kidney of Fissurella. Rémy Perrier (14) describes its histological structure as identical with that of the right kidney, and consequently it has been generally assumed that, like the left kidney of Patella, it is depuratory in function, but this is not certain and the subject requires renewed investigation. All observers agree in describing the left kidney of the Fissurellide as being in a rudimentary con- dition, and it is possibly nearly if not quite functionless. It may even be absent in some species of Fissurella, for I have been unable to find a trace of it in transverse and longitudinal sections of F. greca. In the single male specimen of Incisura of which I have sections the left kidney is larger than in any of the females; the papilla projecting into its cavity are more numerous, are covered with a much more definite layer of epithelial cells, and I could not find any trace of phagocytosis in the latter. Whether this is a constant sexual difference I cannot say, as I was unable to find another male. A section through this kidney is represented in fig. 15, which also shows the left reno- pericardial canal. The last-named structure is found in the same position in both male and female. It opens into the kidney close to the external aperture of the latter, and runs towards the left as a very fine canal which traverses the floor of the kidney and opens into the left-hand corner of the peri- cardium, as indicated in the figure. ‘lhe cells lining the nephric end of the canal appear to bear very fine cilia, but I am unable to speak with certainty on this point. The right kidney of Scissurella and Incisura has been very briefly described by Pelseneer (12), who figures it as a very small tube lying below the rectum in 8. costata and to the right of the rectum in Incisura. He describes it as being rather narrow in its anterior portion and says further: “Il s’etend partiellement sous le rectum, comme chez Trochus, et pénétre dans la masse viscerale, au cOté droit de ce corps, sur et entre les convolutions de l’intestin.” INCISURA (SCISSURELLA) LYTTELTONENSIS. 15 I may amplify this account by saying that the right kidney of Incisura is a structure of considerable size and importance which may be described as consisting of three lobes. The most anterior lobe varies considerably in size: it lies in the roof of the mantle-cavity to the right of the rectum (figs. 4 and 10) and somewhat posterior to the left kidney. It opens byasimple slit-shaped aperture (fig. 10, /.7.0.) into the mantle- cavity, and a few sections further back than the one figured it extends over to the right, forming a considerable projection into the posterior part of the mantle-cavity. Posteriorly it gives off two lobes. That on the right runs nearly vertically downwards close to the right side of the vertical loop of the intestine and passes inward among the viscera, curving round the floor of the middle part of the stomach and eventually coming in contact with the gonad, but it does not effect any communication with this organ. The left posterior lobe passes below the rectum and overlies the anterior cecal end of the stomach. The excretory cells of the depuratory kidney of Gastropods are notoriously difficult to preserve, and in my specimens were too much macerated to admit of a satisfactory study of their structure. For the same reason I have been unable to satisfy myself completely as to the relations of the right reno- pericardial canal. For some time I was uncertain whether any communication existed between the right kidney and the pericardium, but the series of sections represented in figs. 22 to 26 demonstrate that this connection does exist, and that, as in Trochus, there is an intimate connection between the right reno-pericardial canal and the gonaduct. In fig. 22 the ovi- duct (od.) is seen lying close to the right side of the anterior lobe of the kidney, and from it a narrow canal leads upwards and inwards. The histological features of this canal are not well preserved in any of my specimens, but its walls appear to be formed by cubical epithelial cells containing small, deeply staining nuclei, whose characters as shown in figs. 22 and 23, suggest that they bear cilia and form a ciliated funnel open- ing into the pericardium. The connection between the canal 16 GILBERT C. BOURNE. and the pericardium is clearly shown in fig. 24, and figs. 28 to 25 show that the lower end of the canal is, in fact, con- tinuous with the gonaduct, and opens along with it into the kidney, close to the external orifice of the latter. It should be noted as a peculiar feature in Incisura that there is no distinct duct to the right kidney ; its simple slit-like opening into the mantle-cavity is a Pectinibranch character. The gonad, in both sexes, is a simple tubular structure lying to the left side of the stomach, and in the case of the ovary partly embracing this organ. The anterior end of the gonad extends as far forward as the posterior limit of the mantle- cavity and ends blindly below the first bend of the rectum. The cavity of the ovary, in all my specimens, is filled with ova in all stages of development, the ripe ova being very large relatively to the size of the animal, and abundantly supplied with yolk-granules. The testis, in the single male I have been able to examine, is very small, and I think the individual must have been a spent one, as the cavity of the testis only contained a few free spermatozoa and I could find no trace of spermatogenesis. The course of the gonaduct and its connection with the right kidney has been correctly but all too briefly described and insufficiently figured by Pelseneer. He only says of it: ‘Ta glande genitale est unique et occupe le sommet de la masse viscerale. Elle n’a pas d’orifice exterieure ; son con- duit arrive au rein droit.”” But it would be difficult for any- body to guess the course of the gonaduct before its arrival at the right kidney by an inspection of his fig. 115, perfectly correct as it is. As shown in fig. 4 the ovary, which in the more anterior and broader part of the visceral mass is on the left side of the stomach, extends into the narrow commencement of the terminal whorl of the spire, and here its posterior end is produced from left to right imto a fairly spacious thin-walled sac which hes between the upper and lower of the two posterior lobes of the liver extend- ing into the spire. The walls of this sac are not lined by a germinal epithelium but its cavity often contains a ripe ovum. INCISURA (SCISSURELLA) LYTTELTONENSIS. NG It is the commencement of the oviduct. Rapidly narrowing in diameter it passes forward to the right of the posterior end of the stomach and the liver lobes originating from it, and maintaining a position close below the external body-wall, it passes as a very much flattened and very thin-walled duct along the right side of the visceral mass, gradually mounting from a more ventral to a more dorsal position till it arrives above the right-hand loop of the intestine. All this while it has laid close to the outer body-wall, and it is extremely difficult to follow its course, owing to its being flattened between the liver and the external integuments. It turns in- ward just above and in front of the right visceral ganglion and runs in the roof of the posterior end of the mantle-cavity towards the right kidney (fig. 11). Here its walls become thicker and are lined by a distinct cubical epithelium. The duct does not at once enter the kidney but runs along its outer wall and opens into it in close proximity to the renal orifice. As stated above the gonaduct opens into the kidney coincidently with a reno-pericardial canal, the relations being very similar to those in Trochus. The vas deferens takes the same course as the oviduct. The alimentary tract.—The buccal bulb is relatively of enormous size. There are two large odontophoral cartilages on either side whose shape, as seen in section, is very similar to that of the cartilages of Fissurella as figured by Boutan. As shown in figs. 6, 7, and 8, the anterior and dorsal cartilages are the larger, and support the radula ; the posterior cartilages lie ventrad of the hinder ends of the anterior cartilages and have concave upper surfaces, with which the hinder ends of the latter articulate. A similar arrangement obtains in Trochus, and has been well described by Randles (15). The musculature of the buccal bulb is powerful, but I have not attempted to follow it out in detail. It is noticeable, how- ever, that the cross-striation, both of the intrinsic and ex- trinsic muscles of the odontophore, is very well marked. Though it is well known that these muscles are cross-striped in Gastropods, I am not aware that the character of the stria- VOL. 55, PART 1.—NEW SERIES. 2 18 GILBERT ©. BOURNE. tions has been carefully studied, and I take this opportunity of giving a drawing (fig. 15) of three fibres of the extrinsic muscles attaching the anterior end of the odontophore to the integuments of the snout. ‘These were specially well-stained, and it is obvious that the ends of the fibres nearest the snout are in a state of contraction, while their odontophoral ends— the lower ends in the figure—are relaxed. The fibres are not round but elongate oval in cross-section. That on the right has been cut through its long axis; in the two fibres on the left the section passes through the shorter axis, near the edge of the fibres. It can be seen that each fibre is a single metamorphosed cell, with a single nucleus situated near its broader end. The central portion of the cell, in which lies the nucleus, is composed of but little-altered cytoplasm, exhibiting an alveolar or reticular structure, differing from the normal only in the fact that the meshes of the reticulum are very regularly disposed in rectangular fashion. This cytoplasmic core of the fibre is invested by a sheath of con- tractile substance, which is thickest at the two ends of the long axis of the oval, and therefore appears as two bands in the right-hand fibre in the figure, while in the two left-hand fibres only the contractile substance is cut through. The whole is invested by a delicate sarcolemma. ‘The most interesting thing about these fibres is that the reticular arrangement of the cytoplasmic core corresponds exactly with the striations of the contractile substance in the upper part of the fibre on the right side of the figure, and in the left-hand fibre the cross-striations are very obvious and close together in the uppermost contracted part of the fibre, but lower down as the fibre becomes more relaxed, the dark transverse lines become progressively broader and fainter, and each may be seen to be made up of a number of dark longitudinal strie, which may well be interpreted as nodal thickenings of a reticulum. It is, of course, possible that the difference between the two ends of the fibres is due, not to a difference in the state of contraction, but to a greater specialisation of the broader end. Whichever interpretation is correct, the INCISURA (SCISSURELLA) LYTTELTONENSIS. 19 appearances lend support to the reticular theory of the con- stitution of striped muscle-fibre, and are inconsistent with the opposing theory of sarcomeres. The mandibles occupy the usual position at the sides of the mouth, and are composed of a number of plates or “tesserae” as described by Vayssiére for Scissurella costata. Randles has shown that each tessera is the product of a single epithe- lial cell in Trochus, and the same is evidently the case in Incisura. The radular sac occupies the usnal position. Lying at first between the upper horns of the odontophoral cartilages it maintains a median position to the posterior end of the buccal bulb, and then curves to the right between the right cesophageal pouch and the pedal ganglia and soon ter- minates in a swollen bilobed extremity lying on the right side of the hemocele. The radular teeth are represented in fig. 20. The centrals are squarish, with an expanded basal plate ; their anterior margins decurved, and furnished with five very distinct and sharp-pointed denticulations. The next three teeth (medio-laterals) are oblong, with decurved denticulate margins ; they decrease somewhat in size from within out- wards. The next tooth is much smaller, has a somewhat sigmoid curvature, a thickened base, a narrow neck, and a single recurved marginal denticulation. The next tooth is very large, shaped somewhat like a rake with a crooked handle, its expanded margin decurved and bearing about a dozen denticulations. Then follow the marginals or uncini, which are numerous, curved, slender, with expanded and re- curved denticulate margins. The radular formula may be written : Oi1(4+1+44) 10 Vayssiére has given a good figure of the radula of Scis- surella costata, which is similar to but differs in small details from that of Incisura. The radula of the Scissurellidee is usually described as resembling that of 'Trochus, but it is much more nearly like that of the Fissurellide. A reference to Thiele’s figures in the concluding chapter of Tréschel’s ‘Gebiss der Schnecken’ shows that the radula of Incisura 20 GILBERT C. BOURNE. very closely resembles that of Subemarginula picta, the shape and relative size of the outer medio-lateral tooth bemg almost identical, as also the characters of the centrals and uncini. ‘The large and specialised outer lateral tooth, though it differs widely in detail in different species, is characteristic of the Fissurellide. A close resemblance also exists between the radula of Incisura and that of Emarginula pileolus, and a less clearly marked resemblance can be seen in the radulee of various species of Fissurella. On the other hand, no comparison with the radula of Pleurotomaria is possible. A general view of the alimentary tract, as determined by reconstruction from sections, is given in fig. 3, which so far explains itself that little description is necessary. The ceso- phagus is enormously dilated in the anterior part of its course, forming in addition to the wide lateral diverticula or cesopha- geal pouches (figs. 9 and 10, w. p.) a spacious ventral pocket or “‘jabot.” These are all lined by a soft-looking glandular epithelium. Behind the level of the pedal ganglia the posterior section of the cesophagus leaves the jabot as a narrow tube with thick, longitudinally ridged walls formed by a long ciliated columnar epithelium. It runs back below the stomach and opens into the latter near its posterior end. Near the cesophageal opening numerous liver czeca open into the posterior end of the stomach. There is no spiral cecum connected with the entry of the liver-ducts as in Pleuroto- maria, Haliotis, and Trochus, but there is a deep ciliated ventral groove, the lips of which are bordered by specially long ciliated columnar cells, extending along the floor of the stomach from the cesophageal opening to the pylorus. A precisely similar groove occurs in the stomach of Fissurella, and has been well described and figured by Boutan (2). Randles has shown that in Trochus a cxcal groove, bounded by two conspicuous folds, extends into the spiral caecum from the cesophageal opening, and that the larger of the two bile- ducts opens into this groove. Though the spiral cecum is absent there can be little doubt that the ventral groove of the Fissurellidz and Incisura corresponds in function to the cecal INCISURA (SCISSURELLA) LYTTELTONENSIS. 21 groove of the Trochide, and it has the same relation to the liver-ducts. It should be noted in this place that Incisura, in the possession of numerous biliary apertures, resembles Fissurella and differs from Trochus, which has two, and Pleurotomaria, which has only one bile-duct. The intestine leaves the stomach on the ventral side of the anterior third of the stomach in Incisura. Beyond it the stomach narrows rather abruptly, and is continued forward as a small cecal diverticulum, the front end of which is inserted in the loop formed by the left-hand bend of the rectum. The walls of the blind end of this diverticulum are covered internally by a thick chitinoid layer, and thrown into complicated folds and ridges, but the cecum is not spirally coiled, and situated as it is at the end of the stomach furthest from the bile-ducts, it cannot be homologised with the spiral caecum of Pleuroto- maria, Haliotis, or Trochus. It must, however, be the cecum referred to by Pelseneer (12). The walls of the intestinal end of the stomach of Incisura have the columnar cells with striated borders and thick cuticle so fully described by Randles for Trochus. The intestine is provided throughout its length with a single longitudinal ridge or typhlosole. On leaving the stomach it makes a sharp bend from left to right, passes vertically upwards to above the level of the stomach, thence turns sharply to the right, describes a wide loop on the right hand, as shown in fig. 3, and bending sharply again to the left, passes nearly straight across the body till it reaches the left-hand corner of the pericardium, when it turns upward and to the right in the mantle roof, and becoming rectum, traverses the pericardium in its diagonal passage across the roof of the mantle-cavity to end in the anus opposite the mantle-slit. The liver ceca, as may be seen in figs. 3 and 11, are few in number, of relatively large size, with large lumina bordered by large secretory cells. As far as I could determine they do not branch, but have somewhat convoluted courses, and open independently into the cesophageal end of the stomach. 22 / GILBERT C. BOURNE. A few details may be added relative to the structures connected with the buccal cavity and cesophagus. In the mid-dorsal line the roof of the buccal cavity is deeply folded to form a median ridge containing a narrow lumen T-shaped in transverse section. This lumen of course communicates freely below with the buccal cavity. This median fold or ridge is deepest anteriorly over the mouth, and extends backwards for about two thirds of the length of the buccal bulb, gradually shallowing posteriorly till it dies out altogether. The walls of this ridge are composed of simple columnar cells, the internal ends of which have a striated border, and bear short cilia. On either side of the anterior part of this mid-dorsal ciliated groove is a somewhat shallower but still conspicuous groove appearing on the dorsal surface as a pair of folds lying close and parallel to the median ridge. These may be called the salivary grooves, for the small, simple, tubular buccal or anterior salivary glands open into them near their anterior extremities (fig. 21, sg. and s. d.). These anterior salivary glands are simple short ceca lined by an epithelium, consisting mainly of large finely granular secretory cells with a few columnar supporting cells between them. ‘lhe salivary grooves die out posteriorly at the point where the cesophagus leaves the buccal cavity, and at this level a second or posterior pair of salivary glands opens into the roof of the buccal cavity, just to the outside of the salivary grooves. These posterior salivary glands are very small tubalar structures with minute lateral diverticula. They correspond in position to, but are much smaller than, and not so much branched as the second pair of salivary glands in Fissurella. Otherwise the structures just described are identical in the two genera. As soon as the cesophagus is separated from the buccal cavity its right and left walls are produced into the broad and flattened cesophageal pouches, but from the first the right-hand pouch is considerably larger than the left. The T-shaped lumen of the dorsal ciliated groove may be traced for some way along the roof of the esophagus, but presently it dies out, and is replaced by a INCISURA (SCISSURELLA) LYTTELTONENSIS. 23 band of ciliated cells which diverges towards the left, and eventually passes completely over to the left side and passes into the narrow posterior part of the cesophagus. Ventrally, to the right side of the narrow cesophageal tube, the floor of the spacious anterior cesophageal cavity is produced into a capacious pouch or “jabot,” which runs back for some distance alongside of the narrow cesophageal tube (fig. 11, 7), and eventually ends blindly. The deviation of the cesophagus to the left and the preponderant size of the right cesophageal pouch have been noted by Boutan in Fissurella, and it is indeed a common feature in the Rhipidoglossa, indicative, as Amandrut has pointed out, of the larval torsion which brings about the asymmetry of the adult Gastropod. The nervous system.—Fig.5 is adiagram of the prin- cipal ganglia and nerve-trunks, as reconstructed from sections. Pelseneer’s description of this system in Scissurella costata and Incisura lytteltonesis is as follows: “ Dans les deux espéces, les cordons pédieux sont dans la masse musculaire du pied, et s’étendent jusqu’a la partie postérieure. A leur extrémité tout a fait antérieure se trouvent des ganglions pleuraux bien distincts. La commissure viscerale nait de ces derniers ; elle est croiseé et porte un ganglion supra-intes- tinal presque accollé au ganglion branchial ou osphradial gauche, comme dans ‘Trochus. ‘'‘l'out ce systéme nerveux ressemble done beaucoup plus a celui de Trochus qu’aux parties correspondantes connues de Pleurotomaria, telles que les ont décrites Bouvier et Fischer.” Since this was written we have had the more complete account of the anatomy of Pleurotomaria by M. F. Woodward, and the difference between the nervous system of this genus and that of the Scissurellidee is even more apparent than before. As may be seen from the diagram, the nervous system of Incisura is at once typically Rhipidoglossate and specialised. As the nervous systems of various Rhipidoglossa have been described in great detail by sundry authors, it will only be necessary here to mention the more important and pecuhar features. 24, GILBERT C. BOURNE. The cerebral commissure is long and situated far forward in front of the anterior pair of salivary glands. It is a true nerve, not ensheathed by a layer of ganglion cells, differing in this from Pleurotomaria. The cerebral ganglia are of large size, sub-triangular in transverse section, and produced into prominent lobes at the origins of the more important nerves. The tentacular and optic nerves have separate origins from the cerebral ganglia, Incisura agreeing in this point with Trochus and Fissurella but differing from Pleurotomaria. The labial lobe is very large, and forms a long, conical, taper- ing, antero-ventral process of the cerebral ganglion, which curves inward below the odontophore on either side, maintain- ing its thickness for about two thirds of its course towards the middle line. Then it tapers abruptly to form a thin labial nerve, which passes between the muscles of the lower lip, and as far as I can determine is connected by an extremely fine prolongation with its fellow of the opposite side, thus com- pleting the labial commissure. The buccal commissure is given off from the labial lobe about half way between the cerebral ganglion and the mid-ventral line. It passes inwards among the muscles of the odontophore and at once turns abruptly upwards to run between the extrinsic and intrinsic muscles to the top of the buccal bulb. Here it enlarges to form a ganglion of considerable size, lying close to the inside of the cerebral ganglion, and from this a stout nerve—a true nerve without a sheath of ganglion cells—passes inwards and backwards over the top of the odontophore and enlarges below the origin of the cesophagus into a small ganglion, which is connected by a very short commissure with its closely adjacent fellow of the opposite side. Bouvier (8) has figured and described two swellings at the ends of each of the elongated buccal ganglia of Turbo setosus, but I infer from his descrip- tion that they are not separate ganglia, but merely swellings at the ends of a long and ill-defined ganglion. I find precisely the same arrangement in Fissurella greca, but Boutan figures four clearly defined ganglia in F. reticulata. The sub-division of this elongated ganglion into two distinct ——————— ee INCISURA (SCISSURELLA) LYTTELTONENSIS. 25 ganglia isan indication of specialisation and a peculiar feature in Incisura. For the rest the characters of the cerebral ganglia, the size of their labial lobes, and the relations of the buccal ganglia are very similar in Turbo, Fissurella, and Incisura. The cerebro-pleural connective, as is commonly the case, is larger than the cerebro-pedal; both are true nerves, devoid of any sheath or local accumulations of ganglion cells. The pleural ganglia are distinct and that of the right side is relatively large, but both are fused to the dorsal surfaces of the pedal ganglia. The visceral commissure is typically streptoneurous, and for the same reason that the osphradial ganglia are situated far forward, the whole commissure is con- tracted antero-posteriorly as in Patella; on the other hand, it is considerably extended right and left. The sub-intestinal ganglion is distinct, but elongated and rather ill-defined; as Pelseneer remarks it is connected by a very short nerve with the large left osphradial ganglion. The left symmetrical pallial nerve passes straight out from the left pleural ganglion almost immediately below the supra-intestinal ganglion, and traverses the posterior fibres of the left columellar muscle, turning nearly verticaily downwards to enter the thickened border of the mantle. Before turning downwards it gives off a very fine branch, which makes connection with the short nerve uniting the supra-intestinal with the osphradial ganglion, thus establishing a left-hand dialyneury very similar to that of Trochus. The subintestinal nerve is very stout, and crosses over the dorsal surface of the hinder part of the pedal ganglia almost at right angles to the long axis of the body. ‘The sub-intes- tinal ganglion is fairly large and distinctly indicated by an accumulation of nerve-ganglion cells. It is triangular in shape, and from its right-hand lower corner the visceral nerve, and from its right-hand upper corner the osphradial nerve is given off. The latter is a very slender nerve, which passes into the substance of the columellar muscle, and turns verti- cally downward and then forward along the dependent edge of the mantle, running in this part of its course at the base 26 GILBERT ©. BOURNE. of the gill-filaments, which, as has been explained above, run back along this region of the mantle. At the anterior edge of the columellar muscle the nerve expands to form the large right osphradial ganglion. ‘The right symmetrical pallial nerve takes its origin from the ventral side of the right pleural ganglion, just where the latter becomes fused to the pedal ganglion. It runs outward, traverses the columellar muscle some way in front of the osphradial nerve, and takes a direct course to the right osphradial ganglion, which it crosses dorsally, and in so doing enlarges and makes an intimate connection with it. Just in front of the osphradial ganglion the pallial nerve divides into two branches. The posterior branch, which is slender, runs back along the thickened border of the posterior part of the mantle. The anterior branch runs forward to the mantle-slit, where it expands to form a small ganglion, indicated by a distinct accumulation of nerve-gauglion cells, and is here joined by a slender nerve from the anterior end of the osphradial ganglion. This little ganglion at the hinder border of the mantle- slit gives off an external branch supplying the posterior sense- papillz of the mantle-slit, and a stout anterior branch which passes round the mantle-slit and is continued forward as the peripheral pallial nerve, meeting and uniting with its fellow of the opposite side on the anterior border of the mantle. There is thus a very intimate dialyneury on the right side. These relations are very hard to make out, and require careful study with high powers of the microscope, but I can vouch for the correctness of the account here given of them. ‘The rela- tions in Fissurella are somewhat similar, but the proportions of the lengths of the nerves differ greatly, and apparently differ in different species, for in my sections of F. greca the sub- intestinal is close to the right osphradial ganglion, whereas in F. reticulata Boutan figures them as far apart and con- nected by a long slender nerve, as in Incisura. The origin of the right symmetrical pallial nerve from the upper surface of the pedal ganglion rather than from the right pleural ganglion is identical in Incisura and Fissurella. INCISURA (SCISSURELLA) LYTTELTONENSIS. 27 The visceral loop bears three distinct accumulations of ganglion cells, forming as many ganglia. The right ganglion lies close below the gonaduct and gives off a slender nerve to that organ. ‘The pedal ganglia, as may be seen in fig. 5, are very much concentrated. Anteriorly they are rather flat, but in about the middle of their length they increase considerably in thickness, this increase being due to the addition of a considerabie ventral thickening to each ganglion. In this region, in fact, each pedal ganglion consists of a dorsal and a ventral moiety, as is the case in all Rhipidoglossa (fig. 9). Here also the whole of the pedal ganglia lies in the hemoccele, as is the case with the more elongated pedal cords of Fissurella. But in Incisura the dorsal moieties of the pedal ganglia have very little posterior extension. The ventral moieties, on the other hand, extend back behind the dorsal moieties, and, narrowing in diameter, plunge into the muscular substance of the foot (fig. 10). ‘There they are continued backwards for a short distance, giving off nerves from their outer edges, and diminishing rapidly in diameter, partly because of fibrils given off to the different nerves, but also largely because of the thinning out and eventual disappearance of their coating of nerve ganglion cells. Posteriorly the cords become simple nerves, and end some distance in front of the posterior end of the foot. Pelseneer states of Scissurella costata and Incisura lytteltonensis: “Dans les deux espéces, les cordons pedieux sont dans la masse musculaire du pied, et s’étendent jusqu’a la partie posterieure.’ This is certainly not the case in Incisura; the left pedal cord, or rather nerve, dies out at a distance of 125 w from the posterior end of the foot in two specimens in which I calculated its extent, and remembering that the animal is only 1 mm. long this is a considerable distance. In short, one can hardly speak of pedal cords. The pedal centres, particularly the dorsal portions of them, have become concentrated into two clearly defined pedal ganglia, and it is only the ventral portions that are continued backwards to represent in some measure the elongated pedal centres of other Rhipidoglossa. In addition 28 GILBERT C. BOURNE. to the thick anterior commissure connecting the dorsal portions of the ganglia, there is a single anterior thin com- missure connecting the ventral portions, but this is the only trace of the usually numerous cross commissures of other lowly organised Gastropoda. Such a concentration of the pedal centres is very unusual if not unique among Aspido- branchia, and indicates that Incisura, and, if one may judge from the similar relations indicated in Pelseneer’s figures of S. costata, the Scissurellide in general are highly specialised. Much has been written about the significance of the dorsal and ventral moieties of the pedal cords of archaic Gastropods. The French authors hold that the upper moiety is pleural, or, as they say, pallial, the lower moiety pedal in character. Pelseneer and most English and German authors hold that both moieties represent pedal centres. The facts in Incisura seem to uphold the latter view. I have no wish to re-enter upon a controversy which has become almost wearisome by repetition, but may state that in Incisura the cerebro-pedal connectives certainly join the dorsal moieties of the ganglia; that the epipodial nerves are certainly given off from the dorsal moieties, and that whereas the left symmetrical pallial nerve is undoubtedly given off from the left pleural ganglion, the right symmetrical pallial nerve certainly appears to be given off from the dorsal moiety of the right pedal ganglion and not from the right pleural, both in Incisura and Fissurella. Advocates of the French view will take this last fact as evidence in support of their theory. The nervous system of Incisura certainly bears no resemblance to that of Pleuro- tomaria. On the whole it most nearly resembles that of the Fissurellide, in which family the pedal cords, though still elongate and ganglionic, and provided with several cross- commissures, have undergone a considerable reduction in length as with those of other Rhipidoglossa. The sense organs.—The eyes, as already stated, are closed and provided with a distinct lens. Their structure resembles that of the eyes of the Fissurellidz, and differs from the eyes of the Pleurotomariide and Trochidz, which are open. ee ee, Eo INCISURA (SCISSURELLA) LYTTELTONENSIS. 29 The otocysts occupy the usual position on the dorsal surfaces of the pedal ganglia and present no unusual features (fig. 9). The osphradia are strips of modified epithelium running for some little distance along the lower side of the gill-axes in front of the osphradial gangha and just ventral to the osphra- dial or branchial nerve (fig. 16). They are very similar in structure and position to the osphradia of Fissurella greca. Sense-papillee occur not only on the cephalic tentacles but also on the epipodial tentacles, all round the margins of the mantle and on the cirrhi bordering the manile-slit. Those on the cephalic tentacles are by far the largest, those on the margins of the mantle are very minute, but all have essenti- ally the same structure. Fig. 28 represents a longitudinal section through three of the papillee of the cephalic tentacles. Each papilla is a conical projection of the integument of the tentacle and is composed of a number of elongated cells of two kinds, closely packed together like the cells in a taste- bud from the human tongue. ‘I'he larger cells with larger, pale nuclei are evidently supporting cells, their characters being similar to the adjoining epithelial cells. The more slender, finely granular cells with smaller, deeply staining nuclei are the sense-cells, and each ends in a short stiff cilium projecting from a small cup-shaped depression at the end of the cone. According to Vayssiére these cilia are in constant movement in the living animal. The tentacles of Fissurella are clothed with a vast number of minute papille giving a velvety texture to the surface. These papille, though not so highly specialised, have each a single apical sense-bulb, the structure of which is similar to that of the sense-papille of Incisura. Finally, mention may be made of the pedal glands. The anterior pedal gland consists of a mass of unicellular glands lying in the hemoccele below the buccal bulb (fig. 7, p. gl.). It extends back nearly as far as the pedal ganglia. Ante- riorly these glands become more deeply seated and pass into the muscular mass of the foot, where they debouch into a median ciliated duct (fig. 6) which runs forward and opens 30 GILBERT C. BOURNE. on the anterior face of the foot in the groove between it and the lower surface of the snout. The posterior pedal glands are a mass of unicellular glands lying above the epithelial cells of the sole of the whole posterior surface of the foot. Kach unicellular gland has its own duct, which runs between the epithelial cells to open on the surface. The genera Scissurella, Schizotrochus, Incisura and Schis- mope, which have been grouped as a separate family Scissu- rellidee by some few authors, are generally placed in the family Pleurotomariide because they are zygobranchiate Rhipido- gloss, with a labral incision of variable length and position in the shell. ‘There is no frontal veil between the cephalic tentacles, an epipodial ridge is present, and there is a corneous multispiral operculum. Fischer (5) writes: “ Quel- ques auteurs distinguent deux familles, Scissurellide et Pleurotomariide, mais les differences qui existent entre ces deux types n’ont pas plus d’importance que celles qu’on con- state entre les divers groupes de Trochidee. Je les considere comme des sous-familles.”” Pelseneer (18), who had studied their anatomy, retains these forms in the family Pleuroto- mariide. Yet it is obvious, from what precedes, that the Scissurellidee cannot possibly be retained in this position. The differences in the radula alone are sufficient to distinguish the two types. But in addition to this the Scissurellide differ from the Pleurotomariidz in a number of characters, which may be summarised as follows: (1) The Scissurellidz have two columellar muscles ; Pleuro- tomaria has only one. (2) The eyes of Scissurellidee are closed ; thoso of Pleuro- tomaria are open. (3) The subocular tentacles of the Scissurellide are absent in Pleurotomaria. (4) The epipodium of Pleurotomaria is destitute of tentacles, cirrhi, or lappets. (5) he wide distance apart of the ctenidia, the large size of the pericardial cavity, the forward position of the ventricle of the heart, and the more distinct shifting of the organs of ———— INCISURA (SCISSURELLA) LYTTELTONENSIS. 31 the pallial complex into a median position in the roof of the mantle-cavity are all points in which the Scissurellide differ from Pleurotomaria. (6) In Pleurotomaria the right kidney has a distinct duct, with thickened glandular walls in the female; in the Scis- surellidze there is no such duct. (7) There is no spiral caecum to the stomach in the Scis- surellide, and the form of the stomach differs largely from that of Pleurotomaria. (8) The hepatic orifices are numerous in Scissurellide, whereas there is only a single orifice in Pleurotomaria. (9) The nervous system of the Scissurellide differs in detail in almost every point from that of Pleurotomaria, particularly in the concentration of the cerebral ganglia; the extreme fine- ness of the labial commissure; the presence of distinct pleural ganglia; the well-developed symmetrical pallial nerves estab- lishing a right and left dialyneury; the presence of distinct supra- and sub-intestinal ganglia; the shortness of the visceral loop; the concentration and abbreviation of the pedal centres. Not only are the Scissurellide distinct from the Pleuro- tomaride, but they are clearly less closely related to them than the Halitide or even than the Trochide and Turbonide, for the last-named families, though they have lost the labral incision in the shell, as also the right ctenidium and the structures correlated to it, have retained many anatomical features which find their counterpart in Pleurotomaria. Where, then, shall we find the nearest relatives of the Scissurellide ? Though Mr. Hedley was clearly in error in removing Incisura from the Scissurellide, I think he came very near the truth in suggesting the affinity of this genus with the Fissurellidee. His comparison of the adult Incisura with the post-larval stage of Fissurella is a just one. Almost all the differential external features which serve to distinguish the adults disappear on comparison of the adult of the one type with the post-larval stage of the other. In the young Fissurella we see a coiled shell with spiral sculpture, a labral incision of considerable length to the right of the middle line. 32 GILBERT C. BOURNE. There is a pair of ciliated post-ocular tentacles on either side of the head (I find vestiges of these structures in the adult of F. greca), a well-developed pair of ciliated epipodial ten- tacles in the vicinity of the opercular lobe, and a corneous multispiral operculum. Even the gills, if one may judge from Boutan’s figure (PI. xlii, fig. 8), have a close resemblance to those of a Scissurellid. Ifthe animal were sexually mature one would not hesitate to place it among the Scissurellide. In the next or Rimuliform stage the epipodial tentacles are multi- plied ; Boutan figures six in addition to the sub-ocular tentacles in F. reticulata and two in F. gibba. The labral incision has been converted into a foramen by the approximation of its edges at the labrum, but a suture still connects the foramen with the margin of the shell. his condition is exactly paral- leled by the Scissurellid genus Schismope. Subsequent de- velopment leads to the assumption of Fissurellid characters. The visceral spire, and with it the spiral coils of the shell, become obsolete. The foramen in shell and mantle become situated at the summit of the Patelliform shell, the post-ocular and epipodial tentacles (which obviously belong to the same series) degenerate, the operculum is cast off, and the oper- cular lobe disappears. In short, the Fissurellid develops along lines which remove it further and further from the Scissurellid condition of the larva. But, as must be apparent from the preceding pages, there is a considerable number of anatomical features in which the adult Scissurellid more nearly resembles the adult Fissurellid than any other family of the Rhipidoglossa. These features may be shortly recapitulated, Incisura being taken as a type of Scissureilid structure. The jaws of Incisura in position and structure very closely resemble those of a Fissurella. The radula of Incisura lytteltonensis finds its nearest counterpart in the radula of Subemarginula picta, and in general is distinctly Fissurellid in character. In the alimentary tract the characters of the salivary glands and cesophageal pouches, the absence of a spiral caecum in the stomach, the presence of an ceso- EEE INCISURA (SCISSURELLA) LYTTELTONENSIS. 33 phageo-intestinal groove in the capacious stomach, the existence of numerous hepatic ducts, are all points in which Incisura agrees with Fissurella, and differs, to a greater or less degree, from the Pleurotomaride, Haliotide, Trochide, and Turbonide. The presence of a right and left columellar muscle in the Scissurellidz is evidently an antecedent stage of the horse-shoe shaped columellar muscle of the Fissurellide. The eyes, which are open in Pleurotomariide, Haliotide, and 'I'rochide, are closed in both the Scissurellide and the Fissurellide. The subocular and posterior epipodial tentacles of the Scissurellidz are paralleled by the similar larval organs in the Fissurellide, In both the Scissurellidz and Fissurellide the increased size of the last whorl of the shell and the diminution of the visceral spire has led to a broadening of the dorsal part of the body, in consequence of which the bases of the ctenidia are widely separated on the right and left sides of the body, the pericardium is transversely elongated, and the heart and kidneys are shifted towards the mid-dorsal line in the roof of the mantle-cavity. In these respects Incisura is inter- mediate between Fissurella and the other families of Rhipido- glossa enumerated above. The nervous system of Incisura, though much specialised, shows more resemblance to that of the Fissurellidze than to that of any other Rhipidoglossa, as has been explained in detail in the descriptive part of this paper. The corre- spondence in the labial commissure, the buccal ganglia, and the visceral commissure is very exact. The pedal centres of the Scissurellidee have undergone great concentration, but this is foreshadowed in the pedal cords of the Fissurellide, which are much shortened in comparison with the elongated scalariform pedal centres of such families as the Pleuroto- maruidee, Haliotide, and Trochide. There can be little doubt, then, as to the affinity of the Scissurellidz with the Fissurellide, but the exact relationship of the two families remains to be considered. In my opinion VOL. 55, PART 1,—NEW SERIES. 3 34 GILBERT C. BOURNE. it is not exact tosay, as Hedley has, that Incisura represents an arrested stage of development of a Fissurellid. It is a more reasonable inference from the facts that the two families have descended from a common stock, and have diverged in different directions. There are several arguments in favour of this inference. One which in my opinion has great weight is derived from the condition of the left kidney in the two families. In the Scissurellidz, as I have shown, the left kidney is relatively of large size, and is a true ‘papillary sac,”’ phagocytic in function like the left kidney of the Pleurotomariidz, Haliotide, and Trochide. In the Fissurellidz this organ is reduced to a mere rudiment, and may, I believe, disappear altogether in some species, for I have failed to find a trace of it in transverse and horizontal sections of I. greca. , Remy Perrier (14) has stated that the epithelium of the left kidney of Fissurella is identical with that of the right kidney, but there is some doubt about this, and a renewed investigation of the left kidney of several species of the Fissurellide is much to be desired. But there is no doubt that it is a vestigial organ, and that in this respect the Fissurellide have been specialised along a different line to the Scissurellide, which have retained the left kidney in a fully functional state. Per contra, while the Fissurellide retain to a large extent the primitive scalariform character of the pedal centres, the Scissurellide have in this respect sur- passed them in specialisation, for their pedal centres are concentrated to a degree elsewhere unknown among the Rhipidoglossa. The divergence of the two types is obvious, and one may conclude that both have been derived from a stock very nearly represented by the so-called Emarginuliform larva of Fissurella, which had a spirally coiled shell with a large umbilicus, spiral sculpture and a considerable labral incision. A corneous multi-spiral operculum and a well- developed epipodial ridge bearing sub-ocular as well as posterior epipodial tentacles were present. The left kidney was a well-developed papillary sac, and the pedal centres were INCISURA (SCISSURELLA) LYTTHLTONENSIS. 35 elongate and scalariform. Such an ancestral form would not be far removed from a Pleurotomaria, but would differ from it in the development of a double columellar muscle and in the tendency to acquire a secondary symmetry always correlated with the doubling of this muscle. The Scissurellidee have retained most of the features of this parent form, but have undergone considerable specialisation in the nervous system. The Fissurellid branch must early have acquired a “ sessile ” habit, and have been much modified in connection with it, but its members have largely retained the primitive condition of the pedal centres. The Scissurellide, though for the most part constant to the primitive type, are also under- going modification in the same direction as the Fissurellide. In Incisura the visceral spire is reduced, the shell is becom- ing thick and solid, the spiral sculpture is absent, the margins of the aperture are in one plane, the foot is becoming short and broad, and its whole organisation is indicative of a semi- sessile habit. Further specialisation along these lines would give it Fissurelliform or rather Emarginuliform characters. It is interesting to note that another member of the family, Schismope, while retaining its spiral coil and widely open umbilicus, has undergone specialisation in another direction, for the labral shit has been converted into a foramen by the approximation of its edges, so that although distant from the margin it is connected with it by a suture. In this respect it closely resembles Semperia, a sub-genus of Emarginula. Semperia leads on to Rimula, and as we have seen there are Emarginuliform and Rimuliform stages in the development of Fissurella. This is an undoubted example of the develop- mental stages of one form resembling the adult stages of other forms, a phenomenon the occurrence of which some persons are inclined to deny nowadays, though the evidence in favour of it is very large. The parallel stages of evolution among the Scissurellidz and Fissurellide afford interesting examples of the pheno- menon of convergence, and illustrate a principle which, I think, has not been sufficiently attended to in drawing inferences as to 36 GILBERT C. BOURNE, the affinities of animals from morphological evidence, namely, that a similar environment and similar habits of life reacting on a similar organisation may often produce very similar struc- tural results. Not, however, identical, for however similar the results may appear at first sight in all cases of convergence a close analysis will always disclose differences which exclude the idea of direct descent of the animals in question, This instance is particularly instructive; the Haliotide, Scissu- rellidee and Fissurellidz have all inherited the same structure from a presumably Pleurotomariid ancestor, viz. the slit in the mantle and the corresponding labral incision in the shell. It has been variously modified, and similar modifications are displayed independently by different groups, the similarity of the evolutionary series being, as far as one can judge, correlated with the adoption of similar habits, ADDENDUM. It is long since I first read the short but profound essay of Sir Ray Lankester (9) “On the Use of the term Homology in Modern Zoology,and the Distinction between Homogenetic and Homoplastic Agreements.” On referring again to this essay, I find that the conclusions arrived at in the foregoing paragraph, as also similar conclusions arrived at after a detailed study of various members of the Neritide (1), are unconsciously expressed in nearly the same words that he used forty years ago. I have to beg Sir Ray Lankester’s pardon for not making specific reference to his essay in my former paper. But I find a certain satisfaction in not having had the form of his argument clearly in my mind while I was working to the same conclusion from evidence gathered from the study of the probable lines of descent of animals belong- ing to a different class to that which he used to illustrate his original thesis. Had I consciously set out to prove, or even to disprove, his contention, I could not have avoided a certain amount of bias. To have arrived unconsciously—or sub- consciously, for the idea of homoplasy inculcated by him was INCISURA (SCISSURELLA) LYTTELTONENSIS. 37 always present to my mind—at an identical conclusion is to give unequivocal support to the validity of the arguments by which it was sustained. In the essay in question Lankester showed that the term homology, which really belonged to the platonic school of the natural philosophers of the end of the eighteenth and the beginning of the nineteenth century, acquired a new connotation after the publication of the ‘Origin of Species.’ But this new connotation was indefinite. On the one hand structures were said to be homologous which “are genetically related, in so far as they have a single representative in a common ancestor.” For this kind of homology Lankester proposed to substitute the term ‘ homo- geny.” On the other hand, various organs were described as homologous which could not possibly be included under the idea of homogeny, because, over and above general resem- blances such as might be referred to inheritance from a common ancestor, they exhibited a number of detailed resemblances such as could not possibly be supposed to have been represented, in like detail, in a generalised ancestral form. Therefore, Lankester pointed out, there must be a second quantity covered by the term homology, and he described it in the following words: ‘‘ When identical or nearly similar forces or environments. act on two or more parts of an organism which are exactly or nearly alike, the resulting modifications of the various parts will be exactly or nearly alike. Further, if, instead of similar parts in the same organism, we suppose the same forces to act on parts in two organisms, which parts are exactly or nearly alike and some- times homogenetic, the resulting correspondences called forth in the several parts of the two organisms will be nearly or exactly alike. . . . I propose to call this kind of agree- ment homoplasis or homoplasy. . . . What exactly is to be ascribed to homogeny and what to homoplasy in the relations of a series of structures is a matter for careful con- sideration.” Somewhat further on in the essay homoplasy is defined as “‘ depending on a common action of evoking causes or moulding environment on homogenous (= homogenetic) 38 GILBERT C. BOURNE. parts, or on parts which for other reasons offer a likeness of material to begin with.” The term “ homoplasy ” has passed into current use, and the principle expressed by it has been freely used to explain numerous large and general resemblances which have obviously been evolved independently, such as the general resemblances between different kinds of patelliform gastropod shells, e.g. between Patella, Fissurella, Septaria, Capulus, and Siphonaria, or the general resemblances of external mor- phology of fishes and cetacea. But the term homogeny has not been so generally accepted, and many, if not most, zoologists have preferred to retain the old word homology, and in so doing it is clear that many of them have failed to distinguish between the two quantities contained within the single term, of which the differences were so clearly pointed out in Lankester’s essay. For it must be evident to anybody who is well acquainted with the morphological literature of the last thirty years that, so far from attempting to distinguish between homogenetic and homoplastic resemblances, a large number of authors have shown a vast amount of ingenuity in referring the most minute resemblances in the organs of animals, which are certainly not very closely related to one another, to homology. ‘The most extreme instances of this tendency to ascribe every resemblance, however detailed, to inheritance, ignoring the possibility that similar structural changes may be induced by the incidence of similar forces, are to be found in the works of those authors who attempt to derive the lower members of one phylum of the animal kingdom from highly differentiated members of another phylum. It is, of course, true that several of the most thoughtful and best informed among contemporary zoologists have been fully aware of the error lurking in the indiscriminate use of the term “ homology,” notably Gegenbauer and Fiirbringer in Germany ; Cope, W. B. Scott, E. B. Wilson, and Osborn in America. It is not my present intention to enter upon along discussion of this subject, which I hope to return to on a future INCISURA (SCISSURELLA) LYTTELTONENSIS. 39 occasion. But I take the opportunity of dealing with an interesting and suggestive essay by Osborn (10), in which Gegenbauer’s admirable analysis of the different forms of resemblances obtaining among animal structures is largely quoted. In the first place Osborn makes it evident that I, in common with others, have fallen into an error in using the term “ con- vergence” to denote the parallel stages of evolution among the Fissurellid and Scissurellidee. In the common meaning of the word, convergence might appropriately be used to signify that apparent approximation of structural character- istics which not infrequently leads to two forms being classified together in the absence of sufficiently complete information as to their internal anatomy. But it has acquired a special meaning, defined by Osborn as the “independent similar development of unrelated animals, bringing them apparently closer together.” As it has been the purpose of my paper to show that the families of Molluscs treated of are related, and closely related, the term convergence is not applicable to resemblances recurring in those families. But when I come to consider whether other resemblances between various mollusca should be described as due to “ parallelism” or “homoplasy”’ I find myself in a difficulty. Parallelism is defined as the ‘independent similar development of related animals, plants, or organs’’; homoplasy as the “‘ independent similar development of homologous organs or regions giving rise to new parts.’ It is added that homoplasy always involves homology, while parallelism and convergence may or may not involve homology. In Incisura the reduction of the visceral spine, the oblitera- tion of spiral sculpture, the levelling of the margins of the aperture, the alteration in the shape of the foot are changes parallel to those observed in the ontogeny of a Fissurellid, and they involve homogenetic organs; the parallelism in this case involves homology and should be called homoplasy. In Schismope the conversion of the labral slit into a foramen is a change parallel to that observed in the 40 GILBERT C. BOURNE. ontogeny of a Fissurellid and it involves a homogenetic character, therefore it also is due to homoplasy. On the same reasoning the resemblances in the shell, foot, and mantle of more distantly related forms, the Patellidee, Septaria, the Capulidz, and Siphonariidz are homoplastic. But should the pallial branchize of a Patella and the gill of a Siphonaria, be attributed to parallelism or homoplasy? They are certainly not genetically derived from the typical molluscan ctenidium, and to this extent are deficient in the element of homology which Osborn says should always be associated with homology. On the other hand they are vascular outgrowths of the mantle, which is assuredly a homogenetic structure in all the forms in question, and therefore there is an element, though a more remote element, of homology. Im this case it is simply a question of the importance attached to the degree of homo- logy whether these structures should be ascribed to parallel or homoplastic development. But Lankester’s term, homo- plasy, as originally defined, covers all the cases. It appears to me that, while there is a contrast between homoplasy and convergence, there is no such contrast between homoplasy and parallelism, and that for the sake of clarity the last term should be abandoned, homoplasy being retained in the sense originally defined by Lankester. It has the priority over Firbringer’s term homomorphy, which, as Osborn points out, has the same connotation ; and it has the advantage of indicating a resemblance due to the moulding influence of environment, whereas homomorphy only calls attention to similarity of form. In the latter half of his essay Osborn raises a most interest- ing question, which has presented itself with various degrees of insistence to workers in various groups of the animal kingdom. Drawing his evidence from paleontological as well as recent types, he points out that the accessory cusps in the molar teeth of Mammalia arise in the same order and with the same relations to the primary cusps in groups which can be proved to have diverged widely from one another before any complication of the tooth pattern arose. Here, then, are INCISURA (SCISSURELLA) LYTTELTONENSIS. Al examples of detailed resemblances which cannot be due to inheritance nor yet can they be due to external forces acting upon homogenetic parts, for the teeth are formed below the gum and the cusps are in place before any mechanical forces are brought to bear on them. ‘The characters of the teeth are clearly congenital, and the resemblances between the patterns which have arisen independently in different groups cannot be accounted for by the preservation of fortuitous variations by natural selection, for paleontological evidence shows that variation has in each case proceeded along one line and not along several lines, one of which has been selected. Calling to mind Lankester’s suggestion of the “common action of evoking causes . . . on parts which for other reasons (than homogeny) offer a likeness of material to begin with,” Osborn pleads for the recognition of a latent or potential homology, by which term I understand him to mean a tendency or capacity to produce a definite structure, which capacity must have been present in the ancestors of the existing orders of Mammalia, but has only manifested itself in such groups as possessed or were subject to the co-operating factors necessary for evoking the latent capacity, and thus producing the structure in question. The objections to a principle of this kind are that, in the first place, as Osborn himself admits, it leads us on the dangerous ground of teleological speculation; and, in the second place, that it might, if loosely applied, be used to explain anything or everything by a phrase. Nevertheless, I think that some such principle may be admitted, with due caution, in explanation of a large number of difficulties which present themselves, with increasing insistence, to every class of zoological workers. In a recent paper on the Neritide I alluded to the great difficulty of finding a satisfactory theory to account for the distribution of the fresh-water Neritids, described as species of the sub- genera Paranerita and Septaria, in remote oceanic islands. As their general anatomical and conchological characters 42 GILBERT (C. BOURNE. differ in a very small degree from those of the marine species of the genus Nerita, abounding in the seas in which the oceanic islands inhabited by the fresh-water Neritids are placed, it did not seem an unwarrantable assumption that in each locality the marine species had ascended from estuaries into rivers (just as prawns do in so many parts of the tropics), and had been similarly modified as a result of the fresh-water environment. But when I found that the accessory genera- tive organs of the fresh-water species from different localities were always alike, and differed in the same direction and to the same degree from the accessory generative organs of the marine species from the same localities, particularly in the fact that the female gonaducts of the freshwater species are always triaulic, whereas those of the marine species are diaulic, I was no longer able to sustain the opinion that I had first formed as to the possibility of the independent but similar modification of the marine species in different parts of the world. It seemed to me impossible that the triaulic condition should have been evolved several times over. The problem, however, is of the same kind as, though of less magnitude than, that presented by the cusps of mammalian molar teeth. If we can conceive the presence in the germ-plasm of Neritidz of factors competent to produce the triaulic condition of the genital ducts, but that the activity of these factors is only excited by the co-operating action of other factors—in this case by reduction of the salinity of the water—the detailed resemblances between structures existing in animals living so far apart but under similar conditions are susceptible of explanation. A few years ago such an explanation would have been inadmissible. But since Mendelian experiments have shown that definite changes affecting parts of the organism in a similar manner may require the co-operation of two or more factors, and cannot be produced unless those factors are brought together; and since such experiments as those of Stockard on Fundulus have shown that a relatively slight change in the salts dissolved in water may induce profound INCISURA (SCISSURELLA) LYTTHLTONENSIS. 43 changes in certain organs of developing embryos, it is no longer possible to reject such suppositions as fanciful and incredible. Those who have given unprejudiced consideration to the objections raised against the all-sufficiency of natural selec- tion, must have felt that a term is wanting somewhere in the current forms of argument used to explain resemblances between structures which are only doubtfully homogenetic. The missing term may possibly be found when we have a more exact knowledge of the kinds of factors whose co-opera- tion is necessary to produce specific structure. Some of these factors must be germinal, but evidence is accumulating that germinal factors are not simple but compound, and may be split into subordinate factors which, taken alone, do not pro- duce the specific result. There is further evidence that germinal factors react differently to different external factors, and if this be so many kinds of resemblances and differences may be accounted for by laws of interaction of which we are as yet only dimly aware. The evidence on these matters is insufficient to enable us to arrive at definite conclusions, but it is at any rate sufficient to earn respect for a suggestion supported by such a large number of positive facts as that of Osborn. I believe that in the future morphologists, in conjunction with systematists, will be largely occupied in attempting to discriminate between the different kinds of resemblances among animal structures, between similarities due to the “ common action of evoking action or moulding environment,” and similarities due to direct descent, and I venture to think that such morphological studies, carried out with scrupulous attention to detail, are not useless, but will give precision to, and perhaps modify our views on, the causation of modifica- tion of animal structure. Ae GILBERT C., BOURNE. List oF tHE PrincipAL PAPERS REFERRED TO IN THE TEXT. i 2. 3. 13. 14. 15. 16. iv 18. Bourne, G. C.—* Contributions to the Morphology of the Neri- tana,” ‘ Proc. Zool. Soc. Lond.,’ 1908. Boutan, L.—* Recherches sur la Fissurelle,” ‘ Arch. de Zool. Expér. et gén.,’ (2), iii bis, 1885. Bouvier, E. L.—Systéme nerveux, morphologie générale et classification des Gastéropodes Prosobranches,” ‘Ann. des Sci. Nat. (Zool.),’ (7), iii, 1887. and Fischer, H.—* Etude monographique des Pleurotomaires actuels,” ‘ Arch. de Zool. Expér. et gén.,’ (3), vi, 1898. . Fischer, H.— Manuel de Conchyliologie,’ Paris, 1887. - Fleure, H. J., and Gettings, M. M.—** Notes on Common Species of Trochus,” ‘ Quart. Journ. Mier. Sci.,’ 51, 1907. - Haller, B.—“Beitriige zur Kenntniss der Niere der Prosobranchier,” ‘Morph. Jahrb.,’ xi, 1886. . Hedley, C.—* Additions to the Marine Molluscan Fauna of New Zealand,” ‘ Records of the Australian Museum,’ v, 1904. . Lankester, E. R.—‘ On the Use of the term Homology in Modern Zoology, and the distinction between Homogenetic and Homo- plastic Agreements,” ‘ Ann. Mag. Nat. Hist.’ (4), vi, 1870, p. 34. . Osborn, H. F.—* Homoplasy as a Law of Latent or Potential Homology,” ‘ American Naturalist,’ xxxvi, 1902, p. 259. . Pelseneer, P.—* Les reins, les glandes génitales et leurs conduits dans les Mollusques,” ‘ Zool. Anzeiger,’ xix, 1896. “Recherches morphologiques et phylogénétiques sur les Mollusques» Archaiques,” ‘Mém. couronnés de Acad. Roy. de Belgique,’ lvii, 1898-99. “The Mollusca,” Lankester’s ‘Treatise on Zoology,’ part v, London, 1906. Perrier, R.—‘* Recherches sur l’anatomie et l’histologie du rein des Gastéropodes Prosobranches,” ‘Ann. des Sci. Nat. (Zool.),’ (7), viii, 1889. Randles, W. B.—* Observations on the Anatomy and Affinities of the Trochide,” ‘Quart. Journ. Mier. Sci.,’ 48, 1904. Smith, E. A—** A Description of some Shells from New Zealand,” ‘Proc. Malacol. Soe.,’ i, 1894. Thiele, J., in Tréschel’s ‘ Gebiss der Schnecken,’ vol. ii. Berlin, 1891. Vayssiére, A—‘ Etude zoologique de la Scissurella costata, var. levigata,” ‘ Journ. de Conchyliologie,’ (5), xxxiv, 1894, INCISURA (SCISSURELLA) LYTTELTONENSIS. 45 19. Woodward, M. F.—“The Anatomy of Pleurotomaria Beyrichii,” ‘Quart. Journ. Mier, Sci.,’ 44, 1901. EXPLANATION OF PLATES 1—5, Illustrating Mr. Gilbert C. Bourne’s paper “On the Anatomy and Systematic Position of Incisura (Scissurella) lytteltonensis.” LETTERING FOR ALL THE FIGURES. An, Anus. a.g.f. Anterior gill-filaments. aw. 1. Left auricle. au.7. Right auricle. B, Buccal ganglia. 0b. b. buccal bulb. 6. ¢. Buccal cavity, br. x. Branchial nerve. b. sk. Branchial skeleton. b.v. Blood- vessel. C. Cerebral ganglia. car. buccal cartilage. c. ce. cerebral com- missure. cl. Cilia. cdl. c. Ciliated canal of dorsal ridge of gill-filament. eil.l. Lateral ciliated cells. c.pd. Cerebro-pedal connective. c. pl. Cerebro- pleural connective. di.1. Left dialyneurous connection. d.g.7. Dorsal glandular ridge of gill-filament. e. Eye. ep. Epipodium. ep. n. Epipodial nerve. ep.t. Epipodial tentacle. F. Foot. f.c. Frontal cilia. g. Ganglion behind mantle-slit. g.f. Gill-filaments. g.n. Genital nerve. hy.g.l. Left hypobranchial gland. hy.g.r. Right hypobranchial gland. 7. Intestine. z.v. Vertical loop of intestine. j. Jabot. jw. Jaws. k.l. Left kidney. k.r. Right kidney. J. b. Left branchial ganglion, lb. 1. Labial lobe. l.c. Labial commissure. J. c¢.m. Left columellar muscle. J. ct. Left etenidium. Jl. Liver. Ui. d. Liver-ducts. 1. @. p. Left cesophageal pouch. /é. Lateral tooth of radula. m. Mouth. m. c. Mantle-cavity. md. t. Medio-lateral teeth of radula. m. f. Median dorsal fold of buccal cavity. m.s. Mantle-slit. mt. Mantle. mn. Nucleus. od. Oviduct. od.o. Opening of oviduct into right kidney. o@. Csophagus. os. Osphradium. of. Otocysts. ov. Ovary. P.Pedalganglia. pa. Pallial nerve. pc. Pericardium. p.g.f. Posterior gill-filaments. p. gl. Pedal gland. phg. Phagocytic cells of left kidney. pl. 1. Left pleural ganglion. pl. r. Right pleural ganglion. p.n. Pedal nerves. p.v. Pallial blood- vessels. Rk. Rectum. r.b. Right branchial ganglion. +. c. m. Right columellar muscle. +. ct. Right ctenidium. rd, Radula sac. 1. @. p. Right cesophageal pouch. rp. d. Reno-pericardial duct. sb. 7. Sub- intestinal ganglion. sg.! Anterior salivary glands. sg.2 Posterior salivary glands. sn. Snout. s. 0. ¢. Sub-ocular tentacle. sp.z. Supra- intestinal ganglion. sp. 1. Left symmetrical pallial nerve. sp. . Supra- intestinal nerve. sp. 7. Right symmetrical pallial nerve. st. Stomach. 46 GILBERT C. BOURNE. ‘sy. p. Sensory papille. TT. Cephalic tentacle. tn. Tentacular nerve. un. Uncini. V. Ventricle of heart. v. aff. Afferent branchial vessel. v. eff. Efferent branchial vessel. v. g. 1. Left visceral ganglion. v. g. 7. Right visceral ganglion. v.. Visceral nerve. [All the figures are of Incisura lytteltonensis.} Fig. 1.—A specimen viewed from the left side and below to show the size and shape of the foot, the operculum, the epipodial tentacles, ete. x about 40. Fig. 2.—Dorsal view of a female specimen which has been stained and mounted as a transparent object. x 80. Fig. 3.—A reconstruction of the alimentary tract ; semi-diagrammatie. x 80. Fig. 4.—A diagram showing the relations of the right and left kidneys, the heart, pericardium, ovary and oviduct. The extent of the pericardial cavity is indicated by a thick black line. x 80. Fig. 5.—The nervous system as determined by reconstructions from sections. x 80. Fig. 6.—¢. A transverse section through the posterior part of the head, including both eyes. x 135. Fig. 7.—¢. A transverse section taken just behind the mantle-slit, showing the bi-pectinate character of the right ctenidium. x 135. Fig.8.—¢. A transverse section through the anterior ends of the pedal ganglia. x 135. Fig. 9—Q. A transverse section through the hinder ends of the pedal ganglia. Note the position of the left kidney, k./., and its open- ing into the mantle-cavity ; the size and extent of the right and left cesophageal pouches, 7.@.p. and l.@.p.; the size and position of the right and left pleural ganglia, pl.r. and pl.l., and the supra-intestinal ganglion, sp.7.; the pedal ganglia, P., are clearly seen to be composed of a dorsal and a ventral moiety. x 135. Fig. 10.—?. A transverse section passing through the posterior end of the mantle-cavity, showing the large size of the left auricle, au.L.; the orifice of the left kidney, k.r.o.; the pedal nerves, p.n., which are the posterior continuations of the ventral moieties of the pedal ganglia shown in fig. 9. x 135. Fig. 11— 9. A transverse section passing through the posterior end of the foot. Note that the large pedal nerves shown in the previous figure do not extend into the hind part of the foot. x 135. Fig 12— 9. A transverse section taken near the terminal part of INCISURA (SCISSURELLA) LYTTELTONENSIS, A7 the visceral spire, showing the opening of the oviduct, od., into the hinder end of the ovary, ov. x 155. Fig. 13.—. A section through the left kidney showing the reno- pericardial duct, *p.d. Note the band of ciliated cells, ezl., on the floor of the mantle-cavity opposite the opening of the left kidney. x 225. Fig. 14.— 9. A portion of a section through the left kidney showing the rounded phagocytic cells, phy., which have taken up solid rod- shaped bodies from the subjacent blood-vessel, b.v. x 1000. Fig. 15.—Striped muscle-fibres attaching the anterior end of the buccal bulb to the integument. x 1000. Fig. 16.—A transverse section through the anterior filaments of the left ctenidium. Note the osphradium, os., lymg under the branchial nerve, b.n. xX 535. Fig. 17.—A transverse section somewhat posterior to that drawn in fig. 16, passing through the posterior filaments of the left ctenidium. In this and the previous figure note, d.g.7., the dorsal glandular ridges of the gill-filaments. x 535. Fig. 18.—A transverse section through two gill-filaments of the right ctenidium ; cél.c., the ciliated canal traversing the dorsal glandular ridges of the filaments. x 1000. Fig. 19.—The left ctenidium stained and viewed from above as a transparent object. x 225. Fig. 20.—A portion of the radula. x 800. Fig. 21—Part of a transverse section passing through the anterior end of the buccal bulb to show, m.f., the median dorsal fold of the buccal cavity and, s.g.', the anterior salivary glands and their ducts. x 535. Figs. 22-26.—A series of transverse sections through the right-hand posterior corner of the mantle-cavity showing the relations of the ovi- duct and the right reno-pericardial duct to the right kidney and the pericardium. x 225. (These figures are drawn as seen reversed under the microscope.) Fig. 27.—A cephalic tentacle showing the two multiple rows of sensory papille. x 225. Fig. 28.—A longitudinal section through three sensory papille of a cephalic tentacle. n', pale nucleus of a supporting cell; n*, deeply stained nucleus of a sense-cell ; c7l., cilia borne at the ends of the sense- cells. x 1000. Quart. fourm Meer Sev. Vat, 55, NS FEL eS Vy : 4 tae 2 a Oar OSE os: 4A EO Yo) bee yf TSO INOS / ROO aoe LY / ef} eed S st/ Cle FEBs 5. G C. Bourne del. ANATOMY OF INCISURA. 7 ‘ ——" ai = “VYNSIONI 40 AWOLVNVYV ‘TPP euinog 9 #9) ~ Ye A y Ut ») A (Acres cil iG 1 Z 2 = aa in J Nee \Y Oem oN Ne ¥ Me eS ) eer TL ony ns a b, : S ® TAZIND LEME BNL, "4 J we ate XN Ok HONS BE f 7 i \ ° Nae SS SS /, 34 \t xy, Ne st ~~ 2 St SEN ) SEF LE SES er No g SEs 2027 < NS, WEE RRSe Sx. G.C.Bourne del ANATOMY OF INCISURA. G.C.Bourne del ANATOMY OF INCISURA. ey THE EYE OF PECTEN. 49 The Eye of Pecten. By Ww. J. Dakin, M.Sce., Demonstrator and Assistant Lecturer in Zoology, University of Belfast. With Plates 6 and 7, and 2 Text-figs. Tue first reference to the eyes of Pecten that I have been able to find is that of Poli in 1795. Since that date more than a score of investigators have studied these small organs and treated in more or less greater detail the histology. Hach has made new discoveries, which have in very many cases been refuted by their immediate successors, to such an extent, in fact, that it was almost impossible to determine from the literature on the subject the truth in regard to cer- tain parts. One of the last and most reliable papers was that of Hesse, published in 1900 (84). He pointed out that some points were still unsolved (though adding one or two dis- coveries himself), and that the success of the methylene-blue method, if attained, would possibly elucidate all. In 1904 a paper appeared by Miss Hyde (89), embodying the results of a successful employment (according to the author) of the methylene-blue methods for nerve-endings in the retina, but these results were certainly not those expected by Hesse nor probably by other authors, for they stand in striking opposition to the views previously held. Whilst working at a memoir on Pecten in 1907, I came to the con- clusion that this, the latest investigation of the Pecten eye, differed greatly from the preceding ones, and that only one VOL. 55, PART 1.—NEW SERIES. A 50 W. J. DAKIN. more confusing series of results had been added to the already existing multiplicity. I determined therefore to make a complete study of the histology of the eye. The privilege of occupying the British Association Table at Naples enabled me to carry out this investigation on a species previously examined by most writers on the eye—Pecten jacobeus—and this was com- pleted by a considerable stay at the Port Erin Biological Station. The results have been the discovery of several new points, the confirmation and refutation of many discoveries of different workers, and I hope the complete elucidation of the structure of the retina. It has been due to the frequency of occurrence of artefacts and the difficult histological work required for such complicated organs that the structure of these eyes has remained so long a puzzle. By the use, however, of numerous methods it has been possible to eliminate to the greatest extent the artefacts, and incidentally the trial of so many fixatives, etc., has enabled me to obtain practically all the appearances seen and figured by the various investigators. The account of the structure will be given at some length, since a comparison of the various views is necessary, and, with the exception of Hickson’s and Patten’s papers very little has appeared in English. I am indebted to the British Association for permission to use their table at the Zoological Station of Naples, and also to the staff of that well-known institution. My thanks are also due to Professor Herdman and to the Curator of the Port Erin Biological Station for the trouble taken in supplying me with material and apparatus for carrying out detailed work at the latter place, and to Pro- fessor Drew, of Maine, for specimens of P. tenuicostatus. History. Only the history of the references to the Pecten eye before and including the fundamental paper of Hensen will be given in this section, since the other works will be discussed more THE EYE OF .PECTEN. at fully when describing the structures involved, and it will avoid repetition if they be omitted here. In 1795 Poli, in his large work on the Mollusca (1), gave figures illustrating the general anatomy of Pecten, in which the eyes are depicted, and also a view of the mantle-edge showing more clearly the tentacles and eyes, but no details of structure are given what- ever except the external pigmented ring bounding the cornea and the pigment stripe on the tentacles. He recognised a likeness to the human eye, and as usual. applied some of the names given to parts of the latter, a feature followed by his successors, who naturally recognised at once the resemblance to the vertebrate eye, which is such a striking character of the eyes of Pecten. These organs were mentioned, though left practically undescribed by suc- ceeding naturalists. Cuvier refers to them as “globules verdatres,”’ and Lamarck as “tubercules oculiformes.” The next description is to be found in Robert Grant’s “Comparative Anatomy’ (2), where reference is made to the “smooth cornea,” the “iridescent choroidea,” and a “small crystalline lens.” Another English writer, Robert Garner (8), 1837, continued the work. He states that Pecten, Spondylus, and Ostrea (probably Pecten Jacobus, Ostrea jacobeus of Poli) possess “small, brilliant, emerald-like ocelli, which, from their structure, having each a minute nerve, a pupil, a pigmentum, a striated body, and a lens, and from their situa- tion at the edge of the mantle, where alone such organs could be useful, and also placed, as in Gasteropoda, with the tentacles, must be organs of vision.” There are no figures illustrating his shortaccount. Almost simultaneously Krohn (5) and Grube (4) published descriptions of the eye. Grube des- cribed the position and number of the eyes in P. jacobeus, P. varius, and P. opercularis. Krohn gave a much more detailed account. He stated that the eye was a closed spherical vesicle containing two transparent bodies separated by a septum (he was therefore the first observer to see this structure). The hinder of these bodies he described as being of fibrous texture. Krohn was the first investigator to notice 52 W. J. DAKIN. that the nerve in the eye-stalk divided into two branches, one of which ran up to the optic vesicle, where he lost it, whilst the other passed up the side and entered the vesicle, lying on the septum. Will (6) noticed the cellular structure of the lens, and Keferstein (12) recognised the retina in the hinder trans- parent body of Krohn. ‘This brings us to Hensen’s paper (18) published in 1865, which is the first account of the histology of the retina. Hensen divided this part of the eye into five layers : 1. First cell layer. . Second cell layer. . Rods. . Tapetum. . Pigment layer. The cells of the first layer, which may be arranged in a single or double row, are spindle-shaped. The second layer is made up of cylindrical cells (the rod-cells), the third layer is that of the rods, and then follow two others—the tapetum (first demonstrated by Krohn), and the pigment layer. The innervation is described as follows: The proximal branch of the optic nerve does not bore through the optic vesicle below, as Keferstein had assumed, but splits into a number of small branches which enclose the lower part of the optic vesicle, and these branches of the nerve form a plexus in the peripheral region of the retina. Apparently Hensen assumed that they were connected with his second cell-layer (the rod-cells)—“ Der Zellenauslaufer geht so continuirlich in den Nerven iiber, dass man nicht sagen kann, wo der eine anfingt und der andere aufhort.” The other nerve-branch penetrates the septum, and the fibres become connected to the cells of the first layer. Hensen, it will be seen, discovered the different groups of cells in the retina, described the nerve innervation correctly (though since he did not recognise two types of cells in the outer layer and in the rod-cell layer, this was probably more accidental than otherwise), and saw the axial fibre in the Or & © pO THE. EYE OF PECTEN, 53 rods—truly a marked advance in the knowledge of the eye- structure. TECHNIQUE. This investigation of the eye has been carried out by the study of sections (paraffin and paraffin-celloidin), by macera- tion preparations and by the teasing of fixed material. It is impossible to over-estimate the value of macerations in conjunction with section work, and the true shape of many cells could not have been determined without this method. For both fixation and maceration it was found that different reagents were necessary according to the cells to be studied. In the retina alone the various elements reacted very differently to fixatives and macerating fluids, and it was surprising to notice how different the preservation of the different cells might be after treatment with the same fixative. The fixatives giving the best general fixation of all parts were Zenker’s fluid and Carnoy’s mixture. Zenker was used as follows: Fixation lasted for about twelve to twenty-four hours, and was followed by washing first with water and then in alcohol of gradually rising strength. Sections were made after paraffin embedding, the usual thickness being that of the rod-cells, namely 6 u, but others were only 2 uw, and some were 10 » thick. ‘The stains used after Zenker were Mallory (connective-tissue stain), iron hematoxylin (Heidenhain), a modified Weigert, and picric acid—siurefuchsin. Mallory’s connective-tissue stain.—The sections, on slides, were stained in an aqueous solution of saurefuchsin, 0:05 per cent., for ten minutes, then rinsed quickly in water and placed in | per cent. solution of phosphormolybdic acid for three to five minutes. After washing in several changes of water for five to ten minutes the sections were stained in the following solution for eight to fifteen minutes : Aqueous aniline blue (Griibler) ; ‘ 0°5 gr. OrangeG. . : : : : : 2:0 gr. Oxalic acid. : ; ; ; : 2°0 gr. Water -. " : : : : = L00-Olei: 54 , W. J. DAKIN. This was followed by a rapid washing out in water, dehydra- tion in 90 per cent. alcohol to absolute, and mounting after xylol or origanum oil in balsam. Iron hematoxylin.—The sections were mordanted for twenty-four hours in a 4 per cent. solution of iron alum washed in water, and stained in a 0°5 per cent. to 1 per cent. solution (aqueous) of hematoxylin for twelve to twenty-four hours. This was followed by differentiation under microscopic observation with 2 per cent. iron alum solution. ‘Tap-water, alcohol dehydration, etc., as usual. The modified Weigert method was only used after Zenker fixation. It was partly like that used by Schreiner (80), but modified in combination with Zenker. Schreiner used a 10 per cent. alcoholic solution of haema- toxylin (P. Mayer says that the “10 per cent.” must be a misprint). I used a 5 per cent. solution, but did not investigate the effects of a stronger nor of al per cent. solution, which Mayer believes to be the one intended by Schreiner. The sections (on slides) were placed ina 3 per cent. solution of potassium bichromate for twenty-four hours, then rinsed in water and alcohol, and placed in a 5 per cent. solution of hematoxylin (alcoholic) for a time varying from ten minutes toan hour. The sections must be black, and this takes place much quicker after the hematoxylin solution has been used once or twice and is oxidised by contamination with bichro- mate. After staining, the sections were rinsed in water and placed in a saturated aqueous solution of copper acetate, which turns them a steel-blue colour. Differentiation was carried out (under microscopic observation) in the following solution : Borax . ‘ : : : 2°0 gr. Pot. fernioyanite : : ; : 2°5 gr. Distilled water ‘ : i . 100°0 c.c. The sections were then aaahod in tap water and mounted in the usual way, after alcohols and xylol, in Canada-balsam. Picric acid—saurefuchsin (van Gieson).—The sections THE -EYE OF. PEOTEN. 5) were stained in Delafield’s hematoxylin and washed well in tap water. This was followed by staining for five minutes in a mixture of— 1 per cent. solution (aqueous) siurefuchsin . 50 Saturated solution of picric acid in water . 1000 The stained sections were washed in tap-water and taken up to balsam as usual. Carnoy’s fixative was used in the following strength : Chloroform ; ; E ; , ; 10°0 Acetic acid f : : : t ; 30°0 Absolute alcohol : F , t 60:0 This is the best fixative for the retina. Iron hematoxylin and Bethe’s toluidin blue were the stains used on material so fixed. Bouin’s fluid (‘ Lee,’ edit. vi., p. 76) gave excellent results for rod-cells and rods, especially when followed by Mallory’s stain. The axial fibre of the rods was stained better by the siurefuchsin in this method than by any other except the modified Weigert. Zenker’s fluid, Mann’s fluid,' and a mix- ture of equal parts of corrosive sublimate saturated aqueous solution, and Hermann’s platinum-osmic fluid were useful for the lens, especially the latter. Other fixatives used were 4 per cent. formol, corrosive sublimate (aqueous solution and solution in salt water), Mayer’s picronitric mixture (‘ Lee and Mayer,’ ed. vi, p. 68), Flemming, Von Rath’s picro-platinum-osmic mixture, and treatment with pyroligneous acid. The latter did not give particularly good results. There were also special fixing and other processes connected with the following methods—Golgi’s silver process (Cajal’s modification), Bielschowsky-Paton silver method for neurofibrille (41), Apathy’s nachvergoldung and hematein IA methods, Nabias’ gold method, Lists’ eosin method, and methylene blue processes. The latter were failures, though injection methods, staining in aqueous solu- tions, solutions in Pecten serum, and dusting powder over the eye were all tried. The results given by the other and more 1 Mann, ‘ Physiological Histology,’ p. 96 (solution d). 56 W. J. DAKIN. ordinary methods were more complete than by the complicated ones, and there was usually a far greater freedom from artefacts. There remains finally the maceration methods to be referred to, The lens-cells, with all their peculiar processes, were easily isolated after immersion of the eyes directly in a 3 per cent. solution of chloral hydrate in sea- water for about four hours. The same solution was used for the retinal cells, and the eyes were placed, as above, directly into this medium. After two hours the retina was dissected out from the eye, placed in a drop of water on a slide, and a cover-glass supported by wax feet placed above it. Gentle tapping on the cover-glass separated the elements. Chromic acid solutions in sea-water of =, per cent. strength gave very good results for macerations of the rod-cells and rods. This was also used as advised by Patten after fixation of the eyes in } per cent. chromic acid for five minutes. The maceration preparations were examined unstained, and stained with picro-carmine. The chief species examined have been Pecten maximus and P. jacobeus, with the following others: Pecten oper- cularis, P. varius, P. tigrinus, and P. tenuicostatus. Position AND NumBer or Kyzs. The eyes of Pecten occur on the mantle-edges of both valves. ‘The mantle-edge can be said to be divisible into three folds, the periostracal fold, the ophthalmic fold, and the velum (Pl. 7, fig. 2, V.). All three possess tentacles, those situated on the first two being long and mobile sensory structures, well provided with sense-cells for the perception of tactile and olfactory stimuli, whilst those on the velum are short and rather immobile. The eyes are situated. on the median fold, between the periostracal groove and the base of -the velum (PI. 7, fig. 2, Hye), and amongst the long tentacles. Poli in 1795 noticed a certain resemblance of the eye-stalks to the tentacles, and considered them as modifications of the latter. The number of eyes present varies considerably for the THE. EYE OF PECTEN. 57 different species, and there is, further, considerable variation among’ the individual members of any species. Carriére (21) stated that those species with large eyes possessed fewer than those with small eyes; that there were always more on the upper mantle-lobe than on the lower; and that in general, large specimens had more eyes than smaller ones of the same species. This latter sentence was an important assertion, since it implied growth and development of new eyes during life, and certainly it appeared supported by the fact that large and small eyes exist side by side. Patten (22) also pointed out that there were more eyes present on the left valve than on the right, and that they were larger, but he disagreed with Carriére, stating that no new eyes develop after a size of 2 centimetres has been attained. Rawitz (25) found similarly more eyes on the left mantle-lobe than on the right, and agreed with Patten on the development. Schreiner (30) agrees also with reference to the number of eyes on the two mantle-lobes, but states that those of the right are not smaller than those of the left (Patten). Had Schreiner examined P. jacobexus, the chief species investigated by Patten, he would not have made this assertion. The eyes are always more numerous on the left mantle-lobe than on the right, as all observers have found. ‘The exact relations, however, vary in different species. The eyes are situated in three groups, on each mantle-fold, one group on the anterior auricular area (two to seven eyes close together), another on the posterior auricular area, close up against the hinge-line, and the third and largest group along the ventral margin of the mantle. Spaces without eyes separate these three regions. In each series the eyes vary considerably in size. Patten (22) asserted, in fact, that a regular arrangement of small and large eyes existed, and Rawitz (25), though denying the existence of Patten’s arrangement, stated that a large eye was always followed by a small one. I have examined all the species referred to by Patten and Rawitz and find no such arrange- 58 W. J. DAKIN. ment. There is a quite irregular series, and a small eye may be followed by another small one or by two large ones, or a group of large eyes may exist together. The eyes on the left mantle-lobe exceed in number those on the right, in particular in species with the most inequivalve shells (as far as the species I have examined are concerned), that is, in P. maximus and P. jacobeus, and this difference in numbers is greatest in P. jacobeus. The eyes in this species are far less numerous on the right lobe, and are alsovery much smaller (PI. 7, fig. 2) (contra Schreiner). I believe, however, that the greater number of eyes on the left mantle-lobe is due primarily to the fact that this valve is always uppermost, and not to its shape; and if a Pecten is turned over on to the left valve, it very soon rights itself by a peculiar turning movement. Patten (22) connected this numerical superiority of eyes on the left valve with its position but was puzzled to see how this could be an advan- tage to the animal, since the eyes on the lower mantle-fold received the light direct from above, and the eyes on the upper one were apparently directed downwards, Schreiner (30) also figures them as lying pointed to the ground and at an angle of 45° to the valve. Ifa Pecten be watched as it opens the valves, it will be seen that the eyes of the left mantle-lobe project just outside the shell, and their field of view is practically as much above the animal as that of the eyes in the right valve. The upper valve is alsoa little shorter than the lower one, and lies inside it when the shell is closed; the mantle lining the lower valve is retracted accordingly to a greater extent when the shell is closing. ‘The valves of the almost equivalve species meet, however, ventrally, and the conditions appear either more favourable to the eyes of the right mantle-lobe than in P. jacobeus, or else, as will be referred to again, this form is an older and more specialised one, and the eyes have begun to degenerate in the lower valve. Some figures are appended which will give an idea of the number of eyes in the three THE EYE OF PECTEN. 59 groups on the mantle-edge of P. opercularis and also of the individual variation in this species (the specimens were from the Irish Sea). Left mantle-lobe. Right mantle-lobe. Length of ant.- Li faved post. diameter. of shellin cm. | Total | No.on | No. on | Total No. on | No. on No. ant. ear. |post. ear.| No. ant. ear. |post. ear. | 75 59 6 4 42, 0 3 6°4 50 4, 6 41 0 4 5'8 48 4 4 39 0 3 38 37 4 5 Bil 2 5 64 54 2 5 40 ) 4 58 50 4 4 39 1 4, 4-25 44 4 5 41 2 4 51 52 3 6 42 2 4 51 Dee || 4 5 50 2 6 5°25 45 3 4. 39 2 3 5:15 54 5 5 35 il 3 515 53 5 6 48 3 3 5°85 55 5 5 49 3 4, 7:25 58 a 5 47 2 5 4:10 59 4 Se ane 1 5 50 55 4 3 50 2 4 7A 51 2 4 38 i 5 0:25 61 i] 7 47 4 HY 475 55 5 6 43 il 4 4-70 54 5 Sys oli I 4 O38 62 6 a aig) i 4 44, 51 6 Dl Ail 1 5 49 57 5 5 | 43 1 4 It will be seen from these figures that there is no relation between the size of the animal and the number of eyes, though if the first five only had been taken the reverse would have appeared to be the case. Possibly Carriére only examined a few and chanced to get an accidental series. No one appears to have examined the very small eyes occurring with the large ones. I sectioned some of those taken from the right mantle-lobe of Pecten jacobeus and found that they agreed in every respect with the large eyes of the left lobe, all parts being represented and in the normal positions. The 60 W. J. DAKIN. only difference was in the number of cells present; they were apparently as large as usual but fewer in number. ‘These eyes, in fact, appeared to be young ones, or rather, they had been arrested in development and had remained with the small number of component cells characteristic of young eyes, though they were just as old as the large ones. In examining hundreds of eyes one meets some strange abnormalities, though the latter are of rather rare occurrence. In a specimen of P. opercularis two eyes were fused together so that the pupil was oval with a slight constriction indicating the boundary of the separate organs. Often the eyes appeared with very little black pigment—that is, all the eyes of a specimen, even the “iris” cells being almost un- pigmented. I never found any of the eyes completely covered with pigment as stated by Patten, nor has this feature been met with by any of his successors. GENERAL STRUCTURE OF EYE-STALK. The eyes are situated at the ends of short stalks (Pl. 6, fig. 1), which, as already pointed out, were considered by Poli as modified tentacles. This eye-stalk is made up of connective tissue, which is a direct continuation of that of the mantle-edge and is clothed by an epithelial layer, also a direct continuation of the pallial epithelium. The connective tissue is more homogeneous or hyaline in appearance than that of the tentacles, and is not broken up so much by crossing muscle-fibres, which, as might be ex- pected, are a prominent feature of the retractile tentacles. This homogeneous tissue extends also below the eye-stalk for some distance, and the transverse muscle-fibres which raise the velum are absent under the eyes, being arranged in bundles situated between these sense-organs. Large blood- spaces occur irregularly scattered in the stalk, communicating with one another and usually containing blood-corpuscles (Pl. 6, fig. 1, Lac.). There is, however, scarcely such a THE EYE OF PECTEN. 61 defined space as a “ Hauptader des Augenstieles” to which these lacunz belong (Rawitz [25], p. 105). Neither do they always surround the nerve (Schreiner, p. 11). Whilst the long sensory tentacles are, in the living animal, continually in motion, being retracted and again extended, and moved from side to side, the eyes are practically motion- less and point fixedly in one direction only. They contract and may move away from a point of stimulation, this bemg rendered possible by means of muscle-tibres, which lie longi- tudinally arranged, near the epithelium (PI. 6, fig. 1, Mus.). The latter are narrow fibres, and are not striated, as figured by Patten. Striated muscles do occur, though elsewhere, in the mantle-edge of Pecten (45). The muscles occur on all sides of the eye-stalk. They terminate, according to Rawitz, always at the proximal end of the optic vesicle and are never to be found higher ([25], p. 105). Rawitz has presumably taken the finer muscle-fibres, which do extend up to the cornea, for connective-tissue fibres. Schreiner found prac- tically no muscles in small eyes ([30], p. 11),and states that in P. islandicus, where they were exceptionally well developed on the shell side, they could be traced to the entrance of the distal branch of the optic nerve. I have traced them to this point in P. maximus, but more delicate fibrils (Pl. 6, fig. 1, M.f.), staining quite differently from connective-tissue fibres, extend under the epithelium as far as the edge of the cornea, and are, moreover, present between the cornea and the lens (oie; fic. 1, .N.Lf.; Pl. 7, fis. 7, Lf): These are, evidently the “fine smooth fibres” mentioned by Patten in contra- distinction to his “long striated muscle-cells” of the lower part of the eye-stalk. These fibrils do not, however, enter into any connection with the epithelial cells bounding the cornea, and Patten’s “ciliaris ” does not exist. They will be referred to again wher discussing the fibres situated between the lens and cornea. Ganglion cells do not occur scattered in the connective tissue of the eye-stalk, a fact already noted by Patten’s successors, who criticised his observations on this, as on other 62 W. J. DAKIN. details, somewhat severely. The epithelium covering the eye-stalk is a direct continuation of the pallial epithelium, but is modified in various regions of the eye-stalk and becomes a transparent cornea over the free pole. Below the optic vesicle the cells are small and cubical, or rather deeper than wide (Pl. 6, fig. 1). They contain no pigment here, and the nucleus is situated near the base. A distinct cuticle’ is present. Some little distance below the optic vesicle these cells increase in depth and at the same time begin to contain pigment. This pigment extends further down that side of the eye which is uppermost (see fig. 1, Pl. 6; the right-hand side is the shell side of the eye and also the uppermost, since it is an eye from the left valve). At the level of the middle of the optic vesicle, that is, about the plane of the septum, the epithelial cells have attained their greatest depth and are almost filled with dark pigment, occurring in the form of fine - granules. The external portion of the cells is usually less thickly crowded, and if the sections are stained to bring out the nuclei it will be seen that these have moved, with the acquisition of pigment, so that they reside near the surface instead of at the basal end. The statements of Rawitz and Schreiner in regard to the colour of this pigment in the different species appear to me to be of little importance, and in any case I can hardly confirm them. ‘The colour of the granules in Pecten jacobeus, P. maximus, and P. opercularis is dark brown, and the exact shade varies in any one species and according to fixation and preservation; moreover, the cells are completely filled in P. jacobeus, or at least those of the upper side of the eye-stalk. Another point that may be noted here is that the increase in height of the epithelial cells opposite the optic vesicle is common to all the species I have examined, though Rawitz states that in P. jacobeus the epithelium is everywhere the same in height and figures it as such ((25], p. 106). Pecten abyssorum possesses (Schreiner) no pigment in the cells of the mantle-edge or of the eye-stalk. Patten appears to be the only one who has noticed that there is more pigment THE EYE OF PECTEN. 63 present on the upper side of the eye-stalk, and there is really a longitudinal band present, exactly similar (though not so definite) to the one on the corresponding side of the tentacles. The pigmented area bounding the cornea was termed the “iris” by Patten. Since, however, as described above, these pigmented cells extend far down the eye-stalk on both sides, it is difficult to make any division into regions or to define a boundary. If, moreover, the physiological action of the iris were considered solely to be that of a diaphragm, keeping out oblique rays, the name might perhaps be applied, but, as Rawitz pointed out, there is no proof whatever of this area being capable of contraction with diminution of the “ pupil,” and since this region is not to be homologised with the vertebrate structure of the same name it is better to use the term pigment-mantle (Pl. 6, fig. 1, P. man.) if a special one is necessary. Patten considered that the “ pupil” could be diminished to almost half its previous diameter (p. 571), but I have been unable to find any trace of this under natural conditions, nor do any other authors appear to have been more fortunate. The same writer states that on the shell side even in fully formed eyes the pigment may sometimes be absent so that a colourless fissure is left—termed by him the “choroid fissure ” (p. 578).. I have not seen this in any eye examined, and fail to find any references confirming the statement of its existence. The pigmented epithelial cells pass suddenly into the transparent cells of the cornea (PI. 6, fig. 1, Co.), through which is seen in the living specimen the silvery glance of the subretinal structures. In P. maximus the depth of the tall epithelial cells may decrease slightly in one or two cells, and then the next is much lower and completely free from pigment. Sometimes, however, the decrease in height takes place after the pigment becomes absent. The nuclei take up again a central position or a position nearer the base in the corneal cells, but there are certain exceptions which will be considered later. The cells are hexagonal in surface view and are much flatter than those of 64 W. J. DAKIN. the pigment-mantle. ‘They are usually constricted in the middle, so that they appear hour-glass-shaped in section, an intercellular space being left between them (PI. 7, figs. 4 and 10). Externally there is a very distinct striated cuticle (Pl. 7, fig. 10, Cut.) which forms a hexagonal plate over the cell, and if the cornea is carefully focussed down upon from above these hexagonal plates are seen with their edges in close contact forming a definite mosaic (Pl. 7, fig. 3). If the corneal cells are now brought into focus at about the level of the nucleus, they appear still hexagonal in section though rather irregular, and the cell-walls do not touch. ‘he spaces left between the cells on each side are crossed by numerous intercellular bridges (Pl. 7, fig. 4). I have no doubt that these are what Patten took to be interlocking processes of the cells. Carriére (26) was the first to discover their true nature, but asserted that Patten could not have seen them at all, since they were finer than his interlocking processes. Schreiner (30) stated that the intercellular spaces were filled with a prominent cement substance which, through shrinkage during fixation, caused the appearance seen by Patten, and does not mention any intercellular bridges whatever. Rawitz was also of the same opinion and does not refer to Carriére’s statement (Rawitz [25], p. 109). I have seen them quite distinctly in the pigmented cells of the pigment-mantle as well as in the cornea, and they have the same structure in both places. There is another detail to be mentioned here which illustrates the difficulties caused by artefacts. Patten stated that the corneal cells had basal processes like the lateral ones, but which were longer and penetrated the underlying connective tissue, reaching the lens. This has been denied by all investigators since, and J had seen no traces of any such structures in hundreds of sections examined. After using the Bielschowsky-Paton silver method, however, the result figured (Pl. 7, fig. 10) was obtained. The tissues were fixed in 4 per cent. formol and lay in | per cent. silver nitrate solution for three weeks, which one might say was a likely method for artefacts. On the other hand, the structures THE BYE OF PECTEN. 65 appeared well preserved and very little contraction had taken place. The processes were very definite, and had I found them by other confirming methods [I should not have hesitated to describe them as actual cell processes. I have figured, however, the preparation, and prefer to leave the question of their true nature open. The type of cornea just described is that of Pecten jacobeus, P. maximus, and P. opercularis. Rawitz ([25],p. 108) divides the types of cornea into three classes: (1) Cells of cornea considerably smaller than those of the pigment-mantle, ex. P. flexuosus, P. glaber, and P. opercularis; (2) cells of cornea, smaller at periphery against the pigment-mantle, but rapidly increase towards the centre, where they equal the pigment-cells in height, ex. P. jacobeus and P. varius; (3) corneal cells are as high as cells of the pigment-mantle at periphery, but increase rapidly in height towards the centre, the nucleus lying near the base, ex. P. pusio. I hardly think it advisable to make such a division, since, in the first place, the appearance often varies with the size of the eye, and it is difficult to fix a boundary between the two first groups. The corneal cells of P. jacobzus are, moreover, not equal in height to those of the pigment-mantle, though they are much higher in com- parison with the same cells in P. maximus. There is, however, a well-marked division in which Pecten pusio and also P. tigrinus can be placed. ‘The latter is figured (Pl. 7, fig. 12). In this species the corneal cells are very different from those of P. maximus. Those next to the pigment-mantle are of similar size, or smaller than the adjoining pigment-holding cells, but towards the centre the cells increase in height very considerably until they are deeper than the pigment-cells, the height of the corneal cells being double that of the latter. The cell-boundaries are not very distinct, and intercellular bridges are not to be seen. I have been unable to make out any reason for the peculiar difference in these two forms. The connective tissue of the eye-stalk has already been VOL. 55, PART 1.—NEW SERIES. 5 66 W. J. DAKIN. referred to; it is continued around the optic vesicle (PI. 6, fig. 1, Con.) forming the inner wall of this (the outer being formed by the epithelium), and finally persists much diminished in thickness as a thin, transparent, and practically structureless layer underlying the cornea and separating this from the lens (PI. 6, fig. 1, Co. S.). Thisis the ‘‘ pseudo-cornea ” of Patten, and the “ innere Pellucidaschicht” of Rawitz. Nuclei are on rare occasions to be seen in it, but generally it is free from the connective-tissue fibrils and muscle-fibrillz, which appear in that part just outside the corneal area, under the pigment- mantle (Pl. 6, fig. 1, M./.).. This more hyaline character is in all probability due to the fact that light rays have to pass through this layer before entering the optic vesicle. Tue Lens.—The lens (Pl. 6, fig. 1, Z.) is one of the. structures that gave much trouble to the early investigators, but has lately been considered, entirely understood, and passed over somewhat lightly. Hesse made out some new and highly interesting structures, which I have been able to confirm. I find, however, that the shape of the lens-cells has been quite misunderstood, and the cells are certainly of a very peculiar nature. The early authors could not determine the correct shape of the lens itself. Kefersteim believed it to be spherical ; Hensen was uncertain, but believed it to be bi-convex ({13], p- 222) ; Hickson considered it, however, as elliptical ([18], p. 447). _ The confusion was again due to artefacts. It may be taken as definitely proved that the lens is bi-convex. ‘The distal surface is, however, almost flat, whilst the proximal is very convex, and may appear dome-shaped. The actual degree of convexity depends largely on the contraction which has taken place in the eye during fixation, and the lens, dissected free from its limiting elements.in a living specimen, probably alters in shape considerably, since it is not of very firm con- sistency. The lens is suspended from the subcorneal connective tissue (Pl. 6, fig. 1, Co. S.), against which its lesser convex surface is fastened. In surface view this face is circular and THE -EYE OF. PECTEN. 67 not elliptical. Its diameter is a little greater than the cornea, since its periphery extends under the pigment-mantle for a short distance (PI. 6, fig. 1). The space in which the lens is suspended is bounded by the connective-tissue wall of the optic vesicle, the subcorneal extension of the same, and by the septum (PI. 6, fig. 1, Sep.), a membrane separating the dioptric part of the eye from the retina. This space was regarded by Patten as a blood-space. Carriére (21) first saw the blood-corpuscles in this part of the eye, and Patten, though also finding them, was at a loss to account for their presence, since the retina seemed to shut off all communication with the blood-lacunz of the eye-stalk. Rawitz appears to have found a definite vessel running on the outer surface of the optic vesicle and entering the distal part of the eye ([25], p. 113). Schreiner considers these cor- puscles due to pathological conditions, and remarks that the three other authors named above considered them as normal ([80], p.17). This is not strictly correct, since Patten stated that they might be forced into the cavity artificially by the contraction of the connective tissue through the action of reagents. I have only found blood-corpuscles present in this space on extremely few occasions, and on one of these, when there were many, I could trace quite easily a series of spaces in the connective tissue, connecting up the lacune of the eye- stall with the lens-cavity. This may of course have been an abnormal condition, and the lacune may have been produced artificially. These corpuscles had been forced in on the inner side of the eye, and I find no traces of Rawitz’s blood- vessel on the outer side. The blood plays an important part in the extension of the tentacles, and if a small living Pecten is watched under the microscope, the corpuscles can be traced running rapidly along the cavities of the tentacles as they are extended and back in the reverse direction as they contract. I believe their presence in the eye is due to contraction, and that they are forced there from the lacune of the eye-stalk. 68 W. J. DAKIN. There is no membrane covering the lens and helping it to retain its shape. Hensen and Hickson could not find such, but Patten described a “suspensory ligament,” and also stated that the lens was attached to the septum by a con- nective-tissue ligament (P. varius). None of Patten’s successors could find any suspending capsule, neither does the connective-tissue ligament exist. The lens may touch the septum (it very often appears so in sections), but this depends on the contraction during fixation, and usually the retina leaves the posterior wall of the optic vesicle and lies across the middle, coming naturally against the proximal end of the lens. Patten’s connective tissue was in all probability the sheath of the distal nerve-branch (PI. 6, fig. 1, Op. Ds.), which would be touching the lens and lying between this and the septum if the retina had been forced up. Patten’s theories of accommodation as expressed at some length on p- 571 I cannot confirm, and they are somewhat irrational. They have not been referred to at all by his successors. He believed that the contraction of certain muscles supposed to be attached to the suspensory ligament would cause a move- ment of the lens towards the retina. This meant an inward movement of the septal membrane to which the lens (accord- ing to Patten) was attached. ‘The elevation of the lens was to be brought about “by the tendency of the elastic septal membrane to return to its natural position, after the contrac- tion of its peripheral circular fibres has relaxed the tension upon the central portion.” There is, however, no suspensory ligament nor attached muscles, and the lens is not attached to the septum. The septum, moreover, cannot move forward without taking the whole of the retina with it, and if this was the case (rather an absurdity) the recipient elements would always be the same distance behind the lens, whether it had been elevated or otherwise. Accommodation will be referred to later when discussing Hesse’s theory. The lens cells had received little attention until Hesse described them (34). Hensen stated that the lens consisted THE EYE OF PECTEN. 69 of polygonal cells with thick walls. Patten described them as irregular with excentric nuclei, which appear in many cases to have disappeared from the cells near the inner sur- face. Rawitz described them as polygonal and membraneless with small nuclei, and Schreiner terms them “pretty large” vesicular cells, the peripheral ones flattened, with a large nucleus and no cell-membrane. ‘lhe latter writer noticed that in sections of the lens some cells appeared to be without a nucleus (see Pl. 6, fig. 1), but went no further into the question. Hesse says (84) the lens ‘‘besteht wie schon lange bekannt, aus zahlreichen, dicht neben einander gepackten Zellen, deren K6rper sich an einander abplatten und bisweilen eigenthiimliche Formen auf den Durchschnitten zeigen.” Later he adds (p. 395) “. . . da man ferner aus einem Durchschnitt auf die Gesammtgestalt der Zellen nicht schliessen kann, so ist es nicht méglich hier einen Zusam- menhang zwischen Lage des Centralkérperchens und Gestalt der Zelle festzustellen.”” Hesse, however, did not adopt any maceration methods to solve the difficulty presented by sections. In sections through the lens, which is well pre- served in formol-fixed specimens or Hermann-sublimate, the cells only rarely possess a polyhedral shape, in fact it is only here and there that they appear sharply angular. ‘The cell contours are very distinct and appear rounded, so that there are irregular oval, pear-shaped and long band-shaped cells (Pl. 7, figs. 5 and 6). The size, too, varies considerably, and a very small, apparently non-nucleated cell may adjoin a large one. If, however, this small cell be followed through several sections, it will be found to be merely the continua- tion of a cell which is elongated to an extraordinary degree. The true shape of the cells was found after macerating the eye in 23-3 per cent. chloral hydrate solution in sea-water for four to six hours. This medium preserves admirably the delicate processes of the cells, and the preparation gives the lens-cells, separated, uncontracted, and with all details of structure undamaged. 70 W. Ji DAKIN. The cells vary considerably in shape. ‘Those near the surface of the lens, particularly the proximal surface, are flattened and are strap-shaped (PI. 7, fig. 6, ¢.), or are con- stricted in the middle and have two bulging ends. The length may be very considerable. The common appearance is that depicted in fig.5 (PI. 7). The cells are pyriform, with the cell-body drawn out into extraordinary long tapering pro- cesses many times the length of the swollen part. In addition to this, processes are often given off very abruptly from the broad end. Other cells are more rectangular, yet also with rounded contours and the same abrupt fine processes. ‘hese extensions are wedged between adjacent cells (PI. 7, fig. 5), which fit close together, and the result is a mass of great compactness, whose components, though having the most varied shape, fit together without intercellular spaces being left between them. It is often quite difficult to separate some of the cells in macerations. It is now quite obvious why there appears to be no nucleus in many cells in sections, for it may be at one end and the cell be so long that many sections may cut through the latter without touching the nucleus. The cells have a very distinct membrane, and it is difficult to imagine how this could have been missed by Rawitz and Schreiner, especially after Carriére. had asserted its presence. It is easier now to understand why there is no need of a lens- capsule or supporting ligament, for the soft protoplasmic cells are tied together by their processes and the superficial cells are practically converted into fibres or straps. The contour of the lens is, in fact, as even as if formed by a connective-tissue sheath or a layer of pavement epithelial cells. The cell contents are finely granular, with a slight trace of pigment, and stain intensely with eosin. ‘The nuclei are similar in size to those of the epithelial cells, and since the lens-cells are usually somewhat larger than the latter the nuclei can hardly be termed pretty large (Schreiner), though such terms are purely arbitrary. Hesse (84) discovered in the lens-cells of P. jacobeus, which had been fixed in sublimate =e too Oa pa THE. EYE OF PECTEN. 71 and stained in Heidenhain’s iron hematoxylin, a remarkable structure. In addition to the nucleus there was present a dark staining body from which delicate but very distinct fibrils radiated out to the periphery and became attached to the cell-wall. Most of them were straight, some were bent, but all went out from the one point and all could be followed to the cell membrane if their whole length lay in the section. I have found the same structures (PI. 7, fig. 6, 5.), not only in material fixed and stained as above but also after the following treatment : After fixation in Hermann-sublimate mixture and staining in iron hematoxylin, the shape of the cells is well preserved, the contents are homogeneous or very finely granular and stain grey, the nucleus is black, and radiating fibrillee appear distinctly in many cells though not in all. After Zenker fixation and Mallory’s stain the cell contents are very granular in appearance and stained deep red, the nuclei being yellow- red, and there is just a slight trace of the fibrille. They are also to be made out, though not distinctly, after Bouin fixa- tion. Von Rath’s treatment caused the cell contents to appear very granular and vesicular (PI. 7, fig. 6, d.) the radi- ating fibrille were often very distinct, but the central dark staining body did not look exactly like the normal centrosome of dividing cells. This permanent centrosome (PI. 7, fig. 6, Cent.), if it be such, does not appear to have any definite position, but since it cannot be made out in macerations it is almost impossible to determine its true position, for sections cut the cells in all directions. In addition to the species enumerated by Hesse I have found these structures in P. tenuicostatus, and probably they are present in all species. Hesse naturally compared these with the centrosome and astral rays which appear in cells undergoing mitotic division. Such structures have been demonstrated as persisting in the resting stages of certain cells, in pigment-cells of fishes, and more particularly in leucocytes. It has not been possible for Hesse or myself to determine any connection with cell-division. ‘he astral 72 W. J. DAKIN. rays are very fine and remarkably definite, ‘There are three explanations of these structuresthat may be given. ‘The first and most unlikely is that they are artificial productions ; the second, that they are modified astral rays and centrosome kept permanently for another function; the third, that they are entirely different from those functioning in the cell divi- sion, but have arisen ina similar way and are purely supporting fibrilla. The appearance of the structures and their presence after such varied treatment is against the first view. It would only be possible to demonstrate which of the latter were correct if the origin of the aster had been observed. I believe they are supporting fibrillee whatever be their mode of origin, and this is Hesse’s view, he considering they are for the purpose of increasing the elasticity of the cells, This is put forward in an interesting theory of accommodation, and the fibrillee are considered to form the antagonistic appa- ratus to another, to be referred to presently, which alters the shape of the lens, Between the sub-corneal connective tissue and the lens is a layer of peculiar fibres, first seen, though incorrectly described, by Patten. He made out two layers, a series of radiating fibres extending from the centre of the distal surface of the lens to the periphery, superimposed on a layer of strong circular fibres concentrically arranged (p. 581). As such do no fibres exist. Rawitz saw none here whatever, and regarded Patten’s structures as artefacts ( [25], p- 113). Hesse discovered the true conditions, which I can confirm with some slight additional features. There is one layer of fibres only (Pl. 7, fig. 8), and these have a kind of spiral arrangement, so that towards the centre of the lens surface they are running at almost right angles to their previous course. Near the periphery they run more or less concentrically (Pl. 7, fig. 8). They do not terminate at the centre of this surface, but continue across for some distance, and there results a series of fibres crossing one another in all directions. In thin sections cut parallel with the plane of the cornea it is possible to see a number of nuclei here, with very deli- THE EYE OF PECTEN. 73 cate celi-outlines enclosing them (PI. 7, figs. 7 and 9). These cells have their ends drawn out into the long fibres seen in macerations so easily, and which are many times the length of the cell-body (Pl. 7, fig. 2). In some cases, as Hesse pointed out, a number of fine parallel fibrils appear to pass out of and through the cells (Pl. 7, fig. 7). He regards the fibres as muscle-fibres, and the cell-body as containing the remaining myosare and nucleus. This view is based on the reaction to picric acid—saurefuchsin, which stains muscle yellow and connective-tissue red. I was not sure that they were not connective tissue cells, and in fact believed them to be such. For this reason Mallory’s connective-tissue stain was used as recorded on p. 53. The fibres and cells were stained by this process an intense red, against the blue sub- corneal tissue above (Pl. 7, fig. 7). They stain therefore as muscle-fibres. Hesse says ([84], p. 397) that these fibres extend to the edge of the lens but not further. The same fibres, however, are to be found in the connective tissue extending down the sides of the optic vesicle (PI. 6, fig. 1, W.f.) and often quite near or even on the inner surface of the same. I believe they have a far wider distribution than Hesse supposed. ‘This is the apparatus that, aided by the lens-cells, is (according to Hesse) concerned with accom- modation. Through the contraction of these fibres the outer surface of the lens becomes reduced in extent, the lens-cells are compressed together here, and, being plastic, change their shape, the contents swelling towards the inner surface where there is less tension. ‘The result is an alteration in the shape of the lens and hence of the focus. If the muscles are relaxed the elastic cells (aided by the fibrille) return to their previous shape and the focus is adapted for more distant objects. No physiological proof has yet been brought to support this theory, and, as far as experiments go, I could find no evidence of accommodation (see p. 102). Hesse has built up his theory simply to account for the fibres on the lens and the persistent astral rays in the cells. The function of the latter may be simply to give greater 74, W. J: DAKIN. rigidity to the lens, and if the former were accommodation muscles one would expect a more definite and efficient arrangement. ‘I'he same red-staining fibres can be traced, however, down the sides of the optic vesicle in the connec- tive tissue, and those present between the lens and cornea may be simply for the purpose of tying the lens to the sub- corneal layer. Before leaving the lens it will be advisable to refer to another condition seen in some of the lens-cells. This is a peculiar condition of the nucleus (perhaps patho- logical) observed in one or two cells in preparations fixed in von Rath’s fluid and also in Hermann-sublimate mixture (preparations stained with Heidenhain’s iron hematoxylin). The latter specimen was an eye from a small P. opercularis or P. varius. The nucleus (PI. 7, fig. 6, a, wuc.) is perfectly spherical and much larger than the normal ones. The size of the normal nuclei was 5°3 wby 4 (they are oval in shape), whereas the spherical ones attained a diameter of 10°6 p. These nuclei were homogeneous, not staining deep black as the normal ones, but rather grey, slightly darker than the cytoplasm. VA. oe: &3 y ay ee ay ~~“ Section of one of the nuclei of a couple in the second trophic phase. The nucleolus does not project very far, and the surface of the absorptive area is flush with that of the nucleolus, although the area itself is very large. assimilation ceases, returns, as will be seen later, to the interior of the nucleus. (iii) Encysted Phase. The cyst-wall, though always fairly strong, varies a good deal in thickness. It stains bright blue by Mann’s method, bright red with carmine, but not strongly with hematoxylin. From it often project radially inwards curious irregular, branching filaments, never reaching much more than a third of the way to the centre, as to whose nature and function I am quite in the dark (fig. 16). The cytoplasm is reticular, with minute granules on the 168 JULIAN 8S. HUXLEY. threads, and larger, chromatic granules here and there. It always looks denser than in the unencysted forms. The nuclei in what I take to be the earlier cysts are much like those described for the second trophic phase, except that they stain a little deeper, and that the nucleoli do not pro- ject so far out from the surface (text-fig. 5). In the next TrxT-Fia. 5. A cyst found in the gut. The nuclei are not actually touching, but very near to each other. The cyst-wall is very thick in this specimen. stage (fig. 16) the nuclei, bounded only by a very thin mem- brane, stain quite deeply, as they are almost filled with chromatic granules of various sizes. The nucleolus is still in contact with the cytoplasm, but its outer surface is now flush with that of the nucleus. This outer border of the nucleolus is made up of rows of minute vacuoles, while the GANYMEDES ANASPIDIS (NOV. GEN., NOV. spP.). 169 rest is dense, with a clean-drawn boundary towards the interior, and homogeneous except for a few large vacuoles. To this stage probably belongs the cyst in fig. 19, stained by Mann’s method. The nucleolus is blue, having given up most of its chromatin to the nucleus, which is violet with dark purple grains. In fig. 17 we have another state of affairs: The nu- cleolus, now retreated from the surface, seems to be giving off chromatin to the nucleus in the shape of hollow spherules. It is itself formed of a single central vacuole, surrounded by a layer of small ones embedded in a dense chromatic cortex (the lower nucleolus is cut tangentially, and so does not show this condition). The nucleus, apart from the chromatic spherules, appears perfectly homogeneous, with no achro- matic network, and differs also from the nuclei of other stages in being ameebiform, with ‘‘ pseudopodia” that can be very clearly seen on focussing up and down. From what we know of other Gregarines, it is clear that these stages are preliminary to the breakdown of the large trophic nuclear apparatus, and the reconstitution of the idiochromatin to form the gametocyte nucleus. But, as above mentioned, the cysts soon after this pass into the gut and out by the anus, so that their further development must remain for the present unknown. Conciusions: Systematic Postrion. Though here more than ever must we lament the absence of spores, it is still possible to draw some fairly definite conclusions. To start with,Ganymedes is not a Polycystid, nor does it belong to any existing family among the Mono- cystids. ‘hus a new family, the Ganymedida, must be created, whose characters will provisionally be those of the genus: these may be here conveniently summarised as follows : (1) The possession by the motile form of a special exten- sible organ at the front end, which may serve for fixation to the cells of the host. 170 JULIAN 8. HUXLEY. (2) The presence of a special cup-like structure at the posterior end, which co-operates with the epimeritic organ at the anterior end to effect a close union of two individuals during association. Association is thus by dissimilar ends, and lasts for some time. (3) The eventual complete cytoplasmic fusion of the associated couples, and the existence of a second trophic phase, when the animals grow very fast, but are morpho- logically quite degenerate. (4) The position of the nucleolus in this phase, on one side of the nucleus, partly in contact with the cytoplasm. (5) The habitat, in the gut and liver of Syncaridan Crustacea. Considering these characters in relation with other members of the class, we find that no known Gregarines inhabit the liver of any Crustacean; none have the nucleolus in the same position; none go through two trophic phases; none have any special structure for association at the posterior end; and none have a protrusible organ of the same sort at the front end. It is thus at least obvions that Ganymedes is the representative of a very divergent line. The suggestion I would make is that, while nearer to the Monocystid type, Ganymedes is partly intermediate between the two great groups of Hugregarines, as represented diagrammatically in the following tree: Polycystidea. Ganymedes. Monocystidea. Ancestral Eugregarines. GANYMEDES ANASPIDIS (NOV. GEN., Nov. SP.). 171 In the first place, the ball and the cavity containing the secreted fiuid represent with great probability an epimerite and protomerite. ‘'rue, there is no cuticular septum; but the secreting tissue forms a fairly definite barrier between these on one side, and on the other the deutomeritic posterior part. Here alone, it is to be remarked, do we find the true granular endoplasm. Occasionally, too, this latter can be seen ending off with a definite contour within the secreting tissue (fig. 3). The ball itself, when extruded, would pass for a typical epimerite save for the absence of a septum behind it; but in so far asit is protrusible, it is only paralleled by the anterior extremity of Lankesteria ascidiz (Sied- lecki, 4). his, however, seems to be merely a pseudo- podium, or a drop of the hyaline inter-reticular substance of the cytoplasm pressed out through a hole by contraction of the animal, and its extrusibility has obviously been inde- pendently evolved. The fact of its being a parasite of the digestive tract is the second link with the Polycystidea. The only Monocystid eut-parasite whose life-history has been thoroughly worked out is Lankesteria, and this possesses an “ epimeritic” organ. The three or four other genera of this sub-class that live in the gut, such as Callyntrochlamys and Ancora, are very insufficiently known; it is even possible that they may be Polycystid in early stages. Regarding the matter phylogenetically, we find that the early Eugregarine stock must have been motile, Polycystid gut-parasites ; their association was by dissimilar ends, and took place only at the very end of the trophic period ; and they showed well-marked anisogamy. One of the first steps towards the typical Monocystid condition was the change of habitat, due very likely in the first instance to the evagination of the full-grown tropho- zoites from the gut into the coelom—as takes place to-day in certain insect-parasites at the time of the host’s meta- morphosis. For a full discussion of the further stages, leading eventually to complete isogamy, coupled with entirely 172 JULIAN S. HUXLEY. coelomic habitat, precocious ‘association, and degenerate structure, the reader is referred to Woodcock (6). Suffice it here to say that the course of affairs in Ganymedes must have been somewhat different. It is probable that Gany- medes at first associated only at the close of the trophozoite stage. Some of the couples having migrated into the liver, found it (like the ccelom for other Monocystidea) a safe retreat and abounding in soluble food. Here too the Gregarine could afford to dispense with all the structures necessary for a life in the open gut, and devote all its energies to growing. One might have thought then that Ganymedes would have associated in the sporozoite stage, like Cystobia, and migrated at once into the liver; but, whether non-motile couples below a certain size could be expelled from the tubes or be engulfed and digested by the activity of the liver-cells (see Smith, 5, p. 536), or from some other cause, Ganymedes has found it necessary to remain in the gut till it has attained a definite bulk, thus presenting to us the phenomenon of two sharply-distinct trophic phases after the sporozoite stage. As the parasites are non-motile when they are about to sporulate, conjugation must needs be precocious, so that no Gregarine shall migrate alone into the liver, and thus be, from the point of view of the species, wasted. For this fairly lasting association some special mechanism was imperative, hence the cup and ball; while the necessity of remaining some time in the gut has led to Ganymedes retaining more of the original Polycystid structures than is usual in the morphologically degenerate Monocystidea. Finally, although the sporogony remains unknown, it may be confidently prophesied that this Grega- rine will be found to be completely isogamous. Thus it will be seen that the Ganymedide diverged very early from the Monocystid stock, and possess now many new and peculiar characters intermixed with those they have inherited from the common ancestor. For the complete disentangling of these from each other, further work must be done on Ganymedes, and in addition all GANYMEDES ANASPIDIS (NOV. GEN., NOV. SP.). 178 Syncaridan Crustacea should be searched for allied parasites, whose structure would at once give us new standpoints from whence to view the problem. LITERATURE REFERRED TO. _ . Awerinzew.—‘ Studien tiber parasitische Protozoen : I. Ceratomyxa Drepanopsettae,” ‘ Arch. f. Protistenkunde,’ vol. 14, 1909, p. 74. i) . Léger and Duboseq.—‘ L’Evolution schizogonique de Aggregata eberthi,” ‘ Arch. f. Protistenkunde,’ vol. 12, 1908, p. 44. 3. Minchin.—Article ‘‘ Sporozoa,” in Lankester’s ‘ Treatise on Zoology,’ Part I, 2nd fascicle, 1903. 4. Siedlecki.—‘‘ Ueber die geschlechtliche Vermehrung der Mono- cystis ascidiae, R. Lank.,” ‘ Bull. Internat. Acad. Sci. Cracovie,’ Dec., 1899. 5. Smith.—‘“On the Anaspidacea, Living and Fossil,” ‘Quart. Journ. Mier. Sci.,’ vol. 53, 1909, p. 489. 6. Woodcock.—* The Life-cycle of ‘Cystobia’ irregularis, Minch., together with observations on other ‘Neogamous’ Gregarines,” ‘Quart. Journ. Micr. Sci.,’ vol. 50, 1906, p. 1. EXPLANATION OF PLATE 11, Illustrating Mr. Julian Huxley’s paper “On Ganymedes anaspidis (nov. gen., nov. sp.).” REFERENCE LETTERS FOR THE FIGURES. b. Ball-cavity. c¢.s. Cuticular strie. ect. Ectoplasm (probable myo- cyte layer). p. Cytoplasmic pad at anterior end of ball. s.¢. Secreting tissue. v. Vacuoles. w. Secreted fluid that accumulates to drive the ball out. Bor.-car. Borax carmine. Paracarm. Paracarmine. Hem. Hema- toxylin. M. B. EH. Methyl-blue eosin (Mann’s method). 174 JULIAN S. HUXLEY. PoATH, 11. Fig. 1—Large individual at the close of the first trophic stage, with well-formed cup. (Paracarm. x 640.) Fig. 2—Ball end of the same, to show the ball being pushed out through the hole in the cuticle; the edges of this hole stand out markedly as a rim (cut. rim). (x 1300.) Figs. 3-8.—Ball ends of various Gregarines in different condi- tions. Fig. 5.—(Semi-diagrammatic). Very slightly extended. Secreting tissue very large, with the granular endoplasm (e,) ending off within it. Outside is a non-granular layer (e,), and just beneath the cuticle the still paler ectoplasm, extending on the left to touch the ball-vesicle. (Bor. Carm. X 1875.) Fig. 4.—Semi-extended. The secretion of the secreting tissue is fairly dense. The double contour of the hinder part of the ball is well seen. There seems to be no ectoplasm. (Iron Hem. x 1875.) Fig. 5.—(Semi-diagrammatic.) Completely retracted. Very large cytoplasmic pad (p) with dark grains in it. A large dark granule in the secreting tissue. The ectoplasm extends to touch the ball. (Iron Hem. x 1875.) Fig. 6.—Almost extended. The secreted fluid has here a wide- meshed structure. The thick body-cuticle ends abruptly where it touches the ball, which possesses only a thin cuticle. Cuticular striae are seen on the under surface. No well-differentiated ectoplasm. (Iron Hem. x 1875.) Figs. 7 and 8.—(Semi-diagrammatic.) Completely extruded. In fig. 7 the neck of the ball is well seen, also the more delicate nature of the ball’s cuticle. No cytoplasmic pad is visible. In fig. 8 the ball is directed slightly upwards. The cuticle is distended round the secreted fluid, showing that this is under pressure. (Paracarm., fig. 7 x 1300; fig. 8 x 1875.) Fig. 9.—Section (5 ») through the point of junction of an associated couple in the first trophic phase. The cytoplasm of the ball individual (A) is denser than that of the other (B). (M.B. E. x 1500.) Fig. 10.—Section (5) through the point of attachment of a mobile Ganymedes toa cell of the host’s gut. The cuticular striz are well seen. The ball is thrust into the host-cell, and contains a fluid that is not clear, the reticular structure being probably due to the coagulation of absorbed food. There isan open passage through the neck into a GANYMEDES ANASPIDIS (NOV. GEN., NOV. sp.). 175 cavity in the body of the parasite. The cytoplasm contains numerous deeply-staining granules. The nucleus of the host-cell (n) is large, darkly-stained, and homogeneous, except for some dark grains. (Iron hem. x 1340). Figs, 11-15.—(Semi-diagrammatic). Cup-ends. Fig. 11.—Cup-end of the Gregarine whose ball-end is shown in fig. 5; (a) is focussed near the upper surface, and shows how the cup is separated from the body by a circular groove; (b) shows the greatest diameter ot the cup. (Iron Hem. x 1875.) Fig. 12.—Cup-end of another Gregarine, to show absence of all differentiation. The ectoplasm is thicker at the end than elsewhere. (Paracarm. x 13500.) ». 13.—-Section of the cup-end of Gregarine a in fig. 9, to show the numerous vacuolar spaces beneath the cuticle. (M. B. E. x 1300.) Figs. 14 and 15.—Sections (5 ») to show the structure of the nucleus in the first trophic phase. (M.B. E. x 1300.) Figs. 16 and 17.—Sections of cysts. In fig. 16 the filamentous inward projections from the cyst-wall can be seen. Small chromatic granules fill up the nucleus ; there is no sign of an achromatic network. The nucleoli are retreating to the interior of the nucleus. (Iron Hem. 5 p X 970.) In fig. 17 the nuclei are ameeboid, filled with a homogeneous sap in which are hollow chromatic spherules, apparently emanating from the nucleoli. The cyst-wall is crumpled, and in one place a flap of it has got detached so that its surface-structure is seen. (Ehrlich’s hem. + eosin 10 » x 800.) Fig. 18a.—Section (5 1) through an associated couple in the second trophic phase. The reticular nature of the cytoplasm is not indicated. (M. B. E. x 610.) Fig. 18b.—The next section in the series. The nucleolus and the outline of the nucleus are given, more highly magnified. The three areas of the nucleolus and their structures are shown (see text). At x a vacuole projects beyond the general surface, and is seen to be colour- less. (M.B.E. x 870.) Fig. 19.—Section of a cyst, to show the alteration in staining reactions of nucleus and nucleolus in this stage (see text). (M.B.E. x 400.) Ki JQ * : = Saal - © 4 - 5 , - a « a = Iuanrt. PurrMicrici. Vol. 55 NS MY, I/ ee a oie Steck Senses Huth Tith? Londow oy ing, THE F@TAL MEMBRANES OF THE VERTEBRATES. 177 The Foetal Membranes of the Vertebrates. AN ADDRESS REPRINTED FROM THE “ PROCEEDINGS OF THE SEVENTH INTERNATIONAL ZOOLOGICAL CONGRESS ” HELD AT Boston, 1907.! By A. A. W. Hubrecht. I was honoured by the request of the Executive Committee to give an address at the first meeting of the Embryological Section of the Seventh International Zoological Congress. I hope that in choosing for my subject the present state of our knowledge concerning the foetal membranes of vertebrates I can avoid the disadvantages of too much special detail, and can at the same time call your attention to the fact that these foetal membranes offer a very wide field for theoretical specu- lation, that may in its turn influence our views concerning certain important phylogenetic problems. The foetal membranes of vertebrates are known to occur in reptiles, birds, and mammals. The embryological hand-books tell us that they are absent in amphibians and fishes. In consequence, a primary subdivision of the vertebrates has been instituted, those with foetal membranes being classed as Amniota allantoidea, those without them as Anamnia anal- lantoidea. From this nomenclature any close observer, even when he is not a zoologist, may safely conclude that one of the foetal membranes carries the name of amnion, the other 1 At Professor Hubrecht’s request this address is here reprinted. It will assist readers in apprehending the conclusions which Professor Hubrecht holds to be rendered probable by the large memoir published in this Journal in November, 1908. VOL. 55, PART 1.—NEW SERIES. 12 178 A. A. W. HUBRECHT. the name of allantois. An older, now more obsolete, sub- division into Achoria and Choriata reveals the presence of a third membrane, the chorion, about which we will have more to say hereafter, and which will explain how this third mem- brane came to fall—so to say—between two stools, when the division into Amniota and Anamnia was established. If we now take into account that neither chorion nor amnion nor allantois was ever detected in fishes or in am- phibians, then we must recognise that the problem, how these foetal membranes of the vertebrates did arise, is one well worthy of full consideration. Up to now attempts to explain their gradual evolution have utterly failed. So, for example, the suggestion of van Beneden and others that the amnion, as a protective membrane, arose in consequence of the early embryo sinking into the yolk-sac, which closed up above it, has long since been abandoned. Also Haeckel’s idea that the allantois arose by a precocious segregation of the urinary bladder of an early amphibian which took the habit of carrying blood-vessels, at a very early stage, to the outer wall of the blastocyst, must be dropped by all who object to predestination in evolutionary processes. Whenever an explanation offers itself which does afford a clue to a more logical sequence of events, it should be preferred. And turning finally to the outer layer, the chorion, who can be satisfied with the lame explanation that the appearance of this membrane is a necessary sequel to the formation of the amnion, which we find inside of it, and which later, in so many orders of mammals, never even arises by folds, which, however, in their turn are necessary to explain the chorion’s appearance ? The subsidiary explanation of all the three embryonic envelopes, which I am going to offer you on this occasion, seems to me to have the great advantage of simplifying matters ; especially in this sense, that henceforth we can link them all three to one simpler and earlier stage (which must have preceded in the Carboniferous and in earlier geological THE F@TAL MEMBRANES OF THE VERTEBRATES. 179 epochs) without having to look for incipient stages of any of them among our present ichthyopsids. Nay, we may even say that of this earlier, archaic starting-point evident traces have been preserved in the teleostomes, the dipnoi, and the amphibians, so that we have to reconsider most seriously whether it will be wise to go on subdividing the vertebrates into the two subdivisions of those that have and those that have not the fcetal envelopes above mentioned. Now let us consider the facts as they present themselves to us, when we want to test the question whether one single original foetal envelope could not after all be at the bottom of the three complicated involucra we have just mentioned. As far as I can see, we are only in need of this one assumption, that an invertebrate ancestor was possessed of what we call an exterior larval layer (such as are not uncommon among different worms, and as we find them, with certain further complications, in some arthropods), to be able to explain how, in their vertebrate descendants, chorion, amnion, and allantois gradually came into being. Part of this hypothetical assumption we see actually realised under our eyes wherever one of the mammals goes through its normal stages of development. We find that the cell-material out of which the embryo is going to be built up is surrounded by an expanded cell-layer, which takes no part whatever in the composition of the future embryo. Here we actually have our single larval layer that will be stripped off later, and that surrounds what are going to be the formative cells. In all mammals it is this very larval layer which will become the outer wall of the blastocyst, what we have above called the chorion. But before following it in its further transformations, we have to ask ourselves, what can be the reason that this outer larval layer, this trophoblast, is so far away from the formative cells of the embryo which adhere to it only at one point? We have only to recall the fact of the pilidium larva, in which, similarly, the distance between the outer layer and the 12$ 180 A. A. W. HUBRECHT. cell-material, which is going to be the new worm, is also very considerable, to remove the objection that in this respect mammals would stand isolated. And we may go one step further and say that it is easy to understand why this con- siderable extension of the outer larval layer has come into existence. When we look back along the line of phylogenetic descent we can imagine that at the period when, for the first time, aquatic animals became inhabitants of the land, four- footed instead of four-finned, and adapted for aérial breathing in addition to their respiration by the aid of gills, it may have been a great advantage to them to become viviparous at the same time, i.e. to keep their developing eggs inside of them, where they are better protected and can be better nourished than outside of the mother. The atmosphere and the dry land offer less favourable conditions for the development of that small amount of protoplasm that forms the primordium of each new being than does the water, and so viviparity is likely to have been a parallel phenomenon to the exchange of the aquatic for the terrestrial existence. We can see clearly that once an embryonic envelope, one cell-layer thick, being present (on our original assumption, as far back as the invertebrate ancestor), that this one-layered larval envelope could obtain high efficiency for the incipient viviparity if only it bulged out as much as possible, thereby— (1) Preventing the egg from passing through the genital ducts rapidly and being deposited, so to say, accidentally. (2) Enabling the egg to adhere in various ways to the maternal tissues, either as a simple mechanical improvement of what was attained (1), or at the same time inducing phago- cytotic attacks on that maternal tissue. (3) Creating the occasion for individual trophoblast cells of this outer layer to absorb fluids either from the uterine cavity or accessory to the phagocytic processes alluded to under (2), and thus accumulating nutritive material inside the blastocyst. Furthermore, it is equally clear that, once the viviparity having been establised, and the surface extension of the THE FQ@TAL MEMBRANES OF THE VERTEBRATES. 181 trophoblast going parallel with it,a yet more efficient mode of nutrition than the one alluded to above under (3) might be obtained if the embryonic vascular system, which was slowly coming into existence on the hereditary plan of development, succeeded in spreading out, in one way or another, on this outer trophoblastic layer, and would enter into osmotic inter- change with maternal blood. Finally, the protection of the embryonic shield during its further development by some sort of appliance resembling a water cushion would, in these incipient viviparous animals, undoubtedly have been a most efficient variation, for the earliest origin of which we have simply to go back to the early stage in which we noticed the formative cells of the embryo adhering to the larval layer, the trophoblast, in one spot only. Suppose that in further development this sessile attachment to have become converted into a circular adhesion —by fluid accumulating between the trophoblast cells and the formative cells, as we see it happen under our eyes in Hrinaceus and Gymnura—we then find that the water-cushion, in casu the amnion, took its origin in a most simple fashion, whereas the chorion is in no way dependent on it, but has preceded it as as earlier formation. The rapid summary here given shows us that the assumption of a single monodermic larval layer is quite far-reaching enough to allow us to understand how, out of it, chorion, amnion, and allantois (the latter as representing one form of early vascularisation of the trophoblast) have gradually come about. The only change we have to make, in what I might designate the present “ fashion” in comparative embryology, is that we look upon the earliest ancestors of mammals not as oviparous, yolk-laden vertebrates, but that we acknowledge them to have been viviparous animals with blastocysts that obtained vesicular shape from quite other motives than an eventual “loss of yolk,” such as Rabl has attempted to prove, Here, then, is the place for an appeal to paleontologists. They haye no shadow of direct interest in foetal envelopes which are 182 A. A. W. HUBRECHT. never met with in the fossil condition! But they may, never- theless, be all the more impartial jadges when we have to choose between two different assumptions: the one given in the hand-books, according to which mammals must, through the Ornithodelphia, be derived from some oviparous sauropsidian ancestor, or the one here advocated, according to which a viviparous Prototrapod, provided with an adhesive and dis- tending larval layer diverged into various directions, some of the descendants utilising the conditions of growth and develop- ment (such as they find them) with the highest degree of intensity and becoming primates, others applying their tropho- blast to nutritive purposes in more diverse and less direct ways, becoming the ancestors of most of our other Mono- delphia and Didelphia. Others, again, going a certain distance with the preceding, but then acquiring yolk-laden eggs (Ornithodelphia), whilst yet other very effective branchings off in various directions gave rise to the primitive sauropsidian ancestors. The difference between the sauropsidian and the amphibian descendants of the protetrapods need no longer be so incisive —as those zoologists that divide the Vertebrates into Amniota and Anamnia would make it. The hypothesis here brought forward proposes to look upon what we know as the Decks- chicht of the early larval Amphibia and Dipnoi, and even of the teleostomes, as a last remnant of the very larval layer from which we started in trying to explain the foetal membranes of vertebrates according to what seems to me a simple plan. We have now to look a little closer into certain details, by which we may be enabled to judge of the greater or smaller degree of tenability of some of the views here brought forward. We notice that all the Mammalia-monodelphia, that have up to now been observed in very early stages, fully confirm the strong antithesis which in those early stages prevails between the trophoblast and the embryonic cells strictiori sensu. We also notice this in the Didelphia, as far at least as Selenka’s figures for the opossum go, although he himself has not interpreted the facts he brought to light in the same THE FETAL MEMBRANES OF THE VERTEBRATES. 183 way asI do. Similarly, Wilson and Hill, in their latest paper on the development of the duck-bill, give us figures of sections which make it probable that the distinction between tropho- blast and formative cells holds good here, even though the development of yolk has obliterated the sharp outlines of the process. Again, in reptiles and birds traces of the larval layer have, in later years, been unmistakably noticed. Schauinsland, Mitsukuri, and Mehnert were among the foremost to contri- bute facts in th}. direction, although at the same time they failed to see the essential points of comparison with the mammals. ‘This failing on their partis all the more explicable as the bird’s egg, which has always served as the prototype even of mammalian development, does not clearly bring out the fundamental distinction that exists between trophoblast and formative matter of the embryo. The gradual obliteration of this distinction may, perhaps, be ascribed to the fact that in these sauropsids, as in the ornithodelphia, a shell has developed, which naturally tends to relegate any outer larval layer to the pension list. Concerning the yolk accumulation in the sauropsidian ege, there is no trouble at all to suppose that the vesicular blasto- cyst of an early viviparous ancestor has gradually become yolk-laden. The contrary assumption, found in the hand- books, that the mammalian egg, while totally losing its yolk, has yet preserved the identical developmental features as the sauropsid, is, in reality, much more difficult to reconcile with sound evolutionary principles. We have seen that a simple clue to our understanding of the complicated foetal envelopes of the sauropsids and the mammialia is the assumption of a simple larval layer, one cell thick, among the invertebrate ancestors. We must be ready to admit that this one factor has un- doubtedly given rise to an endless number of variations and modifications in those innumerable families, genera, and species which have come and have gone, ever since the time when viviparity and terrestrial life became an established fact 184. A, A. W. HUBRECHT. in the vertebrate kingdom. What is preserved to us in the recent fauna inhabiting this planet is only the faintest echo of the multitudinous and protean changes that have, during the course of time, succeeded one another. And it has been our mistake to attempt to co-ordinate the present stages of de- velopment with each other in such a sense that they were expected to represent, in lineary arrangement, the successive evolutionary stages of those foetal envelopes. How false the conclusions may be to which this method may lead us is best exemplified by what is at present often taught concerning, e.g., placentation, a phenomenon in which the outer larval layer, the trophoblast, plays such a prominent part. You will find in the text-books that this was started by what is called the diffuse placentation as it is at present met with in many ungulates, in the lemurs, and in certain Edentates. It is my conviction that this doctrine is utterly false. The diffuse placentation is no placentation at all! The horse and the lemur are, by birthright, aplacental animals, much more so than marsupials, such as Perameles aud Dasyurus, which have hitherto ranked among the Mammalia aplacentalia. And still, by careful comparison of various data, we can soon discover that the diffuse placentation, and that variety of it which is styled the polycotyledonary, far from being archaic or primitive, is, on the contrary, very largely a secondary modification. Among the living Carnivora we find several intermediate stages, not in the sense that these have been phylogenetic transitions, but in that wider sense that these Carnivora demonstrate the possibility how more intricate placentary structures may finally have led up to a diffuse placentation, as that of the horse and the pig, conse- quent upon an increase in the area of surface contact between mother and foetus. What was originally a small surface of intense interchange (Procavia) has then gradually become an extended surface, along which two epithelial layers, one maternal and one foetal, between the blood of the mother and the blood of the embryo, offered no impediment for a sufficient interchange of nutritive matter and of oxygen. THE F@TAL MEMBRANES OF THE VERTEBRATES. 185 If we do not accept the starting-point in the placentation- process to be represented in the ungulate arrangement, a proposal which the systematic position of the Ungulata would in itself render doubtful, we must then look for another phylo- genetic sequence which will help us to rightly interpret that momentous process of placentation. And here the important results of Hill’s investigation of very intense placental pheno- mena in some marsupials, such as Perameles, have great weight. We may fairly conclude that kangaroos, phalangers, oppos- sums and other marsupials have only gradually become aplacentary, parallel to those other formidable changes which must have accompanied the elaboration of that peculiar type which we call our recent Didelphia, in which the dentition, the lactation, and those adaptations of the new-born animals for nutrition during their life inside the marsupium form such distinctive characters. And so if the Didelphia are in reality erratic Monodelphia secondarily modified and with an allantois that has been thrown out of the line of its normal development, with the exception of Perameles, Dasyurus, and in part Phascolarctos, then we have again to look, not amongst them, but amongst the Monodelphia, for such forms that can give us an indication as to what may have been the primitive stage of placentation. And I may here state that my own researches on the placentation of both primates and of insectivores have led me to the conclusion that we should look in quite another direction than the one alluded to above, which starts from diffuse placentation. In the earlier part of this address I have con- sidered those early phylogenetic stages when, in viviparous, air-breathing tetrapods, the larval layer, the trophoblast, found the most diverse possibilities open to it. I believe that those forms of which the embryonic tropho- blast actually attacked the maternal uterine mucosa phago- cytically were the pioneers towards the formation of what has later become the discoid placenta. In some forms, even among our recent mammals, that phagocytic attack is com- 186 A. A. W. HUBRECHT. bined with a penetration of the whole blastocyst inside the maternal tissue, e.g. man, anthropomorphe, hedgehog, Gym- nura, and many rodents. This was naturally a far higher position of vantage than any peculiar fixation inside the lumen of the uterus, for now, when once the blastocyst was encapsuled inside its mother’s tissues, it could be most thoroughly bathed in maternal blood without any extravasa~ tion into the uterine lumen. ‘To take three examples of this we may allude to the guinea-pig, the hedgehog, and man. Still, all these utilise the favourable conditions offered to them, thanks to their situation inside a capsula or decidua capsularis, in a very different manner. here is a most remarkable amount of similarity between the hedgehog and man, as far as the conditions are concerned, which the mother offers to the young. But then the embryo itself of man has seen its way to much more intense utilisation of these favourable conditions than the hedgehog embryo has. Principally because the vascular system of the hedgehog develops in a sequence of stages, which serve to bring its area vasculosa on the umbilical vesicle in primary contact with the profusion of maternal blood by which the blastocyst is surrounded. On the contrary, in man this area vasculosa on the umbilical vesicle is not in contact at all with the maternal circulation. In man it is more devoted to hematopoietic functions, i. e. to the formation of new blood-corpuscles for the embryonic circulation. But in another respect the human blastocyst has got far ahead of that of the hedgehog, in so far as the de- veloping embryo has succeeded in vascularising its outer larval layer, its trophoblast, at a quite exceptionally early moment, without the aid of any allantoic outgrowth, and simply in consequence of a very early segregation of certain portions of the mesoblast, into which the entoderm sends both blood-vessels and blood-corpuscles. This very early vascula- risation of the trophoblast leads to a most intense osmotic interchange between the blood of mother aud child—far more intense that what obtains in the hedgehog, where an ompha- THE FETAL MEMBRANES OF THE VERTEBRATES. 187 loidean placentation precedes an allantoidean one, the allan- tois being a vesicular outgrowth, as it is in so many mammals and in all sauropsids. I cannot refrain from looking upon the vascularisation of the outer larval layer or trophoblast, such as it occurs in man, in the monkeys, and in 'arsius, as the more primitive arrange- ment of the two. And inthat case the presence of a connect- ing stalk (Haftstiel) and the absence of a free allantois in man, monkeys, and Tarsius is not a secondary simplication, but a primary fact of high importance. What is known as the allantois tube inside the so-called Haftstiel or Bauch- stiel of man, monkeys, and T'arsius, is not the remnant of what was once a vesicular allantois, but a remnant of that part of the entoderm which has served towards the vasculari- sation of the trohpoblast. It is this portion of the entodermal surface which will become the free allantois in those other descendants of the primitive tetrapods, which have not adhered to the very direct line of utilising most fully and as early as possible all favourable circumstances. This most direct line leads up straight to the primates. Less direct lines, in which conditions of different or of slower vascularisation have.come to the foreground, are, however, represented in various orders of monodelphian mammals, and further in the Didelphia, the Ornithodelphia, and in the different subclasses of sauropsids. In the latter the allantois has grown to the dignity of a separate foetal membrane, which co-operates to the further ensheathing of the developing embryo, and which carries the blood-vessels for respiratory purposes to the inner surface of the egg-shell, whereas, in the ancestral viviparous forms, the same vessels were more directly distributed over the inner surface of the outer embryonic larval layer, in order to improve the nutritory conditions which had been inaugu- rated by phagocytic action of the trophoblast cells on the maternal tissues. This, then, is a short sketch and a rapid review of how the foetal membranes of the vertebrates may be looked upon if we make certain changes in the interpretations that have 188 A. A. W. HUBRECHT. been hitherto adhered to, but by which latter nobody has as yet succeeded in clearing up the actual phylogenesis of these foetal membranes. Full and extensive investigations of all those numerous genera of mammals that have not yet been examined will, I hope, in due time give us occasion to complete or to modify the views here advocated. It was a great pleasure to me to offer them, tentatively, in an address which I was invited to give in the section of embryology of this Seventh International Congress—a section which, with good right, has been called into life for the first time at this meeting in Boston. Kmbryological problems have been attacked by American investigators with wonder- ful results, and the lucidity of exposition that is characteristic of so many of your embryological workers is only equalled by the beautiful transparency of the eggs of those marine animals on which so many important researches on cell-lineage have been conducted. ‘hat I have been less clear is not only a congenital defect, but is parallel with the utter hopelessness of our expecting that we shall ever be able to follow the cell-lineage in the deeply hidden and exceedingly small mammalianeggs. Still, a full knowledge of that very cell-lineage would be eminently decisive for many of the questions that have occupied us in the course of this address, to which you have listened with so much patience. STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 189 The Structure and Life-History of Crithidia melophagia (Flu), an Endo-parasite of the Sheep-Ked, Melophagus ovinus. By Annie Porter, B.Sc.Lond., Zoological Research Laboratory, University College, London. With Plates 12 and 15 and 15 Text-figures. ConrENTS. PAGE Introduction ‘ 3 : 5 : 190 Material and Methods J : , 190 Distribution of Parasites in the Tose : : 192 Movements ‘ E , : 194 Morphology: Pre- feselbes Stage . : E 197 Flagellate Stage ‘ 197 Post-flagellate Stage in the Beceun of the Host 2 Longitudinal Division ; 202 Hereditary Infection of Melo Attias us ovinus by C. melophagia : ; : 204 Casual Infection . : : ' ; 207 Environmental Effects 3 ; : 208 General Remarks : : ; : 210 Summary : 211 Appendix 208 the Geemmcaes of a Spicoeiete (S. melophagi, n.sp.)in Melophagus ovinus . 213 Appendix I]—Note on a Fungus found in the Mal- pighian Tubules and Intestine of Melophagus ovinus P 214 Appendix IJI.—On the Oeeneencs of an fake coneultn in the Alimentary Canal of Melophagus ovinus, and its Significance in Relation to Crithidia melophagia : : , 216 References to Literature . : 218 Explanation of Plates. : ‘ ; 220 VOL. 55, PART 2.—NEW SERIES. 13 190 ANNIE PORTER. INTRODUCTION. Tue part played by insects as agents in the transmission of the pathogenic organisms of sleeping sickness and other pro- tozoal diseases gives great importance to the investigation of the parasites found within them. It is necessary for any- one seeking developmental stages of pathogenic flagellate Protozoa to have also a first-hand working knowledge of the possible flagellates that may be purely parasites of the insect involved, for certain stages of insect flagellates may resemble possible developmental phases of such organisms as 'l'rypano- somes. Much useful information regarding stages of flagel- lates can be gained from the study of such a parasite as Crithidia melophagia (Flu), occurring in the alimentary tract, ovaries, and ova of the sheep-‘‘ ked,’ Melophagus ovinus. This insect, which is blood-sucking, is also known as the sheep-‘‘ tick” or sheep-‘‘louse.” It belongs really to the Diptera (Hippoboscide), possessing extremely reduced wings. Crithidia melophagia (Flu) was recorded by E. Pfeiffer in 1905, but not named by him. The parasite is of peculiar interest, for I am able to bring forward evidence of a double mode of infection, both hereditary and casual. Swingle (1909) studied the flagellate stages and briefly described infection in the egg of Melophagus. Flu (1908) found parasites in the gut, ovaries, and larva, but was not clear as to the mode of infection (see p. 211). Owing to conditions of environment it was impossible to conduct the whole of this investigation in a large city. Con- sequently the work has entailed travelling, and I have to thank many friends in agricultural centres for their kindly help. MatertaL AND Meruops. Many specimens of Melophagus ovinus were examined during a long period of investigation, but owing to the effective operation of the dip laws in England there was STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 191 much difficulty in obtaining the “keds.’! Indeed, it seems probable that the sheep-ked may soon become almost extinct in England. Those obtained came in very small numbers from many localities in the south of Hngland, namely, Sussex, Hampshire, Kent, Middlesex, and Gloucestershire. I also received a number of keds from different parts of Scotland, but these never contained the Crithidia. Many of the Melophagus, however, were infected by a fungus (see Appendix II). Where fungus was present Crithidia very rarely occurred. I shall show later, from experimental evidence, that this fungus was fatal to the Crithidia (p. 210). Unlike Swingle (1909), who found that practically every Melophagus he examined in Nebraska was infected with Crithidia, I found that this was very far from being the case. Much depended on the locality from which the Melo- phagus was obtained. The more heavily infected individuals came from the southern districts of England. Often entire stocks of keds from one locality proved to be uninfected. Again, it was impossible to keep keds alive more than three days after their removal from the sheep. Both young and adult Melophagus and many puparia in all stages of development were carefully examined. Raising puparia naturally upon a sheep was tried, but was not an easy matter, and as one could not be sure of having infected keds, there was always a percentage of uninfected puparia. For observations of the living organism two methods of pro- cedure were followed. ‘The alimentary canal was isolated and divided into cesophageal, crop, stomach, intestinal and rectal portions, which were separated one from another. These were either teased with needles, mounted in 0°75 per cent. salt solution, and covered, the cover-slip being carefully vaselined, or the contents of the isolated portions of the gut were expelled by gentle pressure, and these only were examined, being mounted as before. Alkaline methylene ‘In this paper I shall frequently use the term “ked” to denote Melophagus ovinus. 192 ANNIE PORTER. blue and neutral red were occasionally used as intra-vitam stains and were sometimes useful. For fresh preparations used in work on hereditary infection, the ovaries and gut were dissected out very carefully, kept as far as possible relatively in situ, and mounted in 0°75 per cent. NaCl solution. The behaviour of the Crithidia visible through the walls of the gut and their action when they passed out from it were then most carefully watched. I have attached very great importance to the study of the living organism in all its phases. For making permanent preparations the alimentary tract of the Dipteran host was carefully removed and divided into portions as before. ‘These isolated portions were usually teased very finely and fixed wet. Formalin vapour and osmic acid vapour were chiefly used for instantaneous fixation of the hanging-drop preparations, which were then spread. The preparations were subsequently treated with methyl or ethyl alcohol. Corrosive-acetic-alcohol (Schaudinn’s fluid) and Bouin’s fluid (sightly modified and containing a little alcohol) were also used for fixation. Various stains were employed. Giemsa’s stain gave some pretty results; thionin acted rapidly and well; iron-hema- toxylin, carefully differentiated with iron-alum, was very serviceable ; while gentian violet and Delafield’s hematoxylin were of great use, particularly in obtaining details of the membrane and flagellum. In the investigation of Crithidia melophagia, as in all other flagellates on which I have worked, I found that pre- parations mounted in neutral Canada balsam were superior to dry films or to films mounted in any other media. Preparations of ovaries, eggs, and puparia were treated similarly. Special methods adopted are detailed in the section dealing with hereditary infection (p. 204). DISTRIBUTION OF THE PARASITE IN THE Host. The Crithidia parasitic in the alimentary canal of Melo- phagus are often mixed with the blood obtained by the ked STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 193 from the sheep. This blood from the sheep in the cesophagus, crop, and anterior part of the stomach of Melophagus is always fluid, and of an extremely bright red colour. That in the remaining part of the stomach is duller red but fluid, and in the intestine the blood, now semi-digested, is always darker in hue, sometimes brownish or greenish, while in the extreme rectum it is black. ‘The enhanced red colour in the anterior portions of the alimentary canal has been shown experimentally to be associated apparently with the presence of an anti-coagulin in the digestive tract of the sheep-ked (see Appendix III). Crithidia can be found throughout the length of the alimentary canal of Melophagus ovinus. In the anterior parts of the canal they are small, rounded, non-flagellated forms, which, when they come in contact with the blood, rapidly develop and divide, the products of division becoming the typical flagellates found throughout the rest of the canal. The parasites, after this rapid development, pass backwards towards the partly digested blood, which would appear to be a medium more suited to their requirements. In the posterior third of the stomach there are large numbers of young flagellates which form great aggregation rosettes (Pl. 12, fig. 45) and clumps, while true division rosettes are also present (Pl. 12, fig. 56). In the intestine the: same holds good. When many Crithidia are present in a ked, they usually swarm in the fore-part of the intestine. Repeated division occurs in the intestine, so that small flagellates are found in the rectum. Most of these attach themselves to the gut-wall or to débris and encyst, the resting (post-flagellate) stage of the parasite then being found on the walls of the rectum and in the feeces. The ovaries and ova serve as places in which a kind of post-flagellate development occurs, the ova being penetrated by flagellate forms of Crithidia, which rapidly lose their flagella and ultimately round themselves off, and pass through a resting stage (Pl. 15, figs. 57-94). 194 ANNIE PORTER. The Malpighian tubules of Melophagus ovinus are sometimes invaded by Crithidia melophagia, but this is not common. Parasites were more numerous in female than in male keds. Repeated investigation of sheep’s blood failed to show the presence of any flagellate therein. Flu and Swingle obtained similar results. C. melophagia is, then, purely a parasite of Melophagus ovinus, MoveMENTs. The movements of C. melophagia are very vigorous. The parasites are even more active than C. gerridis (see Porter [1909], p. 352). As in C. gerridis, the membrane takes an important share in locomotion, but the movements of the body of C. melophagia are not so noticeable as in the parasite of the water-bug. When C. melophagia was examined under the water immersion (2°5 mm.) objective, the movements of the less active organisms could be analysed. In progression the organism moves with its flagellum foremost, and the latter executes vigorous, slightly spiral, boring movements. The body also aids in progression, for waves pass from the posterior end towards the flagellum, causing a series of peristaltic-like swellings. The body of the parasite seems to become shorter during this period, and then by relaxing to move forwards. The bead-like swellings due to undulatory movements are more noticeable in certain areas, and in the hving organism myonemes could be sometimes seen both on the body and in the membrane in these regions. Flu has also figured myonemes on some of the parasites he drew, and observation of them in life confirms his work, but it was with the greatest difficulty that I could find myonemes in stained specimens (Pl. 12, figs. 17, 18, 40, 42, 45). The body of C. melophagia, compared with that of C. gerridis, is relatively rigid, but slight twisting movements do occur. The previous workers on C. melophagia are agreed STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA, 195 as to this rigidity. The anterior end, to which the flagellum and undulating membrane is attached, is naturally more flexible than the posterior end, and its movements are more marked. Movements of contraction of the posterior end of the body of C. melophagia result ina temporary concentration of the protoplasm around the nucleus of the organism. The body then resembles a short, ‘hick pear, drawn out at its anterior end into a long, narrow stalk. Sometimes the body remains in this condition, which is fairly common in forms about to encyst, and in such forms withdrawal or degeneration of the flagellum, followed by the secretion of a thin gelatinous wall, completes the encystment. In other parasites from the stomach, where no encystment occurs, this concentration of the protoplasm in the nuclear region is not so marked, and when relaxation occurs the organism is propelled forward with a very slight jerk, and repetition of the contraction follows, as has been before described. ‘I'he jerking is never so marked as in Herpetomonas, for the membrane has the effect of producing smoothness of motion. Reversal of the direction of motion occurs and is very rapid. ‘he flagellum swings out, describing a semi-circle, of which the body acts as the diameter for an instant, but the force of the movement of the flagellum is so great that the body also swings outwards in a line with the flagellum, and the organism moves away, not exactly in the same course as before, but in one at a very small angle to it. The path of the organism is frequently parabolic in nature. Many peculiar movements can be observed when C. melo- phagia is endeavouring to free itself from débris in the lumen of the gut. Much writhing, both of the flagellum and body of such a parasite, is then seen, and the organism often swings round and round, the point of attachment serving as the centre of rotation. If the posterior end should be attached, the flagellum executes violent lashings and spiral movements, these latter not being, as a rule, very noticeable in the normal organism. 196 ANNIE PORTER. Occasionally I have seen the flagellum and membrane of specimens of C. melophagia torn away from the body, and for a few seconds after, the flagellum executed intermittent flickers or lashing movements before it finally became still. Ageregation-rosettes (Pl. 12, figs. 41,45; Pl. 15, figs. 95, 96) are common in C. melophagia. Rosettes seem to move fairly as a whole, and I have watched them rotate rather quickly. Each individual of such a rosette is attached by its flagellum to débris, usually epithelial in nature, and moves up and down in a slightly inclined plane. In division the movements of the daughter organisms are very noticeable. I will defer the description of their motion until division is discussed. During encystiment in the rectum of the host, which occurs with some of the parasites, movement of the nucleus towards the flagellar end of the organism occurred. I have also seen the migration of the nucleus from the mid-region of the body to near the flagellum during periods of violent movement of the latter organella. I have never seen migration of the blepharoplast in living organisms under similar conditions, though it may occur at times, since blepharoplasts can occa- sionally be found in the post-nuclear region (PI. 12, figs. 40, 42), as well as by the side of the nucleus (PI. 12, fig. 33) in different stained specimens. By far the commonest position for the blepharoplast is the pre-nuclear one. ‘The other movements occurring during encystment will be described in the section of the paper dealing with that subject (see p. 200 and text-figures 1-10). MorpuHo.oey. The life-cycle of Crithidia melophagia may be con- veniently divided into three stages, which gradually merge into one another. They are—the pre-flagellate, flagellate, and post-flagellate stages. The morphology of these forms may now be described. STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA, 197 The Pre-flagellate Stage. The early pre-flagellate stages of C. melophagia are more or less oval or rounded bodies (PI. 12, figs. 1-6), varying from 4°54 to 6 long, and from ly to 45 broad. They are most abundant in the fore-gut of young Melophagus, but the pre-flagellate stage is passed through with great rapidity and is easily missed, ‘This probably accounts for the very brief references to these small forms by Flu and Swingle. The protoplasm of the pre-flagellate forms is very finely granular (Pl. 12, figs. 1-5). ‘he nucleus is usually round and not quite central in position (Pl. 12, figs. 1, 9-12). The bar-like blepharoplast (kinetonucleus) is very deeply staining, and lies either below (Pl. 12, figs. 2, 10) or to one side of the nucleus (Pl. 12, figs. 1, 6). A chromatophile area with its chromatin in a very diffuse condition is sometimes fairly prominent, and from this a fine thread arises, which grows outwards, forming the flagellum (PI. 12, figs. 9, 10), and appearing to draw out the end of the body with it (PI. 12, figs. 11-13), while the periplast of the body forms the mem- brane (Pl. 12, figs. 14-20). The posterior end elongates at the same time (figs. 16-18) and the flagellate form (PI. 12, figs. 19, 20) is assumed. This.development is in accord with that of C. gerridis and C. tabani, and I have watched these processes in living specimens of both C. gerridis and C. melophagia. Division of pre-flagellate forms can occur before the develop- ment of the flagella (Pl. 12, figs. 5, 4). 'his will be described in the section dealing with division. The Flagellate Stage. The mature flagellates vary very much in size, the variation being due to division and growth. Very large forms (PI. 12, figs. 44, 45) may beas much as 50 uw to 754 long, this measure- ment including the flagellum,’ while short forms just flagel- 1 Tt is almost impossible to differentiate between the limiting areas of the body, the membrane and the free flagellum of C. melophagia, as so much variation occurs in different specimens. 198 ANNIE PORTER. lated (Pl. 12, figs. 18, 19) in the crop, or the small forms produced by division prior to encystment (PI. 12, figs. 20, 21; 99) are very much smaller (12 to 20, long). The breadth of the flagellates varies from 1°5 x to 2°8 yu. The protoplasm of C. melophagia is very slightly alveolar or almost hyaline, differing therein from the more alveolar protoplasm of C. gerridis. ‘There is no suggestion of large permanent vacuoles or of a cyto-pharynx. Occasionally the protoplasm is more granular at the posterior end (PI. 12, figs. 30, 34) and slight alveolation occurs there. At the anterior end, near the origin of the flagellum, the remains of the chromatic area, from which the flagellum arose, sometimes persist. The nucleus (trophonucleus) of C. melophagia is oval (Pl. 12, figs. 21-24) or rounded (figs. 26, 50, 32) and some- what vesicular. There is a fair amount of chromatin present, which may consist of a number of very fine granules, evenly distributed (fig. 32), or the chromatin may be concentrated into about eight masses (fig. 44), or, as is often the case, the chromatin is present in the form of bars (figs. 25-29), which sometimes extend across the whole breadth of the nucleus (figs. 34-37), less frequently across part of its breadth (figs. 24, 42), or in an even more rare condition dots and bars occur in the nucleus of the same organism (figs. 30, 59). In certain cases the chromatin ot the nucleus may be con- centrated into a central mass (fig. 25). The nuclear membrane is fairly distinct in most of the specimens I have examined. | think that such a membrane must be present to keep together the nuclear material during the migrations of the nucleus seen during life. The blepharoplast (kinetonucleus) of C. melophagia is very evident in a stained preparation, for it colours deeply whatever stain be employed. Like the nucleus, it can also be seen in life as a small bright refractile bar. In some cases it is slightly bowed or curved (Pl. 12, fig. 32), or oval (Pl. 12, fig. 34). It is dumb-bell-shaped in forms about to divide (Pl. 12, fig. 44). The blepharoplast, which is typically STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 199 rod-like, usually lies transversely across the organism (PI. 12, figs. 21-28). It is exceptional to find it in any position other than anterior to the nucleus, though on a few occasions the blepharoplast was at the posterior end of the body (PI. 12, figs. 46, 42), but in these cases the flagellum originated in a pre-nuclear position. As a rule the blepharoplast shows no differentiation of structure (PI. 12, figs. 21-39), but sometimes in dividing forms, in which the blepharoplast is dumb-bell shaped, there seems to be a concentration of chromatin in the ends of the dumb-bell (Pl. 12, figs. 40, 44, 45). A clear area (PI. 12, fig. 51) is often present around the blepharoplast. Chromidia are present, scattered in the general proto- plasm (Pl. 12, figs. 25, 37, 39). They stain in the same way as the nucleus, aud less densely than the blepharoplast. The occurrence of such chromatoid granules at division (Pl. 12, fig. 45) suggests that they have been given off from the nucleus into the general protoplasm, and exercise some controlling influence over the same. The undulating membrane and the flagellum.— The flagellum originates from a chromatic area in the pre- flagellate form, and is attached to the body by a narrow mem- brane (PI. 12, figs. 21-46), which is a periplastic outgrowth of the anterior end of the body. There is but one flagellum in any single, undividing individual (Pl. 12, figs. 21-39). The flagellum is thick, but gets thinner towards its free end (Pl. 12, figs. 40, 45). At times it appears to show very fine transverse striations. In stained specimens the membrane sometimes shows myo- nemes (PI. 12, figs. 39, 42, 45), though, curiously enough, the myonemes were much more obvious in some of the living specimens that I examined. Flu described myonemes in C. melophagia, but figured the myonemes as accompanying a central spindle. ‘This latter feature I have never seen. A basal granule (blepharoplast of Minchin) is often present (Pl. 12, figs. 17, 27, 33, 42, 45) between the point of origin of the flagellum and the blepharoplast (kinetonucleus). 200 ANNIE PORTER. The Post-flagellate Stage of C. melophagia in th Rectum of Melophagus ovinus. The preparation of Crithidia melophagia for life outside the body of the host occurs in the rectum of the sheep-ked. Large numbers of small flagellates (Pl. 12, figs. 27-29) are present in the hind gut, also some forms in process of division (Pl. 12, figs. 97, 98). The small forms attach themselves to the wall of the rectum and encyst there, but encystment can be watched when the rectal contents are expressed on to a slide and examined under the microscope. ‘The flagellate (text-fig. 1) at first executes violent lashing movements with its fagellum, and during this motion migration of the nucleus nearer the tlagellar end of the organism frequently occurs (text-fio. 2). At the same time the body of the Crithidia shortens and thickens (text-figs. 3, 4; Pl. 13, fig. 100), waves of contraction passing rhythmically down the body, which gradually may become somewhat triangular (text-fig. 5; Pl. 13, fig. 101). The flagellum meanwhile shortens (text-figs. 5, 6), and the organism may bend on itself (text-figs. 6,7) during this period. Concentration of the protoplasm occurs, the flagellum becomes less wavy (text-fig. 7), and, little by little, it contracts nearer the body (text-figs. 8,9; Pl. 13, figs. 102- 106) and is withdrawn, the parasite becoming oval (text-fig. 10; Pl. 18, figs. 109-112). The organism at this time becomes surrounded by a thin layer of refractile, gelatinous substance, which rapidly hardens to form a closely adherent resistant cyst-wall. The oval bodies (Pl. 15, figs. 109-114) so produced are post-flagellate forms, which become detached from the walls of the rectum, and pass out with the feeces of the ked, from which feces they can be recovered. These cysts, which measure from 2°5 «7 to 5°5 uw by 15 to 3, serve for the infection of other Melophagus ovinus. All Crithidia melophagia do not go through a post- flagellate stage in the gut of their host. Some, after passing a portion of their existence as flagellates in the gut of the ked, STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA, 201 pierce the walls of the alimentary tract and make their way to the ovaries of the ked, where their development is con- tinued. TEXT-FIGURES 1-10. 9 10. Encystment of Crithidia melophagia in the rectum. Text- figs. 1-5.—Parasite rounding off and flagellum disappearing. Text- figs. 6-7.—Show bending of | parasite on “itself. Text- figs. 8-10.—Final stages in loss of flagellum and assumption of typical eye form. Swingle (1909, p. 104) has described thick-walled cysts. I have but rarely seen the thick-walled forms (PI. 13, fig. 114), most of the cysts found being thin-walled. 202 ANNIE PORTER. LONGITUDINAL DIvIsIon. The longitudinal division of the living organism has been frequently watched. While the movements of the dividing flagellates are noticeable, those of the smaller dividing pre- flagellates are far less marked. When a flagellate is about to divide, the protoplasm of the posterior end concentrates somewhat in the nuclear region, and the organism appears to shorten. The protoplasm migrates from the centre of the parasite towards the sides, so that a comparatively clear area is left at the centre (PI. 12, fiz. 46). The greatest change at this stage is seen in the blepharoplast and flagellum. The blepharoplast becomes slightly dumb-bell-shaped (Pl. 12, figs. 44, 45) and gradually constricts into two (Pl. 12, fig. 46). The flagellum splits rapidly at the body end (PI. 12, fig. 46), and then, more slowly, the halves become free. The nucleus meanwhile becomes slightly indented in the median line (PI, 12, fig. 46) and then gradually constricts into two, the halves migrating to the periphery (PI. 12, fig. 47). During this nuclear division the daughter-flagella execute very vigorous lashing movements, and a constriction appears at the flagellar end of the parent organism. A split appears at this end (Pl. 12, figs. 47-49), and, at the same time, vacuoles in the clear median area fuse, and thus the extension of the split is facilitated. The daughter- organisms rapidly separate from one another, their appear- ance at times being suggestive of diverging curved calipers (Pl. 12, figs. 51, 52). At length the two are practically in a straight line (Pl. 12, figs. 53-55), in which condition they remain for a short time and then finally separate. The division of the pre-flagellate forms is initiated by the division of the blepharoplast, and is followed by the division of the nucleus and the appearance of vacuoles. A slight split appears at one end (PI. 12, fig. 3), and the organism remains in this condition until the flagellum of each half has partly grown, when final separation 1s effected by their movements. Sometimes repeated division of a pre-flagellate form occurs STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA, 203 and a rosette (Pl. 12, fig. 4) is produced, but the rapidity of the process of formation of flagella causes short duration of the rosette stage. On the other hand, repeated longitudinal division of flagellated individuals occurs, and as the individuals so produced do not separate immediately, rosettes (PI. 12, fig. 56) are formed. In division, the posterior ends of the daughter-organisms are the last parts to separate. As the daughter-forms remain in proximity and themselves proceed to divide with rapidity, true division-rosettes are formed, in which the posterior ends of the organisms are central, while the flagella radiate out from the common centre. Such division-rosettes (PI. 12, fig. 56) differ from the aggregation- rosettes (Pl. 12, figs. 41, 43; Pl. 15, figs. 95, 96) where the organisms become attached by their flagella. The distinction between the two forms of rosettes has not been shown by previous workers on C. melophagia. Longitudinal division results in the formation of both equal and sub-equal daughter forms. While the occurrence of equal longitudinal fission is the commoner (Pl. 12, figs. 50, 54, 55), I have seen cases of marked inequality in the size of the daughter-parasites, the one being very thin and narrow, the other considerably broader and thicker (Pl. 12, figs. 51, 53). As the entire process of sub-equal division has been watched in living organisms, there is no possibility of it beimg mistaken for anything else. The polymorphism resultant on division is strongly against the idea that there are sexual forms of Crithidia, and I have never seen the slightest indication that there is sexual dimorphism, in C. melophagia, C. gerridis, Herpetomonas jaculum, H. musce domes- tice, H. culicis, and a new Herpetomonas from Vespa crabro, all of which [I have examined in the living con- dition (see Porter [1909] on C. gerridis and H. jaculum). Division, usually twice repeated, is found to occur in parasites destined to encyst, and the resultant forms are very small. The first division is of the usual flagellate type (Pl. 13, fig. 97). The process of the second division rather 204 ANNIE PORTER. resembles that of the pre-flagellate stages, for before it is accomplished the flagella have almost disappeared. Some- times no flagellum is visible at all, and the parasites look hike dividing cysts. On rare occasions the posterior end of a flagellate has divided before the anterior end (PI. 15, fig. 98). Tue Herepirary InFectioN OF MELOPHAGUS OVINUS BY CRITHIDIA MELOPHAGIA. Casual infection of Melophagus ovinus by the ingestion of post-flagellate cysts of Crithidia melophagia is fairly easily observed. The development of the parasite in the egg can only be studied with difficulty. I now wish to give a fuller account than exists up to the present of the processes leading up to the birth of Melophagus infected with Crithidia melophagia. The first point to be determined was the way in which the Crithidia reached the egg. Infected Melophagus were carefully dissected so that no rupture of the gut was made. The ovaries also were dissected out and kept as far as possible in the position beside the gut that they occupied in life. Crithidia could be seen through the gut-wall moving actively about. Suddenly they concentrated in one place and soon began to pass through the wall, their posterior (blunt) end first. They rapidly swam direct to the ovaries and penetrated them in the same way, that is, with the non- flagellar end first. The flagellum was very rarely used as a boring organ to allow of the passage of the organism. Penetration of the ovaries of their host by the parasites occurs in other cases, e.g. C. gerridis, H. jaculum, but the ova are apparently unattacked and the flagellates simply degenerate. But in the case of C. melophagia the organisms (Pl. 13, figs. 57, 59) make their way rapidly to the ova, to which they cling, whether the ova are mature or immature. In some cases one Crithidia only enters the egg (Pl. 13, fig. 58); at other times several penetrate it at once. In STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 205 penetration the blunt end of the flagellate enters the egg first. Occasionally the flagella are cast off as the Crithidia pass into the egg and remain on the outside. | In the case of older ova, the parasites seem to penetrate the egg at a definite spot (Pl. 13, fig. 58), which probably becomes the mouth of the embryo. Parasites invading older embryos enter by the embryonic mouth. Like Swingle I did not find parasites in the milk-glands or milk of Melo- phagus. In investigations of the stages of C. melophagia in the egg and puparia I found that smear preparations were preferable to sections. Greater rapidity of manipula- tion and thinner preparations could be obtained by this means. The method adopted was to prick the egg or open the young puparium and express the contents on to a slide. The contents were at once fixed and then were allowed to flow over the slide, so that no artificial spreading was required, and therefore no mechanical distortion or tearing of the parasites could occur. The preparations so made contained much fatty matter. The slides were treated with ether to remove the fat, and then after washing with absolute alcohol were stained and mounted in the usual manner. Once within the egg the parasite gradually loses its flagel- lum (PI. 13, figs. 61-63). This may be cast off entire, for flagella are found floating freely in the vitellus of eggs that had been treated with the utmost care in the manner pre- viously detailed. In many cases the flagellum appears to be gradually absorbed (PI. 13, figs. 64, 66). Longitudinal division of the flagellates in the egg may occur, though rarely. The protoplasm of the Crithidia then concentrates round the nucleus and blepharoplast (PI. 13, figs. 64-69) and the para- site gradually becomes more or less rounded (PI. 13, figs. 70- 73). Multiple division of both nucleus and blepharoplast then occurs (PI. 13, figs. 74-77), and the daughter-blepharo- plasts appear to pass outwards towards the periphery (PI. 13, VOL. 55, PART 2,—NEW SERIES. 14 206 ANNIE PORTER. figs. 76, 77). A “plasmodial”! form (PI. 13, figs. 75, 77) is thus assumed. The protoplasm collects around the nuclei, and gradually fragmentation of the ‘ plasmodium ” occurs, the result being the formation of a number of small bodies, which rapidly round off, forming definite resting bodies (PI. 15, figs. 78-81). Sometimes these resting bodies remain in proximity to one another, so forming groups (Pl. 13, figs. 80, 81). The parasites now measure only l'5 to 4m long and ly to 2°54 broad. Sometimes one chromatic mass (Pl. 13, fig. 82) only can be distinguished. Often both nucleus and blepharoplast (PI. 13, figs. 80, 81, 83, 84) are present. As the embryo grows the rounded forms of the parasite in the stomach (which is the chief cavity within the young Melo- phagus) also grow (PI. 13, figs. 82-84). The Crithidia then undergo multiple division, small rosettes (Pl. 13, figs. 85-88), analogous to pre-flagellate rosettes, being produced. The division clusters may separate, giving rise to small, pear- shaped or ovoid individuals (PI. 13, figs. 89-94), or they may remain as a rosette (Pl. 13, fig. 88) for some time. Whether the Crithidia remain as groups or become isolated as oval non-flagellated bodies, they undergo no further development for a considerable period. In fact, when the young Melo- phagus is hatched, a month after extrusion of the puparium, there is still no further differentiation in the parasite. Freshly hatched Melophagus do not contain the fully developed flagellates, but the rounded or pear-shaped pre- flagellate forms (Pl. 13, figs. 92-94) and rosettes (Pl. 13, fig. 88) may be present. The parasites appear to lie dormant for a day or two, during which time the young insect does not appear to suck blood. Soon after the first meal of blood is taken, rapid development of the pre-flagellate forms occurs, and the adult flagellate form of the Crithidia is quickly assumed. ‘A plasmodium is really a multinucleate mass of protoplasm formed by fusion of small amcebe. However, the term is sometimes used, as in describing certain Haplosporidia, for a multinucleate mass of protoplasm formed by division. STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 207 Casual INFECTION. The method of cross-infection in many species of Crithidia has not been demonstrated, but in the cases known the casual or contaminative method seems to prevail. The post-flagellate stages of Crithidia gerridis and C. tabani are known, and the cysts of these parasites are shed in the feces of the insectan hosts. ‘The crithidian cysts are swallowed by new hosts when they feed on material accidentally contaminated by the faces of their neighbours. The cysts then develop in the alimentary tracts of the new hosts. Melophagus ovinus also becomes infected with its Crithidia by the casual method. When studying C. melophagia I have noticed that the feeces of Melophagus ovinus are voided near spots on the sheep from which blood has recently been sucked (particularly is this the case at times of extrusion of puparia); that the feeces contain crithidian, post-flagellate cysts, and sometimes active flagellates; and that other Melophagus, feeding at the same spot, have thrust their proboscides into the semi- fluid faeces to reach the blood of the sheep. Ingestion of cysts under such circumstances is easy. The ingestion of feces has been seen particularly well when a number of keds have been kept confined to a small area of the sheep’s body. At shearing a slight injury was caused to one sheep, and the keds seemed to collect round the small bleeding patch. Their habits were carefully observed then, and were similar to those described above. I do not agree with Swingle that casual infection is only a remote possibility ; to my mind itis a certainty. A modified contaminative cross-infection is rendered pos- sible by the cannibalistic habit of Melophagus ovinus. The keds have been seen to attack one another, the point of seizure invariably being at the end of the abdomen near the anus. When a ked so attacked has been freed from its aggressor and then dissected, I have found that the abdominal cavity was almost empty, the viscera having been devoured 208 ANNIE PORTER. by the attacking ked. By this cannibalistic habit it is possible for Melophagus ovinus to acquire practically every stage of Crithidia melophagia direct, and this is probably a subsidiary method of spreading the parasite. ENVIRONMENTAL EFFECTS. Crithidia melophagia shows less response to slight changes of environment than does C. gerridis or Her- petomonas jaculum, both of which I have studied. Never- theless, under certain conditions remarkable effects have been produced by relatively simple means, and these may now be recorded. (1) Response to light.—lIncreased intensity of white light produces increased velocity of movement of Crithidia melophagia. Green light somewhat retards the movements of the organism. This is also the case with Herpetomonas jaculum. Intense light causes aggregation-rosettes of C. melo- phagia to separate. C.melophagia lives very much longer in diffuse light than in bright light. (2) Responseto changes of temperature.—C. melo- phagia can live at a temperature just below that of the blood of the sheep, but the flagellates are killed at a tem- perature above 40° C. At room temperature (15° C.) the parasites will live for several hours. (3) Response to change of medium.—Though the flagellates normally live surrounded by fluid blood (a diseus- sion of which will be given in Appendix III), yet they can live in other media and can resist the effects of such media to varying degrees. (a) Tap-water when added to the parasites in the gut- liquid seemed to have little effect. Though the movements of the flagellate became slightly more active, this was possibly STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 209 due to the greater space in which the parasites could move, the débris being distributed over « greater area than before. (6) 0°75 per cent. NaCl solution increased the activity of the parasites. Five parts of tap-water added to one part of 0°75 per cent. NaCl solution containing Crithidia caused the flagellates to move more rapidly, the spiral boring movements of the flagellum becoming more exaggerated. (c) Caustic potash.—T'wo per cent. solution killed all the Crithidia within a minute; 1 per cent. potash solution killed them in from seven to twelve minutes, but their bodies were not dissolved, this pointing to the chitinoid nature of the thin periplast or ectoplasm. (d) Acetic acid.—One third per cent. aqueous solution had the effect of swelling the parasites, which then died. (e) Grape-sugar.—A most remarkable effect was that produced on C. melophagia by a solution containing a very small amount of grape-sugar. When this was added to the parasites they commenced to divide very rapidly, and many divisions occurred. ‘To ascertain if there were a connection between this division and the occurrence of sugar in the natural medium of the parasites, some experiments were made. ‘The results were as follows: (i) Sheep-serum contains a very small amount of grape- sugar. (ii) The liquid obtained when wool cut from the sheep was ‘boiled with water and then concentrated also showed traces of sugar. There were, then, these two sources from which the ked probably could obtain minute quantities of sugar. It is possible that the traces of sugar may take a small share in stimulating division of C. melophagia, which goes on more rapidly in the stomach of the ked than elsewhere. (f) Fresh blood (human or sheep’s) added to a pre- paration of living Crithidia caused the parasites to move away to areas where the blood was somewhat less concentrated, where they proceeded to divide rapidly. 210 ANNIE PORTER. (gy) Dilute glycerine killed C. melophagia almost at once. Vaseline had the same effect after a very short time. (4) Effect of a parasitic fungus of Melophagus ovinus on C. melophagia.—The presence of a fungus in Melophagus ovinus has already been mentioned. As I very rarely found the fungus and Crithidia co-existing in a ked, it was deemed advisable to find out any possible inter- relation of the two parasites. The Malpighian tubules of the ked—often blocked with fungus—were the most heavily infected organs. Fungus taken from the Malpighian tubes was crushed with a little water. The emulsion, which probably contained an enzyme, was added to a preparation of actively moving C. melophagia. ‘he movements of the flagellates slowed at once, their protoplasm became much more vacuolated, and the parasites appeared to burst. After seven to nine minutes no living Crithidia were to be seen. The fungus-infected Melophagus ovinus seems widely distributed. Specimens from Scotland were practically always heavily infected with it, and some keds from each locality tried in England also were infected. These keds very rarely contained Crithidia. The fungus seems to have a pathogenic action upon the flagellate, and I believe that the co-existence of the fungus and Crithidia for long together is almost impossible. GENERAL REMARKS. Regarding the previous work done on the genus Crithidia, I have already noted most of the memoirs dealing with the subject in my paper on Crithidia gerridis (1909). Conse- quently the remarks now appended relate especially to the flagellate of Melophagus ovinus. E. Pfeiffer (1905) first briefly described a flagellate as occurring in the gut of Melophagus ovinus. He mentions that L. Pfeiffer had seen and recorded the parasite in 1895. The flagellate stage only was described, and no definite name was given to the organism, which was stated to be “like a trypanosome.” STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 211 P. C. Flu (1908) published an account of the flagellate under the name of Crithidia melophagia. Flu stated that he saw parasites in the ovary of Melophagus, and small forms in the larva, but was unable to determine the mode of infection of the host. L. D. Swingle (1909), working in Nebraska, wrote a description of the parasite, calling it C. melophagi. From a private communication I learn that Swingle’s work was completed, but not published, before Flu’s paper appeared, thus accounting for the specific name melophagi (described as new), which cannot stand. ‘he chief value of Swingle’s work lies in the fact that he described rounded and *‘»nlasmodial”’ stages of the parasite as occurring in the egg of the host. While Swingle was working in Nebraska, I was investiga- ting the parasite independently in England. It gives me great pleasure to be able to confirm Swingle’s work, and to add many more details concerning the modes of infection of the parasite and its general life-history. SUMMARY. (1) Crithidia melophagia is a flagellate occurring in the alimentary tract, ovaries, ova, and puparia of Melo- phagus ovinus. (2) The parasite has three stages in its existence, a pre- flagellate stage (PI. 12, figs. 1-20), passed chiefly in the crop and fore-gut of the insect host, a flagellate stage (PI. 12, figs. 21-44), occurring chiefly in the posterior two thirds of the gut, and a post-flagellate stage, occurring either in the rectum and feces (Pl. 13, figs. 97-114) or in the ova and pupe (Pl. 13, figs. 57-94). (3) ‘The pre-flagellate stage is passed through very rapidly. These parasites are small, usually oval bodies, lu to 4°5 uw by 4°5 uw to 6 w, with round nuclei and bar-lke blepharoplasts. The flagellum arises near the blepharoplast from a chromato- phile area. Division of pre-flagellates may occur (Pl. 12, fig. 4). 212 ANNIE PORTER. (4) The flagellate forms are from 12 » to 75 long, and 15 w to 2°8 « broad (including the flagellum). The general protoplasm is slightly alveolar. ‘The nucleus is vesicular. he blepharoplast is well marked, rod-like, usually anterior to the nucleus, and generally homogeneous. Chromidia may occur as isolated granules. (0) The undulating membrane and flagellum are well marked. ‘here are indications of myonemes (PI. 12, figs. 40, 45) in some stained specimens, but the myonemes are more evident in some living specimens. The membrane is of great use in securing smoothness of motion. ‘lhe flagellum is long and forms a chromatic edge to the membrane. A _ basal granule may occur near the root of the flagellum. (6) The post-flagellate stage in the host’s rectum (Pl. 13, figs. 97-114) gives rise to resistant (resting) bodies that are passed out as cysts with the feces and serve to infect new hosts. ‘The cysts measure, on the average, 4 u by 2°5 uw. The flagellates divide, usually twice, and the four small forms thus produced lose their flagella, become round, and then invested with a thin gelatinous wall. (7) The post-flagellate stages in the ova and puparia of Melophagus (Pl. 15, figs. 57-94) serve for the hereditary transmission of C. melophagia. ‘The flagellates pass through the wall of the gut near the anterior ends of the ovaries, swarm towards and enter the ovaries and penetrate the ova—the posterior (aflagellar) end of the parasite being used in penetration. Within the ova each parasite loses its flagellum and becomes ovoid or rounded (PI. 13, figs. 64-73). Nuclear multiplication follows and “plasmodial” forms are produced (Pl. 13, figs. 74-77). These give rise to small, rounded bodies (Pl. 13, figs. 83, 84) about 3 u by 2 « which undergo multiple fission to form rosettes (PI. 13, fig. 88), which give rise to the typical pre- flagellates. (8) The young Melophagus do not show flagellates until after their first feed of blood, the blood stimulating the pre- flagellates to form flagella. (9) Multiplication of C. melophagia by longitudinal STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGTA, 2138 division occurs in both the pre-flagellate and the flagellate stages of the parasite. (10) Infection of Melophagus ovinus with C. melo- phagia is either hereditary or casual. In the case of casual infection the insects ingest the post-flagellates voided with the feces of other Melophagus. (11) A very dilute solution of grape-sugar causes the parasites to divide. There are only traces of sugar in both sheep-serum and wool extract. (12) Sheep’s blood or human blood added to the Crithidia also increased the rapidity of their growth and division. (13) A fungus present in the Malpighian tubules and gut of the ked (see Appendix II) has a rapid, fatal effect on the Crithidia. (14) An anti-coagulin is present in the salivary glands, stomach and intestine of Melophagus ovinus (see Appen- dice TUT). (15) A new spirochete, S. melophagi, was found in the gut, ovaries and puparia of a few of the Melophagus examined (see Appendix I). APprENDIX I. On the Occurrence of a Spirochete, 8S. melophagi, n.sp.,in Melophagus ovtnus. I wish to record the occurrence of a rare spirochete in the gut, ovaries and puparia of Melophagus ovinus. The spirochete was observed in life in the above-mentioned organs of a very few of the Melophagus examined, and at very different periods of the year (February, September, October). Very few spirochetes occurred, and consequently it is impossible to give full details regarding size and structure. The organisms seen were from 10 to 30 long and were narrow. ‘They vary in length, some being practically half the length of others, indicating the probability of transverse division. As some parasites were thicker than others there 214 ANNIE PORTER. is the inference that longitudinal division takes place. This would be in accordance with the behaviour of other spirocheetes, for Fantham (1907-8-9) has shown that both forms of division occur in §. balbianii and S. anodonte. I (1909) also have observed the same, while the joint researches of Fantham and myself (1909) have demonstrated that both directions of division occur in 8. recurrentis and S. duttoni, and that there is a periodicity in the direction of division. The movements of 8. melophagi are fairly active, and are of the typical spirochete nature, namely, of forward progression accompanied by spiral or corkscrew rotation on its course. The occurrence of S. melophagi in the ovaries, ova and puparia of the ked is of much interest, for it indicates that the spirochete is transmitted hereditarily. Hence Melo- phagus ovinus can transmit both Crithidia melophagia and Spirocheta melophagi to its offspring. Aprenpix IT. Noteona Fungus found in the Malpighian Tubules and Intestine of Melophagus ovinus. A fungus was present in many specimens of Melophagus ovinus examined, especially those obtained from Scotland. Crithidia were not seen in the “keds” received from Scotland, and I have shown experimentally that the action of the fungus is fatal to the flagellate. The fungus occurred chiefly in the Malpighian tubules of the insect, and to a lesser extent in the intestine. The Malpighian tubules were frequently blocked by the fungus. A brief description of the fungus may now be given. The hyphe were long and filamentous with few septa. Many spores were produced. At the extremity of some hyphe globular heads were formed, possibly due to sexual processes. ‘The globular bodies contained many nuclei (text- fig. 11) fairly evenly distributed through the protoplasm. STRUCTURE AND LIFE-HISTORY OF OCRITHIDIA MELOPHAGIA. 215 Nuclei and protoplasm then shrank away from the wall of the rounded body—provisionally called a sporangium (text- fig. 12)—so that a space intervened. Segregation of the protoplasm round the nuclei followed (text-fig. 15), and a morula-like body resulted. The morula differentiated into a mass of rounded spores (text-fig. 14), each of which formed a spore coat for itself. The sporangium ultimately ruptured TEXT-FIGURES 11-15. Fungus parasitic in Melophagus ovinus. Text-fig. 11.—Hypha with globular head. Text-fig. 12.—Differentiation of nuclei within the head (sporan- gium). Text-fig. 13.—Spores forming in sporangium. Text-fig. 14.—Mature sporangium. Text-fig. 15.—Dehiscing sporangium. (text-fig. 15), and the numerous small spores were set free. Some spores remained in the Malpighian tubes, others passed out into the intestine and were voided with the feces. Parasitic fungi have been previously recorded in insects, for example, in the house-fly, caterpillar, mosquito. ‘The fungus mentioned by Schaudinn in Culex was probably a member of the Entomophthoree, or related thereto. The 216 ANNIE PORTER. fungus infesting Melophagus ovinus seems to be more nearly allied to the Peronosporee. I learn from a private communication that a similar fungus was found last year by Dr. H. B. Fantham, of Cambridge, in the alimentary tract and Malpighian tubes of the grouse-fly, Ornithomyia lagopodis. From examination of a pre- paration of the fungus of Ornithomyia, kindly lent to me, I believe that the fungi of the grouse-fly and the sheep-ked are very closely related. AppENDIX III. On the Occurrence of an Anti-coagulin in the Ali- mentary Canal of Melophagus ovinus, and its Significance in Relation to Crithidia melo- phagia. ‘The pronounced and peculiar brightness of the blood in the crop and fore-part of the stomach of the keds examined was noticed very early in the investigation. ‘The blood of the sheep in the stomach of keds that had not fed for as long as three days was still practically fluid and had not coagulated much, while twelve to twenty-four hours after feeding the blood had not coagulated at all. This led me to suspect that an anti-coagulin, such as had been described ina tick (Argas persicus) by Nuttall and Strickland (1908), was present here also, and a series of tests were performed at different times which verified this inference. Every test that I performed had the same result—coagulation was delayed. The method of testing was simple. Separate emulsions of the salivary glands, stomach, and intestine of Melophagus ovinus were made with 0°75 per cent. NaCl solution. A known quantity—about 0°5 c.c. of human blood from a pricked finger—was then mixed with the same quantity of organ- emulsion, while for control purposes the same quantity of blood mixed with 0°75 per cent. NaCl solution was used. The test fluid and the control fluid were taken up in small glass capillaries, and the test was applied by blowing out the STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 217 liquid at stated times and noting when coagulation occurred ineach. ‘Typical results of these experiments are tabulated below: A. Adult Melophagus. Coagulation period Coagulation period Experiment. of bloodandorgan- — of blood and °75 emulsion. NaCl solution. (1) Salivary gland . 20min. : 7-8 min. er. y ot 220) Saal. (3) Intestine ; Rome Cre 5 Se a (4) * : SP TAs oOSecs. + Smee Obviously an anti-coagulin was present, for considerable delay of clotting occurred. B. Young Melophagus.—Here the interval between the clotting of the test and control preparations was noted. A few typical results are given: (1) Blood mixed with emulsion of the salivary glands clotted nine minutes after the control. (2) Emulsions of intestine added to blood caused the latter to take three times as long to clot as the control preparations took. Comparing the behaviour of the emulsions of the salivary glands of young and of older keds, the anti-coagulin seems to be more strongly developed in the salivary glands of the older keds, while a similar comparison between the intestinal emulsions would tend to show that the anti-coagulin was more abundant in the intestines of young keds. The temperature at which the anti-coagulin was destroyed was also investigated. It was found that below 50° C. the anti-coagulin would act. At about 55° C. its action was checked. When 60° C. was reached it was destroyed. Human blood mixed with emulsions of any part of the alimentary canal at once assumed the vivid red hue so notice- able in the blood removed from the gut of the keds. The red blood-corpuscles of the sheep, seen en masse, appear far brighter on adding emulsions of the gut of the ked containing the anti-coagulin. When much water was added to normal blood, hemolysis occurred, and the colour 218 ANNIE PORTER. of the solution so obtained was made much brighter when an emulsion of crushed salivary glands of the ked was added to it. The leucocytes of the sheep’s blood occurring in the gut of the ked do not appear to be affected in any way by the anti-coagulin. Anti-coagulin appears to be found in all parts of the alimen- tary canal of the ked and to decrease in amount from before backwards. As before mentioned, I determined experi- mentally that freshly shed, and therefore fluid, blood acted as a stimulant to division of the Crithidia. This artificial condition is the counterpart of the natural condition of the blood within the fore-gut of the ked. here, owing to the action of the anti-coagulin, the freshly ingested sheep’s blood does not clot, but remains fluid. It is probable that Crithidia within the gut are stimulated by this fluid blood, and divide rapidly. I obtained similar results in the case of Herpetomonas jaculum, where “division of the flagellate Herpetomonad takes place rapidly under natural conditions after ingestion of blood by the host” (Porter [1909], p. 382). If the Critsidia are in the pre-flagellate condition the rapid multiplication is followed by the outgrowth of flagella, after which the organisms separate and pass further down the alimentary canal. The presence of anti-coagulin, from the salivary glands, in the contents of the fore-gut of the ked may be the cause of the rapidity with which the pre-flagellate stage of Crithidia melophagia is passed through, the blood, kept fluid by the anti-coagulin, acting as a stimulus to further development. . REFERENCES TO LITERATURE. Further references will be found at the ends of some of the papers quoted. Bruce, Sir David. Hamerton, A. E., Bateman, H. R., and Mackie, F. P. (x, 1909).—** The Development of Trypanosoma gambiense in Glossina palpalis,” ‘Proc. Roy. Soc., ser. B, lxxxi, pp. 405-414, pls. 10, 11. STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 219 Fantham, H. B. (i, 1908)—**Spirocheta (Trypanosoma) balbianii (Certes) and Spirocheta anodonte (Keysselitz) : their Move- ments, Structure and Affinities,” ‘Quart. Journ. Micr. Sci.,’ 52, pp. 1-73, 5 pls. and Porter, Annie (1909).—** The Modes of Division of Spiro- cheta recurrentis and S. duttoni as observed in the Living Organisms,” ‘ Proc. Roy. Soe.,’ ser. B, Ixxxi, pp. 500-505. Flu, P. C. (1908).—* Ueber die Flagellaten im Darm von Melophagus ovinus,’ ‘ Archiv f. Protistenkunde,’ xii, pp. 147-153, 1 pl. Léger, L. (1902).—‘* Sur un flagellé parasite de /Anopheles maculi- pennis,” ‘C.R. Soc. Biol.,’ liv, pp. 354-6, 10 figs. Mackinnon, D. L. (1909).—‘* Note on two New Flagellate Parasites in Fleas—Herpetomonas ctenophthalmi, n. sp., and Cri- thidia hystrichopsylle, n. sp.,” ‘ Parasitology,’ ii, pp. 288- 296, 1 pl. Minchin, E. A. (1908).—* Investigations on the Development of Try- panosomes in Tsetse Flies and other Diptera,” ‘Quart. Journ. Mier. Sci.,’ 52, pp. 159-260, 6 pls. Novy, F. G., MacNeal, W. J., and Torrey, H. N. (1907).—“ The Try- panosomes of Mosquitoes and other Insects,” ‘Journ. Infect. Diseases,’ iv, pp. 225-276, 7 pls. Nuttall, G. H. F., and Strickland, C. (1908).—‘* On the Presence of an Anti-coagulin in the Salivary Glands and Intestines of Argus persicus,” ‘ Parasitology,’ i, pp. 302-310. Patton, W. 8. (1908).—‘* The Life-Cycle of a Species of Cr thidia Parasitic in the Intestinal Tract of Gerris fossarum Fabr.,” ‘Archiv f. Protistenkunde,’ xii, pp. 151-146, 1 pl. —— (1909).—* The Life-Cycle of a Species of Crithidia Parasitic in the Intestinal Tracts of Tabanus hilarius and Tabanus sp.,’ ‘Archiv f. Protistenkunde,’ xv, pp. 333-362, 1 pl. —— (1909).—* A Critical Review of our Present Knowledge of the Hemoflagellates and Allied Forms,” ‘ Parasitology,’ ii, pp. 91-143. and Strickland, C. (1908).—‘* A Critical Review of the Relation of Blood-sucking Invertebrates to the Life-Cycles of the Try- panosomes of Vertebrates, etc.,” ‘ Parasitology,’ i, pp. 322-346. Pfeiffer, E. (1905).—‘* Ueber Trypanosomenihnliche Flagellaten im Darm von Melophagus ovinus,” ‘Zeitschr. f. Hyg., 1, pp. 324-29, 1 pl. Porter, Annie (1909).—“*The Morphology and Life-History of Cri- thidia gerridis, as found in the British Water-Bug, Gerris paludum,” ‘ Parasitology,’ ii, pp. 348-366, 1 pl. 220 ANNIE PORTER. Porter, Annie (1909).—“The Life-Cycle of Herpetomonas jaculum (Léger), Parasitic in the Alimentary Tract of Nepa cinerea,” ‘Parasitology,’ ii, pp. 867-391, 1 pl. Pratt, H. 8. (1895).—* Beitriige zur Kenntnis der Pupiparen (Die Larve von Melophagus ovinus),” ‘Archiv f. Naturgesch.,’ liii, pp. 151-200, 1 pl. (1899).—“*The Anatomy of the Female Genital Tract of the Pupipara as observed in Melophagus ovinus,” ‘ Zeitschr. f. wiss. Zool.,’ xvi, pp. 16-42, 2 pls. Robertson, Muriel (1909).—‘ Studies on Ceylon Hematozoa: I, The Life-Cycle of Trypanosoma vittate,” ‘Quart. Journ. Mier. Sci.,’ 53, pp. 665-695, 2 pls. Schaudinn, F. (1904).—‘* Generations- und Wirtswechsel bei Try pano- soma und Spirochete (Vorl. Mitt.),” ‘Arbeit. a. d. Kaiser. Gesundheitsamte,” xx, pp. 387-430, 20 figs. Swingle, L. D. (1909).—* A Study on the Life-History of a Flagellate (Crithidia melophagi, n. sp.) in the Alimentary Tract of the Sheep-Tick (Melophagus ovinus),” ‘Journ. Infect. Diseases,’ vi, pp. 98-121, 3 pls. Woodeock, H. M. (i, 1909).—‘* The Hiemoflagellates and Allied Forms,” article in ‘ Treatise on Zoology, edited by Sir Ray Lankester, pt. i, fase. i, sect. G., pp. 193-273. EXPLANATION OF PLATES 12 ann 13, Illustrating Miss Annie Porter’s paper on “ Crithidia melophagia.” [All figures were outlined with an Abbé-Zeiss camera-lucida, using a 2 mm. apochromatic (Zeiss), or ~ inch achromatic (Zeiss) objective, and compensating oculars 8 and 12 of Zeiss. The magnification is in all cases approximately 1500 diameters, except where otherwise stated. ] PLATE 12. Figs. 1-20.—Pre-flagellate Stages. Fig. 1.—Pre-flagellate with round nucleus, bar-like blepharoplast. No flagellum. Crop. Giemsa. Fig. 2—Oval pre-flagellate. Blepharoplast slightly constricted. Crop. Delafield’s hematoxylin. Fig. 3.—Dividing pre-flagellate. Crop. Delafield’s hematoxylin. STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 221 Fig. 4.—Division rosette of pre-flagellates. Two individuals again dividing. Crop. Delafield’s hematoxylin. Figs. 5-8.—Elongating pre-flagellates. Crop. Thionin. Fig. 9.—Large preflagellate, with round nucleus, rod-like blepharo- plast, flagellum just differentiating. Crop. Giemsa. Fig. 10.—Rounded form. Flagellum longer than in fig..9. Crop. Giemsa. Fig. 11.—Smaller parasite with large nucleus and long flagellum. Crop. Delafield’s hematoxylin. Fig. 12.—Parasite, showing elongation of flagellar (anterior) end of the body. Crop. Giemsa. Figs. 15 and 14.—Crithidia with elongated posterior ends. Anterior part of stomach. Giemsa. Fig. 15.—Pre-flagellate with posterior blepharoplast. Crop. Giemsa. Fig. 16.—Parasite with anterior end more developed. Crop. Giemsa. Figs. 17 and 18.—Almost mature flagellates, membranes showing myonemes. Crop. Giesma. Figs. 19 and 20.—Practically adult flagellates. Fore-part of stomach. ‘Thionin. Figs. 21-43.—Flagellate Stage. Fig. 21.—Small flagellate. Nucleus with chromatin in granules extending part way across the body. Rod-like blepharoplast. Intestine. Giemsa. Fig. 22.—Flagellate, with well-marked myonemes on the body. Stomach. Gentian violet. x 2250 approximately. Fig. 23.—Parasite, with flagellum almost continuous with the blepharoplast. Nucleus with centralchromatin. Stomach. Delafield’s hematoxylin. Fig. 24.—Crithidia showing blepharoplast posterior to the nucleus —an uncommon condition. Stomach. Giemsa. Figs. 25, 26.—Flagellates showing chromidia in their posterior ends. Chromatin of nucleus in bars. Stomach. Giemsa. Figs. 27-29.—Parasites with somewhat pointed posterior ends. Chromidia present in fig. 29. Intestine. Thionin. Figs. 30, 31.—Crithidia showing somewhat alveolar protoplasm. Stomach. Thionin. Xx 2250 approximately. Fig. 32.—Flagellate with blunt posterior end, round nucleus contain- ing large chromatin granules, and extending across complete breadth of body; blepharoplast curved. Stomach. Thionin. 2250 approxi- mately. VOL. 99, PART 2.—NEW SERIES. 15 222 ANNIE PORTER. Fig. 33.—Parasite with scattered chromidia. Blepharoplast slightly posterior to and to the side of the nucleus. End of crop. Giemsa. Fig. 34.—Crithidia with large oval blepharoplast. Stomach. Giemsa. Fig. 35.—Narrow parasite. Intestine. Giemsa. Figs. 36, 37.—Longer parasites with many chromidia. Stomach. Tron-hvematoxylin. Fig. 38.—Flagellate showing alveolar protoplasm, nucleus and blepharoplast almost in contact. Intestine. Thionin. x 2250 approxi- mately. Fig. 39.—Long form. Nucleus with chromatin arranged in bars. Oval blepharoplast. Membrane distinct. Intestine. Giemsa. Fig. 40.—Long parasite with thick flagellum. Myonemes present on body. Blepharoplast showing constriction, so about to divide. Chromatin of nucleus in large masses. Stomach. Delafield’s hwema- toxylin. Fig. 41.—Small aggregation-rosette, showing entanglement of large and small flagellates. Stomach. Giemsa. Fig. 42.— Flagellate with rounded nucleus and posterior blepharoplast. Basal granule near root of flagellum. Myonemes in membrane. Intestine. Iron-hxematoxylin. Fig. 43.—Large rosette. Many parasites shown aggregated around a piece of débris. The flagella serve as points of attachment, therein differing from a division-rosette. Common in stomach and intestine. Delafield’s hematoxylin. Figs. 44-56—Stages in Division. Fig. 44.—Parasite showing constricted blepharoplast with clear area around it. Chromatin in nucleus arranged in masses at periphery. Intestine. Thionin. xX 2250 approximately. Fig. 45.—Stage similar to fig. 44. Well-marked myonemes on body and membrane. Giemsa. X 2250 approximately. Fig. 46.—Parasite with both nucleus and blepharoplast constricted. Flagellum beginning to split at base. Stomach. Delafield’s hematoxylin. Fig. 47.—Flagellate with anterior end of body, nucleus and blepharo- plast all divided. Stomach. Delafield’s hematoxylin. Figs. 48, 49.—Somewhat rounded parasites ; bodies of daughter-forms not yet diverging from one another. Stomach. Thionin. Fig. 50.—Daughter-organisms forming a V. Stomach. Giemsa. Figs. 51, 52.—Further stages in the divergence of the bodies of the STRUCTURE AND LIFE-HISTORY OF CRITHIDIA MELOPHAGIA. 223 daughter-forms. The flagella have interlocked. Intestine. Delafield’s hematoxylin. The parasites represented in fig. 51 divided sub-equally. Fig. 53.—Sub-equal division. Daughter-organisms are almost separated. Intestine. Delafield’s hematoxylin. Figs. 54, 55.—Parasites about to separate. Stomach. Giemsa. Fig. 56.—True division-rosette. The separation of the daughter- individuals takes place from the flagellar end backwards, so that in a rosette the posterior ends of the organisms are centrally directed. Stomach. Thionin. PLATE 13. Figs. 57-94.—Stages of the Parasite in the Ovary, Eggs, and Puparia. (The eggs in figs. 58, 64, 65 are represented diagrammatically.) Fig. 57.—The flagellate as it penetrates the ovary. Delafield’s hematoxylin. Fig. 58.—Flagellate in the act of penetrating a young egg, the blunt end of the parasite being used. Thionin. The egg of Melophagus ovinus is represented diagrammatically. Figs. 59, 60.—Flagellates from ovary. Flagella somewhat reduced. Giemsa. Figs. 61-63.—F lagellates from within the egg. Giemsa. Figs. 64, 65—Rounding-up forms of C. meloplagia within eggs. Delafield’s hematoxylin and fresh preparations. Eggs of Melophagus represented diagrammatically. Figs. 66-72.—Series of parasites showing successive stages in shorten- ing and rounding-up of flagellates when within the eggs. Delafield’s hematoxylin. Figs. 73, 74.—Parasites showing nuclear division. Very young puparium. Giemsa. Figs. 75-77.—* Plasmodial ” stages of C. melophagia in developing puparia. Peripheral blepharoplasts seen. Giemsa and fresh prepara- tions. Figs. 78-81.—Rounded parasites resulting from plasmodial forms. Delafield’s hematoxylin. Figs. 82-84.—Parasites produced by growth of forms similar to those shown in fig. 81. Giemsa. Figs. 85-87.—Rosettes of somewhat oval parasites from young puparium. Delafield’s hematoxylin. Fig. 88. — Well-defined division-rosette from mature puparium. Giemsa. 224: ANNIE PORTER. Figs. 89-91.—Dividing forms. Mature puparium. Giemsa. Figs. 92-94.—Parasites resembling pre-flagellates produced from cyst. Mature puparium. Delafield’s hematoxylin. Figs. 95, 96.—Small aggregation-rosettes. Intestine. Thionin. Figs. 97-114.—Post-flagellate Stages in Rectum. Fig. 97.—Parasite dividing prior toencystment. Intestine. Thionin, Fig. 98.—Uncommon form of division, occasionally seen in living specimens. Rectum. Giemsa. Fig. 99.—Small form. Flagellum in process of absorption. Rectum. Giemsa. Fig. 100.—Parasite showing concentration of protoplasm in the region of the nucleus. Rectum. Giemsa. Fig. 101.—Form common in rectum. Body much flattened. Flagellum disappearing. Delafield’s hematoxylin. Figs. 102-108.— Parasites showing progressive disappearance of flagellum. Rectum. Thionin. Figs. 109-112.—Post-flagellate cysts from rectum. Giemsa. Fig, 113.—Post-flagellate cyst from feces of Melophagus ovinus. Giemsa. Fig. 114.—Thick-walled cyst. Rectum. Giemsa. eee) 23. A.Porter, del. CRITHIDIA Muth, Lith? London, Iuant.bunn Mécnr8i.Ubl 65 NSA, 12 ELOPHAGIA. Cc * ee A Portes, dal. Quant. Sourn Mier Sci. Ut.58. NS A12 9 6 ,., © ® & 1s, & uw. ® ae 107. 109. Tes Huth Sith’ Londan. GCRITHIDIA MELOPHAGIA. STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 225 Studies in the Experimental Analysis of Sex. By Geofirey Smith, Fellow of New College, Oxford. With Plate 14. 3. FurrHer OBsERVATIONS ON Parasitic CastTRATION. Durine my occupation of the British Association Table at Naples this winter I took the opportunity of re-examining certain points connected with the effect of Sacculina neglecta on the spider-crab Inachus mauretanicus, with the especial purpose of trying to settle the exact way in which the gonad of infected individuals degenerates and is absorbed. The mid-winter months being the most favour- able season at Naples for finding numerous individuals of Inachus very profoundly modified by the presence of the parasite, | was able to re-investigate many crucial stages in the modification of the external and internal sexual organs, with the result that, while certain new facts of interest have come to light, I see no reason whatever for departing in any respect from the statement of facts made in my earlier work, or from the deductions drawn from them (‘Naples Mono- graph,’ No. 29, Chap. V). In this paper, besides giving the results arrived at in respect to the degeneration of the gonad, I propose to describe certain new instances of infected Inachus which afford incontestable proof that male crabs with differentiated though reduced male internal organs can assume all the adult female secondary sexual characters. It will also be shown both for the male and female sex that the effect of parasitic castration can on no account be 226 GEOFFREY SMITH. interpreted as a return to a juvenile undifferen- tiated condition. In Part 2 of these studies it has already been pointed out that this interpretation is ruled out by the facts, and this was also pointed out in my earlier work, but not in so detailed and categorical a form, with the unfortunate result that Professor T. H. Morgan, in a recent paper on ‘‘ Sex Determina- tion” (‘Journal Exper. Zoology,’ vol. vii, 1909, pp. 343, 344), has adopted this very explanation of my observations. Thus he writes: “The broad abdomen of the castrated male might be considered to correspond to the juvenile state. ‘The only external structure cited by Smith that might seem to indicate that the characters of the castrated males are female rather that juvenile ones is the presence of hairs on the abdominal appendages of Inachus, absent in the young crab, but present in the adult female. Such evidence would not in itself be conclusive, since the presence of hairs may be due to increase in size or to a later moult rather than to latent female characters. Smith concludes that the male sex, and pro- bably the male sex alone, can be so radically modified in its sexual nature as to assume a perfect external hermaphroditism. If, on the contrary, we assume that we have here, not herma- phroditism, but an imperfect development of male characters combined with the juvenile condition, we might offer a plausible explanation of the facts.” I am sorry that any want of explicitness on my part should have misled Professor Morgan, but I cannot accept the state- ment that the only characteristically adult female character, cited by me as being assumed by the infected males, is the presence of hairs on the abdominal appendages. I pointed out in my earlier work (‘Naples Monograph,’ xxix, pp. 67, 70 and 71) that in the young stages of the female, before the adult breeding form is assumed, the abdoinen is a com- paratively small flat plate, whereas in the adult it becomes suddenly widened and also takes on a hollowed trough-like shape, so that the two forms of abdomen are absolutely distinct morphological structures, distinguishable from one STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 227 another at a glance (see figs. 1, 2, 3 and 4, PI. VII, ‘ Naples Monograph’ and again ‘ Quart. Journ. Micr. Sci.,’ vol. 54, Pl. 30, figs. 10, 11, 18,14). Now, when the infected males take on the female external characters they have never been found to assume the juvenile flattened form of abdomen which charac- terixes the young stages of both males and females, but they invariably take on the hollow trough-like form characteristic of the adult breeding female and of her alone (see the numerous figures on the plates referred to above). That is the first objection to the view that the alteration of the male is merely towards a juvenile condition, and anyone who will examine the series of specimens exhibited in the South Kensington Museum or in the Oxford Museum, or those deposited by me at the Zoological Station at Naples, will at once perceive the entire morphological difference of the abdomen in the young and adult female, and the identity of the modified male abdomen with that of the adult female. Secondly, with regard to the abdominal appendages. It is not a question of the mere presence or absence of a few hairs, as Professor Morgan has unfortunately been led to suppose. The abdominal appendages of the juvenile and adult indi- viduals differ as radically, if not more radically from one another, than the form of the abdomen. In the young form of the female these appendages are short, stout and rod-lke, and provided with a very few short bristles, as shown in PI. 14, fig. 7 of this paper. In very young males similar appendages are present, but they are lost at a very early stage indeed, only the two anterior appendages being kept as the copulatory styles. ‘he form of these two appendages in the young male is shown in PI. 14, figs. 1 and 2. The adult female, at the same moult at which it acquires the characteristic adult form of abdomen, assumes a totally different kind of appendage of the form shown in PI. 14, fig. 4. Here it is seen that instead of being stout and rod-like with a few stiff hairs, as in the young females, the appendage has become transformed iuto two wisp-like branches, the exopodite being densely clothed with long plumose hairs, the 228 GHOFFREY SMITH. endopodite, now a slender-jointed structure, being furnished with exceedingly long pointed hairs for the attachment of the eggs. The structure of these abdominal appendages in the adult female, adapted as they evidently are for repro- ductive purposes, is as morphologically distinct from that of the young individuals of either sex as anything very well could be. Now let us inquire in what form the infected modified males assume the abdominal appendages. ‘The answer is plainly given by reference to Pl. 14, fig. 4. This figure is an actual camera drawing of the second abdominal appendage of an infected individual, which was proved to be a male by the presence of a copulatory style of a somewhat modified form (Pl. 14, fig. 38), and internally by the presence of testes and vesicule seminales of a typical character on either side. The testis and vesicula seminalis of one side of this individual are shown in Pl. 14, fig. 10. The form of the abdominal ap- pendages (PI. 14, fig. 4), of which there were four on each side in addition to the copulatory styles, is identical with that of a normal adult female; in fact, since this figure serves equally well to depict the abdominal appendage of a normal adult female, [ have not considered it necessary to give another figure, which would simply mean repeating the same structures. ‘he infected male individual to which figs. 3, 4, and 10 on Pl. 14 refer is a particularly favourable type for showing con- clusively thatthe abdominal appendages, when assumed by the infected males, are of the characteristically adult female type. Asamatter of fact acommoner condition is that shown in Pl. 14, fig. 5. : In this infected male the copulatory style was greatly reduced (fig. 6) and the abdominal appen- dages were also developed in an imperfect condition, with almost complete suppression of the endopodites. Neverthe- less, the characteristic plumose hairs are present on the exo- podite, which is of.a slender shape, thus conforming to the adult type of female appendage and not really approaching to the juvenile condition. This figure might equally well refer STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 229 to the abdominal appendage of an infected female, in which the endopodites are very frequently thus reduced. We have now examined in some detail two of the mest im- portant characters in which the infected male Inachus is modified by the presence of the parasite Sacculina, viz. the shape of the abdomen aud the form of the abdominal appen- dages, and we have seen that Professor Morgan’s attempt tu explain the modification of the male as a return to a juvenile condition is quite at variance with the facts. But we have still two more points to consider, which render that explanation still more impossible. The most important of these two points is the fact that in a certain small percentage of cases the infected males, on recovery from the parasitic disease, have been observed to have regenerated the gonad, and to have developed large ova measuring about 1 mm. in diameter and full of the reddish- coloured yolk characteristic of the mature ova of the female Inachus. Professor Morgan himself admits the cogency of this fact, so that I need not labour it here, its significance, indeed, being obvious. The second point is one which I have only been able to settle finally during my recent visit to Naples. In my earlier work (loc. cit., p.68) I inclined to the view that the presence of Sacculina caused the young females under 15 mm. in carapace length to assume prematurely the adult type of abdomen and abdominal appendage, and I emphasised this point as being of importance in precluding the view that the effect of the parasite was merely to arrest development or “ause a return to a juvenile state. By a careful examination of the large amount of material put at my disposal by Dr. Lo Bianco this winter, I have found that this premature assumption of adult characters by infected females undoubtedly occurs. During December and January all the uninfected females of carapace length up to 14 mm. had the immature juvenile form ot abdomen and appendage, but all the intected females measuring froin 6-14 mm. had the fully adult type of both those structures. The real theoretical significance of 230 GEOFFREY SMITH. this fact, which has an important bearing on the whole meaning of parasitic castration, will be discussed later, but it has been introduced here as a final nail in the coffin of the theory which attempts to explain the effects of parasitic castration as due to arrested development or the assumption of juvenile characters. Possibly the use of the term “ parasitic castration” has had something to do with perpetuating this unfortunate error, the analogy between ordinary operative castration or mechanical removal of the gonads and their degeneration owing to the presence of a parasite being, as Professor Sedgwick has pointed out, extremely small. In parasitic ‘ castration” the degeneration of the gonad is not brought about by the parasite mechanically removing or attacking the gonad, but by its setting up a deep-seated Alteration of the metabolism of the host which secondarily reacts on the gonad. We may now enter into the question of the method of degeneration of the gonad. In the above paragraphs I trust that the following conclusion has been thoroughly vindicated. The modification of the male Inachus by the parasite Sacculina consists in the assumption by the male of adult female sexual characters to a greater or less degree of perfection; in neither sex can the modifica- tion be ascribed to arrest of development or the assumption of a juvenile immature condition. As I have shown in my earlier work (loc. cit., pp. 72-74) the degenerate condition of the ovaries and testes with their ducts in infected Inachus is due to two causes: firstly, an arrest of growth, so that the gonad tends to remain in the same condition as it was when infection took hold, and secondly, to an actual absorption of the tissues of the gonad and their final disappearance, a process which was often accompanied by an actual irruption of the roots of the purasite into the germinal tissues. ‘I'he arrest of growth of the gonad and the first stages of degeneration, at any rate in the male, were shown to be independent of the irruption of the Sacculina roots. The method of absorption and disappearance of the gonad STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 231 was not clearly made out, and it was my chief object this year to obtain some idea of how this process takes place, to observe, for example, whether phagocytosis takes any active part in it. The condition of arrested growth without any signs of active degeneration is well exemplified by the testis and vesicula seminalis figured on PI. 14, fig. 10, which was dis- sected out of the perfectly modified male whose appendages are represented in figs. 3 aud 4. In Pl. 14, figs. 8 and 9, are drawn, on the same scale, the vesicula seminalis and a small portion of the coiled testis of an uninfected male of the same size, Showing that the gonad of the infected individual has remained very small and undeveloped. Spermatozoa were entirely absent from the infected individual, whereas the vesicula of the normal individual was crowded with them. There is, however, another point to be observed. Investing the gonad of the normal individual is a thin sheath of con- nective tissue with flattened, darkly staming nuclei (c.s., P]. 14, fig. 9). In the infected individual tlis sheath is seen to be of proportionately greater thickuess. In Pl. 14, fig. 11, is shown a portion of the testis of an infected male in which the process of absorption of the gonad has proceeded to a considerable extent. In three places small disconnected masses of testicular cells (¢.) are seen lying ensheathed by connective tissue; between the disconnected pieces of germinal tissue nothing remains but the connective- tissue sheath. By staining such preparations with a triacid stain, e.g. Ehrlich-Biondi, small globules are seen lying between the germinal nuclei and the sheath, which take up the orange stain. These globules may be looked upon as degeneration products of the germinal tissue in process of absorption. In none of the preparations which I have made of degenerating gonads is there any sign of phagocytosis, the degeneration appearing to take place by some process of auto-digestion. Turning to the degeneration of the ovary, PI. 14, figs. 12 and 13, we find exactly the same process. Fig. 12 represents ot 232 GKOFEREY SMITH. a portion of degenerate ovary of an infected female, in which islets of ovarian tissue containing disintegrating ova are seen encapsuled in the connective-tissue sheath. Fig. 13 is a high power drawing of a small portion of the ovary showing the clear distinction between the germinal nuclei (VV), the nuclei of the connective-tissue sheath (cs) and the degenerating ova. In a very great number of infected crabs dissected no trace could be found of the remains of a gonad; and in these, allowing for .a certain number in which I overlooked the degenerating remains, one must suppose that the process of encapsula- tion by connective tissue and auto-digestion had led to com- plete disappearance. Iam unable to state for certain whether the connective-tissue sheath plays an active part in the absorption of the germinal tissue; the chief part is clearly due to a simple disintegration of the same nature as is now known to occur in the destruction and absorption of the larval organs of insects during metamorphosis. In this latter process it was formerly held that phagocytosis played thie principal part, but it is now kuown that a process of auto- digestion by fluids is at least as active an agent. T'o conelude this part, I will attempt to outline, in a more satisfactory manner than was possible before, an explanation of why it is that the presence of a parasite should bring about such profound physiological and morphological changes in its host. We must clearly define, in the first place, what these changes essentially consist in. It has been shown in my earlier papers, and I trust still more fully brought out in this paper, that the effect of Sacculinaon Inachus is to cause thie infected individuals of botl. sexes to assume adult female characteristics. ‘his results not only in transforming the males into hermaphrodites with preponderating female char- acters, but also in hastening on the assumption of adult female characters by immature females. The problem, there- fore, resolves itself into this, Why should the presence of Sacculina cause the host of either sex to become adult female in nature ? STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 233 Let us examine what the process of becoming adult involves in an ordinary female crab. Plainly the most important change is the rapid elaboration of yolk material which accu- mulates in the ovary; causing the latter to grow to a very great size. This elaboration of food material: in the ovary is the fundamental point in which the adolescence of the female gonad differs from that of the male. In the male gonad at maturity we have an immense multiplication of nuclei and of chromatin but a small development of cytoplasmic material and no deposit of yolk; in the female we have the exact opposite of this process. The most important part, then, in the process of becoming adult female, is the active elaboration of yolk material. We have arrived, therefore, at this point of the argument: that the presence of Sacculina causes the crab of either sex to become adult female in nature, and the most important activity of this state is the elaboration of yolk material. Can we prove that the presence of Sacculina actually causes its host of either sex to produce yolk material? [believe we can. If the roots of Sacculina which fill the body of an infected Inachus be examined, they will be found to be packed with small globules of an oily material, and if the roots are stained with such a mixture as Ehrlich-Biondi’s tri-acid stain it may be observed that the Sacculina roots take up the same consti- tuent in the stain, namely the acid fuchsin, as the yolk of an adult female crab’s ovaries. From the observed contents of the Sacculina roots and from their reaction to stains it is clear that they are elaborating from the blood of the Inachus of both sexes a closely similar yelk material to that which is normally accumulated in the ovary of a healthy adult female Inachus. The effect of Saceculina on Inachus is therefore to force the latter to elaborate yolk material of a similar kind to that which is normally developed in the ovary of the female at maturity. As the Inachus elaborates it the Sacculina abstracts it, so that it does not come to be deposited in the gonad until after recovery from the disease, when, as we have 234 GEOFFREY SMITH. seen, the yolk-containing ova may be formed in the gonad of either sex. Meantime the continued production and circulation in the blood of the infected Inachus, whether male or female, of this yolk material, or rather of the substances from which the yolk is built up, is accompanied by the production of the secondary sexual characters proper to the adult female. These yolk-forming substances, or substance, are therefore identical with the “sexual formative substance,’ whose existence we deduced in Part 2 of these studies. We may summarise the above argument as follows: The Sacculina roots require for their nourishment a substance in the blood of the crab which they can work up into yolk material. This substance is provided for them in the female sexual formative substance, which is circulating in small quantities in normal male crabs as well as, in greater quantities, in female crabs. But the Sacculina roots must have the power, not only of abstracting this material from the crab’s blood, but also of forcing the crab to go on forming this substance in excess. This may seem to be a great assumption; but it is exactly here that a very close parallel can be drawn between the phenomenon we are dealing with and the general processes of immunity to parasites and organic poisons. Immunity has been interpreted, especially by Ehrlich, to mean that when a poison acts upon an organism it combines with and anchors certain organic molecules, which are then regenerated in excess and poured out into the blood- stream as antibody. If we suppose, therefore, that the Sacculina roots anchor the molecules of the female sexual formative substance, and this, from the fact of their forming yolk material, they appear to do, it is in accordance with the facts of immunity to suppose that the molecules of the sexual formative substance, wherever they are formed, will be regenerated in excess. The continued operation of this process, namely, the pro- duction of female sexual formative substance in the blood- stream, and its abstraction by the Sacculina roots, would STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 2380 account for all the observed phenomena, viz. the development of adult female characters, which are dependent on the presence of this substance in quantity in the blood, and the abortion of the gonad owing to the Sacculina roots seizing on its proper nutriment and not permitting it to grow or develop. Nevertheless in the case of the hermit-crab infected by Pelto- gaster, Potts has shown that small eggs may be formed in the gonad, even while the parasite is still vigorous, showing that the excess of sexual formative substance has to some extent been seized on by the gonad. In the above manner it appears to me that we not only “‘ parasitic castra- gain a clear idea of the process involved in tion,”’ but the phenomenon, instead of appearing an isolated curiosity of a wholly inexplicable nature, falls into line with the well-known reactions to parasitic infections which are classed under the category of immunity. The clue to the whole theory rests in the truth of the statement that ‘ para- sitic castration ” consists in the assumption by the infected individuals of adult female characteristics, owing to the development within them of the female sexual formative sub- stance. If this statement of the case is rejected by the reader on the evidence which I have adduced, he will naturally reject the theory proposed to account forit, and if he can succeed in framing a different and more satisfactory theory which will include all the facts I shall be very well pleased. But any attempt to explain “parasitic castration” by vague analogies with the effects of operative castration, or by referring the whole phenomenon to arrested development or appearance of juvenile characters, is certainly foreordained to failure. ‘The explanation here offered of parasitic castration differs from that which I proposed in my first work (‘ Naples Mono- graph,’ xxix, p. 82, et seq.) only in its greater precision, not in its general outline. In my original statement of the theory T ascribed the alteration of the male to an adaptive response of the metabolism in order to make good the drain on the system caused by the presence of a parasite. The metabolism 250° . GEOFFREY SMITH. was represented as changing from the katabolic male con- dition to the more anabolic female, and this change was supposed to be effected by the development in the body of the female sexual formative substance. It is clear that this theory is fundamentally the same as that now proposed, but being couched in rather vague and general language, it seems to have made very little impression even on those who un- reservedly accepted my statement of facts. By showing, firstly, that the assumption of the adult female condition involves an active elaboration of yolk material, and secondly, that the Sacculina roots actually withdraw some substance from the ecrab’s blood from which they manufacture a yolk substance closely similar to that normally deposited in the eggs of the crab, and also by emphasising the fact that in reality both sexes of the host react in exactly the same way to the parasite, it has been possible to express the theory in a far more objective manner. a) Summary oF Parr 3. (1) 'The effect of Sacculina on male Inachus consists in the assumption by the male of adult female characteristics, and can in nowise be ascribed to arrest of development or acqui- sition of juvenile or immature characters, as suggested by Professor ‘I’. H. Morgan. (2) The effect of Sacculina on young immature females of Inachus is to force them to assume prematurely adult female characteristics. (3) The absorption of the gonad of infected Inachus of both sexes is brought about by a process of ensheathment with connective tissue and auto-digestion, phagocytosis ap- parently playing no part. (4) The reason why Sacculina causes theassumption of the adult female state in Inachus is found in the facts (1) that the roots of Sacculina elaborate a yolk-substance from the blood of Inachus of a similar nature to that which is elabo- rated in the ovaries of an adult female Inachus; (2) that in order to elaborate this yolk substance they take up from the STUDIES IN THE. EXPERIMENTAL ANALYSIS OF SEX. 237 blood of Inachus the female sexual formative substance, which is the necessary material for forming the yolk; (8) that the female sexual formative substance, being anchored by the Sacculina roots, is regenerated in excess; (4) that the presence of the female sexual formative substance contin ually circulat- ing in large quantities in the body-fluids of the infected crabs causes the production of adult female secondary sexual characters, and, when the parasite dies, of yolk-containing eges. 4. On a Case oF Parasitic CAsTravION IN A VERTEBRATE. Although numerous cases are now known of .the presence of a parasite causing arrest of development or degeneration of the reproductive organs in various invertebrates, no clear instance of this process has been reported, so far as I am aware, among vertebrate animals as the result of bacterial disease of organs other than the reproductive organs them- selves. Of course, where the reproductive organs themselves are the seat of infection, a certain amount of atrophy or degeneration may naturally result, but we have here to deal with a case of parasitic castration, analogous to the case of Sacculina on Inachus, or of Entoniscus on various -crabs, where the reproductive organs are not themselves necessarily attacked by the parasite, but are secondarily affected by the general disturbance of the metabolism, set up by the presence of a parasite in other parts of the body. During December, 1909, I received a pure-bred Gallus bankiva cockerel for breeding purposes. It belonged to the breed known as the Indian Jungle Fowl, a breed which has departed very little from the wild Gallus bankiva. The bird when it arrived appeared in good health; the ‘plumage was in good condition, the comb and wattles well developed and red, the spurs fully developed, the tail carried -erect, and the bird crowed in the normal manner. Its age was one year and a half. About two weeks after it arrived it showed signs of sickness and a tendency to mope in the straw at the back of its run. These symptoms became gradually worse, and at the beginning of February the whole VOL. 5D, PART 2.—NEW SERIES. 16 238 GEOFFREY SMITH. appearance of the bird was changed: the comb and wattles were greatly shrunken, and instead of being bright red were unhealthy pink patched with grey; the skin round the eyes was bloodless; the tail was carried drooping, and the bird never crowed. The bird was isolated and treated with purgatives, but the illness continued, the comb and wattles having withered by the middle of April to about half their original size. The spurs and plumage were unchanged, save for the fact that the tail was always drooped. The bird was killed and dissected on April 8th. The post-mortem examination showed that it was suffering from very acute avian tuberculosis. The liver was inter- penetrated with whitish calcareous nodules swarming with the characteristic tubercle bacillus, while the whole course of the alimentary canal, pancreas and spleen was covered with similar swellings, some of them of the size of a pea, also full of living bacteria. Only the alimentary and lymphatic organs were infected, the lungs, kidneys, and testes being entirely free of infection. Although the testes were uninfected, it was at once apparent that they were very remarkably reduced in size, measuring only 10 mm. in length by 5 mm. in breadth, whereas in a normal cockerel of the same breed and age, at the same time of year, they measured 40 mm. in length by 25 mm. in breadth. The vasa deferentia were also reduced in size, and this was especially noticeable in the coiled lower part of the tubes where they pass into the vesicule seminales: no spermatozoa were present. Sections of the testes showed the testicular tubes intact, with a regular lining of germinal epithelium cells with nuclei in a resting condition. There was no sign of any mitosis or of any other stages in the process of spermogenesis. The testicular tubes, in fact, presented the appearance charac- teristic of immature birds of a few weeks old. In a certain number of the tubes degenerating germinal cells with abnormal nuclei could be seen. In contrast to this extreme reduction and arrest of develop- STUDIES IN THE EXPERIMENTAL ANALYSIS OF SEX. 239 ment in the germinal part of the glands, the interstitial cells, forming islets everywhere between the testicular tubes, were well marked. There was no trace of infection by the tubercle bacilli in either testis. It is clear from the course of the disease and from the post- mortem examination that the reduction of the comb and wattles and the atrophy of the testes went hand in hand with the acute development of the tuberculosis. We know from numerous experiments that the effect of the removal of the testes in Gallus is to arrest the development of the comb and wattles ; otherwise, except for the loss of the crowing and the drooping of the tail, the other secondary sexual characters are not affected. We have seen that as the bird in question became ill, the principal symptom was the reduction in the comb and wattles, and the post-mortem showed that the testis must have been accompanying these organs in a process of atrophy. We have, therefore, in this case, an instance of parasitic castration caused by a bacterial infection of a vertebrate host, exactly parallel to the cases of parasitic castration in various Invertebrata caused by such various parasites as Crustacea, Sporozoa, and worms of various kinds. Ina great number of these cases the effect of the parasitic castration is to arrest the development or cause the atrophy of the primary and secondary sexual characters without actively calling forth the production of the female sexual characters in the parasitised male. In other cases (as far as we know only in the Crustacea) besides the suppression of the sexual characters both primary and secondary proper to the infected individual, we find the active assumption of female characters by the parasitised male, as described in Parts 2 and 3 of these studies. ‘he particular case just described belongs, as far as the evidence goes, to the former of these two categories, v. e. that in which certain of the male sexual characters atrophy without the active assumption of female characters. The principal interest attaching to this case 24.0 GEOFFREY SMITH. consists, firstly, in establishing a bacterial disease of a verte- brate asa cause of parasitic castration and thus extending the operation of this principle to two new classes of organisms, and secondly, in bringing out the correlation between the activity of the testes and the development of the comb and wattles of Gallus bankiva. Inthe next part this correlation will be dealt with more fully on an experimental basis. LETTERING. C. 8S. Connective tissue sheath. Hn. Endopodite. Ex. Exopodite. N. Germinal nuclei. O. Ovary. TT. Testis. V. S. Vesicula seminalis. EXPLANATION OF PLATE 14, Illustrating Mr. Geoffrey Smith’s paper on “ Studies in the Experimental Analysis of Sex.” All the figures refer to Inachus mauretanicus (Lucas). Fig. 1.—First abdominal appendage (copulatory style) of normal uninfected male. x 5. Fig. 2.—Second abdominal appendage of normal uninfeeted male. x 5. Fig. 3.—First abdominal appendage of infected male “A.” x 5. Fig. 4.—Second abdominal appendage of infected male “A.” x 5. (This figure might serve equally well for the abdominal appendage of an adult female.) Fig. 5.—Second abdominal appendage of infected male“ B.” x 5. = Fig. 6.—First abdominal appendage of infected male * BB.” ee Fig. 7—Second abdominal appendage of normal uninfected female, before adult condition is assumed. x 5. (The adult form of this appendage is practically identical with that given in fig. 4.) Fig. 8—Vesicula seminalis of a small normal male, measuring 14 mm. carapace length. xX 20. Fig. 9.— Coils of testis of the same male. X 20. Fig. 10—Vesicula seminalis, duct, and coils of testis of infected male“ A.” xX 20. Fig. 11.—Portion of testis of an infected male, showing absorption of germinal cells in connective-tissue sheath. x 30. Fig. 12.—Portion of ovary of an infected female, showing absorption of ova and germinal cells in connective-tissue sheath. x 930. Fig. 13.—Another portion, higher magnification, of ovary of infected female. X 60. PEER ELOOR AAO RESIS SA pi kt te Se ea i Se Lies SOOO SR d < ee 7 08 a Siece SS if PES SAE Mcafee ~ ’ Z & = ae sq eet aes wee = ‘ : meng eR Se SO > Rea eee eS cca = ——— Sey =_ Se a Sh Huth Tath? London s OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 241 Some Observations on a Flagellate of the Genus Cercomonas. By Cc. M. Wenyon, M.B., B.S., B.Sc., Protozoologist to the London School of Tropical Medicine. With 19 Text-figures. In the present paper I shall describe a flagellate of the genus Cercomonas, a genus first created by Dujardin, in his ‘Historie Naturelle des Zoophytes Infusoires,’ published in 1841. Since Dujardin’s original description numerous flagel- lates have incorrectly been attributed to this genus, so much so that Klebs, in his ‘ Flagellatenstudien’ (1893), says that this genus has not been defined with sufficient accuracy, that it has been confused with Heteromila and Bodo by the over- looking of the tail flagellum, and that the genus Cercomonas must be rejected. It is undoubtedly true that the genus Cercomonas is very confused, and this confusion has been considerably heightened by the description of Cercomonas from the intestine of man and other animals. Davaine (1854) was the first to record the presence of Cercomonas in the evacuations of a man suffering from cholera. Without going into the question of the correctness or otherwise of Davaine’s conclusions, it is undoubtedly a fact that many observers, noting the presence of active flagellates in the intestinal contents, have attributed them at once to the genus Cerco- monas, and as a result of this various species of 'l'richomonas, Lamblia, and possibly other flagellates have been included in this genus. In the present instance the flagellate to be 242 Cc. M. WENYON. described was found in the feces of a patient in the Albert Dock Hospital at the London School of Tropical Medicine. This patient was infected with Entamceba coli, and in order to observe changes inthe encysted forms of this amceba some of the feces were placed in a clean glass-stoppered bottle. In the course of a few days it was noticed that large numbers of flagellates were present. It is probable they had developed from cysts which must have been present in the feces. On first examination it was seen that these flagellates corre- sponded very closely with the original description of Dujardin for the genus Cercomonas, and for this I took them to be. On more careful examination I found that the tapering posterior end was in reality a second flagellum, and that this could be traced along the surface of the body to which it was attached as far as the insertion of the long anterior flagellum. The presence of this posterior flagellum and its attachment to the body required very careful observation to make out, for it can only be clearly seen in certain portions of the animal, and it ig quite conceivable, as Klebs maintains, that Dujardin overlooked this posterior flagellum. Dujardin’s original description of the genus is as follows: “Genre Cercomonas. “An, arrondi ou discoide, tuberculeux, avec un prolonge- ment postérieur variable, en forme de queue, plus ou moins long, plus ou moins filifornie. Les Cercomonas ne different absolument des Monads que par un prolongemert postérieur, formé par la substance méme du corps qui s’agglutine au porte-objet, et s’étire plus ou moins, de maniére a n’étre tantét qu’un tubercule aminci, tantot une queue allongée transparente, tantot enfin un tila- ment presque aussi fin que le filament antérieur, et suscep- tible @un mouvement oudulatoire; mais bien scuvent j’ai cru voir les Monades passer par degrés |’état de Cercomonas.” A comparison of this description with that now to be given will show how closely the two agree. The occurrence of this flagellate has been described above. By transplanting into other media I have been able to keep OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 243 cultures of this flagellate free from other Protozoa for about a year, and it is only that cireumstances preventing ine from continuing these observations I now describe what results I have already obtained. MerHop or OBSERVATION. I have found the best liquid culture medium to be hay infusion to which a small quantity of faeces has been added. The flagellates will live and multiply in hay infusion alone, but, as in other thin media, the numbers of flagellates are always very small, sothat any observation is difficult to make. In the thicker medium the numbers are not only larger but the movements of the Hagellates are slower and accordingly more easily followed. For keeping stock cultures small test-tubes were used as in bacteriological methods, but for making observations hanging-drops in the moist chambers of Max Schultze were most useful. In these hanging-drop prepara- tions the flagellates would live for weeks, till finally, all nutri- ment being used up, encystment followed. | By the addition of fresh nutriment to the hanging-drop the culture would commence again. In addition to the liquid medium I have found the solid agar medium used for the culture of amcebe most useful. It was first employed for the culture of flagellates by Berliner. This observer, working with Copromonas major, found that on the solid medium the flagellates multiplied rapidly till enormous numbers were present. I can fully confirm this, and for the study of the details of nuclear division the presence of such large numbers of dividing forms is very useful. ‘The medium I employed differed slightly from that used by Berliner. For the culture of amcebe I have used with success the medium first invented by Musgrave and Clegg, and I have found it equally good for the flagel- lates at present under discussion. I have employed it in the ordinary Petri dishes. By unveiling the dishes the progress of the culture may be watched under the low powers of the microscope. A very useful method for the 244 Cc, M. WENYON. use of this medium, and one which will allow observations to be made with high, powers, is the following: A long cover- glass (1} inches) is taken and carefully cleaned, On a clean slide ridges of Czokor’s wax, first recommended to me by Professor Minchin, are so arranged, about an eighth of an inch high, that the cover-glass will form the lid of a box. Some of the medium is melted by placing the test-tube in boiling water, and a small drop of this is allowed to fall on to the cover glass, which is lying on the top of the hot-water oven. By careful tilting of the cover-glass the melted medium will form a very thin layer over the cover-glass, which is then removed so that the medium may solidify. he surface of the medium is then inoculated with a small quantity of material from a previous culture and the cover-glass inverted on the wax ridges. By means of a hot wire and more wax the whole may be completely sealed up. It is most essential that not the smallest opening be left, or it will be found that the medium will quickly dry and the culture end. In this way it is easy to follow the multiplication of the flagellates with the in. objective, and if the film of medium has been made sufficiently thin the oil-immersion may be employed. In every case where the flagellates grow in the solid medium their chief nourishment seems to be the numerous bacteria that grow at the same time. For studying the flagellates in the fixed and stained con- dition the cover-glass method has been mostly used. . Some of the liquid medium or some of the culture scraped from the surface of the agar is spread on a clean cover-glass, and without allowing it to dry it is dropped, film side down, on to the surface of some fixing fluid. Another method of obtain- ing a film from the agar cultures is tlis: A cover-glass is dropped on to the surface of the agar culture in a Petri dish. It is gently pressed down till its surface is seen to have touched the culture. On raising it with a needle it will be found that a layer of the culture is adherent to the cover- glass, and it may be fixed as before. OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 2495 For fixing the flagellates the most useful fixative has been Schaudinn’s mixture of two thirds saturated aqueous solution of sublimate and one third alcohol, slightly acidified with acetic acid. ‘his has been used in the manner just described by Schaudinn or ina slightly modified form. The films are best stained with iron-hematoxylin. DESCRIPTION OF THE Livinac FLAGELLATES. When examined in a drop of liquid medium on a slide the Text-figs. 1-8.—Drawings from life. Text-Fi@. 1. TEXT-FIG. 2. Ameeboid form in early division Ameeboid form. stage. flagellates appear as pear-shaped organisms, with a long flagellum, about twice the length of the body, arising from the blunt end. he posterior end of the body is, as a rule, drawn out to a fine and tapering point. By the constant lashing of this long anterior flagellum the animal is drawn along. Sometimes the flagellum is, as it were, hooked around some distant object, and by its flexion pulls the body towards this point. ‘The posterior end of the body, which, as stated above, is also a flagellum, moves much less vigorously than the anterior. Its movements may be quite passive, being only the accidental changes in position produced by the changes 246 C. M. WENYON. in shape of the body. At other times there is a distinct to-and-fro or lashing movement, but at its maximum it is much less violent than that of the long anterior flagellum. The protoplasm of the body may be continued along this posterior flagellum for a considerable distance. On very careful focussing it can be seen that the posterior flagellum TEXT-FIG. 3. Two ameboid forms with entangled flagella. is attached to one side of the body, and really arises from the insertion of the anterior flagellum.% This is very well shown in some of the figures,e.g.3,5,9. When the body is viewed in certain positions it is seen that it is distinctly flattened aloug the line of attachment of the posterior flagellum (fig. 10), and when the posterior flagellum is moving at its maximum rate this flattened edge of the body shows slight but distinct undulatory movement, reminding one most strikingly of the OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 247 movements of the blood inhabiting Trypanoplasma. Indeed, this flagellate in many respects occupies a position intermediate between the genus Bodo and Trypano- plasma. The nucleus is clearly visible in the living animal. There is a distinct membrane, and at the centre of the nucleus is a large karyosome. The nuclear membrane is drawn out at one pole towards the insertion of the two flagella, and occa- sionally a clear line may be detected connecting the apex of TEXT-FIG. 4. Division-stage of free-swimming form. the nucleus with the base of the two flagella. The details of these structures are much more evident in the fixed and stained films. The protoplasm of the body contains food and other vacuoles, but contractile vacuole is not present. Sometimes the nucleus is surrounded with refractile granules, having the same greenish line and refraction as the karyo- some within the nucleus. ‘hese may be present in sufficient numbers as to completely obscure the nucleus. Similar granules occur in the protoplasm of eucysted forms (fig. 6). These granules stain deeply, and are possibly of a chromatin nature. In the hanging-drop preparations especially this organism 248 C. M. WENYON. exhibited a peculiar polymorphism. In the central part of the hanging drop, where the fluid was deep, the flagellates had the typical pear-shaped appearance, with the long, tapering, posterior extremity. At the sides of the hanging drop, where there was only a thin layer of moisture on the cover-glass, the typical pear shape was lost and the flagellates had the appear- ance of amcebe. When first I observed this I thought my culture had become contaminated with an amoeba, but the TEXT-FIG. 5. Ordinary free-swimming type. presence of the long anterior flagellum and the short posterior one disproved this idea. It was possible to watch a single individual swimming in the deep part towards the edge. On reaching the shallow part the character ot the organism changes at once to the amceboid form. Pseudopodia are pro- truded and withdrawn, and the animal creeps about in a typical amceboid manner. All this while the long anterior flagellum is lashing to and fro, but appears powerless to draw the animal across the surface of the cover-glass. It is only in the deeper part of the hanging drop that the flagellum is useful. The posterior flagellum is often not visible, and its prolonga- tion across the surface of the body is more difficult to detect. OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS, 249 When seen it is inert and only moves in a passive manner. It seems to take little share in movements of progression. On the surface of the agar medium the organism is generally of the amceboid form. At the edge of the hanging-drop preparations or on the surface of the agar it is easy to watch these amceboid forms ingesting food by surrounding objects with pseudopodia. As a rule the amceboid forms contain many more food-vacuoles than those swimming in the deeper layers. = PExXT-ETG. 7. Less regular encysted form. Reproduction is by longitudinal division. ‘There is first multiplication of the flagella, whether by new formation or division of those already existing has not been determined. The nucleus next divides. The karyosome is divided into two parts, and finally the elongated nuclear membrane becomes constricted and two nuclei are formed. After a short time the protoplasm becomes drawn out and finally a constriction appears, which ultimately. ends in complete division. ‘The process of this division is very readily watched on the cover- 250 OC. M. WENYON. elass cultures described above. Both the amceboid and the tree-living forms divide in this manner, but on account of the more sluggish movements of the former they are more readily kept under observation. In the cultures encysted forms commence to appear after a few days. In the liquid cultures they are to be found in the scum on the surface or in the deposit at the bottom. On the agar cultures the cysts appear in the older parts of the culture. On this medium the margin of bacterial growths spreads over the surface, and in this margin the actively reproducing flagellates are to be sought. In the oldest part of the culture no free flagellates can be found, but only the cyst. TEXT-FIG. 6. Text-FIG. 8. Encysted forms showing refrac- Free form with refractile granules. tile granules — surrounding Probable preparation for en- nucleus. cystment. In the fresh condition these cysts appear as slightly brownish spherical bodies, with a wall of double contour. At the centre of the cyst is the spherical nucleus, which has similar characters to that of the free form, except for the prolongation towards the flagella. ‘The nucleus is surrounded by the bright refractile granules, which were described as occurring in some of the free forms. It is probable these granules are of a chromatin nature, and that they arise from chromatin passed out from the nucleus, though this process has not been followed. ‘hough these organisms have been kept under observation for a year or more conjugation has not been seen, nor has any sexual process been detected. it is possible that some sexual process is bound up with the encystment, but as the OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 251 entrance into and emergence from the cyst has not been directly observed and no multiplication within the cyst could be seen nothing definite on this point can be stated. Text-figs. 9-19—Drawings from stained preparations. TEXT-FIG. 9. Free-swimming form with granules round nucleus. FIXED AND STAINED SPECIMENS. In the fixed and stained specimens, in addition to the details which were so clearly visible in the living organism, others could be made out. The: protoplasm of the body has a marked alveolar structure. The anteriorly placed nucleus shows a large, deeply staining Poe CG. M. WENYON. karyosome, while connecting this latter body to the nuclear membrane is a coarse linin network. All the chromatin of the nucleus appears to be concentrated in the karyosome. The prolongation of the nuclear membrane towards the Trxt-Fic. 10. Side view of free form showing the flattened side along which the flagellum runs. flagella is clearly shown, while the base of these organs is connected to the apex of the nucleus by a rod-like rhizoplast. In some cases the drawn-out apex of the nuclear membrane shows longitudinal markings, which converged toward the rhizoplast, while in others there is a connection in the form of a more deeply staining pyramin between this body and the OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 253 karyosome (fig. 11). _Prowazek describes for Cerco- monas longicaudia a “ein Art undeutlichen Zwischen- fibrille,’ which connects the karyosome to the insertion of the flagella. Prowazek figures this:connection as a dark line running from the karyosome to the apex of the nucleus, but I have not been able to detect any structure as definite as the one he figures. This flagellate is a very excellent illustration of the fallacy of relying for detail on the old dry Romanowsky methods of Taxt-rie, 11, Shows connection of karyosome and rhizoplast. staining. The nucleus of this organism is clearly visible in the living condition. There is a definite nuclear membrane. At the centre of the nucleus is a large refractile karyosome, while the space between this body and the nuclear membrane is free from granules. The nuclear membrane is drawn out at one point towards the insertion of the flagella. Now if a film of the material containing this flagellate is allowed to dry as in the usual method for the preparation of blood for staining trypanosomes, and stained by one of the modifications of the Romanowsky method, the result may be very beautiful from the colour point of view, but totally misleading in the structure of the nucleus. This latter organ appears in these VOL. 55, PAR’ 2.—NEW SERIES. il 254 0. M. WENYON. dried films as an irregular clump of red staining granules. In other words, its appearances are like those of the nuclei of trypanosomes in similarly prepared films. In films fixed and stained by the wet method described above the structure of the nucleus is comparable with the appearances to be made out in the living organisms. The details of longitudinal division can be followed in the Taxt-ries, 12, 13. Dividing forms. stained preparations. The large karyosome becomes elongated and constricted, and finally divided into two parts (fig. 15). I was never able to detect within the karyosome a centriole, spindle, and zquitorial plate, as described by Berliner in the division of Copromonas major, but the division takes place in an amitotic manner, resembling that of Copromonas subtilis (Dobell). Most usually the karyosome becomes distinctly dumb-bell shaped as in fig. 12, but at other times the division is along the longitudinal axis of the elongated karyosome, the resulting daughter-karyosomes each being elongated (figs. 15, 16). Following the division of the karyo- some the nuclear membrane elongates while the daughter- OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 255 karyosomes separate. The flagella are duplicated at this stage, but they still have a common rhizoplast, which is inserted into one point of the elongated nuclear membrane, which is drawn out slightly at this point towards the anterior end of Trxt-Fies. 14-16. : re, ots 2, ; ae * {eps a, ~ ge py = Pas a : St : + os er 4 pt < , boa, aren ae ee me LEON en ee ‘ . ving fe : LZ ere F a “re epee wos *G. eae a “a= *, : e.:3 Lee ae Q ¥ wee LS 4 é ; A Psd “n ay rae “5 a i4. len 16 . Dividing forms. the body of the flagellate. Division of the nuclear membrane commences by a constriction at the point opposite the in- sertion of the rhizoplast. ‘The division is completed, and the two nuclei, each with an apex, are connected to the base of the rhizoplast. The rhizoplast finally divides longitudinally, so that there result two nuclei, each with arhizoplast and two flagella. The exact method of origin of the flageila I was 256 Cc. M. WENYON. unable to trace, though some of the appearances seem to indicate the formation of two new ones by outgrowth from the rhizoplast. In fig. 18 is the nuclear apparatus of a flagellate partially broken up on the film. It shows very clearly the single rhizoplast with the duplicated flagella. The last stage in the division process is thus the splitting of the rhizoplast, while the first stage is the multiplication of the flagella and the commencing division of the karyosome. After complete division the nuclei pass to opposite poles of the body (fig. 17), TeExt-FIG. 18. Is. Part of nucleus, rhizoplast, and flagella of partly broken-down individual, to show the multiplication of the flagella before division of the nucleus and rhizoplast. and after a varying interval of time the body is divided into two equal parts. The bright refractile granules which were described above as occurring in the protoplasm around the nucleus in the encysted forms and in some of the free forms appear in the stained specimens as dark-staining granules. Whether these are chromatin granules of the nature of a chromidium or whether they are capable of some other interpretation cannot be definitely stated, since their fate has not been followed. They certainly stain as chromatin, and their presence within the cyst (fig. 19) would seem to suggest the possibility of OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 257 their being nuclei of spores destined to escape from the cyst and ultimately to develop, with or without conjugation, into the adult flagellate form. Though cysts have been constantly kept under observation and every inducement possible to encourage the emergence from the cyst has been tried, I have never been fortunate enough to witness this process. That it does occur is borne out by the experiment of adding dried cysts to fresh medium, resulting in a culture of flagellates. In the stained preparation certain appearances are capable of interpretation as a conjugation of the flagellates, and some of the nuclear appearances as processes of maturation, but as no undoubted conjugation was observed in the living flagel- Trxt-Fic. 19. Cyst showing dark-staining granules surrounding the large central nucleus. lates [refrain from describing these. Without the control of observation on the living forms descriptions of conjugation and the accompanying nuclear changes are of little value, since the possibility of error in interpretation is very great. For Copromonas major Berliner has described from stained preparations such a process of conjugation, but without the necessary controls it is always possible that abnormal or in- volution forms have been mistaken for such stages. In rich cultures of the flagellates there is a very great variation in size. Some individuals are comparatively large, reaching a length of 15 or more, excluding the flagella. Others are very minute, being not more than 2-3 u in longest diameter. All intermediate sizes are to be met with in the cultures. ‘he encysted forms have a diameter of about 6 or 258 C. M. WENYON. fod 7. These cysts will withstand drying at ordinary laboratory temperatures, and are capable of giving rise to fresh cultures when brought into suitable media. NOMENCLATURE. It is certain that Dujardin’s original description of the genus Cercomonas is incomplete, but it seems to me quite clear from his account that he was dealing with flagellates similar to the one described in this paper. ‘Though he did not definitely state that the fine drawn-out posterior extremity of the body was a flagellum, still, he says that it was at times so fine as to resemble the anterior flagellum, and that it was capable of independent movements. Further, in his table of classification he divides the Monads into two groups. In the first he includes forms with “un seul filament flagelliform,” while in the second those with ‘pleusieurs filaments ou appendices.” The genus Cercomonas appears in the second of these groups as a form with ‘‘un second filament ou appendice postérieur.” It is therefore quite evident that Dujardin regarded this posterior termination of the body as of the nature of a flagellum. Stein and Blochmaun describe the genus Cercomonas as having a drawn-out posterior end, though they do not describe a definite flagellum. ‘The genus Cercomonas was not accurately defined by Kent or Biitschli, and to Klebs the confusion seemed so great that he proposed the rejection of this generic name and the substitution of Gruber’s name Dimorpha, which was created for a_bi- flagellate showing at certain stages definite heliozooid characters. In this genus Dimorpha Klebs included forms which he identified. with those described originally by Dujardin as Cercomonas, and he suggests that this observer has overlooked the second flagellum. We have seen how near Dujardin was to definitely describing this second flagellum, so that the action of Klebs in rejecting this genus is hardly sound. It seems to me clear that the forms described by Dujardin really possessed two flagella, though OBSERVATIONS ON A FLAGELLATE OF CERCOMONAS. 259 he failed to see this clearly. On this account I think it safer to retain the genus Cercomonas for flagellates of the character described in this paper, viz. flagellates with an anterior blunt end from which arises a single long flagellum and a posterior tapering end also with a flagellum, trace- able over the surface of the body towards the insertion of the anterior flagellum. This conclusion is come to by Prowazek also, who figures Cercomonas longicauda with two flagella arising from the nucleus. The specific name of this flagellate is difficult to determine. Dujardin named several species of Cercomonas, though he was careful to state that he was far from regarding these as true species, but as a convenient means of distinguishing the forms met with in different infusions. From the figures of Dujardin and Stein it is possible that the flagellate belongs to the species longicauda, so that the flagellate described here may be assumed to be Cercomonas longicauda Dujardin. REFERENCES. Dujardin (1841).—* Histoire naturelle des Zoophytes Infusoires,’ Paris. Dallinger, W. H., and Drysdale, J.—‘* Researches in the Life-history of a Cercomonad, a Lesson in Biogenesis,” ‘ Monthly Micr. Journ.,’ VOlaxa Oe: Davaine, C. (1875).—** Monadiens,” in ‘ Dict. Eneycloped. des Sciences Médie.,’ t. ix. Kent, W. S— A Manual of Infusoria,’ London, 1880-82. Klebs, G. (1893).—* Flagellatenstudien,” ‘ Zeit. wiss. Zool.,’ vol. lv, p- 265. Blochmann, F. (1895).—** Die mikroskopische Tierwelt des Susswassers,” Abt. I, ‘ Protozoa,’ Hamburg. Stein, F.—‘ Der Organismus der Infusionsthiere,’ Abt. III, 2 Halfte. Prowazek, 8S. (1903).—* Flagellatenstudien,’ ‘Arch. fiir Protisten- kunde,’ Bd. ii. Biutschli, O—In Bronn’s ‘ Klassen u. Ordn. d. Tierreichs,’ 1885-87. Dobell, C.—** The Structure and Life-History of Copromonas sub- tilis, nov. gen., nov. sp.,” ‘Quart. Journ. Mier. Sci., vol. 52, 1908. 260 Cc. M. WENYON. Gruber, A. (1881).—‘“Dimorpha mutans,” ‘ Zeit. wiss. Zool.,’ Bd. XXXVI. Berliner, E. (1909).—* Flagellatenstudien,” ‘ Arch. fiir Protistenkunde,’ Bd. xv, H. 3, p. 297. SOME OBSERVATIONS ON A NEW GREGARINE. 261 Some Observations on a New Gregarine (Meta- mera schubergi nov. gen., nov. spec.). By H. Lyndhurst Duke, B.A., B.C.Cantab., With Plates 15 and 16. ConreEn'ts. PAGE Introduction : , : : : » e206 Material and Methods : ‘ : ; . 263 Structure of the Trophozoite ; : . 266 Cyst-formation and Development of the ene es ‘ = 210 Discussion of Some Special Points in the Life-cycle. 28 Diagnosis of Metamera schubergi ; 282 Literature References : : : : . 282 Explanation of Plates : : : ‘ . 284 INTRODUCTION. Wate working at Heidelberg in 1906, under Professors Biitschli and Schuberg, the latter kindly called my attention to a new species of gregarine in the gut of Glossosiphonia complanata L. (Clepsine sexoculata), and suggested its further investigation. The preceding summer, while busied with a recently discovered coccidium occurring in the leech Herpobdella atomaria Car. (= Nephelis vulgaris), Professor Schuberg turned his attention to Glossosiphonia complanata, which occurs in company with Herpobdella in the Neckar and occasional ponds in the Heidelberg district. Deeming it probable that two forms so alike in habit and environment might harbour the same parasites, he dissected 262 H. LYNDHURST DUKE. several specimens of this leech, and, though the results were in the main negative, he found several animals infected with a species of gregarine. Reference to the literature proved the parasite to be identical with a species briefly mentioned by Bolsius in 1895 (2), and the subject of a more detailed but still fragmentary paper in 1896 (3). Beyond a super- ficial study carried on incidentally during his work ou the Glossosiphonia Bolsius seems to have paid no further attention to the parasite, which remained unnoticed until 1900, when Castle (5), in an exhaustive treatise on the N. American Rhynchobdellide and their parasites, mentions having observed the gregarine seen by Bolsius in about half the specimens of Clepsine elongata which he examined. He adds, however, that he only finds the animals in the stomach diverticula, and never in the intestine or crop, as indicated by Bolsius in his diagrams. Castle also mentions encysted protozoa which he found in C, fusca, and suggests the possibility of their relationship to the form in G. com- planata. ‘he cysts he found in the muscle-layers of the body-wall, so that they probably have nothing to do with the gregarine in question. Liihe (14) quotes the parasite as having been mentioned by Bolsius, and suggests that it probably belongs to the tricystid gregarines. The gregarine is thus a new and previously undescribed form, for which I propose the name Metamera schubergi.! In the preparation of the sections and the study of the living animal, during the last few weeks of my stay in Heidel- berg, Professor Schuberg assisted me most kindly in every way in his power; and it is due solely to him that I was able to obtain Bolsius’ principal pamphlet. My thanks are also due to Geheimrat Prof. Biitschli, whose practical suggestions I found of the greatest value. 1 The form which appears most closely allied as regards structure of the trophozoite is Echinomera. A study of the life-history, however, has revealed points of difference which seem to warrant the creating of a new genus for the form under consideration. SOME OBSERVATIONS ON A NEW GREGARINE. 263 By the kindness of Professor Sedgwick, who allowed me a free hand in the laboratory of the Imperial College of Science, 8. Kensington, I was able eventually to complete my study of the sections. And in this connection I must express my indebtedness to Mr. C. C. Dobell, who is at present lecturing at the College. His unrivalled knowledge of protozoan life- history and technique has always been most generously placed at my disposal, and has proved of the greatest value in the preparation of this paper. Mareriat AND MeruHops. The leech which serves as host to Metamera schubergi is Glossosiphonia complanata Linn. A few specimens of Hemiclepsis marginata werealso found infected. The leeches live under stones in shallow water—running by pre- ference—though I have found them in smaller numbers in still pools. The material was collected at Heidelberg from the shallows left by the summer fall of the Neckar in the neigh- bourhood of the electric power station, below the new bridge, and also from the opposite bank, along the wall separating the skating rink from the river itself. The leeches are fairly common, and may be found clinging firmly to the under-side of stones at the water’s edge, especially in the numerous lumps of red sandstone which hitter the shore everywhere. Recently I examined some specimens of Glossosiphonia complanata sent me from the neighbourhood of Cambridge, and found them well infected. These latter were obtained in January, when the leeches are hard to find owing to the scanty vegetation in the ponds in winter. In all the speci- mens I examined from this source I only obtained one cyst, and that a very small and early one. The leeches can be kept for an indefinite period in a good- sized glass jar, provided the water be aérated by passing bubbles of air through it. Food is not necessary, though a ' For this I have to thank Mr. Harding, and also for his kindness in assisting me to determine the species. 264, H. LYNDHURST DUKE. few small water-snails are much appreciated. Owing to the transparent nature of the integument in Glossosiphonia, the parasites are visible in the living leech ; and if the latter be forcibly pressed between two slides provided with wax corners, and examined under a low magnification, the gregarines may sometimes be detected in the stomach diverticula and intes- tine. Unfortunately, however, this method of diagnosis is by no means infallible, as the numerous pigment-cells with their clear nuclei look very like gregarines, and render accurate observation impossible. ‘lhe gregarines occur in the hindermost stomach diverticula and the intestine, just as indicated by Bolsius in his diagram. The cysts are found in the same regions of the alimentary canal, but are especially numerous in the intestine. Kxamination of sections shows that cysts can develop as far as the sporoblast stage in the intestinal canal of the host, though they are often expelled with the feces at a much earlier stage in development. In sections just above the anus no cysts were to be seen. This part of the gut was almost occluded by a mass of cephalonts and some sporonts of a peculiarly blunt outline. The leech from which these sections were cut had previously evacuated feces containing a few very early cysts among a greater number in which sporoblasts could be distinguished. As many as ten cysts have been counted in one section. ‘To obtain the gregarine, the infected leeches were partially dried on blotting-paper and the under-surface opened by three incisions—two parallel and close to the margins, and one at right angles to the long axis of the animal, at about the junction of the middle and anterior thirds. The flap of tissue was then carefully turned backwards towards the anal sucker, the animal being placed in a watch-glass containing normal saline solution. ‘The gut-contents were thus emptied into the saline, together with connective tissue, which is of no account. By the aid of a hand-lens the gregarines could now be seen sticking to the bottom of the glass, or still fixed to fragments of the host-tissue. These latter are useful in SOME OBSERVATIONS ON A NEW GREGARINE. 265 studying the structure of the epimerite, as this organ, in the course of the teasing out, 1s very easily torn away, leaving decapitated individuals which may be confused with true sporonts. By gentle coaxing with a pipette the gregarines can be freed from the bottom of the watch-glass and trans- ferred to a slide for further handling. Preparations in toto were made originally under a cover- slip provided with wax feet, and the various reagents drawn through with blotting-paper. In this way, by fixing the gregarines with alcohol and glacial acetic acid (9 : 1), a large number of animals may be treated under one cover-slip, which is an obvious advantage. More recently I made some pre- parations by fixing the selected gregarines in a watch-glass with picro-acetic acid (3:1) and adding the various fluids by means of a pipette and eventually pickmg out and mounting the stained gregarines under a low magnification. I consider the former method of treatment the more satisfactory and certainly less laborious. As stains for these preparations I used Grenacher’s alcoholic carmine solution and Schuberg’s modification of Mayer’s acid carmine. This latter solution, being acid in reaction and not neutral, has the power of penetrating the cuticle, and in employing it the preparations must be very rapidly washed through with } per cent. solution of HCl to prevent precipitation of the carmine during the further treatment with the alcohols. Leeches destined for sections were fixed either in Gilson’s fluid or in the above- mentioned alcoholand acetic mixture. Gilson’s fluid should act for two or three hours, and the sublimate constituent be most carefully washed out with iodine-alcohol or a solution of KI in 75 per cent. alcohol. As staining reagents hematoxylin (Delafield’s) and eosin, safranin, and Heidenhain’s iron-hema- toxylin were employed. Owing to the paucity of material, the laborious expedient of applying both methods in succession on the same preparation had to be employed. It was found that hematoxylin and eosin were satisfactory for the cepha- lonts and sporonts, but gave very incomplete and misleading results with the nuclear changes of the encysted forms, which 266 H. LYNDHURST DUKE. were defined much more distinctly with the iron-heematoxylin method. All tissues were embedded in paraffin, with chloro- form as the intermediary fluid. Culture of the cysts.—To obtain the ripe spores the cysts were simply placed in the moist chamber, where, in the course of seven or eight days, the spores were developed. The cysts were either placed simply on a slide in a drop of Neckar water or under a cover-slip provided with wax feet. The cysts dehisced by simple rupture after about seven or eight days. Cysts placed in normal NaCl solution in the moist chamber did not develop successfully. STRUCTURE OF THE ‘l'ROPHOZOITE. The body is divided by septa into epi-, proto-, and deutomerite, and is elongated in form (figs. 1-6). Some individuals have a more thick-set appearance than others, especially in the extreme hinder end of the gut, where the eregarines are often crowded together. The animal measures about 150u by 45u. At the posterior end of the deutomerite there are often present indications of further subdivision of the body, and occasionally as many as three complete segments are seen (fig. 4). This segmentation is not confined to eregarines of any pecuhar build, being present in both long and short forms, and it varies in the degree of development of the segments. It was present in about a third of the gregarines examined alive in Heidelberg, and is also very distinct in the preparations of these animals made at the time. The Cambridge gregarines also showed segmentation, though it was distinctly less in evidence, both in the living animal and in carmine preparations of it. It appears to vary greatly —from the very faintest indication to quite definite septa. It must be stated in this connection that no segmented eregarines were seen in the sections of the infected leeches, though constantly found in preparations made by teasing out the host-tissues. This compels one to consider the possibility of injury during extraction being the cause of this segmenta- SOME OBSERVATIONS ON A NEW GREGARINE. 267 tion, although the stained preparations do not in the least degree support this suggestion. The epimerite is a dome-shaped structure. It is provided with short club-like processes, recalling those of Kchino- mera, but often branched, arranged in a dense ring around the line of junction with the protomerite, and also on the roof of the dome (figs. 4 and 5). These latter processes are markedly shorter than those of the ring, and decrease in size as the apex of the epimerite is approached. The processes are perforated at their somewhat clubbed ends by small pores, clearly to be seen in the freshly mounted living eregarine by the aid of a ;4; in. oil-immersion lens. Judging from analogy with such forms as Echinomera and Ptero- cephalus (Nina), and also from the appearance seen in sections across the point of fixation to the host, there is no doubt that fine pseudopodia are protruded through these pores, which fix the gregarine to the intestinal mucous mem- brane of the host. The fixing apparatus is by no means easy to identify, as, owing to the unavoidable roughness of the dissection, the gregarines are rudely torn from their moorings, and almost invariably carry away with them a crown-like fringe—derived from the host-cells—which surrounds the epimerite in the zone of the processes, and obscures all details of its structure (fig. 3). When kept under observation for some time—say an hour or so—in NaCl solution, a curious phenomenon ensues. Just at the line of junction between the protomerite and epimerite a bubble-like vacuole appears, which gradually increases in size, and carries with it the fringe of host tissue with the embedded processes till they sit lke a crown on its upper pole, sometimes symmetrically, sometimes displaced to one side. Having reached a diameter about equal to that of the protomerite the vacuole bursts, and the gregarine is suddenly deprived of its epimerite (fig. 2). This vacuole formation has been seen by Léger and Duboseq to occur in Pyxinia (14), and in my opinion has a probable bearing on the mooted question regarding the fate of the gregarine epimerite, in the 268 H. LYNDHURST DUKE. transition from cephalont to sporont. Frenzel (14) believed the epimerite to be absorbed in a manner similar to the assimilation of a tadpole’s tail. He found among numerous cephalonts with large epimerites individuals with but a minute projection from the protomerite, and he regarded this as a scene in the gradual absorption of the epimerite. The sudden disappearance he regarded as pathological, and due to changes in the surrounding medium. My own observations point to the same conclusion. ‘The vacuole formation quoted above is plainly due to plasmoptysis, which can be followed under the microscope from its earliest onset to the bursting of the bubble. Further, when the gregarines were examined in a special solution of egg-albumen, NaCl and camphor, as prepared by Professor Biitschli, the vacuole formation was considerably delayed; a fact explicable on the ground that the solution more nearly resembles the natural environment of the gregarine. The behaviour of the finger-shaped processes also points to the epimerite being absorbed rather than directly thrown off when the cephalont becomes free. In gregarines which are normally lying free in the gut the processes are never to be seen (figs. 1 and 6). The epimerite is still present, but the processes have been withdrawn during the process of separa- tion from the mucous membrane ; just as they are absorbed in Echinomera when the cephalont becomes free in the gut (17). This applies to all the free-lying specimens seen in sections, and to a solitary living form which, together with several cysts and some feces, was pressed out through the anus during examination of a leech between two slides (fies Di In the living sporont (fig. 1) the extreme anterior end of the animal is quite transparent and devoid of granules, a few of which, separate from the main endoplasmic mass of the epimerite, may be seen showing Brownian movement along its anterior border, After some time the whole granular body of the gregarine appears to shrink back somewhat into the cuticular sheath which envelopes it, and this clear area SOME OBSERVATIONS ON A NEW GREGARINE. 269 enlarges proportionally until almost the whole of the conical knob which forms the epimerite is clear of granules. During this process all three divisions of the endoplasm are still quite distinct. By the time this stage has been reached osmosis asserts itself, and the vacuole formation mentioned above commences (fig. 2). In sections, however, the free-lying sporonts all showa curious thickening of the extreme anterior end of the epimerite, which behaves towards stains in the same way as the rest of the cuticle, being, in fact, a thickening of the latter anteriorly (fig. 6). It seems a feasible explana- tion of this structure to say that it represents the cuticular constituents of the numerous processes of the epimerite, which have been retracted on the animal becoming free. It may here be mentioned that Liithe (14), in his review of the gregarines generally, pronounces in favour of the casting off of the epimerite as the typical way in which the cephalonts become free. The nucleus lies in the deutomerite. It consists of a nuclear membrane enclosing a clear ground substance, in which lie a large vacuolated karyosome and a number of masses of chromatic substance (fig. 7). The specimens from which figs. 3 and 4 were drawn were very faintly stained owing to excessive washing out, but some other preparations stained with Grenacher’s carmine confirm the appearancesseen in sections, especially as regards the vacuolated nature of the karyosome. ‘The nuclear area is about 18 in diameter; the karyosome measures about 8 u, and as a rule contains one very large vacuole and several small ones. The large chromatin masses are scattered irregularly throughout the nucleus, and are of varying shape. The nuclear membrane is well marked, and in common with the karyosome and the chromatin masses stains deeply with both Delafield’s hama- toxylin and Heidenhain’s iron-hematoxylin. The ground substance takes on a very faint blue tinge with iron-hema- toxylin. In some of the sections the karyosome has yielded almost completely to the differentiating iron alum, and appears grey by contrast with the black chromatin masses. In VOL. 55, PART 2.—NEW SERIES. 18 270 H. LYNDHURST DUKE. these cases its vacuolated structure is very plain (fig. 7). As a rule, however, the karyosome shows very deeply stained in the adult nucleus. Besides the nucleus there are usually to be seen scattered throughout the body patches of a substance which stains deeply with chromatin stains. These patches have been described by Berndt (1) and others, and are espe- cially numerous in the protomerite. Comes (7) has recently shown that these appearances in Stenophora are probably due to metabolic products, and are not nuclear. There are also deeply stained granules in connection with the epimerite processes in sections stained with iron-hzematoxylin, as described by Schellack in Echinomera hispida (17). Cyst-FORMATION AND DEVELOPMENT OF THE SPORES. The act of association of two animals to form a cyst has not been observed inthe livinganimals. As indicated above, in the sporont the epimerite tends to become less prominent, while a pad of cuticle forms anteriorly. Simultaneously with this shortening of the long axis of the body the protomerite increases in breadth and bulges, particularly around the edges of the apical cuticular pad. From sections it would seem that the two animals come together with their epi- merites in contact. A ring of cuticle now arises around the base of the terminal pad in one animal. Into the cup formed by this ring the cuticular pad of the other gregarine is inserted, while external to, and dovetailing with the ring of the cup, a similar ring of cuticle arises in the second animal (fig. 37). In very young cysts in which the nuclei of the two animals are still unaltered the above arrangement of the parts is very clear; but as development proceeds the septum of cuticle dividing the encysted sporonts becomes increasingly irregular. In this region in the earlier cysts there are patches of deeply stained material suggestive of membrane, which are probably the remains of the cuticle of the contiguous epimerites (fig. 15). Behaviour of the nucleus preparatory to the SOME OBSERVATIONS ON A NEW GREGARINE. paral formation of the first two daughter-nuclei.— Although the material which I was able to collect was very limited, I was fortunate in obtaining one leech very heavily infected. In the intestine of this animal I found numerous cysts, and also an enormous number of adult gregarines mostly fixed to the gut-wall. A study of these sections has revealed several phases of the first division of the nucleus, though to elaborate all the stages is impossible without further examples, which I hope shortly to procure. In order, therefore, to make the most of this limited material, Iemployed first hematoxylin (Delafield’s) and eosin, and then after decolorisation with acid alcohol, re-stained by Heidenhain’s method. This latter method revealed numerous important facts quite indiscernible with the original staining. My thanks are due to Dr. Pembrey, of Guy’s Hospital, who very kindly provided me with all the necessary apparatus for staining. For some time at any rate after a definite cyst-wall has formed, the nuclei of the encysted gregarines remain appa- rently unaltered. Then the chromatin masses begin to frag- ment, with the result that chromidia are formed within the limits of the nuclear membrane. Simultaneously, this mem- brane becomes increasingly thin, and the karyosome throws out masses of substance from its interior, becoming in con- sequence markedly reduced in size. ‘These masses are more or less spherical and of distinct outline; they stain very deeply, showing black with iron-hematoxylin. Their number and size vary greatly (figs. 9-14). At times one large mass is present, almost equal in size to the original karyosome; at others, numbers of small massesare seen. The actual process of extrusion of one of these masses is shown in fig. 36. After their extrusion, the main karyosome-relic shows a blue colour with hematoxylin and eosin, as contrasted with the more purple hue shown by the intact karyosome and the chromatin masses of the trophozoite nucleus. The extruded masses on the other hand behave throughout, as regards stains, like the chromatin masses. After the fragmentation of the 272 H. LYNDHURST DUKE. chromatin masses and the breaking up of the karyosome have proceeded for some time, a new structure appears in the nucleus. In close proximity to the main karyosome residue, which is seen lying near the periphery of the nucleus, an ill-defined mass appears which takes up nuclear stains very definitely. The earliest appearance of this mass is shown in fig. 9 before the chromidia formation has progressed very far. Fie. 2.—Same sporont as Fie. ih showing bub] vle-formation. 5 5 D5 Fig. 3.—Cephalont with epimerite embedded in fragment of host- Fig. 4.—Showing optical section of epimerite. Fig. 5.—Diagram of structure of epimerite, ete. Fig. 6.—Diagram of sporont with cuticular pad on epimerite. Fig. 7.—Nucleus of trophozoite. Fig. 8—Nucleus showing fragmentation of chromatin masses and extrusion process of karyosome. Fig. 9.—Sporont nucleus showing earliest appearance of the “ achro- matic mass,” with fragmentation of the karyosome. Figs. 10 and 11.—Successive sections of another nucleus showing slightly later stage than fig. 9. These three figures (9, 10 and 11) are drawn from same cyst. Fig. 12.—First division of sporont nucleus showing at (a) the ring arrangement beginning at the pole; also the streaks of chromatin and the spindle-fibres in optical section. The two poles are respectively at the extreme upper and lower focus. One of the chromatin streaks is seen running into the polar aggregation. Fig. 13.—An early cyst, containing two associated individuals, with remains of epimerites seen at the centre. Nuclei at stage of first division. In upper animal the polar aggregation and the chromatin streaks are very marked. (Combined from two successive sections.) SOME OBSERVATIONS ON A NEW GREGARINE. 285 Fig. 14.—First division of the sporont nucleus at a somewhat later stage than figs. 12 and 15. Shows polar vesicles more distinct. Also the distinct fibres running down into neighbourhood of original nucleus and karyosome. Fig. 15.—First division of sporont nucleus at a later stage than fig. 14. Vesicles fully formed and fibres running down towards karyosome. The vesicles here shown were 6, apart, lying respectively at top and bottom focus. Fig. 16.—Diagram of first spindle just before final separation of first two daughter-nuclei. Fig. 17, a, b and c.—Earliest stage of daughter-nuclei, eight or nine in cyst. a. Shows centrosomes connected by a thick band. b. Shows chromatin bunched as an early spindle figure. c. Shows karyosome. All from same cyst. Fig. 18.—Somewhat later daughter-nuclei at end of division. a. Shows two centrioles at each pole; also one daughter-vesicle. (The section has not passed through the left vesicle.) b. Shows division of the centriole with poorly developed daughter- vesicle. (The vesicle at the right end of the figure lies outside the plane of this section, and is therefore not seen.) c. Shows a separated daughter-vesicle. All from same cyst. PLATE 16. [Figs. 21-34 and 36 were fixed with Gilson’s fluid and stained with Heidenhain’s iron-hematoxylin. Figs. 19 and 20 were stained with Delafield’s hzematoxylin and eosin. Figs. 19-34 were drawn at magnification of 2000. Fig. 35 is not to scale, being relatively too large. | Fig. 19 (a-e)—From same cyst. Somewhat later daughter-nuclet. All show karyosomes. cand d show early stage of spindles, and the karyosomes in a state of activity. Fig. 20.—Showing differentiation of centrosome into centriole, and centrosphere in a daughter-nucleus of same stage as fig. 19. Fig. 21.—Similar daughter-nuclei showing karyosomes: also corre- sponding spindle. Figs. 22 and 23.—Later stages of daughter-nuclei, mostly showing karyosomes; also corresponding spindles. VOL. 00, PART 2.—NEW SERIES. 19 286 H. LYNDHURST DUKE. Fig, 24.—Smaller daughter-nuclei and spindles. Fig. 25.—Shows the Perlenstadium, with «a single free gamete. Notice the clear karyosome remnants (a), and the residual nuclei (5). Fig. 26.—Gametes. Fig. 27.—A zygote with two unfused nuclei. Figs. 28-32.—Sporoblasts. Figs. 29 and 31 show these in transverse section. Fig. 33.—Shows a sporoblast assuming shape of spore. Fig. 34.—Shows a spore coat in process of developing. Fig. 35.—Fully formed spore, with sporozoites in optical section, Fig. 36.—Shows the karyosome in the act of extruding some of its substance. Fig. 37.—Diagram to show method of apposition of associating sporonts in a cyst. u ' ™ e co , vs WG «ot - . . y - ee Zl ¥ bd * i ° ‘ S = Quart. Iounn.MiorSci VAS NETS. 17a. 17e. Hath, Lith? Tendon Quant. Lurn Maier: ek uot. oo NV, ELL IG Huth Lith! Londen ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 287 On the Anatomy of Histriobdella Homari.’ | CO oes joe) co On D 10. By Cresswell Shearer, M.A., Trinity College, Cambridge. With Plates 17-20 and 5 Text-figures. ConrTENTS. . Introduction, Material, and Methods . Review of Literature and Remarks on Habits . Description of the Nephridia A. First Nephridium B. Second Nephridium c. Third Nephridium. p. Fourth Nephridium . Body-cavity and Nephridia . Muscular System A. Longitudinal Muscles B. Special Muscles of the Generative Serine c. Oblique Muscles . Digestive System . Nervous System . Sense-Organs . Reproductive sean. A. In the Male B. In the Female Conclusion and Summary PAGE 288 291 300 303 303 304. 304 305 308 308 309 311 314. 321 327 328 328 334. 346 1 T have tothank the Director and members of the staff of the Marine Biological Association of Plymouth for their kind attention and interest in my work while at Plymouth. 288 ORESSWELL SHEARER. 1. Inrropuction, Mareriat, AND Merruops. Our knowledge of the anatomy of Histriobdella is based on the papers of Van Beneden (1858), Foettinger (1884), and Haswell (1900). Of these, Foettinger’s account is the most extensive, while Haswell’s paper is perhaps the most valuable. Both accounts contain a more or less detailed description of the internal structure and organisation of the adult. Several years ago I described the presence of solenocytes in con- nection with the nephridia of Dinophilus. This dis- covery rendered it probable that these peculiar structures would also be found in Histriobdella, with which Dinophilus shows many relationships. Moreover, the different description of the nephridial system given by Haswell in Stratiodrilus from that of Foettinger for Histriobdella called for a re-investigation of these organs. For these reasons the present work was begun. I was soon led to undertake a detailed examination of the animal. It is some twenty-five years since the publication of Foettinger’s paper, and during this interval the European species of Histriobdella has received no further attention. In the following account I have endeavoured to clear up Foettinger’s description of several of the organs. I have had the advan- tage of having made use of the methyl-blue method of intra vitam impregnation, which has proved most valuable. With its use I have experienced no difficulty in determining the number of the nephridia and their relationship to the segments, and to make out new details in their structure quite impos- sible from ordinary sections of fixed material. Good methyl-blue! preparations of the nephridia can be obtained by placing the lobsters bearing the parasites in small tanks of sea-water, to which sufficient blue has been added to colour the water a light shade. It is necessary for the animals to remain in the blue two or three days before it appears in the nephridia. As the blue is rapidly absorbed by the living tissues of the lobster, an additional quantity has 1 This is * soluble blue,” and not methylene blue. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 289 to be added to the water from time to time. With good air- circulation anda little attention, a medium-sized lobster can be kept alive for several weeks in a tank of four or five litres capacity without change of water. At the end of the second day the blue will have collected in dark granules on the walls of the nephridial canals, so as to outline these clearly. By this time it has been discharged from the nervous system and the sensory cells of the epidermis. About the bases of the legs of the head it shows a tendency to remain some time after it has disappeared from the brain. It is retained alone by the nephridia on the third day. Here it collects in dense masses on the courses and openings of the canals. In the study of these methyl-blue breeaeioms I fave made use of long, thin cover-slips, such as:are used in pre- paring large serial sections instead of ordinary slides, on which to mount my preparations. The use of a thin cover- slip used as a slide allows of the preparation being examined from each surface, as desired, under an oil-immersion lens. It is thus possible to trace a nephridium first. on one side of the preparation, and then turn the slide over and trace it further on the other surface. Histriobdella is a somewhat difficult animal to fix. The only reagent that has given uniform results is a saturated sublimate solution, with 5 per cent. acetic, used boiling hot. Hermann’s solution and Flemming and the osmic: acid mixtures give very irregular results, and are. not to be depended on for their action. One lot of material will be excellent, while the next, fixed with the same solutions and under the same conditions, are useless. Picro-acetic and Bouin’s solutions, used hot, give good results, but not as good as material fixed with sublimate-acetic. Picro-sulphuric was used for preparations to be studied whole, on account of the excellent preservation it gives of the external form. As stains, the following have given satisfaction: Heemacalcium and Benda’s iron-hematoxylin, paracarmine, lithium-carmine follewed by Lyon’s blue for eggs. 290 CRESSWELL SHEARER. The nephridial canals are remarkably difficult to recognise in sections on account of the retraction they invariably undergo during fixing. It is impossible to trace them with any degree of certainty through consecutive sections. For this reason I have relied mainly in my investigation of the nephridia on methy]l-blue impregnation preparations of living material. The figures accompanying the present paper there- fore represent the appearance of the nephridia in living material. It is necessary to use the highest powers of the microscope to determine the structure of the nephridia, and even then the eye requires considerable practice and training to distinguish the motion of their cilia. It is difficult to convey any idea of the extreme delicacy and minuteness of these structures. The necessity of being compelled to use immersion-lenses for their study excludes the use of any of the ordinary dark ground systems of illumination. Doubtless these would offer an excellent means of investigating struc- tures of this nature in an animal so transparent as His- triobdella, if they could be used successfully with the immersion-lens. Of great service in the study of the methyl-blue pre- parations is, I have found, the use of a number of sodium glass screens of different shades, such as are used in ortho- chromatic photography to vary the exposure from five to fifteen times. To obtain a uniformly constant light I have used an ordinary Welsbach gas lamp, with standard screens. ‘This gives a light much superior.to that of ordinary daylight in bringing out the finer structure of the nephridial canals. For sections I have used the ordinary paraffin and the paraffin celloidin method. The sections were cut of the uniform thickness of 7.4. Inthe reconstructions of the nervous system shown in figures I have used a method which is in part a modification! of that described by Woodworth (‘ Zeit. f. wiss. Mik.,’ xiv, 1897, p. 15). Each section, of which there 1 This I owe to my friend, Mr. E. W. Nelson, of the Marine Biological Association, Plymouth. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 291 were about eighty, was first outlined on paper with the aid of a camera lucida, and the nervous system carefully marked in. Each of these drawings was then measured transversely from side to side, and the measurements plotted out on milli- metre paper, allowance being made for the magnification between sections. ‘The nervous system was also measured, and likewise put in, all the distances being doubled to give an axial line. ‘lhe ends of the plotted points were then joined up, and an outline of the external form and the nervous system obtained. ‘The figures were then reduced to their present size, and at the same time transferred to ordinary drawing-paper by means of an eidograph. In the sagittal section shown in fig. 15 the dorso-ventral diameter was taken instead of the transverse. By this means the relationship of the ganglia to the segments can be accurately determined in a way that would be impossible with the ordinary recon- struction methods (figs. 15, 21, 28). _ 2. HisroricaL Review and GENERAL Remarks on Hapsirs, ETC. Histriobdella was discovered and briefly described by J. P. van Beneden (1) in 1853. He found it as a parasite on the eggs of some lobsters obtained from Ostend. He con- sidered it a larval Serpulid, and placed it among the Poly- chete. Subsequently, in 1858, he (2) pointed out that it was an adult form. From its peculiar structure he remarked that it could not be easily classed with any known group of animals, although some of its features he thought were such as to place it among the leeches. He gave a more or less detailed description of both sexes, and’figured the eggs and immature young. To Foettinger (8) we owe the most extensive account of this animal. He describes the nervous system, nephridia, repro- ductive organs, and, in fact, was the first to give a detailed account of its anatomy based on sections. He supported the conclusions of Edouard van Beneden that it was an Archi- annelid, placing it near Polygordius, but separate from it, 292 CRESSWELL SHEARER. in the family Histriodrilides. In his opinion many of its characters show its inferiority in organisation to Proto- drilus. Among these the absence of any trace of the circulatory system, the feeble internal segmentation, marked by the complete absence of dissepiments and the small number of segments. On the other hand, the presence of well- developed ganglia points towards a higher organisation than that possessed by any known Archiannelid. Again, the com- plicated sexual apparatus of the male is different from any- thing at present found in this class. ‘lhe presence also of chitinous jaws with striated muscles and the anterior and posterior feet he considered as distinguishing it as a type superior to Polygordius. More recently Haswell (13) has obtained, as already men- tioned, a freshwater species from the branchial chamber of a Tasmanian crawfish. In the possession of cirri it differs externally slightly from Histriobdella. Haswell pointed out, among other new features, that the lateral organs which Foettinger considered penes are in reality organs that func- tion as claspers, while the penis, as in Dinophilus, is a median unpaired structure. While Foettinger described the seminal vesicles he was unable to trace their ducts to acommon receptaculum as Haswell has done in Stratiodrilus. The nervous system of Stratiodrilus seems to be different how- ever from that of Histriobdella, in being more highly differentiated. In Histriobdella the ventral nerve-cord is still in complete continuity with the epidermal layer, while in Stratiodrilus it is situated much deeper. ‘This difference may be in great part due to the close union of the epidermal and sub-epidermal tissues in contrast to those of Stratio- drilus. A more important difference is the separation shown by the two component halves of the ventral nerve-cord in Histriobdella, and the somewhat different position and number of the ganglia. The greatest difference, however, is shown in the excretory system. It is impossible to reduce this to a common type. In Stratiodrilus the crossing and branching of the canals ON THE ANATOMY OF HISTRIOBDELILA HOMARI. 293 in the anterior region, and their course in some instances through more than one segment, seems to preclude any com- parison with Histriobdella.' Again, in Stratiodrilus the interior feet are retractile, and can be completely drawn into the head. ‘his is not the case in Histriobdella, where the distal joint alone is retractile. The main mass of the foot is incapable of retraction, even under the action of strong reagents. Histriobdella was found by van Beneden and Foettinger on the eggs of the European lobster, and was considered by them a parasite on these alone. It is, however, like Stratio- drilus, normally an inhabitant of the branchial chamber and gills. It passes to the eggs of the female from the gill- chamber when these happen to be present,” returning to the same situation when the eggs are hatched and the egg-mem- branes shed. In the branchial chamber it is quite difficult to detect at first,on account of its almost colourless condition and the fact that in this situation it does not show the excitable movements exhibited while on the eggs, but crawls slowly, keeping close to the mucous membrane. Examination of the branchial surtace of the carapace, however, once the eye has become accustomed to distinguishing them, seldom fails to show their presence in this situation in either of the sexes. They prefer the carapace to the gill surface, as it affords a better footing, and the long hairs under which they move prevent their being readily brushed off. ‘To the bases of these they attach their eggs in great numbers, especially towards the margin of the carapace, where the hairs are long and numerous. Comparison of the parasites from the “ berry ” with those from the chamber shows no difference between them, except that the jaws of the parasites from the chamber 1 Professor Haswell informs me that since the publication of his account of Stratiodrilus he has re-examined the nephridia and has re-confirmed his statements regarding them. * According to Herrick this takes place once in two years. “The Reproductive Period in the Lobster,” ‘ Bull. of U.S. Fish Commission,’ vol. xxi, 1901, p. 161. 294, CRESSWELL SHEARER. seem a little better developed than those of the “berry.” The parasites are evidently able to migrate rapidly from one situation to the other. On female lobsters whose eggs are about to hatch, many of them have already migrated to the gill-chamber. A certain number, however, are always to be found on the old egg-membranes, although the eggs have been hatched and the membranes are much discoloured with age, showing that the breeding period had passed some time. I have taken females in this condition, and placed thein in tanks with air circulation and kept them under observation. In the course of several. weeks the membranes drop off, but no parasites are found about the tank, showing that they have all taken refuge in the gill-chamber. In the gill-chamber and on the eggs both sexes are present in equal numbers. When the lobster ova are well advanced and about to: hatch, the male Histriobdellid would seem to preponderate over the female. On the ova the immature young are found in greater numbers than in the gill-chamber. Frequently a large female can be seen carrying a male attached to its back by means of its claspers. ‘These would seem to throw out some sticky secretion, for once the male has taken hold of the female it is unable readily to let go, and gets carried about by the female although it makes violent efforts to free itself. In the gill-chamber, as on the eggs, the parasites show the same tendency: to collect in small groups, huddling close together and crawling over and over one another. When disturbed they separate, to re-collect shortly in another group. Why they do this is not obvious, as the individuals are some- times all males or immature young, in which the sexual organs have not yet developed. ‘This habit of collecting in groups therefore can hardly be for the purpose of the impregnation of the females. I have examined a considerable number of ‘“ berried” crabs and rock lobsters, both at Plymouth and Naples, without finding Histriobdella. They would seem to be exclusively confined to the lobster. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 295 Nothing is known regarding geographical distribution beyond the fact that Histriobdella is common on the lobster of the Channel region. J. P. van Beneden (2), as already mentioned, obtained it at Ostend. He also states in his second paper that he had observed it on the lobster of the Norwegian coast. I have been unable to find it on the lobster at Naples. ‘My observations were, however, limited by the rarity of this animal in the Bay of Naples. I only had the opportunity of examining a few adults. So far it has not been reported as occurring on the American lobster.!. It is remarkable that an animal of such peculiar structure should be represented in Europe bya single species, while its nearest ally should be found in fresh-water streams of Tasmania. Little is also known of its life-history and habits. If a small mass of lobster ova with the parasites is placed in a watch-glass of sea-water, it will be noticed that they never crawl on any foreign body brought in contact with them. When left to themselves they collect in groups, twisting their bodies together, and remaining quiet for long periods. On being disturbed they show singular excitement, twisting themselves violently and throwing their heads rapidly from side to side, all the time remaining firmly attached by their powerful hind legs. From time to time they can be seen to bite one another with their strong jaws. While the parasite can be obtained from the branchial chamber or “‘ berry” of almost any lobster on the coast of England, the manner in which it gains access and passes from one host to another has not been determined. Like most parasites, it has limited powers of locomotion, being unable to swim, and crawling very slowly. It has no larval stage that might assist in its distribution. The eggs are attached in capsules to the lobster ova, and the young undergo their entire development within this capsule, emerg- ing in almost the adult condition. ‘There can be no larval 1-Professor Herrick informs me that he has never found it on the American lobster. 296 CRESSWELL SHEARER. stage during which it can live, either internally or externally, on some other host. The parasites are able, however, to pass from one host to another without apparent difficulty. ‘his can be readily demonstrated by placing a lobster in a solution of neutral rose in sea-water until the parasites it bears are stained, and then placing it in company with a number of normal unstained lobsters. In the course of a day many of the stained parasites will be found to have gained access to the normal lobsters, while many unstained parasites will be found on the stained lobster. This takes place readily in large tanks where the animals have room to keep well apart. How this passage is accomplished under these conditions I have been unable to observe, as the female lobster is very shy when ‘in berry,” and unsociable, strictly avoiding its mates and companions. Both Foettinger and Haswell have drawn attention to the remarkable chitinous jaws with which Histriobdella is furnished. Haswell has made a careful study of these in Stratiodrilus, and has shown how the movements of the component parts of the mechanism are brought about. In Histriobdella the jaws are almost identical, as far as I can determine, with those of Stratiodriius. Foettinger repre- sents them as furnished with many more teeth than I can find to be the case. Their use is not known, as neither Foettinger nor Haswell have made any observations on this head. Unfortunately the intestinal contents are reduced to such a fine amorphous condition as to afford no evidence as to the animal’s food. It is probable that the parasites feed on small alge to a certain extent, as the intestinal contents are usually of a greenish tint. Diatoms occasionally are present, and in some instances would seem to compose the greater portion of the food. This is so in the case of the parasites living on the “berry.” In the parasites of the gill-chamber they seem absent, and the intestinal contents consist of a fine brownish mass, among which reddish granules are seen. It is certain that the jaws are not used for tearing the membranes of the lobster’s ova as has been supposed, ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 297 When the animals are excited they have a rapid way of © opening and closing the jaw teeth, but they are never seen to use them to tear open the lobster ova. When suddenly disturbed they sometimes secure themselves by means of their jaws to the egg membranes. They possess the power of protruding the jaw apparatus considerably beyond the mouth orifice, and in crawling they are sometimes seen to seize some object in front of them by protruding their jaws in this manner, after the fashion of many Polychets. Histriobdella is remarkably sensitive to any changes in the sea-water. The circulation of water through the branchial chamber of the lobster insures their receiving a continual change of water under normal conditions. Likewise, on the “berry ” the water is kept in constant circulation round them by the ceaseless motion of the lobster’s swimmerets. With any slight impurity of the water they fall off their hosts, and are found on the bottom.of the tank in a half paralysed condition. They are quickly killed by the addition of small quantities of fresh water, and die very readily when exposed to bright light. This is of interest when it is recalled that Stratiodrilus is found in fresh water. Fertilisation takes place internally. ‘The male drives its penis through any portion of the body-wall of the female. In one instance I saw a male drive its penis into the head and discharge a considerable quantity of spermatozoa. These could be seen under the microscope working their way down into the generative segment. In many cases the males fertilise young females without eggs, and the spermatozoa apparently remain in the body till the ova develop. Many females can be observed carrying spermatozoa but no eggs. The female exercises apparently little choice in the selec- tion of a site in which to deposit her eggs. On the “ berry ” these are usually attached to the membranes of the lobster ova, while in the branchial chamber the carapace side is the one selected. They are usually deposited in groups of four or five, and this would indicate that these are all deposited at one time by the female. The eggs are all of one size, and it is 298 , CRESSWELL SHEARER. impossible to distinguish the sex of the immature young. They develop at once on being laid, showing that they have already been fertilised within the body. Even when the female is isolated in pasteurised sea-water the eggs develop immediately on being deposited; no sperm can therefore fertilise them in the sea-water. The egg-laying is done at night, as every morning fresh capsules are to be seen adherent to the coatings of the lobster “berry.” ‘The eggs within these are always in segmentation or gastrulation stages. ‘hey are laid in great numbers, so that it is easily possible to obtain all the stages of develop- ment up to the time the young worm leaves the capsule. Development is direct and would seem to be rapid, for by the end of the fortieth hour the young are fully formed and appear.ready to quit the capsule. On leaving this they move about the gill-chamber or pass immediately to the lobster ova, where they soon attain maturity. They are readily distin- guishable at this time by their small size and the undeveloped condition of the generative segment. The young of both sexes resemble the female in shape. Van Beneden (2) has figured a number of the young stages, and Haswell (18) mentions that he has obtained a number of the stages in the development of Stratiodrilus. Regarding the nephridia, Foettinger (8) stated that in the male there were five pairs of these organs, while in the female there were four. Each nephridium consisted of an intra- cellular tube running backwards on the border of the longi- tudinal muscle-strands. They turn in sharply towards the median line, to terminate ventrally, on the surface of the succeeding segment to that in which they arise, in a small pore. He could observe no internal openings or funnels, Their heads at their point of origin are on the dorsal surface; since they terminate on the ventral surface they run back- wards in an oblique plane between the dorsal and ventral muscle-bands. The first pair arise in the neck segment close to the head, and run backwards to terminate on the ventral surface of the second segment. The second pair arise in the ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 299 posterior portion of this segment, and terminate in a similar manner in the third segment, ‘he third pair arise in the third segment to terminate in the fourth. In the female the third and fourth pairs overlap, while in the male the fourth pair arises much farther back between the posterior portion of the fourth and the anterior border of the fifth segment. In Stratiodrilus, onthe contrary, according to Haswell, the nephridial system would seem to extend into the head region. Hach nephridium at its anterior end divides into an external and an internal branch. The external branch runs forward into the head, while the internal crosses over to join the internal branch of the opposite side. From the fact that the motion of the cilia of this pair of organs is always from behind forward, their openings are probably in the head. The other nephridia are not branched. ‘In the female an apparently continuous line of cilia is traceable backwards on each side from the Lead canals to a point some little distance behind the second cirrus, where a canal is clearly traceable, which, after bending round in a loop, opens on-the exterior on the ventral side. But as the direction of the movement of the cilia is from before backwards in the posterior part of this line, it would appear probable that there are two pairs of canals in this anterior region in the female. In the male, on the other hand, there is no such evidence of division, the pair of nephridia which branch in the head being traceable back- wards, without change in the direction of the cilia, nearly as far as the bases of the second cirri, at which point they bend in and terminate in the ccelom in the middle line.” In the fourth segment, according to Haswell, it is probable that the oviducis represent the nephridia, while in the male they are represented by the vasa deferentia. In both sexes, in the fifth segment there is a pair of organs (beginning in a loop in the male) which run back in the caudal region to terminate near the anus. ‘I'he direction of the movement of the cilia in these organs is from behind forwards. ‘Thus, in the male there are three pairs of organs, while in the female there are four ; so that the nephridia do not partake of the metamerism 300 GRESSWELL SHEARER. of the body, Stratiodrilus having the same number of segments as Histriobdella. In no part of the canals were ciliary flames observed. 3. GENERAL DESCRIPTION OF THE NEPHRIDIA. From the inspection of figs. 1, 7, and 9, it will be seen that the nephridia have much the same positions as those assigned them by Foettinger (8). Apparently in the male the fourth pair, figured by him in the genital region, have no existence. Like the female, the male has only four pairs of organs. It will be seen that they are the narrow, delicate, §-shaped structures he has described (figs. 4, 5, 6, 10, 14), running in the mesodermic tissue of the body-wall. Their position in sections can be seen in figs. 37 and 43. Each organ takes its origin in a small space—a prolongation or part of the general blastoccelic cavity that surrounds the gut—in the anterior portion of the segment to which it properly belongs, and runs back to terminate on the ventral surface of the following segment near the median line. It arises in a knob- like process that projects slightly into the space. This process is thick-walled, and sometimes contains refractive granules. It is shown in fig. 14. Its structure is difficult to determine, and especially the relationship it bears to the space. What I take to be the real head of the organ is shown in section in fig. 42. Here the space into which it projects is surrounded by darkly staining nuclei. These are not seen in the living condition. It bears no cells that have any resemblance to solenocytes, and these structures would seem to be entirely absent in Histriobdella. In a number of preparations it was obvious that the internal ends of the canals were closed, and that they did not open into the space into which they project. The main portion of the nephridial canal is a thin-walled intra-cellular tube, the anterior end of which contains a few refractive granules and nuclei. It runs directly backwards in an oblique plane, and is much longer than the terminal ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 301 portion. It reaches its greatest length in the case of the second nephridium (fig. 14). Frequently the lumen can be seen to be enlarged into small spaces or lacune, These would seem similar to the spaces I have described on the nephridial canals of Dinophilus. A number of these are usually seen on the course of the second organ (fig. 6), One large one is often found on the posterior part of the third. From the terminal portion of the canal they seem to be absent. The nephridial flagella pass through their centre while their walls themselves are unciliated. It is possible they are due to the somewhat abnormal conditions under which the parasites are kept in the process of their impregnation with methyl-blue, as I have never been able to observe their presence in the unstained living Histriobdella, although something like their appearance can be detected in sections. The terminal portion of the nephridial canal turns towards the median line, close to which it ends in a darkly staining pore (fig. 14). About this the blue usually collects in thick granules, which can sometimes be seen vibrating to and fro in the fluid escaping from the canal. The lumen of this terminal portion is greatly restricted in size. Throughout the length of the nephridial canal the ciliary action of the flagella in their interior can be plainly observed during life. ‘The movement of this is always in the one direction—from before backwards—and I have never observed any reversal of this motion as Haswell has described in Stratiodrilus. Despite numerous observations, I have been unable to determine whether the flagella are derived from the walls of the canal or from the knob-like head of the organ. In some preparations they seemed derived from the wall, in others they seemed derived from the nephridial heads. In sections they would seem to be derived from the walls. In no case can the canals be seen dividing, as Haswell has described in the case of the first pair in the male of Stratiodrilus. I am quite positive in saying no such division takes place in Histriobdella, They run through veL. 55, PART 2.—NEW SERIES. 20 302 CRESSWELL SHEARER. only one segment, in every case terminating in the next segment to that in which they arise in the manner similar to the nephridia in Annelids. In both sexes the first three pairs of nephridia hold the same position, but the fourth varies according to sex. In the male it is situated much farther back—at the junction of the generative with that of the following segment, while in the female it is much more forward—in the anterior portion of this segment. In the female the third and fourth pairs overlap and cross one another in different planes. The most posterior nephridium in Stratiodrilus is in that segment that would correspond to the fifth in Histrio- bdella. In neither of the sexes are organs found in this region in Histriobdella. Foettinger, in figuring a pair of nephridia in the region of the penis, evidently mistook the motion of the cilia in the vas deferens, or the slit in the vesicule, for the ciliary motion of excretory organs. The slit in the vesicule was first described by Haswell in Stratiodrilus, and as such was evidently overlooked by Foettinger. ‘They are even better developed in Histriobdella than in Stratiodrilus. They are edged with very stout cilia, that could readily be mistaken for nephridial flagella. With methyl-blue it is easy to determine, however, that no excretory organs exist at this point in Histriobdella. In Histriobdella, unlike Stratiodrilus, the nephridia partake to some extent of the metamerism of the body. In the third and fourth segments this is masked in the male by the great. development of the reproductive organs. In the case of the nephridia there has resulted a pushing forward in the female of the fourth organ, while in the male this has been reduced in size and moved backwards. Unlike Dinophilus, we do not find the sharp specialisation of the different parts of the nephridial canal into a thick- walled anterior excretory portion and a thin-walled posterior part. It is more uniform throughout in structure. 9 ON THE ANATOMY OF. HISTRIOBDELLA HOMARI. 303 A. The First Nephridium (figs. 1, 7, 9, 37, and 42). The first nephridium arises in the segment immediately behind the head, and opens on the exterior in the second segment. The head of the organ I have never been able to see plainly in the living condition, as it is hidden by the muscle-bands. These are always undergoing contraction during life ; the lumen of the canal in its anterior portion is continually compressed, and thus the action of its cilia rendered very intermittent. The head of the organ at its point of origin is very close under the epidermis; in one case seeming to be almost under the limiting membrane of the epidermis. In section the head of the organ appears as shown in fig. 42, which is taken in a horizontal plane in the dorsal region of the first segment. The space into which the nephridial head projects is shown surrounded by a number of darkly staining nuclei. ‘The canal with its flagella is shown cut in section in the body-cavity. The actual projec- tion of the head into the space is not seen in this section. The neck segment is very clear, and were the canal pro- longed into the head, asin Stratiodrilus, it could easily be seen at this point passing into the head. As this can never be done, it is apparent that the organ takes its origin in the neck segment and is not prolonged into the head. It is also certain that it does not divide and send a branch to join one from the opposite side, as in Stratiodrilus. Throughout its course it is a simple, unbranched, intra-cellular tube, being in the same position in both sexes. B. The Second Nephridium (figs. 1, 6, 7,9, 14, and 37). The second nephridium arises in the anterior portion of the second seement, and runs back to terminate in the anterior part of the third. It is much the longest, being twice the length of the first. Its course is straight backwards along the border of the muscle-bands. ‘The action of its cia is much more constant than that of the others, and for this reason it 304 CRESSWELL SHEARER, is the one most readily observed. The main portion of its canal is slender and thin-walled. The general course of the organ is shown in figs. 1, 6, 7, 9, 14, and 37. It will be seen from these figures that Foettinger observed the organ only at the pomt where it passes from the second to the third segment, and that he was unaware of its considerable exten- sion into the anterior region of the second segment. In the female the segments through which it stretches are somewhat more compressed, and for this reason it appears in the female shorter than in the male. c. The Third Nephridium (figs. 1, 3, 4, 10, and 13). The third nephridium has much the same position in both sexes. It arises in the anterior part of the third segment and runs back to bend outwards in the male and slightly in- wards in the female, and terminates in the anterior part of the generative region. When the body is retracted it overlaps the posterior third of the second. The head of the organ, as already mentioned, at its point of origin is in the normal con- dition ona level with the opening of the second. It is situated close under the epidermis, as in the case of the first nephridium, and on the dorsal surface. It runs backwards, and about the middle of its course makes a sharp turn ventralwards (fig. 5), In the female it overlaps considerably the fourth, its opening on the exterior being internal to the course of this organ, While in the male it makes only one turn outwards, in the female it is S-shaped, the terminal portion running inwards (fig. 3). p. The Fourth Nephridium (figs. 1, 3, 10, and 13). In the male the fourth nephridium arises in the posterior part of the generative segment in the region immediately behind the clasper, It runs backwards and terminates in the anterior part of the caudal segment. Its course is short and somewhat difficult to observe. It is much the smallest of all ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 305 the nephridia, and its action more feeble than that of the others, as the motion of its flagella is the first to stop when the animal is compressed under a cover-slip. It, however, assumes a much darker colour on impregnation with methyl- blue than do the other nephridia, and for this reason seems to play a considerable part in the excretion of waste products, Its position in the male has been correctly indicated by Foettinger, who remarks that no organ is to be found in. this position in the female. Unlike Stratiodrilus, the organ does not begin in a loop or run back so far in the tail region, but opens on the exterior just over the line of separation between the generative and the caudal segments (fig. 12). 4. Bopy-Caviry AND NEPHRIDIA. As in Dinophilus, there is an extensive blastoccelic cavity surrounding the gut, which sends prolongations into the head region, and also into the anterior and posterior feet. It has been described by Foettinger as lined by a more or less definite coelomic epithelium, I cannot find that this is strictly the case. The gut surface of the cavity is covered by a delicate cuticle, in which at rare intervals are seen small flat nuclei. It is difficult to say if this membrane is a definite structure or a mere secretion from the blastoccelic ends of the cells of the gut-wall. ‘he somatopleuric side of the cavity is not lined by any such membrane. The longitudinal muscles, as in Stratiodrilus, are surrounded by a similar delicate cuticle, but no nuclei are to be seen in it as in the gut membrane. I believe in neither of these cases can this membrane be considered a true peritoneal or coelomic epithe- lium. No mesenteries are present, nor can I observe the fusion of the gut to the dorsal ectoderm as mentioned by Foettinger. The gut is more or less closely applied to the dorsal wall, but I cannot find that any true fusion takes place. In the head the blastoccelic space sends prolongations into 306 CRESSWELL SHEARER. the feet, and forward on the under side of the brain. It is more or less separated from the cavity of the trunk by the neck muscles and the narrow constricted condition of this region. Its extension into the posterior feet is in free communication with the trunk, so that in living preparations the eggs in the female can sometimes be forced into the leg portion of the cavity by slight compression of the cover- glass. They slip back, however, to the main blastoccelic space surrounding the gut when this pressure is removed. The whole of the blastoccelic cavity is very irregular in out- line, and is divided, as has been described, in the trunk region by the oblique muscle strands into two lateral chambers. In every respect it corresponds with the same cavity sur- rounding the gut in Dinophilus. There is this difference, however, that the numerous brown granules seen in it in Dinophilus are wanting in Histriobdella, although Histriobdella, like Dinophilus, has no specialised vascular system. It is sharply divided from the sae of the ovary, there being no communication between the two. When the ova are forced into the blastoccelic space of the hind limbs the sac of the ovary is either pushed with them, or is definitely ruptured, and the ova pass directly into the blastoccele. Both at the anterior and posterior regions the wall of the ovary is considerably thickened where it crosses the blastoccelic space between the body-wall and the gut. In the male the sac of the testis is likewise sharply cut off from the blastoccelic space in the anterior and posterior part of the generative region. Histriobdella, like Dinophilus, shows the primary and secondary body-cavity existing together, but sharply divided from one another. The nephridia, as in Dinoaphilus, are in relation with the blastoccelic cavity alone. From the fact that we get two nephridia in the generative region in the female, there is considerable reason for concluding that the oviduct and its funnel can hardly represent a trans- formed nephridium as Haswell has suggested. The arrange- ment of the ganglia and the external appearance of the ON THE ANATOMY OF HISTRIOBDELLA HOMARI, 307 segmentation bear out the conclusion that in the male and female this region is composed of two segments. Haswell, in Stratiodrilus, states that “in the fourth segment the nephridia are probably represented in the female by the ovi- ducts, in the male by the vasa deferentia.” I have shown in the male and female that two nephridia are present in the generative region, although holding slightly different positions in the two sexes. It is therefore impossible that the ovi- duct and vasa deferentia represent transformed nephridia, unless we consider the generative region to be composed of three segments, for which there is no evidence. In my paper on the nephridia of Dinophilus teniatus I have given some reasons for opposing the view brought forward by Schimkewitsch (28) and Harmer (12), that the oviducts and vasa deferentia in the male of this animal represent modified nephridia. Here there are four pairs of close solenocyte-bearing nephridia in the male and five in the female. ‘They Show the same primitive relationship with the blastoccelic cavity as do those of Histriobdella. Harmer’s suggestion is that in the male the fifth nephridium has been modified into the vesicule seminales and vasa deferentia, while it remains unmodified in the female as the fifth nephridium. In the male he holds that one of the pairs of nephridia has lost its primitive relationship with the blastoccelic cavity, and here becomes highly modified into the large ciliated apparatus of the vesicula seminalis and the vasa deferentia. The principal evidence relied on by Harmer in making this comparison is the resemblance of the funnel-like opening of the vasa deferentia into the cavity of the testis, to the funnels with which he thought the nephridia were furnished. | have shown that these do not exist, and that the nephridia of D. teniatus are definitely closed. ‘Therefore the funnels of the vasa deferentia cannot be derived through modification from those of the nephridia, In Histriobdella and Dinophilus, I believe the ovi- ducts, funnels, and vasa deferentia represent structures 308 ORESSWELL SHEARER. belonging to an entirely different set of organs from those of the nephridia, viz. the ccelomoducts of Lankester’s nomen- clature. 5. Muscunar System. The muscular system has been described by Foettinger, whose account is correct in its main particulars. ‘he muscles of the trunk region, as described by him, consist of two groups, the dorsal and ventral longitudinal, and the irregular oblique or transverse muscles. It is to these last that I wish to call particular attention in the present account, as they are only mentioned briefly by Foettinger. In addition to this I have been able to add new details in the division and arrangement of the fibres of the longitudinal muscles that escaped Foettinger’s observation. A. Longitudinal Muscles. The chief muscles of the body are these powerful longi- tudinal bands. They have already been described by Foettinger in considerable detail. They consist of two dorsal and two ventral sets. Each band is composed of from twenty to thirty fibres, flattened dorso-ventrally. They are attached by their outer margins to the cuticle, while their free edges project into the body-cavity. In the generative segment their number seems reduced, but this is due to their confinement within a limited space—against the gut dorsally and the nerve-cord ventrally. In the caudal region they spread out, forming a more or Jess complete wall round the segment, only interrupted dorsally by the gut and ventrally by the nerve+cord. They split up in the head and tail regions, sending fibres to the jaws and the anterior and posterior feet. In the head dorsally they converge on one another, uniting in the median plane, and are inserted in the anterior surface of the jaw apparatus. The ventral bands, on the other hand, divide into two sets of fibres, the outer of which split again to supply fibres to the anterior and posterior surfaces of the ON THE ANATOMY: OF HISTRIOBDELLA HOMARI. 309 anterior feet, while the other set run forward and are inserted ventrally into the anterior part of the jaw mechanism. In the posterior region each band splits likewise, the ventral sending fibres to the foot of the same side, other fibres cross- ing to be inserted in the small appendage of the posterior limb. ‘The dorsal send part of their fibres into the leg on the same side, while the internal ones cross over to be inserted in the leg of the opposite side, these fibres thus forming a cross dorsal to the anal part of the gut. The dorsal longitudinal bands give off a few fibres to the two segments of the caudal region, which run towards the median line and are inserted into the cuticle. It is due to the action of these fibres that the contraction of the caudal segment is brought about. As already mentioned, the ventral bands split in the head region into two sets of fibres. One of these runs forward to be inserted in the anterior part of the jaw apparatus, while the other supplies the extensor and flexor surfaces of the anterior feet. The manner of insertion of these last is some- what peculiar. The fibres of the external side of the longi- tudinal band are not inserted immediately into that side of the foot nearest them, but run to the anterior surface of the organ, while those of the inner side of the band cross these to run to the posterior surface. In addition to these there are also other fibres, derived from the bands of the opposite side of the body, that also run to the anterior face of the foot. These fibres form a cross ventral to the anterior end of the stomach. Besides these there are some strands that run from the same side of the foot directly towards the median line, and appear to be inserted into the anterior end of the jaws. All these are inserted into the distal joint of the foot. ‘The course of these different fibres can be readily understood on reference to text-figs. 1 and 2. B. Special Muscles of the Generative Segment. In the anterior and posterior part of the generative segment, inthe intersegmental region, a few transverse fibres are present, running beneath the epidermis. 310 CRESSWELL SHEARER. In the male special muscles are developed in relation with the claspersand the penis. These are similar in their arrange- ment to the same muscles of Stratiodrilus. The penis possesses a pair of protractors and retractors. The retractor muscles also function as the retractors of theclaspers. They TrxtT-FIGS. 1 AND 2. I. Male. rr Female. The muscles seen from the dorsal side in the male and female. The division of the longitudinal bands in the head and caudal regions is also represented. run from the base of the penis to the base of the claspers, and by their contraction at the same time retract the penis and claspers. ‘lhe claspers have also, as in Stratiodrilus, a set of protractor muscles, which run obliquely forwards and inwards in the generative segment, and also a few fibres that run from the bottom of the clasper sheath to the anterior lip of the same. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 31] - It will be seen that there is some difference between the arrangement of the main muscles in Histriobdella as com- pared with Stratiodrilus. In the neck region I cannot find the complicated crossing of fibres shown by Haswell in his fig. 1. Nor in the posterior legs can I distinguish some of the fibres he represents. .The muscular system of Stratio- drilus is much better developed, and the presence of cirri and the retractile condition of the anterior feet give it a more elaborate muscular system than that of Histriobdella. c. Oblique Muscles. If we examine a number of transverse sections we will see the body-cavity traversed occasionally by oblique strands (figs. 39, 40, 41, 43, and 44). Foettinger mentions their resemblance to the oblique muscles of Protodrilus, but he was somewhat uncertain as to their nature. He says, “Je n’ai pu m/assurer si elles étaient de nature musculaire” (p. 457). They divide the body-cavity, as in the Archi- annelids, into a circular portion surrounding the gut and a right and left lateral chamber. In some of my sections they form almost a continual sheet of fibres, and they are much better developed than one might suppose from Foettinger’s remark. They are found as irregular bundles crossing the body- cavity from the head to the tail region. They are well marked in the posterior part of the head; commencing at a point on a line with the chitinous jaws, they are continued back into the neck region in an unbroken succession. Inthe middle of the segment they almost disappear, while they are more prominent in the intersegmental regions. - In the anterior and posterior parts of the generative segment they are also present, but are entirely missing from the middle in the male, being interrupted by the muscles and accessory glands of the penis. Anteriorly they divide the testis in two portions, forming a right and left chamber (fig. 39). In the anterior part of the first segment they are shown in fig. 43. Here, during part of their course, they touch the wall of the gut. 312 CRESSWELL SHEARER. In Stratiodrilus their presence has been observed by Haswell (13), who states: ‘Throughout the body slender oblique bundles occur at fairly regular intervals, running from the cuticle of the lateral surface to that of the ventral near the nerve-cord” (p. 306). Here, however, they would seem to be much less developed. I think there is no doubt that they correspond to the oblique muscles of Polygordius. It is interesting to note that the nephridial canals, as in Polygordius, are always within the limits of the lateral cavities formed by them. Another point of similarity consists in the manner of their insertion into the dorsal body-wall. ‘They spread out in a fan-like manner, as Hempelmann (15) has shown takes place in Polygordius (see his text-fig. 14). This same arrangement of the fibres, it will be seen, is found in Histriobdella (fig. 41). Theanterior and posterior feet, in addition to the fibres they receive from the longitudinal bands, also possess a special musculature of their own. In the anterior foot this consists of a series of parallel fibres that run from its base to the distal, flat, saucer-like pad of the foot. The foot itself is composed of two parts, a distal retractile portion and a larger non-retractile, cone-shaped basal portion. Some of the fibres are applied closely to the cuticle of the outer part of the basal portion, while those of the bands, as already mentioned, are inserted into the retractile distal portion. They surround and run into the basal gland of the foot, There is a collection of granular mucus cells at the base of the foot, abutting internally on the jaw muscula- ture. ‘They stain deeply with carmine, each cell having a darkly granular periphery, with a clear centre hollowed out in a small cavity. ‘The gland gives off a number of straight tubes, that open on the pad surface of the foot. They run up amongst the muscle-fibres, and can be readily distinguished from these by the manner in which they take the stain. The gland pours out on the surface of the pad some sticky secre- tion, by means of which the animal is enabled to obtain a firm hold. In the case of the posterior limbs a similar, but larger, gland is present.. It extends from the wall of the gut out of ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 313 the centre of each leg to the commencement of the outer third, where it gives off a mass of fine, darkly staining tubules, which open on the pyramidal pad of the foot. This gland is able to pour out a copious viscid secretion. Fre- quently, when the animals are irritated, this secretion can be TEXT-FIG. 3. m.d.t------tM | i we --¢p. f. 3. Head showing the muscles in connection with the feet and the jaws. bl.m., Bulb-like muscular organ of the jaws; cl. p., ciliated pit of the head; cv. f., crossed strands of the dorsal longitudinal muscles ; m.d.f.a., dorsal longitudinal muscles running into anterior surface of the foot; m.d.f.p., dorsal longitudinal muscles running into posterior surface of the foot; m.gl., salivary gland of the mouth; m.d.j., dorsal longitudinal muscle running to jaw apparatus ; tr,m., transverse muscle-strands running into the feet. seen pouring out from the ends of the tubules, forming minute round drops on the end of thefoot. Like the anterior limb, the posterior has some muscular fibres apart from those it receives from the longitudinal muscle-bands. These are a delicate set of fibres just under the cuticle on the posterior surface, that run from the extremity to be inserted on either side of the anus. In addition to these there are some oblique 314 CRESSWELL SHEARER. fibres, asin Stratiodrilus, but they are but feebly developed. A considerable prolongation of the blastoccelic cavity takes place into the posterior limbs, running out along each leg between the muscle-fibres and the glands. Into this space the ova in the female are sometimes forced when the animal is compressed under a cover-slip, showing that it is in free com- munication with the cavity surrounding the gut. The movement of the limbs takes place alternately, the head being swung from side to side with the movement of the feet. It is a most remarkable sight to see the animals rear up, as they sometimes do, on their hind feet, and stand executing movements with their head while they remain firmly attached with their powerful hind feet. ‘They also crawl quite readily, by means of the feet, on the underside of the surface-film of the water. In the ordinary movements of crawling the glands do not appear to throw out any secretion on the pads of the feet; only when they are disturbed do they pour out a thick secretion, which firmly attaches the feet to the surface on which they happen to be. While the animal violently twists its head and body, it never moves its feet. This hold is remarkably firm. On the lobster ova the parasites can be seized by the middle of the body by means of a pair of fine forceps, under a dissecting microscope, and the body pulled off, leaving the feet still attached, the limbs having been torn from the body without loosening their hold. As already mentioned, the front limbs in Histriobdella differ from those of Stratiodrilus in that they are non- retractile. I have never been able to observe any retraction of the feet in the living condition, or in preserved specimens treated with different reagents. 6. DicEstIvE SysTEmM. The digestive system is sharply specialised into a number of divisions. These are readily seen in the figure of an immature parasite (fig. 30). - Here they are more marked ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 315 than in the adult. A more or less slender cesophagus leads dorsalwards and backwards from a quadrilateral-shaped mouth (text-fig. 4). ‘This, although small, is capable of considerable expansion. It is completely everted in allowing the jaws and teeth to be protruded in the act of biting. It commences in a slight ciliated depression, which rapidly deepens into a groove in the anterior part of the head. The cesophagus terminates, on a line with the posterior boundary of the jaw musculature, in a narrow constriction leading into the stomach. It is difficult to say where the mouth ends and the cesophagus commences. ‘lhe mouth and cesophagus are lined throughout with fine cilia, those of the cesophagus being much stouter than those of the mouth. TEXT-FIG. 4. si Showing the outline of the mouth when partially closed. The stomach may be defined as that portion of the intestinal tract lying between the first and the third segment. Its wall is composed of a single layer of cubical cells. It is for the most part uniform in thickness. The rounded internal ends of cells project irregularly into the lumen and are ciliated. At the anterior end, near the cesophagus, the cells are very columnar and contain many granules. ‘They have probably to do with the elaboration of the digestive secretions, as they are seen to be very opaque after the animals have taken food. Those of the ventral wall in this part are somewhat larger than the dorsal. The nuclei are always placed at the bottom of the cell, that.is, farthest from the internal ciliated surface. In the anterior region they are long and oval in shape, while in the middle and posterior regions they are spherical, and the cells themselves cubical in outline. In the posterior 316 CRESSWELL SHEARER. region of the stomach the wall is relatively thin in comparison with that of the anterior part, and its cells on the ventral side are furnished with very long, dense cilia. About the middle of the end of the third segment the stomach contracts into a narrow mid-gut, which runs through the generative region to widen somewhat in the caudal region into a more or less straight hind-gut. The lumen of the intestinal tract, from the stomach backwards, is greatly reduced in size, and, in the contracted condition of the animal, somewhat folded on itself. The character of its ciliation is also different from that of the stomach. At the point where the stomach passes into the mid-gut there is a sort of valve formed by the thickening of the stomach-wall. A similar valve is found at the point of union with the hind- gut. The wall of the mid-gut is relatively the thinnest part of the tract, and its cells are not of the marked yellow colour of those of the stomach. ‘The course of the mid-gut is irregular, from its being slightly folded on itself. That of the hind-gut is comparatively straight, bnt its lumen is irregular and wavy in outline, due to the irregular thicken- ing of the wall at different points on its course. Throughout the generative segment the gut is- very closely confined against the dorsal body-wall. The anus is dorsal. The cells of the hind-gut are of a character quite different from those of the other parts of the tract. They are quite irregular in size, and extend into the lumen so as to make its outline very broken, as if thrown into a number of convolutions. In no part of the wall of the stomach or gut are any contractile muscular fibres to be-seen. In the body-cavity, ventral to the anus, and close to the point where the gut joins the body- wall to form the anus, there is usually present a conspicuous cell on either side. The anus itself is an oblong, vertically placed, T-shaped slit placed more towards the dorsal than the ventral side of the animal. It is apparently kept closed by some contractile fibres of the cuticle which function as a sort of sphincter muscle. ; The digestive traet of Stratiodrilus agrees in all essen- ON THE ANATOMY OF HISTRIOBDELLA HOMARI. Sz. tial details with that of Histriobdella as far as can be judged from Haswell’s somewhat brief description. There is the same reduction of the tract in the generative region, this being much greater in the female than in the male, and its expansion into a more or less large hind-gut in the caudal region. Ascompared with Dinophilus there is a greater difference. Yet with the exception of the pecuhar mid-gut portion of the tract, which is a development due to the peculiar condition produced ‘by the presence of a special generative segment, there is considerable resemblance between Histriobdella and Dinophilus, and in many of the finer histological details there is a very close resemblance. In the first place, the appearance of the cells of the stomach, each composed of a single layer of ciliated cells, the yellow vacuolated appear- ance of their protoplasm, and the basal arrangement of the nuclei, are the same in the two. The terminal dorsal position of the anus and the configuration of the cesophagus and pharynx are remarkably the same in both. According to Nelson (25) there is a feeble strand of muscle- fibres that act in Dinophilus as sphincter ani, as in Histriobdella. Throughout the stomach region there is a lack of muscular strands, and the stomach is not supported by mesenteries, but is closely appled to the dorsal wall, as in Histriobdella. The blastoccelic surface of the stomach, as in Histriobdella, is covered with a fine cuticle. The jaw apparatus of Histriobdella is very similar to that of Stratiodrilus. Haswell has given an extensive description of this, so that I need only briefly consider it. As in Stratiodrilus, it consists of two portions—the upper and the lower. The upper consists of a median rod (fig. 36), which Haswell has called the fulcrum. This is slender, round, and slightly curved; it articulates by means of a number of basal pieces with a series of jointed arms, each terminating in a curved tooth (text-fig. 5). It lies in the median plane dorsal to the two blades of the lower jaws, being set at a different angle to these. Its length is somewhat less VOL. 55, PART 2.—NEW SERIES. Pal 318 GRESSWELL SHEARER. than these last. The cubical basal pieces with which it articulates support four arms on each side, each being com- posed of three or four pieces, the last of which is fashioned into a sharply-curved claw-like tooth. This is strongly serrated on its inner edge. One difference between the jaw parts of Histriobdella and Stratiodrilus consists in the length of the middle joint of these arms. ‘They are much TEXT-FIG. 5. Jaw apparatus. Enlarged figure showing the structure of the teeth and the arrangement of the jJaw-muscles. ba.p.j., basal piece of Jaws; bl. m., bulb-like muscular organ of the jaws; f., fulerum; m.d.j., strands of the dorsal longi- tudinal muscle-bands running to the jaws; st.m.j.., striated muscles of the Jaws; th., teeth. longer in Histriobdella, and allow of the teeth being folded back in the mouth or cesophagus to a greater extent than in Stratiodrilus. When at rest in the ordinary position the teeth are not folded back to their full extent. The middle piece of the arm projects at right angles to the jaws, and in this position the most anterior part of the arm is the distal joint, the tooth being strongly flexed. When the ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 319 arms are folded to their full extent the fulcrum is drawn back on a level with the extremity of the blades of the lower jaws. These are paired throughout. They consist of two long wide blades, thickened at their outer margins, and articulating at their basal ends with two curved wedge-like pieces, the pointed end of the wedge being directed forward in the ventral lip of the mouth. Ventrally they articulate with one another in the median line, and turn up dorsally to form a support for the upper jaws. The upper anterior angle of each plate is turned outwards and backwards, some of the fibres of the dorsal longitudinal muscle-bands being inserted into it. The internal interior edges of these plates are finely serrated, and evidently assist the teeth in their action. As far as can be judged from Haswell’s figures, the shape cf these plates differs slightly in Histriobdella from that of Stratiodrilus. They fold up dorsally to a greater degree. Tbe main portion of the lower jaws are the wide blade-like portions which project backwards parallel with one another. They are widest behind, and taper slightly in front, where they articulate with the wedge-like portions. Connecting the upper with the lower jaws are the pieces that Haswell distinguishes as “ bridles.’ Into the posterior extremities of these are inserted the powerful striated muscles. Apart from their action in binding together the jaw-sets I have not satisfactorily determined their function. They would seem to be composed of a single curved piece in Histriobdella, and its chitinous substance is broken up into a number of dark hairs where the muscle is inserted, giving it a furred appearance. The powerful nature of these fibres shows that their action in pulling on the bridles has to do with some essential movement of the jaws. It is likely that the actual process of biting is brought about by their contraction, as Haswell has suggested, while the fulcrum has merely to do with their protrusion and opening. In addition to these there is the peculiar bulbular muscular organ, not unlike the sub-cesophageal muscle pad of Dinophilus. This is attached to the posterior ends of the ventral surfaces of the lower 320 CRESSWELL SHEARER. jaws. Its fibres form an oval mass attached directly to the jaw blades. Into this mass some of the striated muscle-fibres are inserted. Its action is hard to understand. It is well shown in Foettinger’s figures. In the movements of the jaws the lower blades are sometimes seen to separate con- siderably from one another posteriorly, and it is possible this motion is brought about by them. What this movement has to do with the teeth I have been unable to observe. This muscular organ appears to be wanting in Stratiodrilus, as it is not shown in Haswell’s figures. On either side of the jaws about their middle there pro- trudes laterally a small pear-shaped gland composed of from three to four large granular cells with conspicuous nuclei (text-figs. 1 and 2). This gland opens into the mouth or the anterior part of the cesophagus, and is evidently of a mucous nature, as it absorbs the methyl-blue colour very strongly when the parasites are placed in it for a short time. The protoplasm of the gland-cells is finely granular, each having a very large, darkly staining nucleus with a prominent nucleolus. The duct of the gland converges and opens on the ventral side of the mouth. The posterior portion of the organ lies against the muscular pad of the ends of the lower jaws, while its dorsal surface touches the cuticle of the dorsal surface of the head. In position and structure it is in all respects similar to the glands occupying the same position in Dinophilus, and undoubtedly answers the same purpose. In Protodrilus, also, similar glands are present. It appears to have been overlooked by Foettinger. In fig. 1 of his paper he shows a mass of tissue on either side of the jaws, which in great part belongs to these salivary glands, and not to the jaw muscles, as he evidently thought. Haswell makes no mention of its presence in Stratiodrilus, although it is probably present here also, for he shows a number of round cells in the position that it occupies in Histriobdella, ON THE ANATOMY OF HISTRIOBDELLA HOMARI. S2t 7. THe Nervous System (figs. 15, 21, and 28). The nervous system extends throughout the body, and is composed of a brain, cesophageal commissures, and ventral nerve-cord, with ganglia at intervals corresponding to the external segmentation. The brain is situated well forward in the head, its main mass being anterior to the oral opening, and close to the dorsal surface. It is composed externally of a mass of nerve-cells surrounding a clear fibrous core. The nerve-cells are distributed over its dorsal surface. Behind, the brain is deeply cleft ventrally, descending in lateral lobes on either side of the anterior part of the mouth. This cleft runs forward, forming a small closed sinus in the anterior end of the organ. The brain terminates rather abruptly at a point about on a line with the anterior third of the jaws; here it gives off two fine commissures that run directly ventralwards and backwards, connecting it with the first ganglion of the ventral nerve-cord. At the point where these come off some fibres go to the anterior legs, and others run directly backwards in the dorsal region. ‘They probably correspond with the “nerfs sympathiques” of Foettinger. In addition to these, the brain supplies nerves to the anterior tentacles. The commissures are closely applied to the cesophagus, and are difficult to follow in sections on account of their small size. At about on a line with the posterior boundary of the brain, and slightly in front of the anterior feet, there is a small ciliated pit on either side of the head. The anterior lip of this protrudes slightly, forming a sort of papilla. This pit is undoubtedly sensory in nature, and appears to have some fine nerve-fibres running to it from the brain. ‘The nerve-cells of the dorsal surface of the brain are distinctly differentiated from the cells of the ectoderm. ‘hey are recognisable by the elliptical outline of their nuclei, and the marked way in which they take the stain when treated with the hematoxylin mixtures. As compared with the ectoderm 322) CRESSWELL SHEARER. « cells, their nuclei are rich in chromatin. ‘This peculiarity renders them distinguishable from the supporting cells of the surrounding tissues. Some of the ganglion cells are clearly multipolar, but axons and dendrites are not recognisable. At the base of the tentacles the cells are bipolar, one process going into the tentacle while the other enters the neuropile. They form a dense mass of cells on the anterior dorsal surface of the brain-core. They are, however, quite distinct from it, only sending a few fine threads into its substance. In the median plane a small space, a prolongation of the general blastoceelic space, extends up under the brain, and separates them from the core, dividing them into two lateral masses. The central core of the brain is composed of a dense mass of interwoven nerve-fibres. It is distinguishable by its yellow colour and its non-nucleated character. It is remark- able that both in relation with the brain and the ventral cord the nerve-cells seem quite apart, and outside the fibrillar part of the nervous system. ‘Their relationship seems closer with the ectodermic tissues of the head and the mesodermic and ectodermic tissues in the trunk than with the fibrillar material of the nervous system in these regions. ‘The fibres of the ventral portion of the neuropile seem to run from side to side, while those of the superficial layers ran more longitudinally. In sagittal sections it is lenticular in outline, and in the median plane is divided by a transverse fissure into an anterior and posterior part. Haswell also shows these divisions in the brain of Stratiodrilus (fig. 8). This division is only limited to the median plane; laterally the neuropile swells out into two large lobes on either side. Thus it consists, as in Dinophilus, in a median and two lateral lobes, the median being in turn divided into an anterior and posterior portion. In the figures of the brain accompanying this paper these divisions do not show, as the brain surface is taken from the ganglion cells and not from the central core. Behind the brain, and dorsal to the muscular apparatus of the jaws, there is a second accumula- tion of nerve-cells. These may possibly have to do with the ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 323 innervation of the jaw muscles; they are dorsal and median to the cesophageal commissures. I have been unable to make out their connection with the muscles. ‘They take up methyl- blue much more readily than do the other cells of the brain, and retain it considerably longer. The ventral nerve-cord, like the brain, consists of a similar central fibrous core, surrounded with nerve-cells. The two halves of the cord are separated in the intersegmental regions, joining up in the middle of the segments to form a ganglion. From what can be judged from Haswell’s draw- ings, in Stratiodrilus this separation is much less than in Histriobdella. Unfortunately, most of the sections drawn _ by Foettinger are taken through the middle of the segments, and do not properly illustrate the extent to which the two portions of the cord separate in the intersegmental regions. The two halves of the cord are crescentic in transverse section, the nerve-cells being imbedded on the ventral surface. Where the cords unite these cells are drawn out laterally to form considerable masses on either side. The main ganglia, as already mentioned, correspond closely with the five main segments into which the trunk is divided. The first is situated in the anterior region of the first segment, and is of considerable size. ‘he second is somewhat smaller, and is situated about the middle of the second segment. It has fewer nerve-cells, and, like Stratiodrilus, it is placed nearer the first than the third ganghon. The third is the largest, taking up the greater part of the length of the cord in the third segment, and having a great number of nerve- cells. Between all the ganglia in the intersegmental regions of the anterior segments the component parts of the cord separate as already mentioned ; between the third and fourth ganglia this is hardly perceptible, and from this point back- wards to the tail region the two portions of the cord are in close union, with the exception of a small area near the end. The fourth ganglion is the largest of all, and occupies the middie of the generative segment. ‘The fifth is in the middle of the caudal segment. The position of these ganglia can be 324 CRESSWELL SHEARER. seen from the reconstructions shown in figs. 15, 21, and 28. In these figures the nervous system is seen from the ventral side. The outlines of the cord and ganglia have been measured from the nerve-cel!s, as far as these could be roughly differentiated from the surrounding tissues.’ From fig. 15 it will be seen that the main mass of the fourth gan- glion lies just in front of the penis, but many of its cells extend backwards in the region dorsal to the penis. Here they would almost seem to form a second division of the ganglion. I have not attempted to determine its structure, which differs considerably from that of the other ganglia, on account of the great size of its lateral parts. A few of its cells are distri- buted on the penis sheath. Past the fourth ganglion the cord diminishes rapidly, but enlarges again rather suddenly in the interior part of the caudal region. It is the second in this segment that is the largest. In the posterior region the cord divides to run into the posterior feet. At this point a number of nerve-cells are arranged, forming quite a mass. Itis diffi- cult to decide whether each of these ganglia is to be considered as representing a segment. If so, then there are three main ganglia in the segment itself, and counting the mass of cells at the termination of the cord, it would be composed of four segments. Foettinger came to the conclusion that it was one segment, formed by the partial fusion of three metameres. In the female there are a number of differences in the con- figuration of the nervous system, due to the somewhat different size of the segments as compared with the male. This is most pronounced in the generative region. ‘lhe absence in the female of the penis and accessory glands results in the almost complete disappearance of the cord and ganglia in the posterior part of the generative region, and throughout this portion of the body the cord and its ganglia are much less prominent than in the male. In the absence of the penis the cord retains its ventral position. In the caudal region, on 1 In the reconstruction of the male nervous system shown in fig. 15 no allowance has been made for the dorsal curvature of the cord in the generative region. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. O20 the contrary, the cord and ganglia are much the same as in the male (fig. 15). In Stratiodrilus the cord and ganglia are much the same asin Histriobdella. In the male the fourth ganglion is opposite the claspers. After this the cord is very much reduced where it passes dorsalwards over the penis. In Histriobdella this reduction is not so marked. In the caudal region also the ganglia are smaller. Haswell remarks, regarding the nervous system of the caudal region of Stratiodrilus, that “the ventral chain may be described either as represented by a single elongated ganglion imper- fectly divided into five or six portions, or as consisting of five or six imperfectly separated ganglia” (p. 315). The nerve- cells are arranged on the ventral surface of the cord, and the lateral ganglia are much better developed in Histrio- bdella, especially those of the generative region. They send fibres into the cirri. The second ganglion would seem to be double. Haswell has drawn attention to the fact that the nervous system in Stratiodrilus does not show the complete union with the body-wall tissues as does that of Histriobdella. 1 think, however, no great importance can be attached to this point. The separation shown by the nervous system in Stratiodrilus is due in great part to the thinness of the body-wall as compared with Histriobdella, and not to a more highly differentiated condition of the system itself. It is of considerable interest to compare the nervous system of Histriobdella with that of the Archiannelid it resembles most, that is, Dinophilus. From the study of a species closely allied to D. gyrociliatus, Nelson (25) has deter- mined the main structure of the central nervous system in considerable detail. In the first place there is a marked separation of the two parts of the ventral nerve-cord in the intersegmental regions, much more so than in Histrio- bdella. Unhke Histriobdella they do not unite to form the ganglia, but are joined by commissures, the two portions of cord remaining separated throughout their course. There 326 CRESSWELL SHEARER. are four well-marked ganglia corresponding to the four main segments of the trunk. In addition to this, there are a few cells that probably form a fifth, corresponding with the somewhat reduced caudal segment. If we compare the reconstruction figure he gives of the nervous system with that of either the male or female Histriobdella given in the present paper, it will be seen that, with the exception of this greater separation of the cords, there is a remarkable resem- blance in the general configuration of the nervous system of the two forms. ‘lhe brain and the cesophageal commissures are much the same. In transverse sections the cords hold similar positions in the ectoderm. The ventral sinus found in the brain of Histriobdella, it would seem, is also present in Dinophilus asa small closed cavity in the brain substance itself, In minor histological details they bear a striking resem- blance to one another. The brain is clothed dorsally and laterally with a mass of nerve-cells, having the peculiar eranular nuclei so characteristic of these cells in Histrio- bdella. ‘They are similarly differentiated from the supporting tissue cells. ‘lhe circum-cesophageal commissures are better developed, however, in Dinophilus, and pass backwards round the cesophagus just below the dorsal longitudinal muscle-strands. ‘he centre of the brain is composed of a mass of clear fibrillar material that stains with difficulty. As in Histriobdella, tibres are given off by the cesopha- geal commissures at the point where these leave the brain. They are much bigger in the case of Dinophilus, and are more easily traced through consecutive sections. In His- triobdella there are no pre-oral commissures, and the ganglia are more circumscribed and definite than in Dino- philus. As compared with the nervous system of Protodrilus there is a greater difference than in the case of Dinophilus. This is due to the lack of ganglia on the ventral cord. In Protodrilus the ventral cord shows no ganglionic divisions corresponding to the external segmentation. This is very ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 327 slight, being shown ouly by the ciliated rings. Internally it is better marked by the dissepiments and the nephridia. he two halves of the cord remain separate throughout their course, uniting at their ends in a small ganglion. Such a nervous system can hardly be compared with that of His- triobdella. According to Pierantoni (26), the nerve-cells in Proto- drilus are equally as difficult to distinguish from the sur- rounding tissues as in Histriobdella. While retaining their primitive position in the ectoderm, they send fibrils to the tentacles and the digestive system. In the ventral cord there would seem to be no localisation whatever of the nerve- cells corresponding to the segmentation. 8. SEnsr-OrGANS. Among the sense-organs of Huistriobdella are to be classed the five tentacles of the head and the palps of the posterior legs. All these receive nerve-fibres from the central nervous system, and are armed with short, stiff, sensory hairs. The most essential of the tentacles appears to be the median one of the head. In the larva this is the first to appear, and its nerve supply in the adult would seem to be greater than that of the others. In addition to the tentacles, scattered over the cuticle of the body are a number of cells of a sensory nature that stain readily with methylene blue. On the dorsal lateral parts of the head are the sensory pits described by Foettinger. These, as already mentioned, are very small, and placed a short distance in front of the anterior feet. Foettinger has sought to compare them with the ciliated grooves of Archiannelids. They measure about 14 in their longest diameter, and are oval in outline. They are therefore much smaller than the long grooves of Protodrilus and Polygordius. In the bottom of the pit are placed a few fine sensory hairs. As described by Foettinger, the anterior edge of the pit is developed into a slight lip or ridge that is capable of being folded completely over the pit and of 328 CRESSWELL SHEARER. obliterating it. From the way in which this lip is protruded and the pit opened when the animal is feeling its way or examining any small object it may come across in crawling on the bottom of a watch-glass, it is evident that the pit functions in some way as an organ of taste or smell. It appears to receive a set of nerve-fibres from the brain. ‘There is no doubt that these pits correspond to the ciliated pits of the Archiannelids, despite their small size. They are present in both sexes. According to Haswell they are not present in Stratiodrilus. 9. THe Rupropucrive System. The reproductive organs in the male consist of a testis, paired in its anterior part, two vesicule seminales, two vasa deferentia, and a median penis. Dorsal to each vesicle is the so-called granule gland (fig. 11). In relation with the penis there is a gland of unknown function, as in Stratio- drilus. In the female the organs consist of a large sac or ovary filling the whole of the generative region. On its ventral surface this is furnished with a paired oviduct, armed with a large funnel, the dorsal lip of which only is ciliated. On the course of the oviducts and close to their external openings are the ampulle or shell-glands. A. In the Male. The testis in the male when fully developed fills the anterior and middle third of the generative region. Its extreme anterior end is separated into a right and left portion, its middle portion is fused in the median line. Behind it ends somewhat abruptly in front of the penis. ‘lhe remaining pos- terior third of the generative region is taken up with the penis and its accessory glands. This portion is sharply divided from the anterior two thirds by the limiting membrane of the testis. This fact has not been clearly shown by Foettinger. He seems to have overlooked the well-defined ON THE ANATOMY OF HISTRIOBDELLA HOMARL. 329 nature of the limiting membrane, and fails to show the sharp manner in which the testis is shut off from the general blasto- coelic cavity surrounding the gut. He states that the testis takes up the whole of the generative region, which is not the case, for the penis and its glands take up the posterior third as I have mentioned. ‘The anterior paired portion of the testis is shown in section in fig. 39, while the main unpaired portion is Shown in section in fig. 35. Internally the testis is filled with a number of oval bodies, the spermatidia (figs. 27 and 35). These consist of a number of nuclei with granular chromatin, arranged round the circumference of a small mass of cyto- plasm. In the region close to the anterior end of the testis they form a solid mass, while in the middle they crowd its cavity as a number of oval bodies. The mature spermatozoa are found in the spaces of the testis cavity between them. If we regard the generative region as due to the fusion of two segments, then this conclusion is supported by the arrange- ment of the nephridia and the ganglia. ‘he testis itself takes up the first and largest of these, while the penis and accessory olands take up the second. The division between the testis and penis portion comes at just that point we should naturally conclude that it should from the position of the ganglia. In the female the double nature of the generative region is not so clear asin the male, and the metamerism is masked by the extensive prolongation backwards of the ovarian sac. In the young female, however, the ovary is confined to the anterior two-thirds. The double nature of the generative region then is almost as distinct in the female as in the male. The vesicule seminales are found in the posterior part of the testis, and are pear-shaped bodies with their pointed ends directed forwards. They are readily recognised on account of the large quantities of sperm with which they are always crowded. Leading into the lateral surface of each vesicle is a fine duct from the granule gland. These are a mass of large mucus-like cells that lie against the inner surface of the cuticle of the body-wall of the genera- tive region. They secrete a granular mucous substance which 330 CRESSWELL SHEARER. they discharge into the vesicule. Hach gland is composed of about twenty cells, arranged in a single layer, laterally, against the wall of the segment. They fill up the greater part of the middle third of the region. They commence anteriorly, just behind the orifice of the retracted claspers, and stretch back to a point, on a line behind the vesicule on either side. Dorso-ventrally they extend from the border of the dorsal longitudinal muscles round the sides of the segment to the border of the ventral bands. ‘Their cells have a waxy appearance, and their cytoplasm, which is rela- tively large in amount, is very finely granular. Each cell possesses a round nucleus and a dark karyosome. On a line with the vesicule the dorsal cell of each group gives off a fine duct, that crosses the space of the testis cavity and runs into the ventral external surface of the vesicle of the same side. The wall of this tube is also, hke the protoplasm of the cells of the gland, finely granular. About its middle there are usually two large nuclei embedded in the wall. Where the tube runs round the outer surface of the vesicule it is much thickened, and this appears to be due to the accumulation of drops of the gland secretion in its Jumen (fig. 31). The vesiculx are roundish bodies with thin walls. The lateral and ventral third of their cavities is taken up with the mucous secretion derived from the granule glands. This, in sections of fixed material, projects upwards into the cavity ina mass of finger-like digitations. On the outer ventral surface of each vesicula there is a small slit. Its edges are armed with short stout cilia. It was the motion of these that Foettinger evidently mistook for the presence of a pair of excretory organs in this region. Through this slit the spermatozoa gain an entrance into the vesicule. The vas deferens leads out from the posterior ventral portion of each vesicle and turns in towards the median line, and is continued as a small tube to the base of the penis. It is of considerable diameter, and forms a‘sac-like canal on either side. At the base of the penis the vasa deferentia of both sides meet, forming a smallreceptaculum seminis, ON THE ANATOMY OF HISTRIOBDELLA HOMARI. oon which lies between the two lateral halves of the organ. During life this is always full of very actively moving spermatozoa. The penis is a firm, semi-solid, pear-shaped body, the pointed end being directed backwards. It is always carried retracted within the sheath. Unlike Stratiodrilus, it is not composed of black chitinous material similar to that of the jaws, but of some transparent substance, sufficiently rigid, however, to enable its being driven through the firm cuticle of the female in the act of copulation. It is pro- truded through the quadrilateral-shaped mouth of the penis- sheath by the action of the strong protractor muscles. The organ itself is composed of two lateral blades, the spermatozoa being ejected through the median canal between them during copulation. In relation with the dorsal surface of the penis on either side, and taking up the lateral posterior corners of the generative region, are the so-called accessory glands of the penis (figs. 11 and 40). ‘These are large vacuolated groups of cells forming oval masses running up to the dorsal surface on either side of the gut. From each gland a small duct leads down to the penis, and is inserted laterally about its middle. This opens into the canal on the penis on its ventral side. The gland-cells are divided into an anterior and pos- terior group. In horizontal sections the gland appears as a four-lobed structure, posterior and dorsal to the base of the penis. The anterior and smaller of these groups is composed of numerous cells, while the posterior, although larger, consists of fewer cells. The cytoplasm is granular and very vacuolar. This is shown in fig. 40, where their anterior ends come in the section on either side of the gut. In transverse section the gland will be seen to be composed of two groups of cells, one of which is much smaller and more dorsal than the other. This is wedged in against the gut on either side. Towards the posterior region of the gland the cells are somewhat larger. ‘The largest of these contains a vacuole of considerable size. This probably acts as a receptacle for 332 CRESSWELL SHEARER. the gland secretion. It is connected with the penis by a strand of cytoplasm that runs to its ventral side, and is con- tinuous across the median line with a strand from a similar cell from the opposite side. Posterior and ventral to this are a number of small, darkly staining cells. They are lenticular in shape, with prominent nuclei. They fill up the corners between the large cells. The largest cell of the gland is placed about the middle or slightly towards its posterior end. ‘The section shown in fig. 24 passes just behind its posterior border. The nuclei of the smaller cells are rod-shaped, and frequently bent in a semi-circular form. On the inner wall of the gland, close to where it abuts against the penis-sheath, are a number of darkly staining masses of nuclear material. The ends of all the gland-cells converge on the penis. When the cells are charged with secretion their nuclei are seen to be large and round, with a well- marked karyotheca. The karyoplasm is collected into a darkly staining karyosome. In the cells that have dis- charged their secretion, on the other hand, the nuclei are invariably long and rod-shaped, with a uniformly staining karyoplasm, and no karyosome. In fig. 23 are represented some of the cells of the posterior group under high magnification. The cytoplasm forms a superficial layer which throws threads across the vacuolar interior of the cell. The nucleus is always situated about the middle of the cell and is of considerable size, and contains a darkly staining karyosome. The compartment of the generative region holding the glands is sharply separated from the anterior part of the segment, which contains the testis, as already explained. This is clearly separated from the granule cells and the vesicule, which are within the limits of the testis proper, and enclosed by its membrane. The region of the accessory gland is often seen distended with the accumulation of secretion within the gland. With dark ground illumination this appears opaque and whitish in colour. In the surface view of a living preparation the two ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 3338 portions of the gland appear somewhat as shown in fig. 11. The anterior lobe seems distinctly separated from the pos- terior. The function of these glands is problematical. They doubtless pour some secretion into the canal of the penis during copulation, which assists in this act in some way. They were first described by Haswell in Stratiodrilus, where they are much larger and somewhat different in appear- ance from those in Histriobdella. They seem to have been overlooked by Foettinger, although he plainly figures them in his sections. He evidently mistook them for a portion of the testis. That they are separate structures from this can be easily seen in horizontal sections. They correspond to the similar glands found in connection with the male organs in so many Turbellaria, as in Proxenetes, Provortex, and Plagiostoma. Under the heading of the male reproductive organs come the claspers. These are usually carried retracted, only being protruded when the males are impregnating the females. Under the action of strong reagents during fixation they are sometimes extended, in which case they are always seen pro- jecting ventralwards and never laterally. Each clasper is furnished with a protractor and a retractor muscle that runs to the base of the penis, as already explained. At the base of each organ there is a large mucous cell with a large nucleus. This, in the retracted condition, occupies the anterior wall of the clasper-sheath, and is a conspicuous feature in a trans- verse section through the anterior region of the generative segment. In a full-grown male the cell is very large. A fine duct leads from it to the tip of the organ and pours some adhesive secretion on the surface of the clasper, similar to that poured on the surface of the feet. This cell is shown in fig. 2. ‘The anterior lip of the orifice formed by the retrac- tion of the organ forms a marked projection which overlaps the orifice (fig. 13). When the organ is extended this lip is obliterated, as shown in fig. 9. ‘The gland cell then occupies the middle of the clasper. At the top of the organ there are a few short, stiff hairs. I have already mentioned that once VOL. 95D, PART 2,—NEW SERIES. 22 334 CRESSWELL SHEARER. the male has seized the female by means of the claspers its grip is immediately rendered secure by the gland secretion, and then the male is only able to free itself from the female with difficulty. Sometimes the male can be seen being carried about by the female, making violent efforts to free itself. The claspers never seem to be used for any other purpose than that of seizing the female, and are never extended to enable the animal to hold more securely when an attempt is made to brush them off the lobster ova. B. In the Female. The ovary in the female holds the same position in the generative region as the testis in the male. It has a more sac-like appearance, however, and its lining membrane is thicker than in the case of the testis. In the anterior and posterior regions of the segment there is not the great thickening of the wall seen in the male. It is more uniform in thickness, and the contour of the limiting membrane throughout more distinct. In sagittal sections in the median line it appears as a long chamber lying ventral to the gut (fig. 22). Foettinger’s account of the oviduct and funnel is correct, the funnel being large and collapsible, ciliated on its dorsal side only. It projects downwards into the ventral region of the middle third of the generative segment. Its ventral lip is a short distance from the nerve-cord on either side; its dorsal lip is the longest, and almost meets that of the opposite side in the median line. The funnel is composed of a large number of flattened cells, a conspicuous one being usually seen in the edge of the dorsal lip. The cilia are remarkably stiff and short. It leads into a small, round ampulla which is usually crowded with spermatozoa. This leads into a still larger one, the walls of which are drawn out in a number of digitations. This functions as a sort of shell- gland. Its lumen is filled with a granular secretion that forms the egg-capsule. A short canal Jeads from the second ON THE ANATOMY OF HISTRIOBDELLA HOMARI. oo ampulla to the exterior. When the ovary is full of ova it is sometimes difficult to see the funnel and oviduct, as its lumen is obliterated by compression against the body-wall. The inner surface of the ovary is closely invested with a thin layer, of nucleated cells—the true coelomic epithelium. It is from this in the anterior region that the primitive ova arise. ‘his takes place close to where the sac abuts against the end of the third segment. Here certain of the nuclei are much larger than the rest. They are the odgonial cells. They have relatively little cytoplasm and large, transparent nuclei. A considerable number of them are seen at this point in different stages of development. The fact that the o6gonia arise from a small, circumscribed portion of the anterior end of the ovary, and not from its epithelial surface in general, recalls the condition described by Nelson (25) in Dinophilus conklini, which differs from the other species of this group, D. vorticoides, D. teniatus,and D. gigas, in that only a small portion of the ovary likewise gives rise to the odgonial cells. It is evident that the epithelium of the middle and posterior portions of the ovarian cavity play no part in their formation. As they pass backwards and become the primary odcytes, the epithelium of this part of the cavity throws out processes that attach themselves to the growing oécytes, folding up round them and forming a supporting matrix crowded with small nuclei. ‘They furnish them with the material for their growth, but beyond this take no part in their formation. As the odcyte grows these follicle cells diminish rapidly in size, and their nuclei undergo degenera- tion, becoming long and granular. ‘They appear to have something to do with the formation of the yolk-granules, but how this is accomplished is not plain. These arise in situ, as nothing similar to them can be distinguished in the follicle cells, which are always clear and transparent. At the time of their formation the granules are also clear and transparent, and only acquire their dark appearance after they have been formed some time. For this reason the {small odcytes, although highly granular, are almost as transparent as the 336 CRESSWELL SHEARER. odgonial cells. By the time the odcytes reach the middle of the generative region they turn dark brown in colour. In a few days they increase greatly in size. Their outline becomes regular, and the superficial layer of their cytoplasm seems to stain much more intensely than the deeper portion. Their nuclei become large, round, and transparent, and are readily distinguishable in the living animal. ‘There appears to be no yolk-nucleus present, but the germinal nucleus goes through a number of changes during the formation of the deutoplasm, that probably has to do with the great elaboration of this material. The mature eggs are found in the posterior region, where they take up the greater part of the ovarian chamber. They measure from 80-200, in their longest diameter, according to the size of the female. They are oval in shape and somewhat flattened. They are highly granular, the granules being very uniform in size. Unlike Stratiodrilus, there may be a number of ripe eges within the chamber at one time, although one usually predominates in size over the others. In the violent move- meuts of the animal small fragments of the egg are some- times broken off by compression against the gut, or from friction against the other eggs of the cavity. These are seen to move about the cavity quite freely, and, by some peculiar cohesive process, are capable of joining up with the egg again. This can be seen taking place under the microscope. The fragments have a membrane of their own, and may be seen lying against the egg from which they have separated. The membrane between them breaks down, and they flow together rapidly. Normally the ripe ovum is almost divided in two portions by its compression against the gut. When a ripe ovum is discharged its place is immediately taken by the next in size. I have never actually observed the female in the act of depositing her eggs; as I have mentioned, this takes place usually at night. From the fact that the funnel in the female is well forward at the generative region and the ripe ova are. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 337 sometimes far back at the caudal end of the ovary, they have to move some considerable distance forward before they can find exit through the oviducts. In passing through the second ampulla the egg is surrounded by its capsule, which binds it firmly to the membranes of the lobster’s ova, or the hairs of the carapace surface of the branchial chamber. I have already drawn attention to the fact that the odcyte commences to prepare for the first maturation division and the extrusion of the polar bodies when it has acquired only a portion of its yolk material. It is noteworthy that only one of the odcytes undergoes this change at a time. It is the most advanced and the largest. While the amphiaster is seen in this egg, I have never observed it in any of the younger ones, although some of these to all external appear- ances are as large and as mature as the one in which it has appeared. As the odcyte prepares for maturation its staining reaction changes. Up till this time the superficial layer of its cyto- plasm stains darkly, while the deeper portions surrounding the nucleus do not take the stain. With the appearance of the maturation spindle the staining reaction of the cytoplasm becomes uniform throughout the cell. The first sign of approaching maturation is announced by the changes undergone by the nucleus. It is distinguishable in the living egg as a clear spot in the middle of the dark granular cytoplasm. By a number of changes, which I have not followed in detail, the chromosomes form, the germinal vesicle breaks down, and the amphiaster of the first polar body forms. This at the moment it appears is very small, but erows rapidly with the growth of the egg. From the time it appears to the time it reaches its full dimensions it at least trebles its length, while the egg grows considerably in size. From measurements made of the length of the central spindle, from centrosome to centrosome, and the diameter of the egg in its longest axis, it was found that from the time the central spindle was clearly visible to the time it ceased to grow it trebled its length, while the egg a little more than doubled 338 CRESSWELL SHEARER. its longest diameter. The spindle seems to grow with the egg. ‘lhe size of the amphiaster is always proportional to that of the ovum. In the large female, where the eggs are almost double the size of those of the small ones, the spindle is correspondingly larger. ‘he size of the spindle is appa- rently determined by that of the cell. In Limulus, according to Munson (23), the growing centre of the egg is the vitaline body. ‘This, in the early stages, presents all the appearances and features of the centrosome and sphere, and, in fact, is the centrosome of the dividing oogonia. In later stages it remains as the definite centrosome in the cytoplasm. ‘Thus it appears as the primitive basis or centre of growth of the cytoplasm, building this in part from the granules supplied by the follicle-cells. In Histrio- bdella growth does not seem confined to the region near the umphiasters, but seems to take place generally throughout the cytoplasm of the egg. No yolk-nucleus or vitaline body is present. In sections of fixed eggs the cytoplasmic material in the immediate vicinity of the spindle is markedly less dense than in the peripheral region of the ovum. In some sections the middle of the ovum appears as a space, in the middle of which is the spindle with its chromosomes. The ovum goes through a portion of maturation during the time it is still adding material to its cytoplasm. While the achromatic threads of the amphiaster can be readily seen in the living egg, the chromosomes cannot be detected without staining. At the end of the prophase eight chromosomes are found in the equatorial plate of the spindle. The astral rays are much less definite than the strands of the central spindle. While the former seem in the living egg as if due to the arrangement of the yolk-granules in definite lines, the latter appear as actual threads running between the granules themselves. In speaking of the astral rays Wilson (82) says: ‘A careful study of their relation to the meshwork in the Echinoderm, and in many other forms (especially in Nereis, Thalassema, Lamellidoris, and Asterias), leaves no doubt in my opinion that they are actual ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 339 fibrille, that thread their way among the crowded alveolar spheres. In my best preparations the astral rays appear like wires bending to and fro among the alveoli” (p.13). “From a study of 'oxopneustes one would be led to the conclusion that they ,arise in rows of granules or microsomes, held together by the continuous substance” (p. 15). These words exactly describe the appearance of the astral rays in the living ege of Histriobdella. Towards the centre of the astral figures the rays appear as continuous fibres, while peripherally they break up into rows of granules. I believe in both the asters and the central spindle the granules do not build up the achromatic figure, but are merely incidental] to it. This is borne out by the fact that they are less numerous within it than in the surrounding cytoplasm. For this reason the area of the amphiaster in the living egg is always the most transparent. The archoplasm ean be distinctly seen as a clear substance running between the microsomes. The less dense nature of the astral rays, as compared with the fibres of the spindle, has been clearly demonstrated recently by Lillie (18) on centrifugalised eggs, where the egg-granules are readily driven through the substance of the astral rays, while they are stopped and forced to go round that of the spindle. The chromosomes in Histriobdella are arranged round the periphery of the equatorial plate. Hach chromosome lies directly against one of the spindle-fibres. These run from one centrosome to the other without any break in their con- tinuity. It is obvious that the chromosomes have no proper mantle-fibres, and that the number of fibres composing the spindle is in excess of that of the chromosomes. In sections the number of fibres can be counted. There are twenty, while there are only eight chromosomes. The centrosome itself is not distinguishable as a distinct point or granule in the living egg, but its position is indicated by a small area where the fibres of the astral rays and those of the spindle all converge on one another. No sphere can be distinguished. 340 CRESSWELL SHEARER. In the early stages, during the formation of the central spindle, its fibres in part appear to arise outside the area of the nucleus. In one instance I was able to distinguish the spindle-fibres beyond the still evident remains of the nuclear wall. ‘Ihe centrosome clearly arises beyond the limits of the nucleus, and from the reticulum of the cytoplasm, and its presence can be clearly detected before the dissolution of the nuclear wall. Much has been written on the origin of the spindle and the centrosomes as to whether they are of nuclear or cytoplasmic origin. It has been established that the spindle-fibres may arise from either. In the case of the mantle-fibres they arise almost invariably from the nucleus, while the spindle sub- stance proper arises from the cytoplasm, as has been shown by Meves (22) in Salamandra, Calkins (8) and Ishikawa (17) in Noctiluca, Flemming and Heidenhain (14) in leuco- eytes. In cases where no central spindle is present the astral rays seem to arise from the cytoplasm, as in a number of plants, some worms, as ‘l'halassema, according to Griffin (11), and in a number of Annelids as described by Mead (21). In other cases from the nucleus, according to Flemming (7), Rickert (27), Wilson (83), and Korschelt (18). According to Watase (81) the centre of the aster is merely the point where the greatest number of cytoplasmic filaments meet, the centrosome thus produced giving rise in turn to the spindle filaments. Thus the spindle-tibres originate from the centre of the aster, and not from the nucleus. This is clearly shown in the case he instances of the blastomeres of Loligo, where the nucleus remains a clear area in the middle of the central spindle. ‘There is a short period in the formation of the spindle in Histriobdella when almost the same con- ditions are shown. Again, the observations on eggs that have been artificially fertilised by salt solutions clearly point to the origin of the spindle quite independent of the nucleus. According to Wilson (84) all degrees exist between the asters that lie remote from the nucleus and of undoubted cyto- plasmic origin, and those close beside it. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 341 When the amphiaster attains the prophase, it remains in this stage until the egg is fertilised and deposited in the sea- water. If this does not take place, or if the conditions for egg-laying are unfavourable, it apparently remains in this state indefinitely, not making any further progress. In one instance I was able to keep a large female under observation for the greater part of a week with the amphi- aster of its largest egg in the prophase. At the end of this time the fibres of the central spindle and the astral rays were as distinct as at first, and showed no evidence of dissolution. It is evidently contact with the sea-water that 1s necessary to cause the completion of maturation and the extrusion of the polar body. The spindle is of considerable size, measuring from 50- 60 x from centrosome to centrosome. It can be readily seen in the living egg with the aid of a good hand-lens. As the animal moves and the egg outline is changed by compression against the body-wall, it does not change the position of its main axis with regard to that of the egg. According to Hertwig’s well-known law, as the result of the interaction of the nucleus and protoplasm the spindle comes to lie in such a position that its longitudinal axis corresponds with the axis that passes through the greatest protoplasmic mass. In figs. 18-20 are shown the position of the amphiasters in the egg as it has undergone change. The axis of the spindle, it will be seen, does not always correspond with that of the main axis of the egg, but on the whole it lies very close to this, and the cytoplasm always shows a tendency to group itself symmetrically about the spindle. I have made a number of observations that seemed to show that the form of the egg does not greatly affect the direction of the spindle-axis. In fig. 32 is shown the egg when it has undergone con- siderable pressure in its long axis through contraction of the animal. The spindle shows no appreciable shortening as the result of this pressure. In fig. 18 the egg shows the com- mencement of two furrows running into the cytoplasm, due to compression against the gut. In fig. 32 a small portion 342 CRESSWELL SHEARER, has been broken off the posterior end. This subsequently joined up with the egg again. No polar body is given off by the egg while it remains within the cavity of the ovary. Ihave had a female under observation for several days, and have been able to follow the growth and maturation of a particular egg from the first without seeing the formation of any polar body taking place. I have mentioned that the male is often seen to fertilise the female while she is without eges and still immature and in the larval state. In these females the sperm can be seen working their way through the tissues and finally collecting im the oviduct. I believe this invariably takes place. Whether the sperm, once in the oviducts, retain their vitality till the female reaches maturity aud bears eggs I have been unable to determine. It would seem that it is immaterial whether this does or does not take place. ‘he female is usually fertilised over and over again before she reaches maturity and bears eggs, so that fertilisation is probably effected by the last supply of sperm she may happen to receive. It is clear that the presence or absence of ova in the female play no part as a factor in fertilisation. No matter where the sperm are injected into the body of the female and the male exercises no choice in this respect— they seem to collect ultimately in the ampulle of the eviducts. It would seem as if some substance in this situation exerted a chemotactic influence over their movements, causing them to collect here from all parts of the body. The sperm are frequently seen in the blastoccelic cavity in small masses beneath the gut. In this situation they are still shut off from the cavity of the ovary and the eggs. In the anterior end of the ovary, crowded among the small odgonial cells, are frequently seen small masses of sperm. ‘hese appear to have undergone considerable change and to have partially lost their tails. I[t is probable that these sperm have gained access to the ovary by way of the oviducts. It is remarkable, however, that in the posterior region of the ovarian cavity no sperm are seen free among the ova, but they ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 343 would seem to be confined to its anterior region. Fertilisa- tion takes place within the ovarian chamber, as a large oblong sperm-nucleus is always found in the ovarian egg, in which the amphiaster has appeared. This always lies at some distance from the spindle and close to the egg-membrane, while the spindle is centrally placed. I have been unable to determine at just what stage in the growth of the ovarian egg fertilisation takes place. As the egg is seen to increase considerably in size after the amphiaster has appeared, and as the sperm nucleus is always found in the ovum when this 1s present, it is possible that the egg is fertilised at a stage in which the yolk-granules are first beginning to appear, hie fusion of the pro-nuclei takes place only after the polar body is extruded, and this takes place when the egg has been deposited in the sea-water. In Stratiodrilus Haswell has observed the fertilisation of the egg taking place within the ovarian cavity. In Dinophilus teniatus, according to Harmer (12), the same conditions hold regarding impregnation and fertilisa- tion asin Histriobdella. The penis is inserted anywhere under the skin, the act of copulation taking place repeatedly with the same female. He says, “ the act of copulation has no relation to the maturity of the ova of the female, nor is it prevented by the fact that the female has already received an ample supply of spermatozoa by a preceding operation” (p. 13). The spermatozoa can be seen collected in small masses beneath the gut. Fertilisation is therefore internal. The polar bodies are given off apparently when the eggs reach the exterior, or shortly after they are deposited in the sea- water. In the ripe egg, after the amphiaster of the first polar body has been formed in this manner, a remarkable occur- rence can be brought about, which demonstrates most clearly the semi-solid nature of the spindle itself. In compress- ing the cover-glass on a preparation of a living parasite I happened in several instances to rupture the body-wall in the vicinity of the ovum. The egg-envelope was also broken B44 CRESSWELL SHEARER. at the samé point. The yolk-granules then rapidly poured through the opening into the sea-water, and carried the amphiaster with them. It held together asa semi-solid body, and could be seen turning over and over as it was pushed along by the granules. Once in the water outside the body the granules tended to disperse, while the amphiaster remained with its immediate surrounding granules, apparently a solid body. It remained like this for several minutes until it finally dissolved and disintegrated. I have tried to represent this taking place in fig. 8. The asters go first, while the central spindle still remains intact. This seems to show that the substance of the spindle is of firmer texture than that of the asters and centrosphere ; and this is borne out, as I have mentioned, by the actual appearance of the archoplasmic substance of the asters as compared with the sharp, definite structure of the spindle. The yolk-granules adhere and seem almost a part of the archoplasmic substance of both asters and spindle, the amphiaster really appearing as a mass of brown yolk-granules held together by the thread-like archoplasmic substance. As the spindle begins to dissolve the yolk-granules can be seen being liberated from the trans- parent substance of the archoplasm and moving away in the sea-water. In fact the whole process of the dissolution of the amphiaster, as seen under an oil-immersion lens, is similar to that of some gelatinous substance slowly dissolved by the action of sea-water. That the spindle has some considerable rigidity is borne out by the fact that it keeps its shape, and can be seen rolling over and over as it is drawn along in the sea-water. It shows no tendency at first to flatten under the pressure of the cover-glass. ‘This is always considerable, although its corners are supported as much as possible by wax feet, as the capillary attraction invariably draws the cover- glass down somewhat in the middle. It is not till the spindle has begun to dissolve that this rigidity is lost, when it under- goes flattening. It at the same time becomes more trans- parent, the archoplasmic threads appearing as if actually undergoing dissolution by the sea-water, leaving the dark ON THE ANATOMY. OF HISTRIOBDELLA HOMARI. 345 yolk-granules behind them arranged in positions that had previously been held by the archoplasm. There is a short period during which the spindle almost remains alone, the asters having completely disappeared from either end. It is at this time that the spindle can be seen to roll over as it is pushed farther and farther away from the point of rupture in the body-wall by the escape from the egg of fresh cyto- plasm. In a uumber of experiments I subsequently ascertained that this rupture of the egg and extrusion of the spindle will not occur if the egg is far back in the body-cavity. The body-wall ruptures at its thinnest part, which is well forward in the generative region. If the egg has to move forward some considerable distance under pressure, before it can begin to flow through the rupture the amphiaster is usually broken and destroyed. It takes place most satisfactorily when the egg is only a short distance from the point of rupture of the body-wall. It can only be observed to occur when the amphiaster itself is fully mature. When not fully formed it dissolves immediately any movement of the yolk- granules takes place. Unless, moreover, the rupture in the body-wall is fairly large, the amphiaster is usually broken in the act of being forced through, being destroyed by the granules pushing it through from behind. In one instance the central spindle had the appearance of being composed of a mass of distinct threads, some of which on one side of the spindle had been injured and broken, the yolk-granules appearing as small grains entangled in these fibres. I think this observation clearly demonstrates the truth of a suggestion that has been put forward, that the achromatic threads and amphiasters are firm structures, or at least more rigid than the reticulum of the cytoplasm. Gardiner (10), in his paper on the egg of Polychcrus caudatus, states (p. 89), “That the amphiaster is much more rigid than the surrounding cytoplasm is shown by two instructive prepara- tions which were the result of accident. Ova containing 346 GRESSWELL SHEARER. amphiasters in the stage now under discussion were ruptured just before the worm containing them was placed in Hermann’s fluid. The cytoplasm had flowed or been pressed out of the ovum, carrying with it the amphiaster. In both cases the cytoplasmic network had been completely bent and twisted into a confused snarl. The achromatic rays were somewhat, but not nearly so much distorted, but the centrospheres were almost unchanged. From this I infer that the amphiaster and the rays are, on the whole, much more rigid than the cytoplasmic network or the cytoplasm from which they are formed.” Evidently the same thing took place in this instance as I have observed in Histriobdella, where the large size of the spindle and the granular nature of the egg renders the various steps in the process clearly visible under the microscope. By pricking the egg-membrane of Allolobophora, Foot and Strobell (9) have been able to get the egg contents on the slide, and there photograph it after fixation. “ By this method the germinal vesicle, and sometimes even the spindle, flow out of the egg-membrane intact” (p. 201). Some excellent photographs are shown of these in figs. 125-130 of this paper. In Allolobophora, as in Histriobdella, the early stages of the first maturation division are gone through by the egg while it is still within the receptacula ovorum. 10. ConcLUSION AND SUMMARY. Harmer (12) was the first to point out that Histriobdella was more closely related to Dinophilus than to any other Archiannelid, although Pierantoni (26), in his recent mono- graph, has placed Histriobdella and Dinophilus as an appendix to the Polygordide (including Protodrilus). Schimkewitsch (28) has contended that Dinophilus is closely related to the Rotifers, and Haswell (18) has put forward a similar claim for Histriobdella. In Histriobdella it is certain that the parasitic mode of lite has resulted in a peculiar specialisation, which, combined with its direct mode ON THE ANATOMY OF HISTRIOBDELLA HOMART. 347 of development, renders its relationship hard to determine, and hides the primitive characteristics of its organisation. That the Rotifers themselves are likewise a highly specialised class of somewhat uncertain affinities is an opinion that is gaining ground, since so much doubt has been thrown on their supposed relation to the Annelid trochophore. The work of Wesenberg-Lund (20) has shown that the most simple and trochophore-like of the Rotifers are probably the most highly specialised and the farthest removed from the Annelids. Yet the clearly segmented plan of both Dinophilus and His- triobdella, it must be admitted, is essentially similar to that of a Chetopod. This, combined with the clearly Polychet nature of egg-segmentation in Dinophilus, is sufficient to place these forms in direct connection with the Annelids, quite apart from either Protodrilus or Polygordius. Under the heading of the various organs I have already gone into a more or less detailed comparison of Histrio- bdella with Dinophilus, so that it is only necessary to review the subject here from a more general standpoint. In both forms the animal consists of a distinct head and trunk, the latter composed of relatively few segments. In both the nervous system consists of a well-defined brain or neuropile, and a double ventral nerve-cord, with metameri- cally arranged ganglia. In Dinophilus these are formed by transverse commissures, while in Histriobdella the two parts of the cord unite directly to form the ganglia. The external segmentation corresponds with that of the nervous system. Dinophilus does not possess the feet, cirri, or tentacles that so clearly mark segmentation in Histrio- bdella. But the metamerism is less definitely shown by the ciated bands, mucus glands, and the ring-like constriction of the body into a series of segments. On the other hand the nephridia show a more matamerically placed arrangement than they do in Histriobdella. In both (with the exception of Stratiodrilus) the nephridia open to the exterior in the seg- ment followimg that in which they arise, asin Annelids. In Histriobdella the muscular system shows a very high 348 GRESSWELL SHEARER. degree of development, and for this reason can hardly be compared with that of Dinophilus; in both, however, the main musculature consists in a series of longitudinal ventro- lateral and dorso-lateral muscles. The alimentary canal shows the same divisions, although differing considerably in the relative proportion of its parts. The strong chitinous jaws are wanting in Dinophilus. In each the cavity surround- ing the gutis a primitive blastocele with no definite epithelial lining. This cavity sends prolongations into the head. The equivalent of the ccelom in both is represented by the cavity of the reproductive glands. In the male these consist of a more or less paired testis, vesicule, vas deferens, and median penis, and in the female a large ovarian cavity, paired or unpaired, with oviducts. With Annelids Dinophilus shows a closer relationship than Histriobdella, mainly due to its less direct develop- ment. In fact the development of Dinophilus brings it into line with that large group of animals such as the Poly- chete, Echiuride, Gephyrea, Lamellibranchs, and the Gasteropoda, in having the ectoderm arising from the first three quartettes, mesoderm from the left posterior cell of the fourth quartette (4 p.), and the endoderm from the remain- ing cells. In the derivation of a large part of the ectoderm of the trunk from the posterior cell of the second quartette the resemblance to the Polychzt Annelids is most pronounced. This is further enforced in the origin of the bilateral cleavages in the cross cells and in the products of 2p. ‘The transition from the spiral type of cleavage to the more specialised bilateral type occurs in precisely the same directions as in the Polychets. Moreover the second bilateral divisions of the cells of the posterior arms of the cross continue this resemblance. All these characters, if such they may be called, when viewed as a whole point in no uncertain way - to the descent from the Annelid stem, and at a point not far from that at which the Polycheta arose ” (Nelson, p. 728). The weight of our evidence, furnished by recent work on the morphology and embryology of Dinophilus, is strongly ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 349 in favour, therefore, of a close relationship with Annelids. The general ciliation, the caudal appendage, ciliated rings, nervous system, general configuration of the head, trunk, and alimentary canal are what are found in a number of Annelids, and most clearly in such a form as Ophyotrocha. Nelson (25) has even suggested that the pre-oral nerve commissures can be satisfactorily explained by deriving them from the nerve-ring of the Trochopore. He comes to the conclusion : “ On the whole, Dinophilus can best be considered as a very young Polychet worm, retaining some of its larval features, with sete and parapodia undeveloped, and whose peritoneum and ccelom have been transformed into a generative organ” (p. 135). The relationships of Histriobdella to Polygordius and Protodrilus have been gone into fully by Foettinger (8), Harmer (12), and Haswell (13), so I need not repeat their arguments for this relationship here. It seems to me, from the Archiannelid point of view, it is important to determine what features of Histriobdella are primitive, and what have been derived from its peculiar mode of life. Hisig (5) has gone so far as to suggest that in Histriobdella we have to do witha highly modified, possibly degenerate animal, and not an Archiannelid at all. If Histriobdella is a degenerate form then it must be a degenerate Chetopod as Haswell (18) has pointed out. “If we are to take this view, we must at the same time acknowledge that side by side with the supposed degeneration, there must have gone on a special development in certain directions; that, while the definite characters of the segmentation became lost, a special set of locomotor organs with an elaborate musculature became evolved.” “'This view appears to me to involve difficulties so great that they render the degeneration theory extremely improbable, and it seems to me more in accordance with the facts of the case to conclude that the Histriobdellide are really primitive Annulates, and that the rudiments of their specialised features have been inherited from forms lower in the scale” (p. 327). VOL. 55, PART 2.—NEW SERIES. 25 3590 CRESSWELL SHEARER. Apart from any degeneration I agree with Haswell (18) that the relationship of Histriobdella with Polygordius “is extremely remote, and not such as to justify their inclu- sion in the same class.” The absence in Histriobdella of a blood-vascular system, a distinct prostomium and _ peri- stomium, the presence of mouth opening well forward in the head, chitinous jaws, and complicated generative apparatus in the male, paired limbs, and mucous glands, clearly separate it from Polygordius and Protodrilus, placing it quite apart from these forms, With the Rotifers, on the other hand the relationship is undoubtedly more pronounced, Haswell has pointed out that all the main features of His- triobdella can be traced to this class, although in general features the resemblance is greater perhaps with the Gastrorichia than with the Rotifers proper. The chitinous jaws of Histriobdella can be readily homo- logised with the mastix of Rotifers. In the absence of solenocytes and the general similarity of the nephridia of Histriobdella to the flame-cell type nephridia of Rotifers, we have a further resemblance, In both the cuticle is firm and shows a tendency to contract into ring-like folds. In both, also, the generative organs, especially in the male, can be reduced to the same plan. In Paraseison we have a Rotifer not unlike Histri- obdella in many of its features. The body is elongated and worm-like, with a distinct head bearing the mouth at its anterior extremity. In the middle of a very rudimentary coronal disc which bears no ciliated apparatus are four small bundles of hairs, placed in two pairs. Behind the mouth are found the orifices of two glands, similar to those found on the anterior feet of Histriobdella, On the top of the head is a small tubercle representing the dorsal median tentacle of Histriobdella, There is a narrow cesophagus, which leads into a large cylindrical stomach. There is no gut, and the stomach, which is not ciliated, is definitely closed. But this condition has plainly been evolved within the limits of the genus, as it is not characteristic of other Rotifers, It is ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 351 noteworthy that Paraseison, like Histriobdella, is para- sitic, beimg found on the gills of the Crustacean Nebalia. It is undoubtedly with such forms as Paraseison among the Rotifers that Histriobdella must be compared, The oreatest objection to the comparison of Histriobdella with the Rotifer is encountered in regard, to the nervous system. This in Histriobdella is already so elaborated, and of that type found among the higher Annelids, as to be hardly com- parable to the diffuse, and less differentiated, and centralised system of Rotifers. I cannot agree with Haswell that Zelinka’s (86) discovery of a sub-cesophageal ganglion in Callidina and Discopus renders this comparison more easy. A further difficulty is found in the absence of any true metamerism in the Rotifers. This difficulty is possibly not so great when we consider the arrangement cf the transverse muscle-cells in such a rotifer as Discopus synaptex. Leaving aside any comparison, therefore, of the nervous system, it nevertheless remains a fact that Histriobdella undoubtedly resembles the Rotifers more closely than any other group of animals, If Histriobdella is related to the Rotifers it becomes necessary to determine the relationship of Dinophilus to the same class. Schimkewitsch (28) was the first to point out the similarity of the caudal appendage in Dinophilus to the foot of the Rotifer. In Dinophilus, as in the Rotifer, this is used in attaching the animal. In both forms there is a marked sexual dimorphism. But as Nelson (25) has pointed out, the caudal appendage in Dinophilus resembles more that of some of the polytrochal annelid larve than the foot of the Rotatoria, and the sexual dimorphism can have arisen within the genus, as it is found in other groups of the Annelida besides the Rotifers. One striking difference between the Rotifers and Dinophilus is the apparent total absence of a definite mesoblast in the Rotifers, while it is clearly present in Dinophilus, where it has the same cell- origin as in Polychets. In Rotifers the mesoblast would seem to be represented by the germ-cells alone, and it is 852 CRESSWELL SHEARER. necessary to suppose that the Rotifers separated from the main stem of the Annelida at a stage earlier than that of the formation of a definite mesoderm, while Dinophilus arose only after the ccelo-mesoblast had definitely appeared. On the whole, Dinophilus is not so closely allied to the Rotifers as Histriobdella. Unfortunately our lack of information with regard to the development of the ccelo-meso- blast in Histriobdella prevents our forming any opinion as to how much it resembles the Rotifers in this respect. It is remarkable with regard to the Rotifers that, despite their wide distribution and their great number of species, so comparatively few marine forms should be known. What has become of these if they have ever existed? Are forms like Belatro and Hemidasys (Claparéde, 4), Turbanella (Schultz, 29), or the Echinoderes (Zelinka, 37) to be looked upon as the modified descendants of a marine branch of these animals? Here we have a marked metamerism coupled with the main features that characterise both Histriobdella and the Rotifers. It is possible that it is with some of these somewhat obscure groups that the relationship of Histriobdella really lies. In conclusion, it may be stated that our present knowledge does not warrant us farther than to conclude that Histrio- bdella is a highly specialised form, retaining many Rotiferan features, and that it is to be grouped with Dinophilus asa primitive Annulate, but not directly related to Polygordins and Protodrilus. SUMMARY. (1) Histriobdella homari is a normal inhabitant of the branchial chamber of the European lobster. It is found in equal numbers throughout the year, on both the male and female. (2) he anterior feet of the head, unlike those of Stratio- drilus, are non-retractile. (3) There are four pairs of nephridia in both sexes. They are closed, and are of the primitive flame-cell type similar to ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 353 those of Rotiters. Unlike those of Dinophilus, they bear no solenocytes. (4) There is a pair of salivary glands in connection with the mouth. — (5) There are fewer teeth in the jaw-apparatus than, Foettinger has represented. (6) The ventral nerve-cord is composed of two portions, which separate in the intersegmental to unite in the segmental regions, in prominent ganglia. The metamerism of the nervous system corresponds with that of the external form. (7) In the male there isa complicated generative apparatus. It is similar in all respects to that of the male generative apparatus in Stratiodrilus, (8) Fertilisation takes place internally. The largest egg is usually seen in the prophase stage of the first maturation division. ‘he amphiaster and the spindle can be seen to pass out through the body-wall with the cytoplasm, when the egg is ruptured by pressure. It remains for some seconds intact in the sea-water surrounded with yolk-granules. (9) In the equatorial plate there are eight chromosomes in the first maturation division. (10) Histriobdella is to be placed close to Dinophilus. It retains many Rotiferan features, and is more closely con- nected with this group than Dinophilus. Histriobdella and Dinophilus show distantrelationship with Polygordius and Protodrilus, but cannot be classed with them as true Archiannelids. LITERATURE. 1. Van Beneden, J. P._—‘ Note sur une larve d’annélide d'une forme tout particuliere, rapportée avec doute aux Serpules,’ * Bull. Acad. Roy. Belgique,’ tome xx, 2nd pte., 1853, p. 69. “Histoire naturelle d’un animal nouveau, désigné sous le nom d@Histriobdella,” ‘Bull. Acad. Roy. de Belgique,’ 2nd serie, tome v, 1858, p. 270. 354 CRESSWELL SHEARER. 3. 10. iH 12. 13. 14. 15. 16. is 18. Calkins, G. N.—* Mitosis in Notiluca miliaris and its Bearing on the Nuclear Relations of the Protozoa and Metazoa,” ‘Journ. of Morph.,’ vol. xv, 1899, p. 711. . Claparede, E.—* Observations sur les Rotateurs,” ‘Annals de Sci. Nat. Zool.,’ V ser., t. viii, 1867, p. 5. . Hisig, H.—* Die Entwicklungsgeschichte der Capitelliden,”’ ‘ Mitt. a. d. Zool. Stat. Neapel,’ Bd. xiii, 1898, p. 1. . Flemming, W.— Zellsubstance Kern, und Zellteilung,’ Leipzig, 1882. ‘Zur Mechanik der Zelltheilung,” ‘ Arch. f. Mik. Anat.,’ vol. xlvi, 1895, p. 696. . Foettinger, A.—‘* Recherches sur l’organisation de Histriobdella homari,” ‘ Arch. de Biol.,’ vol. v, 1884, p. 435. . Foot, K., and Strobell, E, C.—*‘ Prophases and Metaphase of the First Maturation Spindle of Allolobophora fwtida,” ‘Amer. Journ. Anat.,’ vol. iv, 1905, p. 199. Gardiner, E. G.—* The Growth of the Ovum, Formation of the Polar Bodies, and the Fertilisation in Polycherus caudatus,” * Journ. of Morph.,’ vol. xv, 1898, p. 73. Griffin, B. B.—‘ Studies on the Maturation, Fertilisation, and Cleavage of Thalassema and Zirphea,” ‘Journ. of Morph.’ vol, xv, 1899, p. 583. Harmer, 8. F.—* Notes on the Anatomy of Dinophilus,” ‘ Journ. of Marine Biol. Assoc.,’ N.S., vol. i, 1889, p. 1. Haswell, William A.—* On a New Histriobdellid,” ‘Quart. Journ. Mier. Sci.,’ vol. 43, 1900, p. 299. Heidenhain, M.—** Neue Untersuchungen iiber die Centralkorper und ihre Beziehungen zum Kern und Zellenprotoplasma,” ‘Arch. f. Mik. Anat.,’ Bd. xliii, 1894, p. 423. Hempelmann, F.—* Zur Morphologie von Polygordius lacteus Schn. und P. triestinus,” ‘ Zeit. f. wiss. Zool., vol. Ixxxiv, 1906, p. 527. Hermann, F.—* Beitrag zur Lehre von der Entstehung der karyo- kinetischen Spindel,” ‘ Arch. f. Mik. Anat., Bd. xxxvii, 1891, p. 569. Ishikawa, C.—“ Studies on Reproductive Elements: II, Notiluca miliaris,” ‘Journ. Coll. Sci. Imp. Univ. Japan,’ vol. vi, 1894, poze: Korschelt, E.—* Ueber Kernteilung, Eireifung und Befruchtung bei Ophryotrocha puerilis,” ‘Zeit. Wiss. Zool., Bd. Ix, 1895, p. 543. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 3590 Lillie, Frank B.— Karyokinetic Figures of Centrifuged Eggs,” ‘Biological Bull.,’ vol. xvii, 1909, p. 101. Lund, C. Wesenberg.—* Danmarks Rotifera: I, Grundtraekkene i Rotiferernes Okologi,” ‘Morfologi og Systemstik.” Kobenhavn, 1899. Mead, A. D.—*‘ The Origin of the Egg Centrosomes,” ‘Journ. of Morph.,’ vol. xii, 1897, p. 391. Meves, F.—‘* Uber eine Metamorphose der Attractionsphire in den Spermatogonien von Salamandra maculosa,” ‘ Arch. f. Mik. Anat.,’ Bd. xliv, 1894, p. 119. Munson, J. P.—‘ The Ovarian Egg of Limulus: A Contribution to the Problem of the Centrosome and Yolk Nucleus,” ‘ Journ. of Morph.,’ vol. xv, 1898, p. 111. Nelson, J. A.—‘* The Early Development of Dinophilus,” ‘ Proc. Acad. Nat. Sci. Phil.,’ 1904, p. 687. “The Morphology of Dinophilus Conklini n. sp.,” ‘ Proce. Acad. Nat. Sci. Phil.,’.1907, p. 82. Pierantoni, U.—‘‘ Protodrilus,” ‘Fauna und Flora des Golfes von Naple,’ 31 Monographie, 1908. Rickert, J—‘ Zur Hireifung bei Copopoden,” ‘ Anat. Heft,’ Bd. iv, 1894, p. 261. Schimkewitsch, W.—‘‘ Zur Kenntniss das Baues und der Entwick- lung des Dinophilus vom Weissen Meere,” ‘Zeit. f. Wiss. Zool.,’ Bd. lix, 1895, p. 46. Schultze, M.—‘* Uber Chetonotus und Ichthydium Ehrb. und eine neue verwandte Gattung Turbanella,’ ‘Arch. f. Anat. u. Phys.,’ Jabrg. 1853, p. 241. Shearer, C.—‘‘ On the Structure of the Nephridia of Dinophilus,” ‘Quart. Journ. Micr. Sci.,’ vol. 50, 1906, p. 517. Watasé, S.—* Homology of the Centrosome,” ‘Journ. of Morph.,’ vol. viii, 1893, p. 433. Wilson, E. B.—‘On Protoplasmic Structure in the Eggs of Echinoderms and some other Animals,” ‘Journ. of Morph..,’ vol. xv (Suppl.), 1899, p. I. —— ‘The Cell,’ New York, 1900. “ Experimental Studies in Cytology : I,” ‘Arch. f. Entwick.,” vol. xii, 1901, p. 529. 356 CRESSWELL SHEARER. 85. Zelinka, C.— Die Gastroctrichen,” ‘ Zeit. f. Wiss, Zool., Bd. xlix, 1890, p. 209. 36. “Studien iiber Raderthiere,”’ ‘ Zeit. f. Wiss. Zool.,’ xliv, p. 396, Bd, xlvii, p. 353, Bd. liii, p. 1, 1885-1892. 37. “Uber Echinoderes,” ‘Verh. d. deutschen Zool. Gesell.,’ 4th Jahrssam., 1894, p. 46, EXPLANATION OF PLATES 17—20. Illustrating Mr. Cresswell Shearer’s paper ‘‘ On the Anatomy of Histriobdella Homari,” LETTERING. ac. Accessory glands of the male reproductive apparatus. an. Anal aperture. ap. p. Appendage of the posterior leg. ble. Blastoccelie cavity. bl. m. Muscular organs of jaws. br. Bridle piece of jaws. brn. brain. ed. g. l-ed. g. 3. Ganglia of the caudal region. el. Clasper. cl. p. Ciliated pit of the head. ca. Celom. ce. ep. Celomic epithelium. com. Nerve commissures. f. Fulcrum of jaws. fol. Follicle cells. jgl. Flagella of the nephridial canals. g. 1-g. 5. Ganglia of the ventral nerve- cord. gl.cl. Gland-cell of clasper. gr.g. Granule gland. nt. Intestine. int.2. Intestine, posterior part. j. Jaws. j.1. Upper ramus of jaws, j. 2. Lower ramus of jaws. Jl. a. Anterior legs or feet. 1. p. Posterior legs. m. Mouth. m.d. Dorsal longitudinal muscles. m.d.p. Median duct of the penis. m. gl. Salivary glands of the mouth. m. ob. Oblique muscles. m.v. Ventrallongitudinal muscles. n.c. Ventral nerve-cord. neph. 1-neph.4. Nephridia. neph.c. Nephridial canals. neph.h. Head of the nephridium. neph. 0. Opening of the nephridial canal on the external surface. neph.s. Spaces on the course of the nephridial canals. es. G@sophagus. o.im. Immature ova. ov.p. Orifice of the penis sheath. ov. Ovary. ovd. Oviduct. p. Penis. 7. Ramus of upper jaw. sprm. Spermatidia. spe. Spermatocyte. st. Stomach. ¢. 1. Median tentacle. ¢.2 and ¢.3. Lateral tentacles. te. Testis. th. Teeth. v. def. Vas deferens. ves. Vesicula seminalis. vn.c. Ventral nerve-cord. vit. Vitellarium or shell-gland. PLATE 17. Fig. 1—Female Histriobdella with eggs. The largest egg shows the presence of a maturation amphiaster x 300. Fig. 2.—Clasper extended. ON THE ANATOMY OF HISTRIOBDELLA HOMARI. 357 Fig. 3.—Third and fourth nephridium in the female. This and all the subsequent figures of the nephridia have been drawn from living preparations impregnated with methyl-blue; 2mm. oil-immer., comp. oes. 4and 6, x 500 and x 1000. Fig. 4.—Third nephridium in the male. Fig. 5.—Third nephridium in the male. Fig. 6.—Second nephridium in the male. Fig. 7.—Male Histriobdella with claspers retracted, x 300. PLATE 18. Fig. 8.—Rupture of an egg through the body-wall in a living prepara- tion by compression of the cover-glass. The first maturation amphi- aster is seen outside the body-wall in the sea-water. x 500. Fig. 9.—Male with claspers extended. x 300. Fig. 10.—Third and fourth nephridium in the male. x 800, Fig. 11.—Generative segment in the male. Taken from a living preparation, showing the reproductive organs. Fig. 12.—Fourth nephridium in the male. x 800. Fig. 13.—Fourth nephridium in the male. Segment contracted. x 800. Fig. 14.—Second nephridium in the female. x 800. PLATE 19. Fig. 15.—Reconstruction of the nervous system in the male, showing the dorsal curve taken by the ventral nerve-cord in the region of the penis. Lateral view. x 300. Fig. 16.—Young in egg-capsule. Fig. 17.—Young in egg-capsule. Harlier stage than that shown in fig. 16. Fig. 18.—Odcyte with first maturation amphiaster. This, with the subsequent figures, 19, 20, 26, and 32, are all drawn from the same egg- cell. They show the changes of shape assumed by the egg in the move- ments of the animal. They were drawn at intervals of from ten to twenty minutes. Fig. 19.—Odcyte, same as that shown in fig. 18, drawn twenty minutes later. Fig. 20.—Odcyte, same as that of fig. 19, fifteen minutes later. 358 CRESSWELL SHEARER. Fig 21.—Reconstruction of thejnervous system in the female. The brain surface is measured from the ganglion cells and not from the fibrous core. Ventral view. x 300, Fig. 22.—Sagittal section in the female showing the sac-like nature of the ovarian cavity. Fig. 25.—A cell of the accessory gland of the male. Fig. 24.—Transverse section in the male in the region of the penis. x 400. Fig. 25.— Eggs attached to the membranes of the lobster “ berry.” Fig. 26.—Odcyte twenty minutes after that shown in fig. 20. Fig. 27.—Spermatidia. Fig. 28. — Reconstruction of the nervous system of the male. Ventral view. x 300. Fig. 29.—Longitudinal section of the wall of the intestine in the posterior region. Fig. 30.—Young, a short time after hatching. x 300. Fig. 31.—Section through the generative region in the male showing the granule glands. 5 5 Fig, 32.—Odcyte twenty minutes later than fig, 26. PLATE 20. Fig. 33.—Transverse section in the male through the region of the vesicule seminales. Fig. 34.—The same. Ina region a little posterior to the last. Fig. 35.—Transverse section through the middle of the generative region in the male. Fig. 36.—Chitinous jaws. Fig. 57.—Horizontal section in the male. Fig. 388.—Transverse section through the middle of the second segment. Fig. 39.—Transverse section through the anterior region of the generative segment in the male, showing the divided nature of the anterior portion of the testis. Fig. 40.—Transverse section through the posterior region of the generative segment in the male. Fig. 41.—Transverse section through the dorsal region of the body- wall in the caudal segment, showing the insertion of the oblique muscle- fibres. ON THE ANATOMY OF HISTRIOBDELLA HOMARL. 359 Fig. 42.—Horizontal section through the region of the first segment in the male, showing the head of the first nephridium. Fig. 43.—Transverse section through the neck region. Fig. 44.—Transverse section through the caudal region, showing the oblique muscles. : —_ ° 3 1 - * S = | os . - 7) . 4 & i _ » - a: : 4 - : & 7 _ y . = f mf , “a Jf i ' - = at = 4 =" cr * 5 ; ; caer a i 8 , : a P ‘ : o . Qua rb. umn clio Dek. Vol. I NS HM. £4 neph Bf ; nepl £ Hath, Lith? London “A reph..¢, Cae burn Mier Sc. Vol.5INS ALIS. Duart. aheanegens mda. Huth, Lith™ _ - Sourn Mucr Sci. lo b5, NS. Te i é ? ! id a ww ° a # . ’ : Tyo - So ee eee ee elie eR sr Ny Pe Conenge Pee Ss, C.S. & G.B.R, del. es Quart. Fournr Mice. Vb SSNS, 20 3 ? oe Bre aN Huth, Lith? London. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 361 On the Artificial Culture of Marine Plankton Organisms. By E. J. Allen, D.Sc., Director of Laboratories and Secretary of the Marine Biological Association ; and E. W. Nelson! Assistant Naturalist. Con'rEN'TS. Introduction . ; : I. Culture of Plankton Diatoms A. Practical Culture Methods . Miquel’s Method : : . Houghton Gill’s Method , F : . (A) Modification of Miquel’s Methods: “ Miquel Sea-water ” : ; , (8) English Channel Water . (c) Tank-Water ( ( — bo ey) p) Animal Charcoal Water . E) Peroxide of Hydrogen Water (r) Cultures in these Media . : ‘ p. Experiments with a view to Determining the Con- ditions which underlie the Successful Culture of Diatoms ; Methods. : The Sea-water employed : The Constituents of Miquel’s Solutions . Animal Charcoal and Peroxide of Hydrogen Reviving Exhausted Cultures . Silica Organic Infusions Co co ~I bo 2 O89 GW OD ~I sI NI <1 Oe Ww “Ss eo) 1 Owing to pressure on our space, this memoir could not be pubiished when first in type. It has in the meantime been issued in the ‘ Journal of the Marine Biological Association,’ vol. vili, No. 5.—E. R. L. 862 KE. J. ALLEN AND E. W. NELSON, PAGE Artificial Sea-water . : , : 395 Alkalinity . P . , . 395 Salinity : : 3 : . 402 Light ; : : f . 403 Temperature ' ‘ ; . 404 General Conclusions . : : . 405 II. Mixed Cultures, : : : . 407 III. Notes on Particular Species of Diatoms, on their Methods of Reproduction, and on other Alge occurring in Cultures ‘ ; j : . 412 IV. Rearing of Marine Larve . é 5 aie Methods A ‘ ; : | AZ Echinus acutus : : : . 419 E.esculentus . : : : . 420 E. miliaris . ; ; : _ wer Cucumaria saxicola, 5 . 422 Pomatoceros triqueter : ; . 422 Chetopterus variopedatus ; ; . 423 Sabellaria alveolata . ; ‘ . 423 Archidoris tuberculata : . . 423 Calanus finmarchicus. : . . 424 Echinus hybrid . - ; ’ . 425 Sacculina carecini : ; ; ; 25 Summary of Method for Rearing Larve : . 426 Bibliography . 5 ‘ . ; . 427 INTRODUCTION, THe observations to be recorded in this paper were com- menced in March, 1905, They originated in an attempt to find a general method for rearing marine larval forms. Several investigators had previously succeeded in rearing echinoderms, molluscs, and polychztes from artificially fertilised eggs, under laboratory conditions, but the process was generally difficult and the results more or less uncertain. The most promising method seemed to be that adopted by Caswell Grave (26), who was able to rear his larve by feed- ing them on diatoms. Grave obtained his diatoms by placing sand, collected from the sea bottom, in aquaria, and using such diatoms as developed from this material. All the methods, however, suffered from the uncertainty of not ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS, 363 knowing what organisms were introduced into the aquaria in which the larve were to be reared, either in the original sea- water or along with the food supply. It appeared, therefore, at an early stage of the work, worth while to make an attempt to carry out rearing experiments on amore definite and precise plan, to endeavour, in fact, to introduce the larve to be reared into sterile sea-water, and to feed them with pure cultures of a suitable food. his was the-ideal to be aimed at. As a matter of fact it has seldom, if ever, been attained in practice; nevertheless, a consider- able measure of success has been achieved by working upon these lines, and during the course of the work innumerable problems relating to the physical conditions under which plankton organisms can best flourish have presented them- selves. Some account of the experiments made may be of interest to other workers, although imany of the problems raised are not yet solved, notwithstanding the fact that some 1500 cultural expernnents have been under observation, It is rather with a view of stimulating other work upon similar lines than of bringing forward conclusive results, that this paper is being published. In the summer of 1907 Mr. E. W. Nelson became associated with the investigation, and since that date the experimental work has been carried out by him. ‘The discussions in this paper of a more chemical character, particularly the section on alkalinity, are almost entirely the work of Mr, Nelson, and we have both had throughout the advantage of the constant advice and help of Mr. D. J. Matthews on all such matters. I. Curture or Pranxton Dria'roms\ (aA) Practical Culture Methods. 1, Miquel’s Method.—Attention was first directed to the culture of plankton diatoms; and the methods, which had been elaborated by Miquel (11) for fresh-water diatoms and had been found by him to succeed with marine bottom diatoms, were tried. 364 kK. J. ALLEN AND E. W. NELSON. The essential features of Miquel’s method, as applied to marine diatoms, are as follows : Two solutions are prepared : Solution A. Magnesium sulphate . ‘ . 10 grm. Sodium chloride , : > LOOP Se Sodium sulphate 3 ‘ : (eS Ammonium nitrate L. Potassium nitrate, ; «.\ Gea Sodium nitrate . BA ie Potassium bromide . , >, (eae Potassium iodide i : . ts Water ‘ , : ; + 0G te Solution B.! Sodium phosphate ee Calcium chloride (dry). ‘ee Hydrochloric acid. : . eer Ferric chloride . : : . > Shea Water : ‘ ; , . ‘SO0'ce Forty drops of solution A and ten to twenty drops of solu- tion B are added to each 1000 c.c. of sea-water, and the sea- water is sterilised by keeping it at 70° C. for about twenty minutes. According to Miquel it is also necessary to add “ organic nutritive material in the form of bran, straw, or filaments of 1 «The preparation of solution A presents no difficulty. Solution B should be made up as follows: To the sodium phosphate dissolved in 40 ¢.c. of water are added first the 2 ¢.c. of hydrochloric acid, then the 2 cc. of hydrous ferric chloride, and then the 4 grm. of calcium chloride dissolved in 40 c.c. of water, taking care to shake the mixture, which I call phospho-ferro-calcic solution. The addition of this last solution to the maceration throws down a slight brownish flocculent precipitate, formed for the most part of ferric oxide, which should be carefully separated from the liquid used for cultivations.” > * Acid chlorhydrique pur a 22°.” Presumably meaning degrees Baumé = sp. gr. 1169. 3“ Perchlorure de fer liquide 4 45°." As above = sp. gr. 1421. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS, 865 weeds, such as Zostera. Macerations of these should be made up separately some time before they are required for use, and should be carefully filtered and sterilised. Organic matter must, however, be used very sparingly, or else putre- faction will set in arid the cultures will be irrevocably lost.” As a matter of fact we have found that such organic infusions are unnecessary when dealing with plankton diatoms, and it has not been our practice to employ them (cf., however, p. 392). Miquel obtained cultures of single species of diatoms either by picking out individual diatoms under the microscope and introducing them into the prepared water, or by adding a small quantity of water containing a mixture of diatoms and other organisms to some prepared water, and subdividing this into'a number of tubes. If the subdivision has been carried out sufficiently some of the tubes may contain one kind of diatom only, from which fresh cultures can be made. In this way, by repeated subdivision, cultures can be obtained which, by inoculating fresh quantities of prepared water from time to time, may, with care, be maintained indefinitely. Such cultures, however, must practically always contain bacteria, and Miquel distinguishes them from bacteria-free cultures, which he terms “ cultures des diatomées a l’état de pureté absolue.” The latter he found very difficult to obtain, but through repeated washing in sterile water, followed by frac- tional subdivision, he succeeded in getting some in which he could -find no trace of bacteria by ordinary bacteriological methods (Miquel [11], p. 155; ef. also Richter [16-181). We propose to call any diatom culture which can be carried on practically mae by inoculating fresh supplies of prepared water a “per sistent” culture, the term “pure” culture being reserved for cultures which can be proved to contain not more than one organism. We are not satisfied that we have yet succeeded in obtaining’ cultures of the latter kind. For the most part our persistent cultures contain one species of diatom only, and are free eu all organisms larger than small flagellates. In our Pacer experiments with plankton diatoms we VOL. 55, PART 2.—NEW SERIES. 24 366 BE. J. ALLEN AND E. W. NELSON. obtained persistent cultures, containing a single species of diatom, by both of the methods recommended by Miquel. We, however, have rarely succeeded by picking out single diatoms or chains of diatoms, for although we have passed the selected diatom through several changes of sterilised sea-water, the resulting cultures, even when the diatoms have multiplied to some extent, have generally shown evidence of contamination by harmful organisms, and have soon died down, Only in one of the earliest experiments, and in one more recent, has complete success resulted, In the first case a small chain of six or eight frustules of Skeletonema costatum, picked out in April, 1905, gave rise to a culture which still persists (November, 1909). Subcultures can still be obtained even from the original flask inoculated in April, 1905. In the second case a chain of eight or nine cells of Chetoceras densum, picked out from a Petri dish culture, has given a particularly good growth, The method of dilution and subdivision has been more suc- cessful, and persistent cultures of a number of species have been obtained in this way. A more ready method of obtaining the cultures is, we have found, to add one or two drops of plankton to, say, 250 c.c. of a suitable sterile culture medium, and to pour this into shallow glass dishes (Petri dishes). The dishes should be placed in a position as free as possible from vibration, and where they can be easily examined with a lensin situ. The temperature should be kept as constant as possible and the dishes exposed to light of moderate intensity, direct sunlight being avoided. In the course of a few days, colonies of diatoms of different species will be seen at different spots on the bottom of the Petri dishes. These can be picked out with a fine pipette and transferred to flasks containing fresh culture medium. The colonies should be picked out from the Petri dishes at as early a stage as possible, because if left too long some one organism, a diatom or a flagellate, may have multiplied so rapidly that the whole of the water in the dish becomes infected with it. In this case persistent cultures of ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 867 a single species would not be obtained. The above method is similar to one described by Miquel, excepting that he placed gelatinous silica at the bottom of the vessel. Some very successful persistent cultures were obtained from the follow- ing experiment, which will serve to illustrate the method: A sample of plankton, from a very fine-mesh bolting-silk tow-net, was diluted down with sterile sea-water, until a single drop examined under a 2 in. objective contained on an average ten organisms, chiefly diatoms, of various species. Petri dishes (4 in.), containing 60 c.c, each of Miquel sea-water, were then inoculated with various numbers of drops of the diluted plankton. The two dishes, to which two and three drops respectively were added, gave the best results, and from these persistent cultures of several species of diatoms were obtained. Hence we may conclude that the most advan- tageous number of single cells or short chains of cells to be added to a 4 in. Petri dish, containing 60 c.c. culture medium, is about twenty to thirty. We have succeeded in obtaining the following species of plankton diatoms in persistent cultures: . Asterionella japonica Cleve, Biddulphia mobiliensis (Bail.) Grun. Biddulphia regia (M. Schultze).! Chetoceras densum Cleve. Chetoceras decipiens Cleve. Chetoceras constrictum Gran. Cocconeis scutellum EKhr. var. minutissima Grun, | Coscinodiscus excentricus Ehr.’ Coscinodiscus Granii Gough. Ditylium Brightwellii (West) Grun. Lanuderia borealis Gran. Nitzschia closterium W. Sm. Nitzschia closterium W. Sm. forma minutissima. Nitzschia seriata Cleve. Rhizosolenia stolterfothii H. Perag. 1 See p. 413. 7 See p, 412, 368 E, J. ALLEN AND E. W. NELSON. Skeletonema costatum (Grev.). Streptotheca thamensis Shrubs. Thalassiosira decipiens Grun.! It is hardly necessary to add that in dealing with these cultures similar precautions to those used in bacteriological work must be taken, all vessels and instruments being care- fully sterilised before they are brought into contact with the prepared sea-water. ‘lhe cultures are best made in small, wide-mouthed flasks, which may be plugged with cotton-wool, or simply covered with watch-glasses. ‘The flasks should be kept at as uniform a temperature as possible (from 12°-17° C.) and should be exposed to strong daylight, direct sunlight being avoided. A flask should not be more than half filled with culture fluid, so that the surface exposed to the air may be large in proportion to the volume of fluid. Other Methods.—The addition of the solutions devised by Miquel to sea-water has in all cases given us good cultures of diatoms, and the method is certain in its action. We have, however, made numerous experiments by treating sea-water in other ways, with a view to finding ont what are the best conditions under which plankton diatoms will grow, and of arriving at some explanation of the action of the different salts contained in Miquel’s solutions. 2. Houghton Gill’s Method.—H. Houghton Gill (5), a contemporary of Miquel, made use of a culture medium not essentially different from that employed by the latter. Unfor- tunately he died before publishing his work, but an account of his principal results is given by Van Heurck. In his final method Houghton Gill made use of four distinct solutions, as follows : Solution 1. Crystallised sodium phosphate 2 Calcium chloride’. . . ee Syrup of iron chloride © . 0 Strong hydrochloric acid ‘ sea Water . ; : - 100,48 : ve p- 412, ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS, 9869 Solution 2. Crystallised magnesium sulphate 4 germ Crystallised sodium sulphate . ey: eee Crystallised potassium nitrate 4 ,, Common salt (sodium chloride) 4 ole Potassium bromide . : i ae Ue Potassium iodide. e : Or ew: Water. ‘ : ‘ : NOW! -& Solution 3. Crystallised sodium carbonate . Agrm, Weater . . . : : F 25LOOE Solution 4. Well-washed, precipitated calcium silicate . ; : : + 2 ST Water Loge All the salts employed must be chemically pure. Three c.c. of each of these liquids are added to 1000 c.c. of fresh water or sea-water (according to circumstances), and the whole sterilised. In his earlier work Houghton Gill added a sterilised infusion of grass or of diatoms, but it is not-clear from the accounts whether this was still employed with the above solutions. We have obtained very good cultures with the above solutions, to which we did not add any organic infusion. 3. (A) Modification of Miquel’s Method: “ Miquel Sea-water.”’—Since several of the components in Miquel’s formula for solution A (p. 363) are obviously unnecessary when sea-water is being used as the basis of the culture medium, we adopted for our own work the following modifica- tions: After some preliminary experiments it was found, as would be expected from the composition of sea-water, that the only salts of value to the medium are the three nitrates KNO;, NaNO;, NH,NO,, and possibly KBr and KI. The omission of the two latter was soon found to make no 370 E. J. ALLEN AND E. W. NELSON, difference. Experiments also showed that the formula for solution A could, without any appreciable detriment to results, be further simplified to the one salt KNO, or NaNOs, but not NH,NO,. At first the amount of KNO, dissolved in 100 c¢.c. distilled water, used to make the modified solution A, was the same as the sum of the weights of the nitrates in Miquel’s own formula, viz. 5 grm. But later experiments showed that a considerably greater concentration of KNOs than this gave more lasting cultures; the strength of solu- tion and amount to be added to a litre of sea-water in order to obtain the best results being 2 c.c. 2 M KNOs3. In the case of solution B no modification has been adopted, but it has been found that small variations in the amounts of the ingredients used do not affect the results. A convenient method for measuring the right amount of FeCl, is to warm the salt until it just melts in its own water of crystallisation, and to pipette out 2 c.c. with a previously warmed pipette. No temperature corrections need be considered. Also 2 ¢.c. of the ordinary pure concentrated hydrochloric acid at room- temperature will suffice. Our own formula for preparing Miquel sea-water is now : Solution A.! Potassium nitrate, 20 2 grm. t= 2 M KNO,. Distilled water, 100 eS Solution B.? Sodium phosphate (Na,HPO,12H,0) + : Calcium chloride (CaCl,6H,O) . :- Ferric chloride (melted) ee Hydrochloric acid (pure concentrated) 2 ,, Distilled water . : : : : SOnee ‘ This strength has only been used in the most recent experiments ; and solution A in this paper, unless otherwise stated, means the 5 % solution of KNO3. * For preparing this solution see p. 564. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 371 To each 1000 c¢.c. of sea-water! add 2 c.c. solution A and 1 c.c. solution Band sterilise by heating to 70°C. When cool, decant off the clear liquid from the precipitate, which will have formed when solution B is added to the sea-water. As a rule our cultures were made in 60 c.c. of this medium contained in short-necked, wide-mouthed flasks of 125 c.c. capacity, so that the proportion of air-surface to volume of liquid was large. The medium was found to give constantly satisfactory results. On inoculation from a persistent culture of such diatoms as Thalassiosira, Skeletonema, Chetoceras, etc., a growth visible to the eye is obtained in about ten days, and then multiplication takes place very rapidly. In from three weeks’ to a month’s timea very considerable growth will be seen making a brown, flocculent mass at the bottom and back of the vessel containing the culture. In from two to four months the culture begins to show signs of exhaustion and the frustules lose colour, but they do not, as in the case of sterilised outside and tank-water, com- pletely die off. A great number certainly do die, but some remain in a resting condition, and often, after a period of six months or so, these begin to multiply again and the culture regains its former vigour. his is probably due to the food- stuffs contained in the dead frustules going into solution again, possibly by means of bacterial action. This periodicity in cultures is interesting in that it resembles what takes place in the ocean. Cultures in this medium will persist indefinitely, so far as our experience goes. The oldest culture in our possession is one of Skeletonema costatum made at the very commencement of this work, dated April, 1905. Although the frustules in this culture are quite unrecognisable as any diatom now, on making a subculture in fresh Miquel a normal and healthy growth can always be obtained. In old cultures the diatoms are nearly always found to be very much deformed, and often appear to be only a mass of 1 * Miquel water” seems to succeed equally well, whether it is made by adding Miquel’s solutions to “outside water” or to “tank-water.” 372 HE. J. ALLEN AND EK. W. NELSON. broken-down chromatophores. Whether regeneration can be successfully obtained from a single chromatophore, which must presumably be contained within a cell-wall of some kind, has not been definitely decided, but results seem to point in this direction, At the start of a culture a tendency to teratological forms is often exhibited, but when the growth is well advanced, the shape of the frustules is usually quite normal, (8) English Channel Water (“Outside Water”).—In a large number of our experiments sea-water brought in from outside the Plymouth breakwater, and therefore taken at some distance from the shore, has been used. This is referred to as “outside water.” and the temperature range for the year is from 8° to 16° C. 2 It has an average salinity of about 35°0 °/,, If a sample of ‘‘ outside water ” is inoculated from a persis- tent culture of a plankton diatom, a small growth is obtained in from five to fifteen days. . But soon minute bottom forms of diatoms, other algze, flagellates, infusoria, etc., appear, and the inoculated species is lost. ‘lhe total growth of any form is never large. If the growth of these foreign forms is pre- vented by sterilising the water before inoculation, a consider- ably better growth of the plankton form is obtained. The water was, as a rule, sterilised by simply heating to 70° C., which temperature was found to be quite adequate. Boiling gave equally good results, but the former was preferred, as less concentration due to evaporation took place. Even under these conditions no permanent culture can be obtained, the diatoms soon beginning to lose colour and getting into an exhausted condition, Death takes place in from two to three months after the culture has been started, and in many cases considerably sooner. Long before inability to start new cultures, the test of death, has been established, the valves appear on examination quite colourless and practically empty. Samples of eutside water, taken at times when the quantity of plankton was widely different, gave no appreciable varia- tion in the results obtained by culture methods. It is, how- ever, doubtful whether differences in the amounts of growth ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 3873 in cultures, proportional to the seasonal variation in the quantity of phytoplankton, would be sufficiently marked to be appreciable. The total growth under cultural conditions, although small for a culture, is very much greater than any natural plankton that has come within our experience. (c) Tank-Water.— Tank-water,” or water taken from the supply of sea-water circulating through the tanks of the Aquarium at Plymouth, shows some striking and interesting differences from “ outside water.”’? This water is pumped up from the sea, just below the Laboratory, into two large, covered-in, settling reservoirs, with a capacity of 60,000 gallons each. Pumping is only done at high-water spring tides, so as to get the least contaminated water, and no water is pumped that does not show a specific gravity, measured with a hydrometer, of 0!” = 26:00 (S = 34:00) or over. The water is allowed to settle for about a fortnight before being used for the general circulation. The tanks themselves are made of slate and glass, and the pipes which convey the sea-water to them are of vulcanite, so that the water does not come in contact with metal, excepting in the pumps, which are of cast iron. The two settling reservoirs are used alternately for about a week each. From time to time, tide and water allowing, waste is re- plenished, and about twice a year each reservoir is emptied, cleaned out, and refilled. The aquarium takes about 20,000 gallons, and this is in circulation with one of the two 50,000 gallon reservoirs. An estimate of the amount of life in the tanks of the aquarium must be exceedingly rough, but the intensity of the larger forms of life is far greater than any- thing met with in natural waters. About 500 fish and 2000 invertebrates, including all forms as large as an Actinia equina, might be somewhere near the mark. So it will be seen that the accumulation of excretory products must be a by no means negligible factor. The flora of the tanks is very restricted, and is chiefly composed of minute forms of alge. Minute naviculoid diatoms, Eetocarpus, Cladophora, 374 E. J. ALLEN AND E. W. NELSON, Knteromorpha, Vaucheria, and unicellular alge are the commonest forms. ‘The large seaweeds, such as Fucus and Laminaria, do not live long if introduced. Plankton diatoms, although a great number must be pumped up when the reservoirs are being filled, are not represented. As in the case of outside water, a sample of ‘‘ tank-water,” inoculated from a persistent culture, will only give a very small growth, minute forms, etc., soon multiplying and choking out the plankton form. The ultimate growth of minute unicellular alge other than diatoms is often con- siderable, and many quite unknown and unidentified forms have been obtained. ‘The total growth of vegetable forms is always found to be greater than in the case of outside water. In cultures of plankton diatoms made with sterilised tank- water, a very great improvement on outside sterilised water was always noted. ‘lhe culture of the diatom used to inocu- late this medium persists for a considerable period, and the colom of the frustules remains normal for two to three months, (Dp) Animal-Charcoal Water.—The use of animal charcoal, as a means of purifying the water in small aquaria, has for a long time been known and practised by those who have kept such aquaria in inland places. At an early stage in Our experiments, water from a tank, which was not in a satisfactory condition, was treated with some powdered animal charcoal and filtered. It was noticed that a good growth of diatoms took place in this water. Systematic experi- ments with the use of animal charcoal were then commenced, and these have resulted in a method of great value, both for the culture of diatoms and for the rearing of pelagic larve. Animal charcoal is made by the carbonisation of bones,! 1 Analysis of animal charcoal, from Thorpe’s * Dictionary of Applied Chemistry “— Carbon, , ; ; ; . Ist Ca., Mg. phosphates, Ca. fluoride, ete. . . 8021 Calcium carbonate : : . "SSB Other mineral matter. ; : SS 100°00 ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 375 and is sold in two grades knownas “pure” and “‘ commercial.” Our earlier experiments were all made with “ pure” animal charcoal, but subsequently the ‘‘ commercial” animal char- coal was largely used, and appears to give equally good, if not better results. In both cases the animal charcoal is used in the powdered form. Animal-charcoal water is prepared as follows: (1) A quantity of sea-water is sterilised by heating it in a flask to 70°C., at which temperature it should be kept for about twenty minutes. At the same time some animal charcoal is heated sufficiently to sterilise but not to burn it, covered over, and allowed to cool. When both are quite cold the charcoal is added to the water (ca. 15 grm. to 1000 ¢.c.), and well shaken up in it several times. After an interval of half an hour or more the water is filtered through fine filter-cloth,! the whole filter having been first sterilised with boiling sea-water, and is received in a sterile flask. It is then ready for use. (2) For many experiments, where larger quantities of water were required, the sea-water was not sterilised before being treated with animal charcoal. In this case, if the first part of the filtrate be rejected, the subsequent water will generally be practically sterile, and few, if any, extraneous organisms will develop in it. (3) At a later date an automatic apparatus was set up in the Plymouth Laboratory, by which very considerable quanti- ties of sea-water could be treated with animal charcoal, and subsequently filtered through a “ Berkefeld” filter; water treated in this manner we call “ Berkefeld water.” Tank- water was always used in this apparatus, and was mixed with animal charcoal,” in a clean sulphuric acid carboy, by blowing air through with a pair of bellows. ‘lhe mixture was allowed to settle for at least twenty-four hours, and then syphoned 1 The filter-cloth used for this purpose is the same as is made for use in filter presses, and is known as extra-super swansdown. To prevent this becoming clogged another cloth, known as hydraulic twill, was, as a rule, used over it. 2 Ca. 300 grm. to 20 litres of water. 376 E. J. ALLEN AND £. W. NELSON. over into an inverted bell-jar, with a tubulure at the bottom, into which the Berkefeld candle was fitted. Filtration under these conditions was found to be rather slow, so in order to increase its rate an apparatus was devised by which the pressure on the filter was considerably augmented. This apparatus (see Fig. 1) consists of a glazed earthenware “tobacco jar,” with two tubulures, one at the side, the other at the bottom, and a lid which can be screwed down tightly on to a rubber washer, by means of a triangular metal arrangement fitting into grooves above the lid.! The internal Fia. 1.—Diagram of apparatus for preparing sterile sea-water by filtration, without contact with metal. dimensions of our jars are 11 in. by 6 in., and the diameter of the opening at the top is 3Lin. ‘he tubulures are coned, with the smaller diameter external, and make a good fit for a No. 8 rubber bung. When setting up this apparatus a bung, through which a short glass tube bent at right angles is passed, is fitted into the side tubulure. This tube is con- nected, by means of rubber pressure-tubing, to another glass 1 These jars were made to our specification by Messrs. Price, Powell, and Company, Bristol. The clamps usually supplied with such jars are not strong enough to obtain a tight joint, but these are easily repiaced by stronger ones. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS, 377 tube leading down from the bottom of a small inverted bell- jar, placed some height above (in our case 14 ft., which gives a pressure of ca. 6 lb. to the square inch inside the jar). A serew pinch-cock on this connection serves as a tap. ‘The carboy containing the treated water stands just above the bell-jar, and is fitted with a tightly fitting rubber bung, through which two tubes pass. One is an ordinary syphon, the other the only air-inlet into the carboy. This latter auto- matically keeps the level of the water in the bell-jar constant, by closing the air-inlet as soon as the water covers the end of the tube. When filtering water the modus operandi is as follows: The carboy is filled with tank water, treated, and allowed to settle as before. The Berkefeld candle,’ bung, delivery tube, and connections (see fig. 1) are sterilised by boiling for half an hour, and fitted into place from within. (The delivery tube is shaped so that any drops of water, accidentally running down outside it, do not enter the vessel receiving the filtrate ; and the jar should be large enough to allow the hand to fit the filter into place without much trouble.) The pinch-cock is closed, and the syphon from the carboy started, which will automatically stop if the bung has been properly fitted. This should be watched to avoid acci- dents. The pinch-cock is then opened until the water rises in the jar well above the top of the candle, but still leaving some air-space. ‘he lid can now be fitted into place and screwed down. The tightness of this joint can be tested by pouring a little water into the crack round the lid, and observing if any bubbles are formed when the pinch-cock is opened. If all is right, no bubbles will be seen, and a good stream of water will flow out from the delivery tube. Our apparatus will filter about 20 litres an hour, and the filtrate is exceptionally bright and clear. ‘The candle should be sterilised every three or four days that the apparatus is in use to avoid indirect contamination by growths of organisms 9 through the substance of the filter.2, The water while passing 1 No. 5 porcelain mount, length 8 in., diameter 2 in. 2 See Bulloch and Craw, ‘ Journ. of Hygiene,’ vi, No. 3 (1906), p. 409. 878 E. J. ALLEN AND E. W. NELSON. through this apparatus only comes into contact with glass, earthenware, and rubber, the use of metal having been pur- posely avoided. (r) Peroxide of Hydrogen Water.—As it seemed probable that the action of animal charcoal was due to contact oxidation with the oxygen occluded in the charcoal, experi- ments were made to determine whether a similar effect could be produced by the use of hydrogen peroxide (H,O;), This was used in two ways. In the first method a sufficient quantity of H,O, was added to the sea-water to ensure complete sterilisation (1 c.c. of H,O, of twenty vols. strength per 1000 c.c, of tank-water was found to be satisfactory), and the excess of H,0, was decomposed by adding manganese dioxide, The water was then filtered through filter-cloth, and the filtrate appeared to remain quite sterile. Good cultures of Chetoceras constrictum, Biddulphia mobiliensis, and Skeletonema costatum were made in this water, which seemed to be as good as water treated by the animal charcoal method. The second way of using the peroxide of hydrogen was to start with water sterilised by heating to 70° C. and to add to this H,O,, in small quantities at a time, until its presence could just be detected on testing the sea-water with perman- ganate of potash. In these circumstances, the first amounts of H,O, are decomposed in the oxidation of organic substances in the water, and a very slight excess of H,O, persists. For tank-water 1 c.c. of one vol. H,O, per 1000 c.c. was found to give the best general effect. Cultures grown in water prepared in this way developed satisfactorily, being practi- cally equal to those made in animal-charcoal water, but they became exhausted rather quickly. The treatment of aquarium water with ozone was also tried, as this seems to offer a possibility of treating large quantities of water,! such as the whole bulk of water in an aquarium 1 The use of ozonised air for the purification of fresh water for town water supplies has been adopted in some localities. (See Bridge, J. H., paper read before Franklin Institute, reprinted in ‘ English Mechanic,’ 1907, pp. 369 and 392.) ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 379 circulation, without very considerable expense. Experiments on a small scale, which we were able to make, unfortunately only with imperfect apparatus, showed that water treated with ozonised oxygen gave distinctly better cultures than untreated water. Although the sea-water was not absolutely sterilised by the treatment to which we actually subjected it, a sample of water which was visibly clonded with bacteria became quite clear and bright. (F) Cultures in these Media.—In order to make clear the different results which are obtained by using these different waters, we will describe the probable result which would be got from a series of flasks set up with the following media, and each inoculated with a persistent culture of a true plankton diatom, such as Thalassiosira, Skeleto- nema, or Chetoceras. A. “ Outside water” untreated. Small growth in from five to fifteen days, almost immediately swamped by growths of foreign forms ; the latter, however, will never be large. B. Ditto, sterilised. Shghtly larger growth, very soon becoming ex- hausted. c. * Tank-water ” untreated. Same result as in a, but growths will be much larger and healthier, and will last longer. p. ‘'Tank-water”’ sterilised. A fair growth of the inoculated species, but the total growth will not be as great as in c; the diatoms will retain their normal appearance for some time. E. ‘Outside water” + Miquel’s solutions A and B, sterilised. Best culture in series, both in quantity and quality. The diatoms will remain normal and healthy for a very long period. F. Outside water” sterilised and treated with animal charcoal. Fair growth, especially at first; diatoms will soon grow pale and become exhausted ; better than p. 380 E. J. ALLEN AND E. W. NELSON. a. “Tank water” sterilised and treated with animal charcoal. As ¥, only growth will be slightly greater and will last considerably longer. Third best in series. u. ‘Tank-water ” treated with animal charcoal and filtered through Berkefeld filter. This will usually be the second best culture in the series, but the difference between this and G will only be slight. k. “ Outside water ” treated with H,O,. This will most resemble F, but will not be quite so good. L. “ Tank-water” treated with H,0O,. A distinct improvement over k. This medium is rather variable, and in some cases the growth obtained has been quite equal to r, if not better. zB. Experiments with a View to Determining the Conditions which underlie. the Successful Culture of Diatoms. The attempt to make cultures of diatoms for use as food when rearing pelagic larve, led naturally to an effort to determine the best culture medium and the most favourable conditions for the: rapid and continuous growth of diatoms. Before success can be attained in this direction exact know- ledge as to the nature of the essential food-stuffs, and, -in fact, as to the general physiology of the Diatomacez, is necessary.! Numerous experiments. extending over the last three years, have been carried out, with a view to obtaimng such knowledge, and the results, though still by no means complete or conclusive, are perhaps worth recording. A great difficulty which has to be metin carrying out such investigations on marine diatoms, is caused by the fact that when sea-water is used as a basis for the culture media, we 1 For general references to literature see ‘* Bibliography,” especially Miquel (12), Richter (18). ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 381 are dealing with a solution of a very complex and very variable character, the exact nature of which it is extremely difficult to determine. The most direct method of research, namely, chemical analysis, has not proved of much service, owing to the uncertainty, and in many cases impossibility, of accurate determinations, in sea-water, of such minute quan- tities of substances as those upon which the growth of plankton diatoms has been found to depend. We have had, therefore, to rely, for the most part, on the lengthy and tedious process of analysis by “trial and error,” the experiments being largely conducted on lines suggested by Liebig’s well-known “law of minimums” (Pfeffer, vol. 1, p. 413). The ideal at which we aim is to find a culture medium with artificially prepared sea-water as its basis, such that the absence, or diminution in quantity, of any one of its constituents would have a profound effect upon the growth of diatoms in it. Whether the conditions regulating growth in such a medium would be at all comparable to the natural conditions of life in the sea is a question that would have to be decided by experiment, but in any case this could be made a starting point for much more definite research than has yet been attempted. Up to the present time we have not, unfor- tunately, succeeded in finding such a culture medium. Throughout the work we have had very great difficulty, in spite of much care and many precautions, in obtaining consistent results. It may even happen that in two flasks containing the same culture medium, inoculated with the same culture of diatom and standing side by side, under exactly identical conditions, as far as can be recognised, quite different degrees of growth will be observed. All experiments must therefore be frequently repeated before entire confidence can be felt in any conclusions which they seem to indicate. It must be remembered, also, that in all the persistent cultures of diatoms that we have used, bacteria have pro- bably been present, and this fact has probably had some influence on the result. Unfortunately our attempts to obtain absolutely pure cultures have not met with success. VOL. 05, PART 2.—NEW SERIES. 25 382 E. J. ALLEN AND E. W. NELSON. Methods.—In carrying out the experiments to be described in this section the procedure has been as follows: All media have been prepared from sterile sea-water, and sterile vessels and instruments have always been used. ‘The cultures have usually been made in 60 c.c. of liquid, in short-necked, wide- mouthed flasks of 125 c.c. capacity. When a number of cultures were to be compared, the flasks were kept standing in a row together in such a way as to keep the physical conditions as similar as possible. Control cultures in standard media were included in each series, so that results from different series could be compared by reference to the controls. The various media were inoculated from a persisteut culture of a species of plankton diatom, which in the great majority of cases was Thalassiosira decipiens (p. 412). When preparing the different media the methods used were, as far as possible, identical, and although only about 60 c.c. was needed for a culture, a litre was made up, so that errors due to-measuring very minute quantities might be avoided. The media were all freshly prepared for each comparative series of cultures, the same sample of sea-water being used, when the basis of any two or more was the same. Comparative estimates of the amount of growth in the different cultures were made by eye alone. Any difference between amounts of growth that has been described here as appreciable has always been accompanied by a marked difference in appearance to the eye on holding the cultures up to the light. A few drops from each culture were also, from time to time, examined micro- scopically, as a test of the quality and purity of the growth. The Sea-water Employed.—tThe sea-water employed as a basis for the culture media has been either (1) “ outside water” or (2) “tank-water.” A general description of these will be found on pp. 372-374. An accurate chemical analysis of both types of water would probably make clear many difficult points, but, as already pointed out, no chemical methods of sufficient delicacy have yet been devised. We have seen that if we compare “‘ tank-water,” i. e. water from the closed circulation of the Plymouth Aquarium, with ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 383 off-shore sea-water in situ, a most obvious difference is the much increased density of the larger forms of animal life in the former, combined with the almost complete absence of plant life. Hence the concentration of excretory products in the tank-water must be very much higher than in outside water. Other factors, such as increased bacterial action, artificial aération, etc., in tank-water, must also be taken into account (cf. Vernon [58], Smith [561). There seems to be direct. evidence to show that the concentration of nitrates, possibly due to the action of nitrifying bacteria on the products of excretion, such as urea, ammonia, etc., is considerably higher in the tank-water, and the presence of soluble organic matter in concentrations never met with in the sea, can almost certainly be assumed. It is probably due to the presence of these nitrates and soluble organic substances that sterilised tank-water is a much better medium in which to grow diatoms than sterilised outside water (see p. 379). The Constituents of Miquel’s Solutions.—It has been already stated that no better medium for the culture of plankton diatoms has been found by us than the solutions recommended by Miquel, although these solutions may be modified and simplified in various ways with equally good results. The formulee recommended by Houghton Gill give very similar cultures. The essential features of Miquel’s and Houghton Gill’s methods, when adapted to sea-water, are the same. Miquel’s solution A and Gill’s solution 2, can both be replaced by a solution of potassium nitrate (p. 369). Again, Miquel’s solution B and Gill’s solution 1 only differ in the proportionate amounts in which the various constituents are prescribed. The formule are: Miquel’s sol. B. Hi. Gall’sisol.- 1. Na,HPO,,12H,0 : 4 germ. . : 2 grm. CaCl, 3 : : 4 ,, 2 5 Aye FeCl, (syrupus) : PAO ; : 0°. 5; HCl (concentrated) . raat : : oe Water ; 80, 5 LOOM 2 Use 1 c.c. per 1000. Use 3 c.c. rer 1000. 384, HK. J. ALLKN AND E. W. NELSON. The proportionate amounts added to equal volumes of sea water are : Miquel’s sol. B. H. Gill’ssol. 1, Na,HPO, . ; se lO > : _. ae CaCl, . ' ; AO) ag . 24 FeCl, . ; ; ‘ 5 : : : 3 HE 0G) ee : : 5) ; ; ; 6 Since cultures can be obtained with no appreciable difference by using media prepared by adding either of these solutions, together with Miquel’s solution A, to sea-water, a con- siderable latitude in the proportions of the salts present is tolerated. We must now consider what is the rédle of the various constituents in Miquel sea-water. The part played by any salt of a culture medium may be considered as being either, firstly, “nutritive,” or secondly, “ protective.”! Under the first heading, any direct addition of food material must be included; under the second, any removal or neutralisation of harmful substances, such as toxins and possibly bacteria, and any more remote effects, which, although influencing growth, do not directly enter into the metabolism of the plant. Our experiments have proved that solution A can be reduced to a simple solution of potassium nitrate” without detriment (cf. p. 369), and that the amount of growth is, within limits, roughly proportional to the amount of KNO, added, as the following experiment shows :— Inoculated from persistent culture of Thalassiosira decipiens: A. Normal Miquel sea-water. Growth as usual. B. Ditto, but only half amount of solution A. 1 Loeb, ‘ The Dynamics of Living Matter,’ New York, 1906, p. 77. 2 For the sake of convenience the expression solution A will be used throughout the rest of this paper to indicate a simple solution of potas- sium nitrate (5 per cent.), and solution B to indicate Miquel’s phospho- ferri-calcic solution. Unless otherwise stated the amounts of each added to 1000 ¢.c. sea-water will be normal, i,e. 2 ¢.c. solution A and 1 cc. solution B, ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 385 Good growth at first, but exhausted sooner than A. c. Ditto, but two and a half times amount of solution A. Was slower than either A or B at start, but after- wards was better than a or B, and lasted longer. p. Ditto, but five times amount of solution A. As c, but in greater degree. Considering the nature of the substance added, and its already well-known action in piant metabolism, these results, coupled with the fact that exhausted cultures can often be regenerated by the simple addition of nitrates (see below, p. 390), are quite consistent with the assumption that solution A is simply nutritive in action. The concentration of nitrates in natural sea-water is so low (Brandt [47]) that the amount available in a culture of untreated water very soon becomes completely exhausted, and it is this deficiency that solution A probably corrects. Considering now the action of solution B, it must first be observed that increased concentration of nitrates alone will not explain the whole action of Miquel’s solutions, for no increase in growth is obtained when nitrates or solution A only are added to sea-water. ‘To illustrate this point an account of an actual experiment may be given :— Inoculated with Thalassiosira decipiens: A. Normal Miquel sea-water. Good strong culture, in every way normal. B. Outside water sterilised. Small growth at first, very soon exhausted. c. Ditto + solution A. No improvement over B. p. Ditto + solution B. Fair growth. Great improvement on B and ¢, but exhausted considerably before a. E. ‘l'ank-water sterilised. Appreciably better than 8, but growth not large. F. Ditto + solution A. Not even as good as £. G. Ditto + solution B. 386 E. J. ALLEN AND E. W. NELSON. Next best in series to a; lasted longer than p, and had better colour. _ To generalise, no improved culture is obtained with solution A alone, but a fair, though not very lasting, growth can result from using solution B only. The action of solution B is to some extent obscured by the fact that, when this solution is added to the alkaline sea- water, a precipitate is formed. ‘This precipitate is at first white, but, on heating or standing for some time, it becomes greenish-yellow. We are indebted to Mr, D. J. Matthews for the following analyses. Ten litres of normal Miquel sea-water were prepared, and the precipitate was collected on a filter-paper, washed, and dried at 100° C, Weight of dry precipitate from 10 litres = 0°2949 grm. Analysis of Dry Precipitate. P.O. ' : ; ‘ . 26°36 per cent. Wes. oe ‘ " ; . 41°31 © CaO : : : . ,; ee ES H,O ; : : 4 . 24°86 100°16. . = Or the precipitate from 1 litre of normal Miquel sea-water contains— P.O; oes ‘ ; : . . °00777 grm. FeO, « ‘ ; ; . °01218 CaO . ‘ ; : ~ *OO22R ae An analysis of 1 ¢c.c. Miquel solution B, the amount added to 1 litre Miquel sea-water, gave— 12710 ae : : ; ; - *00825 FeO... - - : : . °0105 CaOi, : ‘ : : .! 2014S Comparing these figures it seems probable that, when added to sea-water, all the iron in solution B is precipitated, and a certain amount also of the phosphate and calcium. The 3) 2) 33 ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS; 387 additive effect on the sea-water is, therefore, a slightly in- creased concentration of phosphate and calcium. “An analysis of a sample of tank-water for phosphorus, before and after treatment with solution B (1 c.c. per 1000), gave the following figures : Tank-water ‘5 mgrm. P per litre = ‘00163 grm. P.O. Tank-water + solution B (without precipitate) 1:5 mgrm. P per litre = ‘00488 grm. P,O;. It will be noticed that the figures from the different analyses do not agree very well. This is probably due to the fact that different samples were used for analysis in each case, and also to the fact that the solutions, were made up in the ordinary way, without any special precautions, volumes, for instance, being measured in cylindrical glasses, pipettes, etc. Cultures’ were tried in sea-water containing the normal amount of solution A, plus the normal constituents of solution B, less all the iron and less the amount of phosphate that would combine with the iron to form basic ferric phosphate (P.O, 2Fe,03;12H,O). This solution should have very nearly the same chemical composition as normal Miquel sea-water from which the precipitate has been removed. Successful cultures could not, however, be obtained init. Neither could cultures be grown in sea-water to which had been added the normal amount of solution A and 1 mgrm. P (as sodium phosphate) per litre. To ascertain the effects of the different constituents of solution B, experiments were carried out with separate solu- tions of these constituents, each of the same strength, as in Miquel’s formula. Different combinations of these solutions were added, together with solution A, to sterilised sea-water, and the resulting media were inoculated in the usual way. It was found necessary to repeat these experiments a great number of times, as the results obtained were rather contra- dictory. ‘To illustrate the methods used a list of the different media, and notes of the cultures obtained in them, are given below. These media were inoculated from cultures of Thalassiosira decipiens, and the cultures were kept 3888 ir. J. ALLEN AND E. W. NELSON. under observation for at least four months. Series were made as uniformly as possible, and controls in standard media were included in each. The strength of the various solutions used in these experiments was the same as in Miquel’s formula. A. Outside water + solution A + solution B (normal Miquel sea-water. First control. s. Outside water + solution A + Na,HPO, solution + FeCl, solution + CaCl, solution. Second control. Good normal cultures were always obtained in these two controls. c. Outside water + solution A + Na,HPO, solution. A very uncertain medium. Sometimes no growth has been recorded, and at other times a fair growth results, but these cultures are never equal to normal Miquel. p. Outside water + solution A + FeCl. solution. Occasionally a very small growth has been obtained, but at the best it is very poor. gr. Outside water + solution A + CaCl, solution. About equal to D. r. Outside water + solution A + NasHPO, solution + FeCl; solution. Uncertain as c. No cultures have been obtained equal to the best in c, g. Outside water + solution A+ Na,HPQ, solution + CaCl, solution. Some cultures very nearly equal to the controls have been obtained in this medium. H. Outside water + solution A + FeCl; solution + CaCl, solution. Poor, about equal to D. Analysing the above results we see that— (1) None of these modifications of solution B give results equal to solution B itself. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 3889 (2) The best result is obtained from the combination of the phosphate and calcium chloride solutions. (3) Of the solutions used singly the phosphate is the best, the iron and calcium chloride being about equal. (4) The addition of FeCl, to Na,HPO,, or the addition of CaCl, to FeCl,, does not improve the medium to any extent. Experiments were also made to determine whether the precipitate thrown down in sea-water by Miquel’s solution B, itself had any influence on culture media. A quantity of this precipitate was prepared, filtered off, and then added to outside sea-water + solution A (nitrates). A small growth was obtained, which was a distinct improvement on the control without the precipitate, but exhaustion soon set in. Further discussion of the mode of action of solution B, and as to whether that action is purely nutritive, or partly nutritive and partly protective, is better postponed until a later section, after the action of animal charcoal and other substances has been considered (see p. 405). Animal Charcoal and Peroxide of Hydrogen.—The most successful culture medium for plankton diatoms, next to Miquel sea-water, is that prepared from animal charcoal (cf. p. 379). Animal charcoal water gives at first almost as good cultures of plankton diatoms as Miquel sea-water, but the tendency to paleness and exhaustion appears much sooner. ‘he best cultures were obtained in “ Berkefeld water,” that is, tank-water from the Plymouth Aquarium treated with powdered commercial animal charcoal and filtered through a Berkefeld filter. ‘Tank-water as a basis for animal charcoal water is very much better than outside water, probably on account of the higher concentration of nitrates, etc. There is a very striking resemblance between the effect of animal charcoal and of Miquel’s solution B upon sea-water used for diatom cultures, and the growths obtained by using tank-water + solution B and tank animal-charcoal water are very similar in character. If Miquel’s solution A is added to animal-charcoal water there is a great improvement, both in 390 E. J. ALLEN AND E. W. NELSON. the colour and quantity of diatom growth, and in the case of Thalassiosira decipiens the chains are long and well formed. With animal-charcoal water + solution B, on the other hand, practically no growth was obtained. It is possible that a certain amount of phosphate, and perhaps of calcium, from the animal charcoal, goes into solution and serves as a ‘‘ nutritive ” material for the diatoms. But we are inclined to think that its chief action is “ protec- tive,” and due to its power of occluding gases, such gases being in a state of higher chemical activity than under normal conditions.! As was explained in a previous section (p. 378), the possibility that the action of animal charcoal might - have some sort of effect comparable to oxidation, led us to experi- ment with hydrogen peroxide. Fair growths of diatom could be obtained in sea-water prepared in the manner described, but they always showed a tendency to rather rapid exhaustion. As in the case of animal-charcoal water, tank-water proved a much better basis for treatment with H.O, than outside water. Reviving Exhausted Cultures.—Several experiments. were carried out with water from old, exhausted cultures. The sediment was filtered off, the filtrate was sterilised bs heat, and then treated by various methods. In one typical experiment the following was the result :— Water from an exhausted -culture of Skeletonema costatum in Miquel sea-water, reinoculated with the same diatom : , A. Filtered and sterilised. No growth obtained. B. Ditto + solution A (nitrates only). Good culture, but did not last very long; further addition of nitrates made no improvement. c. Ditto + solution B. 1 Against this view would seem to be the fact that when powdered cocoa-nut charcoal, which has a still higher power of occluding gases, was used in place of animal charcoal, very poor cultures were obtained. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 391 No growth. p. Ditto + solution A + solution B. Very good growth, lasting considerably longer than B. g. Ditto + animal charcoal. No growth. Exhausted cultures in animal charcoal water gave. the same general results on treatment and reinoculation. In an old culture of Biddulphia mobiliensis in outside water + solution B only, which was in a very exhausted condition (nine months old), the addition of KNO; gave a very rapid regeneration, and the diatoms became of normal colour and form. This renewed growth, however, did not last very long, and a further addition of KNOs did not give any result. The addition of sodium phosphate also failed to stimulate growth, The same rapid regeneration, on the addition of potassium nitrate, has been obtained with almost every medium, but a second attempt has always failed. Silica.—A very noticeable character of the true plankton species of marine diatoms is that their skeletons are very markedly less siliceous than the great majority of other forms. Their valves are only feebly marked, if at all, and they will not stand the vigorous treatment of cleaning with acids and heat that is commonly used in the case of fresh-water diatoms. In cultural forms this absence of silica is still more obvious, and no marking can usually be seen on even those forms, which, under natural conditions, are the most siliceous, e. g. Coseinodiscus excentricus. Deformed and distorted frustules are the rule in certain stages of growth in our cultures, and it is often very hard to make out more than the thinnest coating of silica. It is quite probable that this deformity can be accounted for simply by the absence of a strong siliceous skeleton. As a rule, the more rapid the growth the more -teratological forms will be found. In untreated outside water little deformity will take place, but in normal Miquel, where very rapid growth takes place, the diatoms may assume almost any conceivable shape. The 392 E. J. ALLEN AND E. W. NELSON. form of the frustules tends to come back to the normal again, when the culture is well started, and in old stages the majority will be perfectly formed, although small and pale. It was found that the addition of silica (in early experiments as fragments of potassium silicate) was, as far as could be judged, immaterial, which fact led to the conclusion that a sufficiency dissolved out from the glass flasks in which the cultures were kept. During rapid growth, it is possible that the silica does not dissolve out fast enough to supply the demand, although it is also possible that diatoms, during rapid division, cannot absorb silica and form a perfect skeleton, even when the supply is abundant. Richter (18) has proved the necessity of either CaSi,O, or K,81,0; for the growth of Nitzschia palea, grown in pure cultures. We tried the addition of silica in various forms, and in one instance, in a culture of Coscinodiscus excentricus, to which a little precipitated calcium silicate had been added, the uniformity and markings of the valves were much more regular than in the control. ‘The presence of a trace of pure, dialysed silica also, in one experiment, gave an improved regularity of form, but the quantity or rapidity of growth did not seem to be affected. No sign of regeneration could be obtained in exhausted cultures by the addition of silica. Organic Infusions.—Miqnel recommends the use in culture media of infusions of organic substances, such as bran, straw, diatom broth, etc., in addition to the saline solu- tion. He does not make it quite clear if he ever dispensed with them at all. In his general directions he certainly states that the addition of both saline and organic nutrient material is necessary. As would be expected from the general metabolism of plants, the saline constituents are sufficient for growth. At the same time, excellent cultures have been obtained from dilute organic infusions, both with and without the addition of Miquel’s solutions A and B. About a square inch of Ulva was boiled in 600 c.c. of sea-water for half-an-hour, cooled, and filtered. In this medium an excellent growth of Coscinodiscus excentricus in one case, and Biddulphia ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 393 mobiliensis in another, was obtained, the growth lasting for some considerable time. Infusions, made in the same way from a small piece of fresh fish, gave the same results, and although growth was rather slower at first, the final result was, if anything, slightly better. As Miquel points out, these infusions must be made very dilute, otherwise growths of bacteria, moulds, etc., will completely swamp the diatoms. Karsten (7), in some interest- ing experiments, showed that Nitzschia palea (Kutz) W. Sm. could be made to alter completely its mode of nutri- tion. On placing this diatom in organic nutrient solutions, it lost all chlorophyll and became colourless, but in saline media the chlorophyll would not regenerate, and the nutrition change back from heterotrophic to autotrophic.! Of course, with our infusions, it cannot be said that the diatoms were necessarily feeding on dissolved organic material, as some necessary, saline, nutritive materials conld have dissolved out from the weed or fish. If the former is the case, it might explain the superiority of tank-water over outside water, since the tank-water must contain a much higher percentage of organic substances in solution. If an alternative mode of nutrition, autotrophic or mixotrophic, could be proved, especially in the case of the ‘‘ bottom” forms of diatoms, a great many phenomena could be ex- plained, but the evidence is as yet far too slight to warrant any such assumption. Artificial Sea-water.—As we have explained in a previous section, the ideal aimed at in this part of our work has been to obtain strong growths of Diatomacez in purely artificially prepared solutions of simple salts. If this end could be satisfactorily attained the difficulties due to the unknown and variable composition of natural sea-water at once dis- appear. According to van ’t Hoff (85) sea-water is a solution containing salts in the following molecular concentrations: NaCl 100-0, KCl 2:2, MgCl, 7°8, MgSO, 3°8, CaCl, 1:0 (varies). 1 Cf. Zumstein, ‘Zur Morphologie u. Physiologie d. Huglena gracilis,’ Leipzig, 1899. 394 E. J. ALLEN AND E. W. NELSON. Using these molecular concentrations, a sea-water of any desired salinity can be prepared. ‘The chlorine content of average Atlantic water is about Cl = 19-4, and samples of artificial sea-water were prepared with the same chlorine value, thus : NaCl . , ; : ‘ : . 267a i S| ‘715 MgCl, . 3°42 CaCl, : ; : ; : 5] MesO, ; : : , : : Pg | Double distilled water ; ; . 966°47 1000-00 To make this solution comparable to natural sea-water, the “alkalinity”? must be raised by the addition of an alkali such as Na,CO;. After the importance of “ alkalinity ” asa factor had come before our notice, 2°4 c.c. M/, Na,CO, was always added to the above solution in order to make the amount of base in equilibrium with CO, equivalent to the usual 40 mgrm. OH vree The only success we attained with artificial sea-water as a basis for culture media was with four isolated cultures in one of our earlier experiments. ‘I'wo of these were cultures of Coscinodiscus excentricus in artificial sea-water + Miquel’s solutions A and B. The two cultures were identical except that one was in an ordinary bohemian glass flask and the other ina “ resistance glass” flask. No difference between these two could be seen. The growth obtained in both was in every way equal to normal Miquel sea-water, and is still fair, although over two years old. The other two successful cultures were growths of the same diatom in the same media, plus a small quantity of weed infusion, made by boiling up a small piece of Ulva in artificial sea-water. These gave just as good results, but the addition of unknown factors from the weed detracts from their general interest. In spite of frequent attempts, over fifty in number, we have not been able to repeat this experiment, which may possibly be due to ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 395 some accidental impurity in the salts or distilled water from which the successful media were prepared. Alkalinity.—Tornée (48) and Dittmar (33) were the first to investigate the fact that sea-water showed on analysis an apparent excess of base over acid, which excess they termed “the alkalinity of sea-water.” Dittmar defines the alkalinity of sea-water as “‘a measure of its potential carbonate of lime,” but this definition and his supposition that this excess of base combines directly with dissolved CO, to form car- bonates and, further, but only in very small proportion, bicar- bonates, is hable to give a quite erroneous idea of the state of equilibrium actually occurring in the ocean. For, as Fox (84) has shown, “ sea-water reacts in situ very nearly neutral, and actually just slightly more acid than distilled water.” This is due to the fact that sea-water always contains a consider- able quantity of dissolved COQ. If a salt solution with neutral reaction, that is, containing H: and OH! ions in concentrations equal to one another and the same as for pure water, be exposed to an atmosphere con- taining CO,, a definite amount, depending on pressure, tem- perature, and salinity, would go into solution. This CO, would combine with water and form the very weak acid H,CO;, which would ionise with the formation of free H- ions, thus : HCO. 2 Be ECOG pbs (HCO! 2 - CO”,) The second stage of dissociation is so small as to be negligible. The concentration of H: being now increased and OH! decreased, the solution would have an acid reaction. The actual amount of CO, thus dissolved would always be small; for instance, a salt solution of strength Cl = 20-00 (average Atlantic water Cl = 19:4) will at 10° C. dissolve about 3 c.c. CO, per litre from an atmosphere containing 3 loo CO, (about normal). But the ocean is found to contain very much greater quantities than this, 60 c.c. or 200 times this amount being a not unusual figure for the total CO. The difference between this amount and the ‘3 c.c. or so dis- 396 KE. J. ALLEN AND E. W. NELSON. solved by the neutral salt solution, as above, is kept in equilibrium with the 3 °/,,, CO, of the atmosphere by the amount of “excess” base equivalent to the amount of acid neutralised when an acid such as HCl is added to sea-water in excess, If a solution identical with sea-water but abso- lutely free from CO, (a practical chemical impossibility) could be obtained, then there would be present an excess of base over acid, and consequently an excess of OH! ions over H, ions, and an alkaline reaction. On exposing such a solution to the atmosphere, CO, would go into solution, ionise, and the H: ions thus set free would react with the OH! ions, due to the excess base, to form water. And this reaction would continue to take place, on more CO, dissolving, until all the excess OH! ions were neutralised, at which point the solution would react neutral. Now, as before with the neutral salt solution, a further small amount of CO, would go into solu- tion, bringing the solution into equilibriam with the atmos- phere, and the excess H- ious thus formed would give an acid reaction. ‘The final result would be a ‘solution exactly identical with natural sea-water. ‘he total CO, found in sea- water can be considered as existing in two parts: the larger part in equilibrium with free base, its amount depending on temperature, pressure, and alkalivity; the smaller in equili- briuin withthe partial pressure of CO, in the atmosphere, its amount depending on temperature, pressure, and salinity. Although sea-water in situ has an acid reaction, it still main- tains the property of being able to neutralise a certain amount of any acid stronger than H,CO,, that is, any acid which, on dissociation, forms a higher concentration of H° ions ; for the stronger acid will turn out the H,CO, in equilibrium with the “excess base” and CO, will be evolved. In consideration of these points, a less confusing definition > would perhaps be a measure of its potential capability of neutralising a strong acid! with the evolution of CO,. This can be conveniently expressed, as is usual, in mgrm. OH °/... of the “alkalinity of sea-water’ 1 Such as HCl, with a high degree of ionization. ARTIFICIAL CULTURE OF MARINE PLANK'TON ORGANISMS, 397 Some of our earlier experiments seemed to show that “alkalinity ”’ was a factor of considerable importance for the successful growth of cultures of plankton diatoms ; so an attempt was made to analyse the various samples of water both before and after treatment as culture media. The method adopted was a modification of that used by Tornde and Dittmar. Solutions of NaOH and H,SO, of strength N/;), by intention, were made up and stored in five-litre “aspirator” bottles. Two accurately graduated burettes standing side by side were connected to these by tubes, so that they could be readily filled by gravity. All air inlets to burettes and stock bottles were fitted with tubes of soda lime. A standard solution of Na,CO, of exactly known alkalinity, approximately that of average sea-water (40°00 mgrm. OH °/..), was prepared by diluting down from a N/,,) solution, all operations being performed by weighing. These standards were stored in stoppered bottles of the fairly insoluble dark green glass, but those that had been kept for any length of time were not trusted, fresh standards being prepared, On analysis these standards agreed with one another to well within ‘1 mgrm. OH °/,,. The water used for diluting the standards was distilled water from the laboratory still, re-distilled in all-glass apparatus with potasstum bichro- mate and sulphuric acid. When ‘carrying out an analysis, equal volumes (about 100 c¢.c.) of sample and standard were measured out into Jena glass Erlenmeyer flasks with a Knudsen automatic pipette. The specific gravity of each was determined by weighing in a 25 c.c. pyknometer. Sampleand standard were then titrated by running in acid from the burette and back titrating with alkali, using a 1 per cent. alcoholic solution of aurine as an indicator and keeping the liquid boiling. The acid to alkali equivalent was determined by titrating a pipetteful of double distilled water in the same manner, The mean of at least four readings was always used. Let N and n be number of burette divisions of alkali equivalent to standard and sample respec- tively, and D and d their density at the time of pipetting out. VOL. 55, PART 2.—NEW SERIES. 26 398 EK. J. ALLEN AND E. W. NELSON. Then if A is the alkalinity of the standard and X the required alkalinity of sample: Dn x= Nd Since all operations were carried out at the same room temperature, no corrections for temperature are necessary. In spite of the greatest care consistent results could not be obtained by this method of analysis. A sample analysed against the same standard would sometimes give results varying as much as 0°5 mgrm. and occasionally 1°0 mgrm. OH °/,,... The work on indicators by Salm (42) and its application to this question has only recently come to our notice, and it is our intention to experiment on this in future research. ‘lhe figures quoted below as the results of analyses have been rounded off as whole numbers, since their interest lies in their comparative rather than their absolute value. For convenience they are quoted as “ alkalinities,” although we are fully conscious that the methods used do not warrant this assumption, and that their actual chemical significance is still obscure. The mean value for ‘outside water’? was found to be fairly constant at 40°0 mgrm. OH °/,., which figure agrees with results obtained by others for average ocean water. Samples from the aquarium tanks never gave as high figures as this, the average being approximately 37°5 mgrm. OH °/... From this it seems that the amount of base in equilibrium with CO, in tank-water is appreciably less than in outside water. A series of thirteen samples taken from seven miles beyond the Eddystone to well inside the Cattewater (an inner tidal harbour near Plymouth) showed a gradual lowering of the alkalinity from the normal 40, to 388 mgms. OH °/,, as the water became more estuarine and polluted. The addition of Miquel’s solution B to sea-water was found, on analysis, to reduce the “alkalinity ” by an amount equiva- lent to 10 mgrm. OH °/,, or more. The 1 c.c. solution B added to a litre of sea-water in itself contains a certain amount of free acid, equivalent to less than 4 mgrm. OH °/,.. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 399 But this reduction of alkalinity cannot be accounted for by the addition of free acid alone, because if only one quarter the amount of solution B is added, the alkalinity of the sample will be found to be, if anything, only very slightly higher. Also, if the various constituents of solution B are added as separate solutions, thus obviating any addition of free acid, a reduction equivalent to about 6 mgrm. OH °/,, is still obtained. The presence of ferric chloride in solution B gives a possible explanation of this phenomenon. Ifa solution of ferric chloride is added to a solution of a soluble carbonate, a reaction, which can be expressed by the following equation, takes place: 3R,CO;Aq. + Fe,Cl,Aq. = 6 RC].Aqg. + Fe,0,Aq..+ 3 CO,. When the ferric chloride is added to sea-water, the final result will be that a certain amount of the “excess base,” which was in equilibrium with CO,, will then be in equilibrium with the chlorine, available on the precipitation of hydrated ferric oxide, with a consequent liberation of CO, and a reduction in ‘alkalinity ” will, therefore, take place. An analogy between the actions of Miquel’s solution B and animal charcoal can be seen in the fact that water treated with animal charcoal also shows a reduced “ alkalinity,” the amount being very variable in different samples. Sea-water treated with H,O, also showed a lowering of the alkalinity, but in a much less degree when, as usual, minimal quantities were used. Control experiments on double distilled water, which had been treated with these substances, were tried, but great difficulty was found in obtaining an end point with the indicator. As far as could be judged, distilled water treated with solution B (quantities as with sea-water) showed a negative “alkalinity,” equivalent to about 8 mgrm. OH°/,,, and in the case of animal charcoal a positive alkalinity equivalent to 6 mgrm. OH°/,,, but the colour change was so slow that these results are only the roughest estimates. The possibility that the above results are due to some effect on the indicator, which entirely cloaks the true alkalinity, must always be taken into consideration. 4.00 E. J. ALLEN AND FE. W. NELSON. Before any attempts at analysis had been made, the proba- bility that considerable differences might be found in the alkalimity of the various media had presented itself, Im- provement in the growth of diatom cultures was found to result from the purely empirical addition of NaHCO,, this result being most marked in normal Miquel sea-water, outside water + solution B only, and Berkefeld water. No growth could be obtained in either ‘‘tank-water” or Miquel sea- water to which had been added 1 c.c. HCl (pure, concentrated) per litre, but on again raising the alkalinity of the latter by the addition of NaHCO, or KOH, good normal growths resulted. Richter (18) and H. Gill (5), also, both state that a weak alkaline reaction is necessary for the growth of diatoms, In our most recent experiments, all the media have been analysed for alkalinity, and those given in detail below illustrate the importance of determining this factor. Cultures of Thalassiosira decipiens were made in the following media ; A, ‘l'ank-water. Control. Poor growth, hardly normal. Later, good growth of minute forms, etc. B. Tank-water, treated with cold commercial animal char- coal, and filtered. Very good growth indeed. c. Tank-water treated with cold, pure animal charcoal, and filtered. Very poor growth, comparable to a without minute forms. p. ‘T'ank-water treated with pure animal charcoal as in ¢, but the animal charcoal was added red-hot. Fair growth, much superior to c, but not up to B. The sample of pure animal charcoal used here had been previously found to give very poor results, and it was also quite contrary to our experience that any improvement in growth should be obtained by adding it hot. But if we examine the results of analysis of these media for alkalinity ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 401 a probable explanation presents itself. The following figures are only comparative : A, 38 mgrm. OH °/,, (used as standard), Bora dF 3, » (higher than-usual). cr ON as » (very low indeed). dD. 34 90s ” Tt will be seen that the amount of Beatie in each treated sample follows the alkalinity very closely. Solutions of Na,CO,, NaHCO; and HCl were made up, so that 4.c.c. of any one contained an amount of acid or alkali equiva- lent to 10 mgrm. OH. From these a series of normal Miquel sea-waters of different alkalinities were prepared. Cultures of Thalassiosira decipiens were grown in these media. A. Normal Miquel sea-water. Control, A = 32°7 mgrm. OF */co- Perfectly normal growth. B. Ditto + 4 c.c. Na,CO; per litre. A = 41°7 megrm. OH °/,, (= + 90). | No difference between this culture and a. c, Ditto +.8 c.c. Na,CO3. per. litre, A = ;50'2 mgrm, OH = /5.. (==: 15), Best culture in series in quality and quantity. p. Ditto + 4 cc. NaHCO; per litre. A = 42:4 merm, OH °/.. (= + 9°7). Shghtly better than control. E. Ditto +. 8 c.c, NaHCOs, per litre. A = 51°5 mgrm, OE (S17 18i8). PAS! D. F, Ditto. +. 4 c.c. HCl per, litre, :- A = 22°2 mgrm. OH too (= = 10°5). Fair growth, but never up to control; exhausted much sooner. ¢. Ditto + 8 c.c. HCl per litres A = 11:1 mgrm. OH fog (= — 21:6). Poorest in series. 1 Figures in parentheses are difference in alkalinity from control, in mgrm, OH °/,,. 402. , kK. J.. ALLEN "AND &£.’ W. NELSON. Except in the cases where the alkalinity was lowered by the addition of HCl, the results obtained from this series were not up to expectation. Nevertheless the majority showed a distinct improvement from increased “alkalinity,” and in c, where the alkalinity had been raised 17°5 mgrm. OH °/ooo this improvement was very marked. Another point illustrated by cultural experiment is that in two samples of animal-charcoal water, one with “ outside” and the other with “tank-water ” as a basis, the amount of growth in the latter considerably exceeded that in the former, and at the same time it was found that, with the tank-water, the alkalinity had not been reduced to the same extent as in the case of the outside water. How far apparently anomalous results, which have so frequently occurred in our experimental work, could be explained by unforeseen changes in “ alkalinity,” can only be answered by future research. Salinity.—The salinity (or amount of salts dissolved in 1000 grm. sea-water) of the outside water used in these experi- ments only varied between small limits, S = 34°5 to 35°5 “loo: ~=6' The salinity of “tank-water” is also fairly constant, the average being about S = 34°9°/,,; water is only pumped up into the reservoirs at high water, spring tides, and unless the salinity on analysis is weli above S = 345 °/., no water is taken. Experiments to show what effect salinity pure and simple had on the growth of diatoms were undertaken. Samples of sea-water of various salinities were prepared by diluting down “outside water” with double distilled water, and by concentrating “outside water” by slow evaporation. Two litres of “outside water,’ S = 349, were evaporated down to the bulk of one litre, giving a 50°/,! concentra- tion. Miquel solutions 4 c.c. A, 2 c.c. B, were now added, and the solution was divided into ten culture vessels, 20 c.c. in each. Double distilled water was added, 2 c.c. to the first, 4 c.c. to the second, 20 c.c: to the last, so that a series of media were obtained, varying in salinity from 1 i.e. from every 100 c.c. sea-water 50 c.c. H,O had been subtracted. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 403 normal to nearly 50°/, concentration, each containing the same amount of Miquel’s nutrient solutions. These were inoculated from a mixed culture of Skeletonema costa- tum, Biddulphia mobiliensis, and Coscinodiscus excentricus. A good growth took place in all except the two with highest concentration. Of these two, the last remained praetically sterile and the growth in the other was very poor. ‘lhe limit of concentration, therefore, seems to lie between 35 and 40°/,. In the same way series of lowered salinities were prepared, and cultures of the same diatoms were grown in these. Dilution up to 100°/, did not seem to make any difference at all in the quantity or quality of growth. In a series extending the dilution to 200°/,, even in the cultures of lowest salinity a fair quantity of growth took place. ‘The range of salinities covered by the various series was S = 12 °/,, toS = 60 °,,, and within these limits no effect on growth could be observed, except in the very highest, where a distinct deterioration was noted. An attempt to grow Coscinodiscus excentricus in tap-water + Miquel’s solutions was tried, and it was thought that some slight multiplication took place, although it was certainly not at all considerable. Inoculating a culture of normal Miquel sea-water from this after six weeks gave no growth. Light.—Of all the factors controlling the rate of growth ofa culture, light seems to be by far the most important. With- out light a culture soon dies off completely, showing marked signs of malnutrition very soon after having been placed in the dark, the brown pigment being the first to go and later the chlorophyll. A culture (Thalassiosira) placed in the dark for five months was found to be completely killed, the diatoms being quite colourless. In cultures kept in bulbous flasks or in any spherical vessel, the strongest and earliest growth always takes place at the side of the vessel away from the source of light, where the light will be found to be con- centrated owing to the lens effect of a sphere of water. By 404 E. J. ALLEN AND E. W. NELSON, painting a flask black onthe outside up to the water-line of the medium, a very marked diminution in the rate of growth was obtained. The total growth was not affected, but depends on the available quantity of food-stuffs present. a Experiments on the reaction of cultures to different rays of the spectrum, obtained by coloured glass, were tried, but no results obtained. Miquel obtained marked results with yellow light, but in our experiments, with plankton diatoms, satisfac- tory cultures could not be obtained under these conditions. Temperature.—The highest temperature which diatoms and allied forms can stand was about uniform for all the species tested, and lay between 35°-40° C. Cultures of the following species, viz. Asterionella japonica, Nitzschia closterium, minute naviculoid diatom, Pleurococecus mucosus, Chilomonas sp., were slowly heated in a water bath, and at every rise of 5° C. from 15° C. to 45° C. a few drops of the culture were pipetted out and a fresh flask inoculated. In all the flasks cultures were obtained where the heating process had not been carried above 35° C., but none in those where the temperature had exceeded this. In the earlier stages of experimentation the cultures of diatoms were kept in various places about the laboratory, and so were under quite different temperature conditions. Those placed in the warmer situations, i. e. near hot-water pipes, as arule gave the most satisfactory results. In all the later work the cultures have been kept in one room, and an attempt has been made to keep the temperature of this room as nearly as pos- sible constant at 15° C. A continuous record of its temperature has been kept by means of a recording thermograph, and no very great change of temperature has been noted. In a few isolated cases the temperature has dropped as low as 9° C., and in hot weather has risen just above 20° C., but these have only been for very short periods, the average temperature having kept remarkably constant. An apparatus in which flasks could be kept at different uniform temperatures from 10° to 25° C., by means of hot air, was used, but no really satisfactory result could be obtained. About 17° C. seemed ARVLIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 405 to give the maximum growth, and the cultures below this temperature were usually superior to those above. General Conclusions.—The general conclusions to be drawn from the experiments described in this section, which were made with a view. to determining the conditions that underlie the successful culture of diatoms, may now be dis- cussed. Although the experiments have involved the making of some 750 different cultures, our conclusions on many of the questions raised are still indefinite, and much further work will be necessary before a satisfactory answer can be given to them. If we wish to obtain the maximum quantity of healthy growth of a plankton diatom, the diatom must first be obtained as free as possible from all other organisms, if not in a “pure” culture, at least in a “persistent” culture. All culture media should be sterilised either by heat or filtration, and the experiments should be conducted under sterile condi- tions. Starting with normal sea-water as the basis for the culture medium, it seems to be first necessary to raise the concentration of the nitrates, and possibly also of the phos- phates, in solution. But this simple addition of nutrient materials will not in itself suffice. Some other action, such as that exerted by Miquel’s solution B, by animal charcoal, or by peroxide of hydrogen, seems to be imperative in nearly every case. ‘I'he exact nature of this action we have not been able conclusively to determine. If the substances con- tained in solution B were purely nutritive in character, we should expect that, when alterations in the amounts of the different ingredients were made, or when any one of the ingredients was omitted altogether, the differences in the quantity of growth would show a direct relation to the kind of modification introduced. But our usual experience has been that solution B can be modified within certain limits, without producing any appreciable effect upon the resulting cultures, whilst, if these limits are exceeded, there is an almost complete inhibition of growth. In supplying a neces- sary increase of phosphates, both Miquel’s solution B and 406. HE. J.. ALLEN AND E. W. NELSON. animal charcoal may, and probably do, act as ‘‘ nutritive ” sub- stances ; but, since the addition of phosphates alone does not yield cultures comparable with those produced by either of them, and since, excepting phosphates, there is no possible common nutritive substance in their composition, we are led to conclude that, in addition to any nutritive effect, they must exert some other action. This view is supported by the results obtained by using H,O,. This substance cannot be directly ‘‘nutritive,” although it may be so indirectly, by oxidising into useful food-material substances which the diatoms are incapable of using in their metabolism, e.g. nitrites into nitrates. The absence of any increase in phos- phates, when using H,0,, may possibly be the reason why better results were not obtained with this medium. The action, which, in addition to any nutritive value, we must assume that solution B, animal charcoal, and H,O, can all effect, would appear to fall into the class of “ protective ”’ actions (p. 884). It is quite conceivable that, with different samples of sea-water, this “ protective” action is not neces- sary in every case, and this wouid account for the anomalous results met with when using sea-water + nitrates + phos- phates only, in which medium sometimes good cultures, but more often the reverse, are obtained. The effect of Miquel’s solution B, animal charcoal, and H,O, on the “ alkalinity ” of the sea-water, also points to some chemical change, which does not directly enter into the metabolism of the plants. It may be pointed out that the action of such substances as finely powdered carbon, and ferric oxide precipitates, have been shown to produce a favourable effect on nutrient solu- tions used for the culture of certain higher plants, and it has been suggested that the beneficial action of these substances is the removal of toxic elements from the media (Breazeale [3]). Such removal of toxins would fall under our definition of ** protective” action. Of nutritive substances, other than those already mentioned, we have still to consider, (1) silica, and (2) dissolved oxygen and carbonic acid. Having regard to the conditions under ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 407 which our cultures have been grown, i.e. in glass flasks, the question of silica does not seem to enter into the problems which we have discussed.. A few words must, however, be said as to the dissolved gases. Whipple (62) and Baldwin (44) have drawn attention to the observed relations, which are found in natural waters, between algal growths and the amounts of dissolved oxygen and carbonic acid. That these factors are of great importance cannot be doubted, but in our cultures it seems reasonable to suppose that the conditions of saturation of these gases are the same in all, since series of cultures in standard media, such as Miquel sea-water or Berkefeld water, can be set up with the certainty that, if not every one, at least a very high percentage, will give normal results. Of the purely physical factors, ight is by far the most important. Within limits, the rate of growth in a suitable medium seems to depend directly on the intensity of the light (Whipple [60]). Absence of light, as would be expected, soon completely kills the diatoms. Temperature also seems to affect the rate of growth to a certain extent, but for those temperatures at which we have experimented it does not appear to alter the quantity of growth. Salinity, apart from the quantities of available nutrient materials, can be varied within large limits without appreci- able effect on the diatoms. Il, Mixep CuLtuREs; In what has been said up to the present, we have been dealing with persistent cultures containing a single species of diatom, which are comparatively, if not entirely, free from admixture of other organisms. The study of cultures which contain a considerable mixture of organisms is not without interest. -A number of experiments have been made on the following lines: About 10,000 ¢c.c. of water, taken at some distance 408 EK. J. ALLEN AND E. W. NELSON. from shore, was placed in a tall bell-jar fitted with a “plunger,” which keeps the water in constant movement (‘Journ. Mar. Biol. Assoce.,’ vol. v, p. 176). The water was treated with Miquel’s solutions in normal proportions, and a considerable quantity of plankton taken with a fine-meshed net (150 meshes to the inch) was added, say 10 or 20 c.c. of a moderately rich sample of tow-netting. ‘The experiments were made during the spring and summer months, and the general course of events has been the same, with a certain amount of difference in detail according to the nature of the plankton present at the time. During the first two days the water often became cloudy, owing to the rapid multiplication of small flagellate infusoria, though this was not always the case. Plankton copepods and other animals gradually died off, though some survived for as long as a week or ten days. ‘The plankton diatoms, on the other hand, generally multiplied rapidly during the early days of the experiments, the first to become abundant in the body of the water being usually Skeletonema costatum, which at the end of a week might be so thick that a number of chains could be seen in every drop of water examined with the microscope. Along with the Skeletonema were found, other plankton diatoms, such as Lauderia borealis, Chetoceras (two or three species), Biddulphia mobi- liensis, Ditylium Brightwellii, and in nearly every case Thalassiosira decipiens. These latter diatoms were pre- sentin moderate numbers only, when the Skeletonema was at its height, butas the Skeletonema died down they increased in quantity. At the same time Nitzschia closterium com- menced to appear, both amongst the precipitate on the bottom of the jar and in the general body of the water. Small green flagellates often began to get numerous also at this stage. The true plankton diatoms were at their height about a fort- night after the experiments were started. At this time a great many diatoms of all kinds were to be found amongst the precipitate at the bottom of the jar, Asterionella japonica and Coscinodiscus excentricus being often ARTIFICIAL GULTURE OF MARINE PLANKTON ORGANISMS, 409 numerous here. During the course of the next week, how- ever, Nitzschia closterium rapidly increased in quantity, until not only the sides of the jar were coated with it, but the whole mass of the water became thick and opaque. By this time the plankton diatoms had all disappeared, with the exception of those which may survive for a considerable period amongst the precipitate at the bottom of the jar. Bottom diatoms (Navicula, etc.) had begun to grow on the sides of the jar, and small green and brown alge (Pleuro- coccus mucosus, Hctocarpus, etc.) also appeared. Infusoria (Euplotes and other smaller forms) then became numerous, and as the Nitzschia and bottom diatoms in- creased on the glass, large numbers of Amcebe made their appearance among them. ‘The jars continued in this con- dition for many months, the algze becoming more and more predominant. From these experiments, as well as from instances of mixed cultures obtained in the course of our attempts to secure persistent cultures of single species of diatoms, it seems usnal that, in a culture obtained by inoculating Miquel sea-water with plankton taken freshly from the sea, the true plankton diatoms are the first to develop in considerable numbers. Subsequently bottom diatoms and alge of various kinds become abundant, and the true plankton forms die out. A complete explanation of this sequence of events would probably be of a very complicated character, and we have practically no evidence from our experiments which bears very directly on the question. It would seem, however, that the early predominance of the plankton forms in the cultures would naturally follow from the fact ‘that, in the plankton material used for inoculation, these plankton forms are numerous, whilst bottom diatoms and spores of alge are rare. ‘he subsequent very great predominance of such a species as Nitzschia closterium may be due simply to a very much more rapid growth rate, though it is difficult to avoid the impression that the organisms, which finally take possession of the cultures, are in some way directly inimical to those 410 E. J. ALLEN AND E. W. NELSON. which they supersede, not merely by robbing them of their food supply, but perhaps, also, by the production of toxic substances. This suggestion does not, however, give an adequate explanation of the essential facts concerning these organisms. We have to consider two sets of species—(1) the true plankton forms, which flourish in the open sea and can be grown quite easily in the laboratory, provided the cultures yemain pure, and (2) what we may call “aquarium” or “‘hottom forms,” which under experimental conditions invari- ably take possession, when present in mixed cultures, whilst the plankton forms are killed off. Why is it that, although species of the second class are always present in small numbers in plankton taken from the sea, they are there alto- gether outnumbered by the true plankton forms, whereas under conditions such as those of our experiments they invariably succeed in gaining the upper hand? What are the factors which determine the difference in behaviour of these two sets of organisms in the sea and in the culture vessels? The whole question offers a very fruitful field for further experiment. ‘The evidence at present available is so slight that further discussion of it here is not likely to be of much service. The details of two experiments which we have made bearing on the subject of mixed cultures may, however, be recorded. A flask, containing about 1000 ¢.c. of sea-water treated with Miquel’s solutions, was inoculated with approximately equal amounts of certain persistent cultures of diatoms, which we possessed at the time. The following diatoms were in this way, introduced: Chetoceras constrictum, Bid- dulphia mobiliensis, Skeletonema costatum, Cos- cinodiscus excentricus, Streptotheca thamensis. 'The flagellate (Chilomonas sp.) was also introduced, since it was present in the culture of Coscinodiscus. The ex- periment was started on August 26th, 1907. On September 6th (11 days) Biddulphia, Coscinodiscus and Cheto- ceras were increasing rapidly and were very healthy, ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 411] Skeletonema was not so good, and no Streptotheca was found, On October 2nd (37 days) Biddulphia was numerous and healthy, Coscinodiscus was healthy but not so numerous, Skeletonema was poor, and Chetoceras was not seen. Flagellates (Chilomonas) had become very numerous. On October 31st (66 days) all the diatoms were in very poor condition, Coscinodiscus being slightly better than the others. The flagellates (Chilomonas) were extremely thick, giving the water a deep red colour. Subsequently a small green alga (Pleurococcus mucosus) appeared, having probably been derived from the Coscinodiscus culture. ‘This increased very greatly in quantity, whilst the flagellates become inconspicuous. On July 28th, 1909 (1 year 11 months) some Coscino- discus, which were still in a healthy condition, were seen in a sample examined from the flask. A great quantity of Pleurococcus, in a healthy condition, was also present, but no other oganisms were noted. On tlis date a subculture was made from the flask in normal outside Miquel. The subculture gave a considerable growth of Skeletonema, the cells being, however, of a very abnormal character, and a good many normal and healthy Coscinodiscus were found in each sample examined. ‘lhe whole culture was crowded with Chilomonas in a very active state, which gave the whole contents of the flask a deep red-brown colour. Upto August 24th the green alga (Pleurococcus) had not become suffi- ciently abundant to be detected by the naked-eye appearance of the flask, though it could be seen in samples examined with the microscope. In another experiment a flask of Miquel sea-water was inoculated (May 4th, 1908) from two cultures, one containing the green alga (Pleurococcus mucosus) and the other Thalassiosira decipiens. At first both did well, and on May 20th (16 days) there was a very good crop both of the diatom and the alga. Gradually, however, the alga became predominant, and on October 14th (163 days) only quite empty 412 E. J. ALLEN AND FE. W. NELSON. frustules of Thalassiosira could be found, whilst the growth of Pleurococcus was abundant and healthy, The only case where a diatom was observed to flourish in the presence of this green alga was in a culture of Nitzschia, a bottom form, In this case a very abundant growth of the diatom was obtained, but the Pleurococcus did not multiply to any extent although it could always be found on microscopic examination. III. Nores on Parricutar Species or Diaroms, ON THEIR Meruops or Repropucrion, AND ON OTHER ALGm OccuR- RING IN CULTURES. A list has been already given (p. 367) of those species of diatoms which we have obtained in “ persistent”? cultures. Of these a species belonging to the genus Thalassiosira has been used for experimental work in the great majority of cases. We are not quite certain as to the identity of the species, but since it most resembles ‘I’. decipiens Grun. we have called it by that name, although it does not exactly con- form to the published descriptions of that form. The most characteristic feature of this particular species is the eccentric markings on the valves, which are also seen on the valves of the diatom Coscinodiscus excentricus Ehr., and, as is typical of the genus, the frustules are united into chains by a delicate filament. Jérgensen (50, p. 96) describes the valves as “decidedly convex,” Gran (49) as “‘ flat,” and both agree that there are marginal spines and a single asymmetrical spine. Our cultural forms are united together by a filament into chains, some of which are made up of 500 cells and more, but the distance between each is considerably smaller than that figured by Gran. The valves are quite flat and the marginal spines are often present, although this is not always the case. The odd, asymmetrical apiculus can nearly always be seen. The eccentric markings have only been observed in a few isolated cases, and are then usually very indistinct. In one culture these markings on the vatves were yery distinct, ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 413 and were also easily seen on the megafrustules (cf. below), which developed in it, but in none of the several generations of cultures started from this one have we been able to find any traces of marking at all. ‘he genus seems to be in con- siderable confusion, and it is probable that the conflicting descriptions given by different observers are due to variations in what is really one species. Persistent cultures of Coscinodiscus excentricus Khr. have also been obtained, and it is interesting to note that this diatom sometimes forms chains, but they are rather excep- tional. ‘hese chains are never as long as those commonly found with Thalassiosira, two or four cells only being the rule. ‘lhe filament joining the valves is also finer and more easily broken. The two species are quite distinct, and cultures of them can be discriminated by a practised eye. _ Two species of the genus Biddulphia are commonly met with in our cultures, namely Biddulphia mobiliensis (Bail.) Grun. and Biddulphia regia M. Schultze. These two forms are generally regarded as one species, but Osten- feld (54) has recently shown that they are really distinct. We have obtained persistent cultures of both forms from several different samples of plankton, and the two species are easily recognisable, never merging into one another. When Petri dishes, moculated from plankton (see p. 367), contain both species, the colonies can be easily distinguished with a small hand lens. The most generally accepted theory of the reproduction of the Diatomacee is briefly that the cells divide by simple fission, but on account of the rigid character of the cell-walls each division necessitates a decrease in size of the new valve, since this must always be formed inside the old valve. So the frustules gradually get smaller and smaller as multiplication proceeds, thus necessitating some process by which the original size can be re-established. This takes place by the formation of what are known as auxospores, which ultimately form megafrustules, and these in turn multiply by division until the minimum limit of size has again been reached. VOL. 55, PART 2.—NEW SERIES. 27 414 E. J. ALLEN AND E. W. NELSON. ‘There are also several special processes of reproduction, but no occurrence of any of these has been noted in our work (cf. Miquel [14]). The diatoms in our cultures multiply by simple. fission, and although there is, in nearly every case, a considerable diminu- tion in size when compared with specimens from the plankton, this diminution soon seems to reach a limit, where further decrease does not take place. In chains of Thalassiosira, several hundred cells in length, no difference in size between individuals could be made out. Auxospores are commonly formed with every species, but only in cultures of Coscino- discus and Thalassiosira have megafrustules been found, and in these they are very exceptional. These megafrustules seem to divide once or twice and then die or form new auxo- spores. What exactly is the fate of these auxospores, which are often exceedingly numerous, we have not been able to make out. It seems that cultural conditions are not favour- able to this mode of reproduction, and that the auxospores do not further the multiplication of the diatom at all. If this were not the case, stages of the formation of auxospores into frustules must have been seen in some at least of the very numerous samples examined. As it is, what has been seem to take place is, that the cell contents expand and force apart the valves of the diatom and emerge as a spherical body about three or four times the diameter of the parent cell. The chromatophores and diatomin then collect to one side, form- ing a compact cap against the cell-wall. Beyond this point no stages have been found, except in the case of the few cultures where megafrustules were formed. In these the chromatophores, etc., gradually formed into the shape of the diatom (Coscinodiscus); the siliceous coat with plain eccentric markings was easily seen inside the spore; and lastly, the cell-wall of the spore burst, leaving the mega- frustule free. The megafrustule was measured and found to have a diameter three times that of the parent cell. In the case of the diatom we have very largely used for feeding larve, etc., namely Nitzschia closterium, forma ARTIFICIAL CULTURE OF MARINE PLANK'TTON ORGANISMS. 415 minutissima, a great number of cultures have been made, all originating. from the single drop from which -the first persistent culture was obtained. The total amount of growth in all the various cultures has been enormous, and the number of generations must be quite inconceivable. No diminution in size has, however, been appreciable, and no sign of any method of re-establishment of size has been seen, although these cul- tures have been under constant observation for over two years. This seems to prove that the theory of gradual decrease in size with successive generations cannot be generally applied. The following experiment on the rate of multiplication of Thalassiosira in normal Miquel sea-water was carried out. A single drop from a fresh and vigorous culture was kept under a microscope as a hanging-drop preparation in a moist chamber. ‘The number of diatoms in this drop was counted from time to time and the results are given in the following table : Number of Geometric Day. frustules. progression. 11th : : ; 59 : : 63 14th ‘ 7 ; 62 : 68 19th < : : 85 : : 85 27th : : 140 : ‘ 120 34th : : : 170 ; : 160 41st : : : 190 4 ; 220 The curve obtained by plotting the number of diatoms against the number of days approximates the curve of an ordinary geometric progression, where the ratio is 2 and the periods are equal to sixteen days. ‘To show this the figures read off from the curve at the same intervals as the diatoms are appended in the table. From this it will be seen that, after a start had been made and before exhaustion set in, the numbers obtained agree fairly closely with the assumption that every diatom divided once in a period of sixteen days. Probably in normal cultural conditions the rate of multiplication greatly exceeds this figure on account of better lighting, etc. (ef. Miquel 12). 416 EK. J. ALLEN AND E. W. NELSON. Besides diatoms, many other organisms appear in these cultures. We are. indebted to. Mr. G. 8. West for the identification of a form of unicellular alga, which is very common and difficult to avoid when attempting to obtain persistent cultures of the Diatomacex, namely, Pleuro- coccus muvcosus (Kutz.) Rabenh. This small green alga, if once introduced into a culture of a plankton diatom, will soon multiply at the expense of the latter with its ultimate extinc- tion. It is very hardy, and cultures. of it in almost every medium seem to last indefinitely. Multiplication beyond a certain point probably does not occur, but the cells retain their colour and normal shape, and will start active repro- duction if suitable nutrient material is provided. In cultures inoculated from plankton, many other forms of unicellular and filamentous alge thrive. Several species belonging to the classes Rhodophycex and Myxophycew commonly occur, but we have not been able to identify them. The most usual filamentous forms of Chlorophycezx are Enteromorpha, Vauchera, Rhizoclonium, ete. It is interesting to note that it was the unintentional appearance of young plants of Laminaria digitata in some of our Petri dishes that led Mr. Drew (4) to cultivate this alga in Miquel sea-water and so discover its early life-history. Cultivations of marine alge by these methods would without doubt yield many new species, and would also provide rich material for the study of their modes of reproduction. Many forms of flagellates live either together with diatoms or alone. Among: these is an unidentified species of Chilo- monas, which we have obtained in persistent culture. It multiplies very rapidly, colouring the whole medium a deep red-brown. It flourishes in Miquel sea-water and its nutrition is evidently autotrophic. In one culture, in Miquel sea-water inoculated with plankton, a number of coccospheres developed, probably Coccospheraatlantica Ostenf. Other flagellates and ciliated infusoria are very commonly met with, such as Bodo, Euplotes, Euglena, etc., which all seem to depend on the diatoms or other vegetable organisms for their food material. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 417 IV. THe Rearina or Marine Larva. In the rearing of pelagic larval forms of marine animals,! the principle which we have followed has been to introduce into pure, sterile sea-water the larve to be reared, together with a-pure culture of a suitable food. As far as practicable all other organisms have been excluded from the rearing vessels. It should beadded that the food used in all successful experiments has been of a vegetable nature, and has continued to grow actively in the vessels. This is important from the point of view of oxygen supply. Under the above conditions, or rather under the nearest approach to them at which we have-been able to arrive, no change of water has been found necessary. Methods.—It will, perhaps, best make the matter plain if we first of all describe the actual procedure, which we now follow in the case of such ananimalas Hchinus esculentus or K.acutus. ‘The water to be used is first of all prepared by treating water from the ‘aquarium tanks with powdered animal charcoal, filtering it through a Berkefeld filter (p. 375), and collecting it in sterilised glass vessels. All instruments and pipettes are sterilised by baking in an oven, and a fresh sterile pipette is used for each operation during the progress of the work. Specimens of Echinus are then opened until a perfectly ripe female has been found, that is to say, one in which the eggs separate quite freely when a portion of the ovary is shaken in sea-water. Pieces of ovary, taken from a little below the exposed surface, are then placed in sterile sea-water in a shallow glass dish, and shaken with forceps in order to get the eggs well separated, or a number of eggs from the centre of the ovary are drawn up with a pipette and placed in the water. A very small quantity of active sperm from a ripe male is then added, very little being sufficient to fertilise a large number ot eggs. Excess of sperm should be avoided owing to its 1 See ** Bibliography,” especially Grave (26), MacBride (28-80), Don- caster (25), ete. 418 rE. J. ALLEN AND KE. 'W. NELSON. lability to putrefy. After an interval of ten or fifteen minutes the water, containing the eggs, is filtered through bolting silk of 100 meshes per inch, which just allows single eggs to pass through, whilst keeping back clusters of eggs or other large material. The filtrate is divided amongst a number of tall narrow beakers containing sterile sea-water, and the beakers, after being covered with a glass plate, are placed where the temperature will be uniform and not rise much above 15°C. In the course of twenty-four hours the healthy larvee will swim up to the surface and can be easily seen and removed from vessels of this shape. They are transferred by means of sterile pipettes to jars! of sterile sea-water, about fifty to seventy larvae being put in each jar of 2000 c.c. sea-water. At the same time, a good pipetteful of a pure culture of diatom is added to each jar. The small diatom Nitzschia closterium, forma minutissima we have found most useful, as its size is suitable, and it grows well in animal-charcoal tank-water, floating throughout the body of the water, and so being in intimate admixture with the larves. The jars are placed in a moderate light and at as even a temperature as possible.” No further attention is necessary until the larvee have metamorphosed. ‘The meta- morphosis takes place in from six to nine weeks after fertilisation. Larvee may be taken out from time to time and examined to see if they are feeding well. If the diatoms do not grow sufficiently rapidly in the jar more should be added from the culture flasks. We are more often troubled, however, towards the end of an experiment, by an excessive abundance of diatoms. In this case the jar may either be put ina darker place, or some of the water may be drawn off and replaced by a fresh supply of sterile sea-water. Care should 1 The vessels we use are ordinary green-glass sweet-Jars, having a capacity of about 2000 ¢.c., which are kept covered with the glass stoppers provided with such jars, from which the cork band has been removed. 2 In hot weather we often stand the jars in one of the tanks of circu- lating aquarium water, which maintains them at a very uniform tempe- rature. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 419 be taken to have a sufficient supply of food at the beginning of the experiment, so that the larva may be able to feed as soon as they are ready for food. The method just described can be modified in various ways without detriment to the result. Sufficient sterilisation of the water may be effected by heating to 70° C. for fifteen minutes, affer which it should be aérated by violent shaking. ‘Outside water” may be used instead of ‘ tank- water,” and may be treated with Miquel’s solutions in the ordinary way, to ensure a satisfactory growth of the food- diatom. With regard to the food organisms, we have tried to obtain as large a variety of these in pure culture as possible, and then to make trial of a number of them with each batch of larve on which we have experimented. If no suitable pure cultures are available, success can sometimes be obtained by adding a few drops of tow-netting, collected with a fine- meshed net (180 meshes per inch), directly to the treated sterile water containing the larve. In this case one depends on the chance of a suitable food-organism growing in the vessel, unaccompanied by any destructive organism. On several occasions a satisfactory result has been reached by proceeding in this way, and the method is generally worth a trial, seeing that the number of larve obtainable from an ordinary fertilisation is very large and many different experiments are easily made with them. We will now give details of some of the results obtained by making use of the methods described, or of their modifications. Echinus acutus.—The first successful experiment was made with this species. Eggs fertilised on June 13th, 1905, produced healthy larve, fifty to seventy-five of which were placed, three days later, in a glass jar containing 2000 c.c. of ouside sea-water, filtered through animal charcoal, to which modified Miquel solutions were added. ‘hey were fed on a diatom culture, containing a small species of Chetoceras, which did not form chains, a small diatom probably belonging to the genus Melosira, a small naviculoid diatom, two 420. Kk. J.. ALLEN AND KE. W. NELSON. minute flagellates, and a small green organism, probably one of the Pleurococcacew. ‘The vessel stood in a shallow tank, through which a stream of aquarium water was flowing and the temperature was fairly constant at 15° or 16° C., though there is one record of 19° C. at the end of July. The first two young: Echinus were seen on July 25th, forty-two days after fertilisation, and on August Ist twenty were counted. On August 5th (the fifty-third day) a. careful search through the jar gave twenty-one young Echinus of normal size attached to the glass, six minute but fully formed Hchinus, about twenty-three still in the Pluteus. stage, roughly half of which were well advanced. On August 16th some of the water, which had not been changed since the beginning. of the experiment, was replaced by “outside” water. On October 5th (sixteen weeks after fertilisation) twelve Echinus were still alive. Some pieces of red seaweed were placed in the jar, upon which the Echinus fixed them- selves and fed. Several of these specimens lived for over a year, but sufficient attention was not given to finding suitable food for them after the metamorphosis, so that they did not grow very large. Kchinus esculentus.—Three successful experiments have been made with EK. esculentus. In the first (eggs fertilised April 5th, 1907), “outside”? water treated with animal charcoal and filtered through filter-cloth, but not otherwise sterilised, was used. A number of jars of 2000 e.c. capacity containing Jarve were set up, and, to the most of these, various diatom cultures then in our possession were added, none of which, however, gave a satisfactory result. In two jars, on the other hand, to which no culture was added, there was considerable growth of diatoms and of a flagellate, upon which the Plutei fed. The first young Echinus were recorded in both jars on June 8th (sixty-four days),, but may have been present a few days earlier. Eventually from thirty to forty metamorphosed in one jar and about twelve in the other. The temperature varied from 10°5° Cy. to: 12°5° C. ARTIFICIAL CULTURE OF MARINE PLANK'TON ORGANISMS. 421 Inthe second experiment (eggs fertilised June 8th, 1908), made with similar water, the larve were feda on pure culture of Nitzschia closterium var., and six had completely metamorphosed on July 26th (forty-eight days after fertilisa- tion), two more subsequently coming through. ‘The tempera- ture was generally 15° to 16° or 17° C. In the third experiment (eggs fertilised March 29th, 1909) aquarium tank-water treated with animal charcoal and then filtered through a Berkefeld filter was used.. Plutei fed witha pure culture of a small flagellate (probably Chilomonas sp.) grew satisfactorily, and eight young Echinus were found on June 5th (sixty-eight days after fertilisation), which. had probably metamorphosed some days earlier. ‘I'wo other jars, in which Nitzschia closterium var. was used as food, were not successful, probably because the growth of diatoms became too thick towards the end:of the experiment. Echinus miliaris.—In the first experiment with this species animal-charcoal Berkefeld water was used, each jar containing, as usual, 2000 ¢c.c. In one jar the Plutei, from eggs fertilised on August 27th, 1907, were fed on a pure culture of Nitzschia closterium, var. On October 4th, 1. e. thirty- eight days after fertilisation, one Hchinus has just metamor- phosed. On October 29th about a dozen healthy-lookmg Echini were climbing about the jar, and many were still in a healthy condition on January 8th, 1908. ‘Temperatures : September, 15° to 19° C.; October, 16° dropping to 13° C. towards end; November, 12° to 11° C.; December, 15° to: 10°C. To another jar containing larve from the same batch a few drops of fresh Plankton were added as food. The Plutei in this case fed on flagellates and Nitzschia which grew in the jar, and several metamorphosed. In a second experiment with eggs fertilised on September 13th, 1907, the larve were fed with Nitzschia closterium, but although there were a few well-advanced plutei still living on January 8th, 1908, none completed the meta- morphosis. 422 E. J. ALLEN AND E. W. NELSON. Cucumaria saxicola.—A female Cucumaria, one of a number in a dish containing “ outside” water, laid eggs, which were fertilised, and sezmented on May 12th, 1906. A number of these were placed in a flask in 800 c.c. of “ outside”’ water, which had been sterilised by heating and then treated with animal charcoal and filtered. About 1 c.c. of fine plankton, containing diatoms, was added to the flask on May 12th. On May 25th some of the water was poured off and a new supply added. As the amount of food seemed small, some culture of a green alga (Pleurococcus mucosus [Kutz.] Rabenh.) was added, and this continued to grow well in the flask. The larve continued healthy and formed young Cucumaria, of which many were still alive on July 25th, 1907, i. e. fourteen months after fertilisation. Some of the water was changed in this flask on May 30th, 1906, June 18th, 1906, and September 15th, 1906, and July 25th, 1907. Although many of these Cucumaria remained quite healthy they did not grow to any great size. Probably the food which was suitable to the larve and early stages, ought to have been changed as the animals grew older. Pomatoceras triqueter.—The larve of Pomatoceras are perhaps the easiest to rear, and give the most certain results of any with which we have experimented. They do well on the minute variety of Nitzschia closterium, but will feed upon almost any small diatom. Since the adults live in calcareous tubes attached to stones, and the tubes have to be broken open before the eggs can be obtained, it is not easy to get the latter free from infection of other organisms. If, therefore, the eggs are fertilised and placed in sterilised animal- charcoal water with only moderate precautions, sufficient growth of diatoms or other organisms will generally take place in the jar to feed the larve and bring them to the adult state. When once fixed to the glass the worms are very hardy and healthy, and a stream of ordinary aquarium water can be run through the jar. They then grow rapidly and attain a size equal to any found on the shore. The following experiment may be given in detail to illustrate the time ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 423 occupied in development. On August 29th, 1907, eggs of Pomatoceros triqueter were fertilised in animal-charcoal Berkefeld water, and some pure culture of Nitzschia clos- terium var. added. The larve fed well, and on October Ist (i.e. thirty-three days after fertilisation), a great number had fixed on the sides of the jar and made quite normal tubes. A constant stream of the ordinary aquarium water was then allowed to run through the jar, and the worms continued to grow and flourish, reaching a large size, and are still alive and healthy (November, 1909). A similar result was obtained from the same batch of eggs by feeding on a pure culture of a flagellate infusorian. Temperatures during these two experiments were between 15° and 19° C. Chetopterus variopedatus.—Four experiments were made with this species. The food which gave most promise of success was the diatom Nitzschia closterium var. Larve from eggs fertilised on July 20th, 1908, fed on this material lived until October 30th, and reached an advanced stage. They did not, however, adopt the adult habit and form tubes. Two larve were also reared to an advanced stage by using flagellates, and, in later stages, the diatom Skeletonema costatum as food. Sabellaria alveolata.—One experiment only was made with this species, on eggs fertilised on July 19th, 1908. ‘The eggs were fertilised in “outside” water, and the larve subsequently transferred to jars containing animal-charcoal Berkefeld aquarium water. ‘They were fed ona pure culture of Nitzschia closterium var., and kept healthy and active, and developed well until nearly the end of October, when, simultaneously with a sudden drop in temperature from 15° and 16°C. to 12° and 9 C., they sank to the bottom of the vessel, and in about three days were all dead. ‘Temperatures: During July and August the temperature kept fairly constant at about 17° C., with a range from 15° to 19° C. During September it was generally about 15°C., and continued at about this level until the fall in the middle of October. Archidoris tubercnulata.—A good many trials have 4.24, EK. J.. ALLEN AND E, W. NELSON. been made to rear the larve of nudibrunchiate molluscs, but up to the present not much success has been achieved. The best experiment was one made with larve of Archidoris tuberculata. A number of veligers of this species hatched out on May 8th, 1906, from some spawn which had just been collected from the shore. Some of these were put in a flask containing 1000 ¢.c. of sterilised animal-charcoal water, and about 1 c.c. of fine plankton was added. On May 14th a few veligers were transferred to another flask of sterilised animal-charcoal water and some pure culture of the green alga, Pleurococcus mucosus, was added. Whereas the larvee in the original flask did not live long, those provided with the green alga fed well and developed for some con- siderable time. A number of them were active and vigorous on July 4th,i. e. fifty-one days after. hatching, and several were still swimming at the end of July. On August 15th none could be seen moving, but two of those which lay on the bottom, when examined with the microscope, showed no sign of decomposition. The animal was retracted in the shell, but the tissue looked healthy, and the eye-spots and otoliths could be seen. ‘the growth in the flask seemed to be a quite pure culture of Pleurococcus. Larvee were examined again on September 14th, and appeared much as in August, the tissue still showing no sign of disintegration. ‘The flask was not again examined microscopically until July 25th of the following year (1907). No sign of the larvee could then be seen, but the culture of Pleurococcus remained pure and healthy. Subsequent experiments were made with spawn, which was deposited by the females in confinement. Although the spawn hatched and gave apparently healthy larve, these did not live for more than a few days. Calanus finmarchicus.—A single experiment is perhaps worth recording, as showing that it ought to be possible to rear this species without great difficulty. On August 8th, 1905, to a flask containing 1000 ¢.c. of outside water (unsterilised) there was added } ¢.c. of Miquel’s solution B ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 425 and $ c.c. of a 15 per cent. solution of anhydrous sodium carbonate. A few Calanus finmarchicus and some decapod Zoeeas were put in, together with a quantity of a culture containing mixed diatoms. On September 8th all the Zoeas were dead, but three Calanus were alive, and Nitzschia and a number of bottom diatoms were very plentiful. On September 17th the three large Calanus were alive and vigorous, and a considerable number of Nauplii were seen in the flask. By September 22nd two of the Nauplii had developed into young Calanus. ‘These, however, did not live for more than a week or ten days, and the adults also died. ‘The flask was abandoned on November 13th, the water in it not having been changed: since the commencement of the experiment. Hchinus hybrid.—A successful experiment on crossing H. esculentus and EH. acutus was carried out by Mr. W. De Morgan, who was working at the Plymouth Laboratory. We provided him with treated water and diatom cultures for food, and he followed our methods. We are indebted to him for allowing us to publish these results. Some eggs from a ripe EH. esculentus were fertilised by active sperm from an E. acutus, in sterilised water, on March 29th, 1909. Healthy larvee were obtained, and were transferred two days later to tank-water, which had been treated with animal charcoal and filtered through a Berkefeld filter. A culture of Nitzschia closterium var. was added as food, and the larve developed rapidly, feeding well. Several were completely metamor- phosed on May 7th, or thirty-nine days after fertilisation, In all thirty young hybrids were obtained, and:a number of these are still alive and feeding on red weeds. Sacculina carcini.—Mr.. Geoffrey Smith has recorded the fact (‘Quart.. Journ. Micr. Sci,,’ vol. 51, 1907,. p. 625) that he was able to rear the larve of Sacculina up to the Cypris stage, when they attached themselves to their host, Carcinus menas. These larve were kept in aquarium tank-water treated with animal charcoal and filtered through a Berkefeld filter. In this case the question of food did not 426 E. J. ALLEN AND EK. W. NELSON. arise, as the larvae do not feed after hatching. It must be noted, however, that these larve had previously been reared by Miiller and by Delage. Summary of Method for Rearing Larve.—We have found that the best results in rearing marine larvee have been attained by taking the following precautions : (1) The eggs of the female selected must be really ripe, and the spermatozoa of the male active. (2) The smallest quantity of sperm necessary to fertilise the eggs should be used. (3) Sterile sea-water, treated in such a way that diatoms etc., will grow well in it, should be used. No frequent change of water is then necessary. (4) All dishes, jars, instruments, and pipettes, should be carefully sterilised before use. Every possible effort should be made to prevent the introduction into the rearing-jars of any organisms other than the larve to be reared, and organisms on which they feed. The jars should be covered with loosely fitting glass covers. (5) The eggs after fertilisation must be separated from all foreign matter, pieces of ovary, or testis, etc. As soon as the larvee swim up they should be pipetted off into fresh vessels of treated water, so as to leave behind any unseg- mented eggs, etc. (6) The food organisms should be small in size, so that the larvee can draw them into the mouth by ciliary currents. The food should distribute itself through the body of the liquid, and not settle too readily on the bottom of the vessel. (This is one of the great advantages of the diatom Nitzschia closterium, forma minutissima.) (7) The food should be abundant early, so that the larve may commence feeding as soon as they are able to do so. The food, however, must not be allowed to get excessively thick in the water. It can be kept down by diminishing the light, or by changing some of the water. (8) The temperature should be kept as constant as possible. Within limits the actual degree of temperature ARTIFICIAL CULTURE OF MARINE PLANK'TON ORGANISMS. 427 is not so important as the avoidance of rapid changes of temperature. (9) A good north light, not exposed to direct sunlight, is most suitable for the rearing-jars. (10) In determining the amount of water to be used in any particular vessel, regard must be had to the amount of water surface exposed to the air, which should be large in propor- tion to the volume of the water. 10. ii, (11) A change of food is generally required after the meta- morphosis of the larvee. BIBLIOGRAPHY. Cultures. . Beijerinck, M. W.—* Das Assimilationsprodukt der Kohlensaure in der Chromatophoren der Diatomeen,” ‘ Rec. Trav. Bot. Neerland,’ i, 1904, p. 28. . Benecke, W.—‘ Uber farblose Diatomeen der Kieler Féhrde,”’ ‘Pringsh. Jahrb.,’ xxxv, 1900, p. 567. . Breazeale, J. F.—* Effect of Certain Solids upon the Growth of Seedlings in Water Cultures,” * Botanical Gazette,’ xli, 1906, p. 54. . Drew, G. H.—*The Reproduction and Early Development of Laminaria digitata and L. saccharina,” ‘Annals of Botany,’ acim L9LO, p. 177. . Gill, C. Houghton.—See Van Heurck, H., 21 and 57. . Karsten, G.—* Die Formanderungen von Skeletonema costatum Grun und ihre Abhangigheit von dusseren Faktoren,” ‘ Wiss. Meeresunters. Kiel.,’ N.F. iii, 1898, p. 5. “Ueber farblose Diatomeen,” ‘ Flora,’ Ixxxix, 1901, p. 404. . Lockwood, $8. M.—* Raising Diatoms in the Laboratory,” ‘ Journ. New York Micr. Soc.,’ 1886, p. 153. “ Aberrant Forms in Cultivated Diatoms,” ‘ Amer. Monthly Micr. Journ.,’ 1893, p. 259. ——— ‘“ Formes anomales chez les Diatomées cultivées artificielle- ment,’ ‘Le Diatomiste,’ ii, 1893-96, p. 9; ‘ Ann. de Micrographie,’ x, 1898, p. 1. Miquel, P.—‘‘ De la Culture artificielle des Diatomées,” ‘Le Diato- miste,’ i, 1890-93, pp. 73, 93, 121, 149, 165; ‘Le Micro. Pré- parature,’ v, 1897, p. 69. 428 ES J. ALLEN AND ZB. W. NELSON, 12. Miquel, P.—‘ Recherches expérimentales sur la physiologie, la morphologie, et la pathologie des Diatomées,” ‘ Ann. de Micro- graphie,’ iv, 1891-2, pp. 273, 321, 408, 529; v, 1893, pp. 437, 521; x, 1898, pp. 49, 177, 182. ‘Le Micro. Préparature,’ xi, 1903, p. 174; xii, 1904, p. 32. 13. “Du rétablissement de la taille et de la rectification de la forme chez les Diatomées,” ‘ Le Diatomiste,’ ii, 1893-96, pp. 61, 88. 14. -~ “Des Spores des Diatomées,” ‘Le Diatomiste,’ ii, 1895-96, p. 26. 15. “ Du noyau chez les Diatomées,” ‘Le Diatomiste,’ ii, 1893-96, p. 105; ‘Le Micro. Préparature,’ xii, 1904, p. 167; xiii, 1905, p. 83. 16. Richter, O.—“ Reinkulturen von Diatomeen,” * Ber. deut. bot. Gesell.,’ xxi, 1903, p. 493. a7, “Uber Reinkulturen von Diatomeen und die Notwendigkeit der Kieselsiiure fiir Nitzschia palea (Kitz) W. Sm.,” ‘ Verh. d. Gesell. deut. Naturf. u. Arzte, Breslau,’ ii, 1904, p. 249. 18. “Zur physiologie der Diatomeen,” ‘S.B.K. Akad. Wiss. Wien.,’ exv, 1906, p. 935. 19. “Ueber die Notwendigkeit des Natriums fiir eine farblose Meeresdiatomee,” ‘ Wiesner-Festschrift. Wien.,’ 1908, p. 167. 20. Senft, E.—* Ueber die Agar-Agar Diatomeen,” * Zeit. d. Allgem. ost. Apotheker- Vereines,’ 1902, n. 9, figs. 1-9. 21. Van Heurck, H.— Notice biographique sur C. Houghton Gill,” ‘Le Diatomiste,’ ii, 1895-96, p. 125. “Culture des Diatomées,” ‘ Zeit. f. angew. Mikrosk.,’ iii, 1897, pp. 195 and 225. 22. Rearing. 28. Agassiz, A.— Revision of the Echini,’ Cambridge, Mass., 1872-74. 24. Cowles. R. P.—‘t Notes on the Rearing of the Larve of Poly- gordius,” ‘Johns Hopkins Univ. Circulars,’ xxii, No. 161, 1903. 25. Doncaster, L.—‘* On Rearing the Later Stages of Echinoid Larve,” ‘Cambridge Phil. Soe.,’ xii, 1903, p. 48. 26. Grave, C.—* A Method of Rearing Marine Larve,” ‘Science,’ N.s. xv, 1902, p. 579. 27. Lillie, R. S—* The Structure and Development of the Nephridia of Arenicola,” * Mittheil. Zool. Sta. Neapel.,” xvii, 1904-06, p. d41. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 429 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. MacBride, E. W.—* The Rearing of Larve of Echinide,” ‘ Reports Brit. Assoc.,’ Dover, 1899, p. 458. “Notes on the Rearing of Echinoid Larve,” ‘Journ. Mar. Biol. Assoe.,’ N.S., vi, 1900-03, p. 94. “The Development of Echinus esculentus, together with some Points in the Development of E. miliaris and EK. acutus,” ‘ Phil. Trans. Roy. Soc.,’ B. exev, 1903, p. 285. Theél, H.—* On the Development of Echinocyamus pusillus,” ‘Nova Acta R. Soe. Sci.,’ Upsala, 1892. Zeleny, C.—* The Rearing of Serpulid Larvee, with Notes on the Behaviour of the Young Animals,” ‘Biol. Bull. Woods Holl.,’ vill, 1905, p. 308. . Chemistry. Dittmar, W.—“On the Alkalinity of Ocean Water.” ‘Rep. Challenger Expdt.,’ 1873-76, Chem. 1, p. 124, London, 1884. Fox, C. J. J— On the Co-efficients of Absorption of the Atmos- pheric Gases in Distilled Water and Sea-Water.’ Part IT: “Carbonic acid.” ‘Publ. de Circonstance, Conseil Internat. pour l’expl. de la mer.,’ No. 44, 1909, Copenhagen. van ’t Hoff, J. H—‘ Zur Bildung der ozeanischen Salzablager- ungen, Braunschweig, 1905. Knudsen, M.—“ Hydrography,” ‘ Danish Ingolf-Expedtn., I, Part 2, p. 21, 1899, Copenhagen. Krogh, A.—‘‘ On the Tension of Carbonic Acid in Natural Waters and especially in the Sea.” ‘Meddelelser om Grénland,’ xxvi, Raben, E.—* Uber quantitative Bestimmung von Stickstoffverbind- ungen in Meerwasser,” ‘Wissensch. Meeresunts.,’ Kiei, viii, 1905, pp. 81, 279. “Quantitative Bestimmung der im Meerwasser gelosten Kieselsiiure,” ‘ Wissensch. Meeresunts.,’ Kiel, viii, 1905, pp. 99. 286. Ringer, W. E.—* Die Alkalinitiit des Meereswassers,” ‘ Verh. uit. Rijksinstituut v. h. onderzoek d. zee,’ iv, 1909, Helder. and Klingen, F. I. M. P—‘* Uber die Bestimmung von Stick- stoffverbindungen im Meereswasser,” ‘ Verh. uit. Rijkinstituut v. h, onderzoek d. zee,’ i, 1907-08, Helder. Salm, E.—‘‘ Studie tither Indikatoren,” ‘ Zeitsch. f. Phys. Chemie.,’ lvii, 1906, p. 471. VOL. 55, PART 2,—NEW SERIES. 28 430 48. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. E. J. ALLEN AND FE. W. NELSON. Tornée, H.—‘ On the Carbonic Acid in Sea-water,” ‘ Norwegian North-Atlantic Expdt.’ (1876-78), Chem. IT, p. 24, Christiana, 1880. General. Baldwin, H. B.,and Whipple, G. C.—“ Observed Relations between Oxygen, Carbonic Acid, and Algw Growths in Weequahie Lake, Newark, New Jersey,” ‘Papers and Repts. Amer. Public Health Assoc.,’ xxxii, 1906. Brandt, K.—‘ Beitriige zur Kenntniss der chemischen Zusam- mensetzung des Planktons,” ‘ Wissensch. Meeresunters., Kiel, iii, 1898, p. 43. “Ueber den Stoffwechsel im Meere,” ‘ Wissensch. Meere- sunters.,’ Kiel, iv, 1899, p. 215; vi, 1902, p. 25. “Ueber die Bedeutung der Stickstoffverbindungen fiir die Produktion im Meere,” ‘ Beihefte z. Bot. Centralblatt.,’ xvi, 1904. On the Production and Conditions of Production in the Sea,” ‘ Rapports et Proces- Verbaux. Conseil internat. pour l’expl. de la mer.,’ iii, appdix. D, Copenhagen, 1905. Gran, H. H.—‘‘ Diatomeen,” ‘ Nordisches Plankton Kiel u. Leipzig, 1908, Botanischer Teil, xix. Jérgensen, E.—‘* The Protist Plankton and Diatoms in Bottom Samples,” ‘Hydrographical and Biological Investigations,’ Bergen, 1906. Klebahn, H.—* Ein Uberlick iiber die neuere Diatomeenlitteratur,” ‘Arch. fiir Protistenkunde,’ i, 1902, p. 421. Lemmermann, E.—* Flagellate, Chlorophycex, Cocecosphxrales und Silicoflagellatz,’’ ‘Nordisches Plankton, Kiel u. Leipzig, 1908; Botanischer Teil, xxi. Oltmanns, F.—* Morphologie und Biologie der Algen (with biblio- graphy), two vols., Jena, 1905. Ostenfeld, C. H.—* On the Immigration of Biddulphia sinensis Grev., and its Occurrence in the North Sea during 1903-1907, etc.,” ‘ Meddelelser fra Komm. for Havunderségelser.’- Plankton Series, i, No. 6, Copenhagen, 1908. Pfeffer, W.— The Physiology of Plants.’ English edition, three vols., Oxford, 1900. Smith, G. P. Darnell—‘ On the Oxidation of Ammonia in Sea- water,” ‘ Journ. Mar. Biol. Assoc.,’ N.S. iii, 1893-95, p. 304. Van Heurck, H.—‘ A Treatise on the Diatomacez’ (translated by W. E. Baxter), London, 1896. ARTIFICIAL CULTURE OF MARINE PLANKTON ORGANISMS. 491 58. Vernon, H. M.—“‘The Relations between Marine Animal and Vegetable Life,” ‘Mittheil. Zool. Sta. Neapel,’ xiii, 1898-99, p. 341. 59. Whipple, G. C.—‘‘ Some Experiments on the Growth of Diatoms,” ‘Technology Quarterly,’ ix, 1896, p. 145, Boston. 60. ‘The Microscopy of Drinking Water’ (and bibliography), New York, 1908. 61. and Jackson, D. D.—‘Asterionella, its Biology, its Chemistry, and its Effect on Water-supplies,” ‘Journ. N.E. Waterworks Assn.,’ xiv, 1899, p. 1. 62. and Parker, H. N.—*On the Amount of Oxygen and Carbonic Acid dissolved in Natural Waters, and the effect of these Gases upon the Occurrence of Microscopic Organisms,” ‘Trans. Amer. Micr. Soc.,’ xiii, 1901-02, p. 103. ene 8 gy 7 = Ae : 2 a > Z > 7 . i, : ; ae vit! a he iw . @ 7 . _ Se eet Aen Arai ee ee NOTES ON THE FREB-LIVING NEMATODES. 433 Notes on the Free-Living Nematodes. By F. A. Potts, M.A., Fellow of Trinity Hall, Cambridge, and Demonstrator of Comparative Anatomy in the University. With 11 Text-figures. I._The Hermaphrodite Species. Con'rENTS. PAGE Introduction . 433 Summary of Sexual tehenomena in Bie Heivcapht baite Species . : , é , Pie Systematic Part : 437 Biology of the Soil Nem: stodas in Beet in to the Methods of Experiment : : : . 443 The Males of Hermaphrodite ereeres ; : . 446 Structure and Organisation ; . 446 (2) Proportions of Males in Hermaphrodite eres . 452 (3 ) Sexual Instincts of the Males. : . 461 The Hermaphrodites in Hermaphrodite Species — . 462 (1) The Hermaphrodite Glands in Rhabditis andl Diplogaster ‘ : . 462 (2) The Fertility of the Soil- Newmatodes : . 468 (8) Partial Hermaphroditism ; . A473 (4) The Nature of Hermaphroditism in the Wematoda . AUS (5) Self-fertilisation in Animals 479 Summary of Results . ; : : . 483 INTRODUCTION. Orley divided the Nematoda into three groups, roughly corresponding to differences of habitat found in the phylum. (1) Nematozoaembracing all parasitic forms, (2) Rhabditi- forme which live free in “ decomposing organic substances VOL. 55, PART 3.—NEW SERIES. 29 45 ¥F. A, POTTS. or in earth saturated with such substances” ; and (3) Anguil- lulidee, the rest of the free-living nematodes, found in soil or water. Such a classification, grounded on cecology, pays no attention to the facts of morphology, and is naturally out of place in zoological arrangement, which aims at expressing the relationship of animals by descent. ‘The methods of life of an animal are, moreover, largely ruled by the mode of procuring nutriment which has been adopted. The first two groups of Orley are parasites and saprophytes respectively, but in the Anguillulide we have a heterogeneous collec- tion of forms varying greatly in their habits of life. Little is known of their sources of nourishment save in the case of a very definite division (e.g. Tylenchus, Dorylaimus), which live on the juices of plants, and for that end are provided with a small protrusible spear for piercing tissues and suctorial pharynx for absorbing sap thus set free. The vast majority of this family, however, possess an unarmed buccal cavity ; but in all the muscular pharynx is constantly at work, now dilated, now collapsed, constantly pumping fluid through the alimentary canal. ‘There is 10 morpho- logical distinction to be observed between such a free-living nematode as is found in the mud of a lake or amongst the alow of the marine littoralanda Rhabditisor Diplogaster of the soil. But the latter class can be kept in a culture fluid which swarms with bacteria, in which individuals of the former class would speedily succumb. ‘The tissues of a Rhabditis must be resistant to bacterial action and unharmed by the toxins which such organisms produce, and the worm is, In fact, capable of building up protoplasm from the bacteria themselves or from the products of their action. These are the most prominent physiological characteristics of the soil nematodes, Orley’s Rhabditiforme, and account for the peculiarities of their distribution, for they are apparently absent from dry soils and those with a small admixture of organic matter, and even in soils rich in humus are only detected in quantity by allowing some animal or vegetable substance to putrefy on the sample. Sufficient NOTES ON THE FREE-LIVING NEMATODES. 435 attention has not been paid to the part which nematodes play in the economy of the soil,! but an investigation of this problem may well reveal results of as great interest as those which have been put on record by Maupas, working on the sexual organisation. In the present paper it is proposed to confine attention to the reproductive phenomena in certain hermaphrodite species, but it is hoped in a subsequent research to return to the nutrition and distribution of the class. Cultures of free-living nematodes in connection with this work were first started at the Stazione Zoologica, Naples, in 1906, and continued at intervals in the next two years at the Zoological Laboratory, Cambridge, using for the most part Diplogaster linstowi. In 1909 I spent July to September at the Sutton Broad Laboratory, Norfolk, and procured from the neighbourhood the two forms, Rhabditis gurneyi and Diplogaster maupasi, the study of which enables me to amplify in one or two particulars Maupas’ account of the free-living hermaphrodite species of nematodes. I wish here to express my sense of the value of the opportunities for research afforded by the Sutton Broad Laboratory, and to thank Mr. Robert Gurney for his great kindness to me while working there. SUMMARY OF SexuaL PHENOMENA IN THE HERMAPHRODITE SPECIES. Guido Schneider, in his ‘ Monographie der Nematoden’ (1866), first discovered and put beyond doubt the existence of sell-fertilismgy hermaphrodite species of free-living uematodes. ' The importance of the protozoan fauna of soil has but recently been realised. Like that of the nematodes their nutrition is composed of bacteria, and the place they take as a limiting factor in the increase of nitrifying forms has the closest possible bearing on the fertility of the soils they inhabit. It is, however, probable that these protozoa are more widely distributed in soil and so exercise a more important influence. (See E. J. Russell and H. B. Hutchinson, ‘Journ. Agric. Sci.,’ vol. iii, 1909, * The Effect of Partial Sterilisation of Soil in the Production of Plant Food,” especially p. 141.) 436 F, ‘A. POTTS. In 1900 Maupas,! in a brilliant paper, drew attention to many striking features in the reproductive phenomena of such species. A full description of all prior work relating to hermaphroditism in the Nematoda is given by Maupas, and I shall here content myself with a short resumé of his own results, which later will be quoted more in extenso in connection with my own observations. The species of the free-living nematodes Rhabditis and Diplogaster fall into one or other of three categories : (1) Bisexual species, in which male and female individuals are produced in equal numbers. (2) Hermaphrodite species, in which, besides the self- fertilising protandrous hermaphrodites which form the great mass of the species, there are occasional male individuals, perfectly developed apparently, but taking no part in repro- duction. (3) Parthenogenetic species, in which males have not been found. It is reasonably supposed that each hermaphrodite species is derived from a bisexual form by the development of spermatozoa in the ovary of the female individuals, which thus become self-fertilising. The males are now useless, and have even to a large extent lost their sexual instinct. Their number dwindles in most cases to an almost imperceptible figure, but final disappearance does not appear to be reached in any species, and this persistence of apparently useless forms is one of the most curious facts recorded in biology. The hermaphrodite species appear even more numerous than the bisexual. There is, indeed, some evidence that the con- version of females to hermaphrodites in the bisexual species is a present-day process, furnished by the examples of partial hermaphroditism described by Maupas. An intermediate condition is shown in some hermaphrodite species by the occasional occurrence of pure females, or in the production of 1 &. Maupas, “ Modes et Formes de Reproduction des Nématodes,” ‘Arch. de Zool. Exp. et Gen., Sér. 3, it. 8, 1900, pp. 463-624, Pls. XVI-XXVI. NOTES ON THE FREE-LIVING NEMATODES. 437 spermatozoa in one half of the genital gland only, the other producing eggs alone. Maupas emphasises the significant fact that these species with an incipient hermaphroditism yield the highest proportion of males he was able to chronicle, This conclusion that the more complete development of hermaphroditism and the suppression of the male sex neces- sarily proceed closely together is discussed further below. It is also highly characteristic of the hermaphrodite species in general that the sperm each individual produces only suffices for the fertilisation of a limited number of eggs, so that the period of fertility is followed by one even more prolonged, during which unfertilised eggs are laid, which do not develop. Such a phenomenon marks the hermaphroditism of the free-living nematodes as a character comparatively recently acquired and as yet not shaped by natural selection in anything like its final form. Finally, a most interesting result was obtained by experi- ments with hermaphrodites which had exhausted their stock of spermatozoa and supplemental males of the same species. In the rare occasions in which fecundation took place the egos which were afterwards laid produced males and females in equal numbers.! Systematic Parr. Diplogaster M. Schultze. This genus includes representatives both from soil and fresh water. But while the former possess a weakly developed bursa, which indicates the relationship of the genus to Rhabditis, the latter are without this character, and this fact, according to Bitschli, affords a natural distinction between the classes. * A preliminary note published in 1908 (‘Sexual Phenomena in the Free-living Nematodes,’ F, A. Potts, ‘Proc, Camb. Phil. Soc.,’ vol. xiv, Pt. IV, pp. 373-5) gave a general confirmation to Maupas’ results, founded on observations on Diplogaster linstowi which was kept in cultures for over a year and then died out. 438 F. A. POTTS. The soil-nematodes belonging to this genus differ widely among themselves, particularly in respect of such definite characters as the number and arrangement of the papille on the tail of the male. The typical number is nine or ten pairs, but D. gracilis Biitschli and others have eight, and D. robustus Maupas, eleven. The arrangement of the papille is more variable than their number, but in a small group of species, with which | am more specially coucerned here, the relative positions are fairly constant and characteristic. The arrangement of the papilla follows the scheme given below. ‘The numbers correspond to those given in the various diagrams (see Text-fig. 4). (1) A pair of papille opposite the anterior end of the copulatory spicules. D. robustus Maupas possesses au extra pair, situated far in front of the spicules. In D. mau- pasi sp. n., as a frequent variation one of this pair may have been shifted forward to a markedly pre-spicular position. (2, 3) Two pairs of papilla opposite the posterior end of the copulatory spicules. In D. robustus Maupas shows three pairs in this position. (4) One pair slightly post-spicular. (5, 6) Two pairs, the anterior situated about half-way between the root of the tail and the anus, and the posterior at the root of the tail. (7-9) Three small pairs at the root of the tail, more ventral than the last-named. Since, then, there is so much similarity between the members of the group, the species are best distinguished by differences in size, proportions and biology, to which they are remark- ably constant. Common Characters of the Group.—Buccal cavity surrounded by lips with short sete. Within two! chitinous teeth. Vulva situated in middle of body.” ‘Some species of Diplogaster, for instance D. fictor Bastian, possess only one buceal tooth. 2D. gracilis Bitschli has a ‘“ monohysterous” 2 organ with the vulva a short distance anterior to the anus. NOTES ON THE FREE-LIVING NEMATODES. 439 Male with bursa and nine (in one case eleven) pairs of papilla arranged in manner described above. Spicules -slender, with accessory piece. Synopsis of Group. (1) Bursa with nine pairs of papille: D. longicauda Claus. Bisexual species. Length of 2 1000-1500 nw; cesopha- gus fairly long (one sixth to one seventh of whole length) ; tail long (one third to one fourth of whole length). Germany. D. linstowi sp.n. Hermaphrodite species. Length of hermaphrodite 17604; cesophagus short (one ninth of whole length) and tail short (one-seventh). Oviparous at first, but soon became viviparous. Naples. D. maupasi sp.n. Hermaphrodite species. Length of hermaphrodite 1024-1252 4; cesophagus (one seventh to one eighth of whole length), tail short (one sixth to one seventh). Oviparous throughout life ; 150-3500 fertile eggs always laid at early stage of cleavage, and then about as many unfertilised egos. Norfolk Broads. . (2) Bursa with eleven pairs of papillae. D. robustus Maupas. Hermaphrodite species. Length of hermaphrodite 2488 wu; cesophagus short (one ninth body length) ; tail very short (one ninth body length). First oviparous, then vivi- parous, after laying 150-230 fertile eggs. In addition to the summary diagnosis above the following characters are distinctive of the two new species. Diplogaster maupasi sp.n. (Text-figs. 1, 4, 5, 6, 8). Typical measurements of old ¢: Head to end of Total Head to ; SN Anus to Leneth ; second bulb of as oe length. vulva. tail. of ege. cesophagus. Py C08 Gy, 12 ni). Wen @) F6u Buccal cavity small, with three indistinct lips, each with a slender seta, often distinguished with difficulty. Herma- phrodite at first lays eggs at long intervals, more frequently later. Males often fairly common. Spicules short, slender, 44.0 PF, Ai POTTS, and almost colourless; accessory piece small, in lateral view generally a right-angled triangle, but frequent departures from this type by the rounding of the angles. Number and arrangement of the bursal papille strikingly variable. Abiup-caueayites, alp im ff 7] I. dors. vent. It was at first thought that the shape of the buccal cavity was distinctive of species. The accompanying diagram of D.maupasi shows how greatly the state of contraction of the mouth affects the buceal cavity. TEXT-FIG. 2. D. linstowi sp.n. (Text-fig. 2). NOTES ON THE FREE-LIVING NEMATODES. 44,1 Typical measurements of old ¢: Head to end of Total Head to 5) 5) ; Anus to tech alan second bulb of tail = * cesophagus. rs 1760 pe 840 pe (4) 200 w (4) 240 we (+) Buceal cavity large, as broad as deep, with six papillar lips, each with a slight seta not easily seen. Males with long and slender copulatory spicules and stout accessory piece, elongated and pointed distally (contrast triangular piece of D. maupasi). Rhabditis Dujardin. (eh. gurneyi sp.n. (Text-figs. 9, 10). Measurements: Head to end ot Neat second bulb of ; tail. cesophagus. Length. Head Be 5 vulva. Old herma- phrodite 14564 709 (3) 243 u (4) 149, (4-35) Diaguosis.—Hermaphrodite rather long and slender, tail short. Lips of buccal cavity indistinct, with very minute sete; buccal cavity narrow and deep. First division of cesophagus thick. Vulva median. Hermaphrodite gland with alternating production of spermatozoa and ova. Sperma- tozoa of large size. Number of fertile eggs laid up to 800. Male unknown; probably never produced. Locality.—In peaty soil, Longmoor Point, Sutton Broad, Norfolk. (2) R. sechellensis, sp.n. (Text-fig. 3). Measurements . Leneth Head to me a Anus to ore vulva. g : tail. cesophagus. Old herma- phrodite 680u 3844u(4) 128u(2) 120p (4-4) A male measured 496 in total length. Diagnosis.—Small Rhabditis of pale, transparent 442 F. A. POTTS. appearance. Lips of buccal cavity indistinct, surmounted by minute sete, only made out with greatestcare. Buccal cavity narrow and deep. ‘l'ail of moderate length. In herma- phrodite vulva median. Number of eggs produced small TEXT-FIG. 3. ace piece (150 or less), mother dying before exhaustion of sperma- tozoa. Males rare, inert. Copulatory spicules short and thick, accessory spicule small and inconspicuous. Bursa supported by nine rays, arranged as in Text-fig. 3. Locality.—Found in. moss from Seychelles; brought back by Professor J. Stanley Gardiner. NOTES ON THE FREE-LIVING NEMATODES. 44.3 Briotocy in Repnarion to Merruops or EXPERIMENT. To obtain soil-nematodes in large quantities, it is only necessary to place scraps of flesh on samples of rich soil or mould kept moist and warm, and wait till decay has set in. Though the normal nutriment of these animals is presumably associated with the decay of vegetable products rather than decomposing animal matter, the latter prove exceptionally attractive. When once putridity commences, five or six days more suffice for the appearance of very large numbers of rhabdites or diplogasters, generally belonging to one or two species. Before, however, the last remains have vanished, it is probable that other species will have appeared and become dominant, entirely replacing the first kinds, so that an alternation is obtained somewhat similar to the succession of Protozoa in putrefying broth. It seems that the soil contains scattered throughout it numerous encysted larve, for, as Maupas has pointed out, when insufficient nutriment is supplied to soil-nematodes, the young larve envelop themselves in a thick cuticle, and become rigid and immobile. They are capable, however, of violent contortions, as if for the purpose of freeing themselves from the cyst, and by these movements migrate easily through the soil. The cuticular protection enables them to live uninjured in a dry environment, so that soil, etc., which has been subjected for long periods to fairly high temperatures, will yet yield large numbers of nematodes when treated in the way described above. The power of encystment, and consequently of resisting prolonged desiccation, is confined to the larve. Adult worms at once die when a liquid culture in which they are contained is allowed to dry up, and the eggs of these forms are provided only with a thin cuticular envelope, aud are incapable of resisting the vicissitudes to which the eges of parasitic forms like Ascaris are successfully exposed. When, then, animal-matter putrefies on a sample of soil, it is the encysted larvee which are attracted to its neighbourhood, where they emerge from their cysts and commence to feed 4.4.4, EF. A. POTS. and grow rapidly. The rate of increase is very great: a single individual when once it has become mature will in five or six days give rise to one or two hundred, the eldest of which will be beginning to lay eggs. But a short interval then elapses between the migration of encysted larvee toward the putrid meat and the appearance of the swarms of young worms of the second generation. It is perfectly easy to keep free-living nematodes in drops of a nutrient fluid, and observe under the microscope every stage of their growth and reproduction. Each of these drop- cultures is contained in a solid watch-glass and secured against evaporation by a vaselined glass cover. Solutions of peptone were adopted as convenient culture media, and used almost exclusively in these experiments. The solutions were first allowed to putrefy till a cloudy growth of bacteria had developed throughout the liquid. So favourable an environ- ment for growth does a peptone solution in this condition constitute, that in four days the eggs laid by a mature hermaphrodite nematode have themselves produced mature individuals. It is only in the presence of great numbers of bacteria, or the substances formed by them, that the nema- todes thrive so well. In sterile solutions growth is suspended, and eggs are only laid at long intervals, for apparently nematodes find it difficult or impossible to assimilate peptones in an unaltered condition. It has not been discovered whether digestion takes place by the secretion of juices dissolving the protoplasm of the bacteria, or is merely confined to the absorption of soluble substances present in the culture fluid and prepared by the action of bacteria. If the second alternative be correct, then « parallel is established with the parasitic nematodes which nourish themselves on the dissolved and broken-down food of their host. An easily observable phenomenon of nematodes in culture is the rapid pumping action of the second cesophageal bulb and the rectum, and it may be argued from this that the nutriment obtained from the stream of fluid so constantly passing through the alimen- tary canal is in the form of easily abstracted soluble substances. NOTES ON THE FREE-LIVING NEMATODES. 445 The insignificant development of glandular cells (which are found only in the cesophagus) may be cited against an intra-intestinal digestion of the bacteria, and whatever else its significance may be, the chitinous layer which lines the alimentary canal throughout must prevent an ingestion of bacteria by the endoderm cells themselves in such a way as Colpidium preys upon the bacteria of the soil. Besides peptone solutions other culture media have been used in the course of experiment. It was found possible to raise two or three successive generations in a saturated solu- tion of gelatin in water, and free-living nematodes matured from the egg in solutions of amides like tyrosin and leucin, but in these cases the growth was so much retarded and the production of fertile eggs so curtailed that oniy peptone solutions were used for extended experiments. The temperature at which the cultures were kept varied from about 18° C. in the summer to 12° C. in the winter, though at one period it fell within three or four degrees of zero. ‘lhe effect of a temperature approaching freezing-point was very marked, and showed itself in the almost entire suspension of growth. Sterility was not induced, but only a very few eges were laid every day. Experiments were also made to find the highest tempera- tures under which life and reproduction could continue. The cultures were placed in a water-bath which could be kept down to 25-30° C. Several individuals of the sixth genera- tion were isolated with the temperature of the bath at 26° C., going up to 28°C. One of these laid forty-three egos on September 8th. By September 11th these had developed into hermaphrodites of mature size, but although they lived for several days and were apparently in a quite healthy con- dition, they never produced mature eggs or spermatozoa. The ovary was distinctly seen with small nuclei, but there was no aggregation of yolk. Changes of this kind occurred in the other cultures. In addition individuals just ready to lay eggs were isolated from the cultures at the temperature of the room and placed 4.4.6 F. A. POTTS. in a bath at 26-28° C. Under these conditions the ovary con- tinued to produce large-yolked eggs, and at first these were fertilised and laid, but after they had completed a few divi- sions they became disorganised. With eggs which later passed from the ovary into the uterus fertilisation did not apparently take effect. No egg-shell was formed, and the uterus became full of an amorphous, yolky mass. It seems, then, that the limits of reproduction lie in Diplo- gaster maupasi between 19° C. and 25° C., though life may be continued at slightly higher temperatures. It was found impossible, however, to keep cultures at a constant temperature of over 30° C. The individual worms became rigid and after a short exposure died. It is seen that the free-living nematodes are most sensitive to increased tempera- ture in the egg stage, when they can hardly endure high summer heat. The adult is also likely to succumb at temperatures which must be common in tropical countries at least. The encysted larve are probably the most resistant stage, and it must be supposed that these animals depend for their existence in periods of exceptional heat to their ‘apacities for survival in this condition. ‘’ae Mates oF HERMAPHRODITE SPECIES. (1) Structure and Organisation. The male sex in Rhabditis and Diplogaster, as in all nematodes, 1s sharply discriminated by the relation of the vas deferens to the alimentary canal, and by the well-defined secondary sexual characters, including a membranous bursa for adhesion to the female during copulation, and an arrange- ment of spicules for insertion into the vulva to facilitate the transference of the spermatozoa. The males of hermaphrodite species occurring in such small numbers, and apparently taking no part in reproduction, might naturally be expected to show some marked signs of degeneracy in organs other than the reproductive system. NOTES ON THE FREE-LIVING NEMATODES. 4.4.7 In the Cirripedes we have another clear case of the successful establishment of hermaphroditism in a group in which the sexes were originally separate. Here, too, in hermaphrodite species there is a survival of the male sex, but the individuals which represent it are so degenerate in form and structure as to be described as little more than a bag of spermatozoa, and so reduced in size as to well merit the title of ‘ dwarf males.” It is, however, a surprising fact that in no particular of structural organisation do the males of hermaphrodite species appear to fall behind those of bisexual nematodes. ‘The conclusions which Maupas reached on this subjects are summed up in the following quotation : fees males . . . noffront rien de particulier et @anormal. On ne remarque rien dans leur structure et dans leur organisation générale qui puisse ies faire considérer comme des animaux mal venus ou mal constitués. Par leur taille, par les proportions de leur corps et par tous les details de leur organisation, ils répondent de tous points au type male ordinaire des Rhabditides dioique. Leur testicule luiméme est constitué dune fagon absolument normale et, ses produits, les spermatozoides, sont palreur forme, leur volume et leur structure absolument identiques a ceux que la glande génitale des femelles produit pendant sa periode d’activité protérandrique.” My own observations show that there is no imperfection of development in the residual wales of such species as I was able to study. The spermatozoa were always produced in vast quantities and exactly like those formed by the hermaphrodites. When liberated by pressure from the body of the male, they could be observed to put out amoeboid processes like those which Ziegler figures taking up their position in the uterus of Diplogaster lougicauda after fertilisation. This observation tends to show that the spermatozoa are physiologically active though the individual which carries them is prevented from playing its part in reproduction, possibly by a defect in nervous organisation, 4.48 ¥F..A, POTTS. The experiments of Maupas with Rhabditis elegans showed that on the rare occasions when males do fertilise hermaphrodities, the spermatozoa are perfectly efficacious in the production of embryos. The curious change in the sex- proportions of the offspring of such unions may, however, be TEXxtT-FIG. 4. eventually traced back to some essential difference in the spermatozoa of males and hermaphrodites respectively, which might be revealed by a thorough examination of the spermato- genesis in the two cases. But though there is no manifest imperfection of organisa- tion in the males of hermaphrodite species, they appear to be NOTES ON THE FREE-LIVING NEMATODES. 449 sometimes distinguished by extreme variability of the secondary sexual characters. In such specific characters as size and proportions of various parts the males are fairly constant, but the arrangement of the papillae supporting the copulatory bursa and the shape of the accessory piece ot the copulatory spicules show wide differences. When Diplogaster maupasi was first obtaimed from various TEXT-FIG. 4. samples of soil round Sutton Broad, the differences existing between the males found in separate cultures made me conclude that I was dealing with a number of nearly related species. It soon became clear that distinct types of male were not characteristic of each culture, but that even brothers from the same family often exhibited wide differences. The typical arrangement of the bursal papillz in Diplo- VOL. 5D, PART 5.—NEW SERIES. 30 450 F. A. POTTS. gastermaupasi is shown in A, T'ext-figure 4, Departure from this type was found, however, in almost every other specimen examined. Below are given some of the clearest cases of variation observed in dealing with a comparatively small number (about forty) of males. (1) There should be normally a pair of papille situated exactly opposite the anterior end of the copulatory spicules. One of the most frequent and easily demonstrated variations occurs when one of the pair (or very rarely both) is shifted forward a smaller or greater distance. So marked a case as fig. c was observed two or three times. (2) A pair of papille (4-4’) occurs a short distance posterior to the anus. Only small variations in position were recorded here, but on one occasion a duplication of the papilla of one side was observed (fig. p). (The papilla of the other side was seen on altering the focus, so it was quite evident that the twin papillee belonged to the same side.) (3) In the position of papille 5 and 6 there is rather frequent variation; they are sometimes nearer together, sometimes further apart. Occasionally it may be seen (when the animal is lying on its back) that the papille of the two sides (5, 5’, and 6, 6’) have a tendency to alternate in position (fig. B shows this, but not very well). An example like fig. B was observed once, in which one of the papillz, either 5 or 6, was duplicated on both sides, and the twin papille then shifted apart. (4) he three small papille at the root of the tail (7-9) are rarely replaced by two. It is only occasionally on examining these animals that a frontal view is obtained, showing the rays of the bursa on both sides. In side views it is often difficult to correctly observe the position of the papillz. On this account only a few definite cases of variation are referred to above. They were observed in dealing with forty to fifty males. The accessory piece of the spicular apparatus varied in form in nearly every individual. Three types are figured. The first shows the most typical, in the shape of a right- NOTES ON THE FREE-LIVING NEMATODES. 451 angled trangle, with an indentation at the anterior angle. In the other two the angles become more and more rounded. In Rhabditis sechellensis variations in the secondary sexual characters are occasionally found, but are much less numerous than in Diplogaster maupasi. Such variability as was observed was manifested in (1) inequality of the copulatory spicules, and (2) occasional asymmetrical dis- position of the rays of the bursa, The only reference to analogous phenomena which occurs in Maupas’ paper is found in his description of Rhabditis. guignardi (p. 525). He obtained only two males, but in one of these the copulatory bursa possessed on each. side nine TEXT-EIG. 9. A. B. C. supporting rays, in the other only seven. In the latter the remaining rays showed a disposition to fuse with each other, a phenomenon, it may be remarked in passing, which was responsible for the asymmetry of the bursal rays in R, sechellensis. The entire disappearance of two rays is a variation as great as any recorded above for Diplogaster maupasi, The position and number of bursal papillee or rays is looked upon as clearly diagnostic of species of Rhabditis or Diplogaster, and as faras I know no striking variation has ever been observed in the bisexual species. The connection of such a variability in the males with their disappearance from the economy of the species is no doubt significant, but it is impossible to offer any explanation of the facts, 4.52 F. A. POYTS. (2) Proportions of Males in Hermaphrodite Species. Another remarkable feature of the males of hermaphrodite species studied by Maupas is their extreme rarity. In only one out of eleven species investigated was he unable to find a male; but in others males were only discovered by organising cultures of very considerable size, containing several thousand mature worms. So while in the majority of species the males were less than 0°1 per cent. of the whole number of adults, the proportion of 4 per cent. to which they rise in Rhabditis marionis affords quite a striking contrast. In Diplogaster maupasi, one of the species obtained from the Norfolk Broads, the ratio of male to female is very much more notable than anything which Maupas records, and does occasionally approach, though remotely, that equality of the sexes which is characteristic of the majority of animal forms. In one large culture the males reached 10 per cent. of the whole (377¢, 38¢ ¢), and in batches of eggs laid by the -same individual up to 30 per cent. (16 eggs, 11¢,53 9; 29 eggs, 234,699). These instances are, of course, specially favourable, and picked from amongst scores of cultures which did not yield a single male. It is very unlikely that a species will be discovered uniformly consisting of equal numbers of males and hermaphrodites. Southern! supposed that in Rhabditis brassice he had discovered such a species, but in a culture with which he kindly supplied me I have been only able to find males and females, but no hermaphrodites. No illustrate the manner of oceurrence of the males, I give here an analysis of cultures of Diplogaster maupasi carried on over twenty-five generations, from August, 1909, to January, 1910. ‘lhe whole series of cultures commenced with a single individual. In every subsequent generation at least one hermaphrodite was isolated just before maturity to carry on the succession. When such an individual had com- menced to lay eggs it was removed every day to another 1 Rowland Southern, ** On the Anatomy and Life-History of Rhab- ditis brassicx n.sp.,” ‘Journ. Econ. Biol.,’ vol iv, 1909, pp. 90-95. NOTES ON THE FREE-LIVING NEMATODES. 453 watch-glass, so that the batch of eggs laid during the pre- ceding twenty-four hours was kept isolated. Hach batch was carefully counted to compare with the actual number of individuals attaining adolescence, and in this way records of cultures which gave the actual sex-proportions were distin- guished from others in which mortality before maturity obscured the true figures. In any drop culture which con- tained more than about thirty eggs the crowding which ensued was distinctly unfavourable to the chances of survival. Precautions were adopted in these experiments to prevent absolutely an association of mature males and hermaphrodites, and so remove any suspicion of cross-fertilisation in the line of descent here followed out. ‘To this end the individual destined to give rise to the next generation was separated before any male had become mature, or else the males them- selves were removed from the culture before the last moult, when they were perfectly recognisable as males, but had not yet assumed the spicular apparatus necessary for internal fertilisation. Both sexes become easily distinguishable a considerable time before maturity by the position of the developing gonad and its duct. In the majority of species of Rhabditis and Diplogaster, the vulva opens at the middle of the body of the female, and the gonad is paired, so that the immature hermaphrodite may be recognised by the symmetrical disposi- tion of the clear ovarian rudiments round the middle point of the body. In the male the rudiment of the testis is situated in the posterior half of the body, so that with a little experi- ence it is easy to distinguish a male, even among a ceaselessly twisting mass of other individuals, by the clear transparent testis running alongside the posterior part of the gut. Sperm-formation begins, it is true, before the last moult. But though the body of the male may contain mature spermatozoa, these can only be conveyed to the hermaphrodite individual by the co-operation of the copulatory spicules and bursa. A young male just before the last moult, at which 454 FA, POTS, these latter are developed, is shown in Text-fig. 6. The proximal part of the vas deferens leading into the cloaca does not appear to be yet fully formed. The cloaca is spacious, and is produced on its dorsal surface into a pair of definite pouches in which the chitinous copulatory spicules are formed at the time of the last moult. TEXT-FIG. 6, ™ intestine ‘he history of the cultures may be divided into alternating periods, which are distinguished respectively by the frequent occurrence of males and their entire absence. During the first six generations, while these experiments were being prosecuted in Norfolk, the percentage of males was often quite high in batches of twenty or thirty eggs, and the off- spring of the majority of individuals contained at least one or two. In addition, the total number of eggs laid by each parent seldom exceeded 130 (150 in one case), and the spermatozoa were not exhausted before death. The seventh and eighth generations were reared away from a laboratory, NOTES ON THE FREE-LIVING NEMATODES. 455 under conditions which made careful recording difficult. On removing the cultures to Cambridge a new kind of peptone! was used for the preparation of a culture-medium, and the behaviour of the nematodes altered considerably with this change. In five generations, from the ninth to the fourteenth, not a single male was produced. ‘The interval elapsing between the arrivals at maturity of successive generations decreased from seven days to four, and the number of fertile egos laid by each parent rose to between 150 and 300. In every case the life of the individual was prolonged under these more favourable (?) conditions, the period of fertile production being succeeded by another at least as long, during which sterile eggs were laid. Later, in the fifteenth generation, the peptone used in Norfolk was again tried, and at once males appeared sparingly in the cultures. Later the individuals raised from certain batches of eggs showed a fairly high ratio (e.g. in the nineteenth generation [25] 19 ¢ 4 ¢ @), but in general males were rarer than in the early cultures of August. After another removal at Christmas, 1909, the second period of male production was terminated like the first. It may well be supposed that the alteration of conditions, slight or other- wise, which ensues on changing the place of experiment was directly responsible for the disappearance of the males. It is not probable, however, that the proportions are controlled by nutrition, for though at first circumstances seemed to indicate that the use for a culture-medium of white peptone acted as a stimulus to male production, from the fifteenth generation onward four series of cultures were maintained, two in white peptone and two in brown (which is the more favourable medium for growth). As mentioned above, males first appeared in the former medium, but in the seventeenth generation they were also observed in brown peptone, and there was no sufficient difference in the figures to suggest which peptone was the better material for the production of males. ' Tn dark brown crystals completely soluble in water. 456 Hy tes OTS) In the second table a fuller analysis of the experiments lasting over the first six generations is given. An attempt was made to isolate strains, constantly producing high proportions of males, by breeding from a large number of individuals of the same generation. Thus in the third genera- tion a batch of 44 eggs produced 32 4 and 12 gg (about 28 per cent.) did not, with one exception, maintain those high proportions. One, however, though giving at first hermaphrodites only, laid a batch of 16 eggs of which 11 became ¢ and 5 ¢ ¢ (31 percent.). Nearly all these herma- phrodites were kept for an examination of their progeny, but five individuals, whose records were kept separate, furnished strikingly retrograde results, though males occurred in every case but one. ‘he male ratio was greater in a culture con- sisting of the offspring of three individuals, reaching 11 per cent. of the whole number. Further selection for the next generation proved equally indecisive. In the third generation a control series was also established by taking sister individuals from a culture in which only hermaphrodites were represented. ‘The total number of off- spring of the five parents selected was 319, of which 302 were $and 17 gg. This is exactly comparable to the total of 262 ¢ and 15 ¢ ¢ produced by the five individuals from a culture with 28 per cent. of males. The individual details are closely similar in the two series. A brief inspection will serve to show how extraordinarily irregular is the distribution of males in the progeny of any single worm. ‘here is no rule that they should appear at stated intervals or restricting their production to a period or periods of maturity, but on the contrary the appearance of a few males from an early batch of eggs may be followed by a succession of hermaphrodites only and vice-versa; the last eggs may produce males when there have been only hemaphrodites hitherto, or, again, males may occur in several successive batches. TABLE 457 Nore. The figures enclosed in circles represent the number of eggs lad tn each batch: those to the right the ndariduals counted on arrwal amaturty or before. 1$*Generation. Offspring of a single isolated hermaphrodite. no males observed. rd a . 26 7” es 32 ¥ gh @ 269 @ Total 7150 6h in © Ve <2 iq Total 127 Re. Z eo SRK , @Z) gh >, GO: Ga) : . a Ct, Go : @ @ : Total ca 2 do". ows Bet 2 ih (aS These first. stc.generations were bred in the Sutton Broad Laboratory, Norfolk for the first generation anwtisiwon of Beef was used. Ahterwards two or three varveties of Peptone Ory, Albumen, Witte’) supplied by HarringtonBros, all of which had substantially the same value as atood stulf- The 10%-719%generations were bred irthe Loological Laboratory at Cambridge. WO: 593d Oe Ss ee MED ee — : 51 ¢208¢. P— : 26 ¢'. G0) : 31 ¥. 48): 37 J. @ +72 ¢. TABLE 1! (Cont?) 11*” Generation. 6): Sy Ge) +: 4a 5 25 ¢ @ — @: «¢ OM: -—- OB: = 182 : 60g WS: Re Ig (8) : 8 ¥ se ee @ wg @) : 3897 OD :uw¢ @ : 77 @ : 23 @:— @: x7 26 9g 198 : BY jae GO): we : ao @ nr¢(@: e¢ QB: 8¢7|\@Q: wg @): If |: we ©: RY|O: B¢ 250: Bg |@%: 4#¢ 74 @ : 4 ¢ \257 : T59¢ @): 67 BO: we G2) : s777¢@) : Beg Qf @: we > — 39g @: 2¢ : 34 TER? Ga) White Peptone. A I53 BrownPeptone. @2) : 2° Brown Peptone. WhatePeptone. gen @) » 2¢ QD: OL OD: WY The one @ : 27 @: w¢ 8 ft 2) u Oae @o) : a 0) : 50 ¥ entirely sterile. : 70 36 £187 @) 52¢ Gi) car et @ = 36 £ 2) : 30 SO. Be eee 285: Ble WE: WII : Woes TABLE 1} (Cont?) Brown Peptone. White Peptone. BrownPeptone. White Pepione . 16*Generation G6): 249 GB: 359307@) : 249 @ : 6 Vikan 18h kag 2? 6): 29 @:8f DW: #H9 &: —. @ : 54$ GA:19f 188 OD: — Pri ext Ore Qi wi @2¢ QO: @ 8 3 249: 162¢ G6: Be 2e 1G: 234:1309485 292 122 f 27 138 § @:i 2B — Be OQ: — @Qisr¢g @ Te we BD: —. @) eee GD :39 ¢ 7 6) Bg es oe oe 4 G) cs ear WI 9 QO > See Gess® 7 @Q : 239 IW5SIBIIMZ 4 F (254: Wag oat ae G9) :36 ¢§ GD: af @:4f @:age2se (‘Dr ws Cte ws GB) = @B:v¢g @:n¢g @):90 ¢ ‘ORs @ : 2g1ds@s): — GO:% 1odd2 23 G2SS @: ng @& :28~187@:2 ¢ 150 M2 g71s3@) +12 ¢ W7:97FTIS 126 : 96 G1 38 @: 97 @O: we OM: Be Q:iwssse @2): 0 ¢¥ G): 15 ¢ @:e¢ @Q: 6 ¢ iv ee QA :a3gidd @: 39 G:9¢ G8 : 74g 6): s1¢ @): 8 ¢ @) : 13 f73é 136: Weg QD: 2 Ff 145 1378683 199:169 § TABLE 2, 460 26 ¥ 2dd' BS f 2 bid ond ion 2 ee ia Calture of reg! hetniap iron produced os bia 377 $38 i3 inthis generation. Bier Ee GO): 36 YY: 32 f12dd gth @9): 23 ¢68082):30924 ve 2343 29G3):3. gf 2¢¢@) WBS LB) 25 C6): 22g3IKED):22 f(D): 23 f " @3):__. Gd):30¢ 336 @9):14 61 SA @9) 134250 Ga)9 f2t8@3):34¢ (A): 74 G3): 209202 Ga):54¢ @): __. @) 7 AON g586 a Se “haa RE B24 Ft (8:17 F700 A): 37 g Que “@: 19 ¢ (3 hermaphrodites gave G3): 32 g7 0 @3):37 f2¢¢@2): 41 ¢ @i):19 g7 dS 9): 27 G2 S¢ 183 f 2bSS) / @):30¢ Gé:a6g G2): 37 ¢ G6): 26¢ GA): 37¢ Ga): 34% 6): 43 gage (—o): __ G5): 44 @2):16 g1a8 ae ae J aba LEE! WBE :127$538132: 15 F lg Vy g/ FS130: 127 $2 IL ” 4): 20 ¢ OO): 55 ¢38898): 12 G1 od : Qs): 24 gids GD: 56 g1de WE E237 NOTES ON THE FREE-LIVING NEMATODES. 461 Sexual Instincts of the Males. Maupas’ conclusion that the residual males could not take any part in the production of offspring is expressed in the following words: “ Mais si ces animaux examinés dans leur structure et leur morphologie, représentent des males vrais et complets, il n’en plus de méme lorsqu’on les étudie au point de vue de leurs facultés et de leur activité sexuelles . . . ces males ont a peu prés totalement perdu tout instinct et tout appétit sexuels. . . . Nous trouvousen présence d’une decadence psychique non concomitante avec une regres- sion morphologique.” This conclusion is supported by the inert behaviour of the males, the fact that they are never seen in copulation with hermaphrodites, but principally by the results of a fairly full series of experiments which Maupas made with males and hermaphrodites which had exhausted their own stock of spermatozoa. These conclusively showed that the males have almost, but not quite, lost their sexual instinct. One species alone stands as an exception. In Rhabditis marionis at various times cultures containing in the aggregate 28 herma- phrodites and 42 males were kept under observation. Since all the spermatozoa of the hermaphrodites were exhausted, any production of developing eggs was plainly due to the inter- vention of the male, and thus a measure of the activity of this sex was obtamed. Fertile eggs were laid by 15 individuals to the total number of 150-200, and all these produced hermaphrodites. This species is one of those for which Maupas described a partially developed hermaphroditism, and the author himself regarded it as specially significant that in such a form the male should be less degenerate. The most’complete series of experiments was made with Rhabditis elegans. Here, in twelve cultures, a total of 159 hermaphrodites with their own sperm exhausted and males were associated. Only six of the hermaphrodites were actually fertilised, a proportion which illustrates exceedingly 462 KF. A. POTTS. well the sexual inactivity of the males. ‘The chief point of interest lies in the constitution of the offspring of these six individuals. The young produced numbered 274, and of these 147 were hermaphrodites and 127 males. So numerical equality of the sexes is secured in this species by cross- fertilisation, a result in striking contrast to that obtained when R. marionis was the subject of investigation. No permanent effect was produced on the heredity of sex, for when 38 of the hermaphrodites obtained by fertilisation by males were em- ployed as parents for the next generation, 2964 individuals were produced, of which only 7 were males, but the rest hermaphrodites. Further evidence of the psychical decadence of the males was secured in other species. Though nearly 100 males were employed belonging to five species only a single successful case of re-fecundation was observed, and in this (Rhabditis duthiersi) the fertilised eggs gave 70 hermaphrodites and 1 male. ‘HE HERMAPHRODITES IN HERMAPHRODITE SPECIES. (1) The Hermaphrodite Glands in Rhabditis and Diplogaster. In Rhabditis sechellensis the structure and develop- ment of the reproductive glands exactly correspond to the description which Maupas gives of R. elegans and R. dolichura. ‘hough no new details can be given, it will be convenient to summarise the changes which the hermaphrodite gland goes through before oviposition commences in any of the above three species. The three diagrams which illustrate the description are partly after my own drawings for R. sechellensis, but closely follow Maupas’ sketches of R. dolichura in Plate XX1I, figs. 7a, 7B, and c, The hermaphrodite organ is double, its two divisions being of equal development, and joining at the short and indefinite common vagina. Each division is U-shaped, and consists of a uterus, which extends from the vagina to within a short NOTES ON 'THE FREE-LIVING NEMATODES. 463 distance of the bend of the tube, and an ovo-testis, occupy- ing the proximal part of the ventral limb and the whole ot the dorsal limb. In individuals examined some hours betore the first egg is laid the whole of the ovo-testis appears to consist of cellular elements of nearly equal size, which possess definite boundaries near the bend, but merge intoa syncytium distally. The anterior testicular region is indicated by the more regular polygonal form of a comparatively narrow belt of spermatocytes which succeed the uterus. ‘The young egg- = TEXT-FIG. 7. yy . Fair amount; constant. 6. Green E i é é - | Fair amount; constant. 2. Green ‘ ; ‘ : . | Traces in 36 % of larvee examined. 2. Green : : : é . | Absent. 1. Almost colourless (pale pink). Absent. These results are of interestin several ways. They confirm on the whole the earlier conclusion that excessively red colour in the parent is associated generally with excess in the early larvee; but they also show that the offspring of green female varians show two types of coloration, namely, that with some red pigment and that with none; and further that the two THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 557 types may be combined in a single brood in the proportion of 36 per cent. dominant or pigmented forms. This result at once suggests that green in the parent is of -twofold origin; and the facts of earlier experiments support the suggestion. It has been shown (Gamble and Keeble) that green is both an independent stable colour-form aud also a colour assumed by brown specimens on a transfer to green weed. Further experiments are necessary to decide whether the green parents with recessive red colouring are of the former type, and those with more dominant red pigment are of the latter colour-history. ‘lhe new points that emerge are the absence of red pigment in certain broods, and its presence in only a percentage of others. Repeated attempts were made to experiment with broods from an isolated parent under diverse conditions of light, food, and temperature, but without much success after the first week or ten days. ‘The chief .results obtained were (1) that zoeze developed and hatched in darkness (from brown parents that became green under these conditions) possess the normal pigmentation, thus showing that light is not essential to pigment development, and also confirming the suggestion just made, that it is those green parents which had been previously brown that give rise to larve with red pig- ment; and (2) that there is a steady increase in the amount of red pigment in broods of green parents. For example, the tint of zcee of green parents approximated after a few days to the colouring of the larve of red parents. It is, there- fore, doubtful whether the initial differences in pigmentation between the broods of similarly or diversely coloured parents are of any moment in determining the ultimate coloration. Tit. (1) Tse Invtuence or Transmrvrep Monocuromatic Ligur ON THE Formation or Pigments 1x Hiprotyre VARIANS. Previous work on the influence of monochromatic light (1900, p. 619, 1904, p. 356) upon Crustacea concerned itself 558 F, W. GAMBLE. chiefly with short exposures made upon an absorbing or re- flecting background. ‘lhe results showed that the light acted irrespective of its colour according to the nature of the back- ground, almost as though it were white light of low intensity. Moreover, experiments with coloured backgrounds of weed, against which young, transparent,almost colourless Hip polyte were exposed to direct sunlight, showed (1905, see Tables) that in two days, eleven out of fifteen prawns became red on red weed, and eight out of twelve became green on green weed. he coloured backgrounds, when flooded with white light, produced sympathetic colouring. The red was a mixture of red and yellow, the former predominating, the green a mixture of the same two pigments but yellow pre- dominating. In both cases a diffuse blue pigment occurs also. This result appeared to lend some support to the view of Wiener (1895) (which has since undergone elaboration [ Bachmetjew, 1903] ), and to suggest that the dominant rays of the background evoked especially that pigment or that group which agreed in colour with the reflected light. In order to ascertain more fully the effect of monochromatic light, I determined to eliminate, as far as possible, this dominant influence of background, and to ascertain the result of exposure to incident light of one colour, So far as I am aware, the experiment in this form has not hitherto been undertaken. ‘The starting-point for this experiment was furnished by young transparent Hippolyte varians taken by netting over Zostera beds and Laminaria-fringes, These fall into two groups: typical faintly red-lined forms provided with red and yellow chromatophores along the gut and nerve-cord, and with red ones at segmental intervals in the integument; and more uniformly coloured specimens with similar pigments, but with chromatophores more evenly distributed. In both cases the amount of pigment is not enough to give the specimens a decided tinge. They are similar to those used for the weed background experiments quoted above, and are figured on PI. 23, fig. 4. The vessels which were used are described above (p. 554), THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION,. 559 and the conditions of the experiment were such as to flood the animals with monochromatic light on all sides, The weed chiefly used for food was the natural food-plant, Ceramium; a little fine green weed was used in one of the red light experiments. . The vessel was surrounded on three sides by the fluid colour screen and rested on a faintly re- flecting surface, so there was no strong background effect. The light employed was direct and diffused sunlight, and the effects of heat and of ultra-violet rays were largely obviated by the conditions of the experiment. The results of the experiment are given in Table III, pp. 577, 578, and show that whilst the Hippolyte, in white light, developed into brown forms containing both red and yellow pigment in about equal proportions, those in red light passed through a brown stage, but ultimately (three weeks) becaine green, some remaining, however, reddish-yellow in 1909, whilst the survivors in green light became bright carmine, In other words the ultimate colour in this experiment is the complement of that of the incident light. The details of the end-result show clearly that the green Hippolyte produced in red light and the crimson Hippo- lyte produced in green light are peculiar and distinctive. The former possess yellow pigment in a maximally expanded state, and such little red as they possess is of a vermilion tint. Moreover, the yellow is of a distinctly greenish tinge and is accompanied by either very little diffuse blue or none. Thus the green colour in these experimental specimens under red hight is largely due to an increase in the amount and quality ot the yellow pigment accompanied by contraction of the formerly dominant red pigment (PI. 23, figs. 6 and 9), The crimson Hippolyte produced in green light is no less distinctive (Pl. 23, fig. 5). In contrast to the usual type of red forms, the yellow pigment has completely disappeared and the chromatophores are entirely filled with a deep carmine pigment suffused with a bluish tinge. The general deep carmine colour was new to me. Moreover, the chromatophores on the surface of the eye-stalks were abnormally developed. 560 Ir. W. GAMBLE. In view of this very decided complementary colour-change the regrettable mortality that occurred in vessels exposed to green light in 1908 does not seriously diminish the value of this result, though larger numbers would add to its cogency. These were obtained in 1909. The experiments of 1908 and 1909 are compared on Table V and with the other experiments of this paper on ‘lable LV. (2) The Influence of White Backgrounds and of Monochromatic Backgrounds in White Light. The effect of short exposures to the influence of white (porcelain) and of black (cloth or paint on glass) backgrounds on the colouring of young and old Hippolyte has been fully treated ina previous memoir.' It was there shown that what- ever the quality or quantity of the hght employed (within the experimental limits), the background effect dominated, pro- ‘ducing contraction if: white and expansion if black. It occurred to me, however, to see whether the same results would follow a long exposure made with young specimens in which the pigments were rapidly developing. The results of a month’s trial are of considerable interest. The Hippolyte on black surfaces simply followed the usual procedure under such conditions, and deveioped maximal amounts of red and yellow pigments, which gave them a deep reddish tint. On the white background, however, after a first phase of transparency, they began to develop red pigment along the nerve-cord, and finally became uniformly marked with a ventral red stripe, whilst over the rest of the body the pigments were reduced to microscopic dots or disappeared. This remarkably adaptive result was obtained in diffuse light, the top of the deep porcelain vessel being covered with muslin, through which a stream of water was maintained from a tank above (Table IV, p. 579). In 1909 these background effects were extended so as to include the results of red and green. ‘Ihe vessels employed were large museum jars, painted, except for a large rectangular 1 (1904), p. 353. THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION, 561 window, with several coats of pure paint. Spectroscopic tests showed that the red was pure and that the green paint reflected only a trace of blue in addition to the whole of the green light. These vessels were kept under a water-circula- tion and faced a south light. Finely divided pieces of Ceramium were employed as food. . The Hippolyte used were smail, almost colourless specimens, similar to those employed in the other experiments on coloured light. The results of exposure to these monochromatic back- grounds was very decisive (Table IV, p. 579). Uponthe green one the development of pigment was arrested. ‘The Hippo- lyte assumed a semi-nocturnal (green) tint, and remained with the red pigment contracted throughout the experiment. This green colour is, however, not retained if the background is changed. Under these circumstances the animals revert to the pale red-lined colour variety which they exhibited initially. Upon the red background, on the other hand, the red and yellow pigments had considerably developed, and after a month’s exposure gave a bright orange-red tint to the specimens, and this persisted after change to other back- grounds. It would be of interest to know whether Minckie- wicz (1907-8), who has also obtained results of this kind with Hippolyte, tested the permanent or transient nature of the induced colouring. IV. THe Foop or Hrerotyre as A Possiste Source or PIGMENT’. The relation of Hippolyte varians to the alge of its choice is a distinctive one. The peculiar features of this species, the range and cryptic character of its variable colora- tion, its choice of, and tenacity of hold upon its weed, its distribution, and its food are all bound up with the presence of these plants. It is possible that Idothea and some Amphipods are equally intimately related to their habitat, but among macrurous Decapods Hippolyte varians is probably unique in this dependence upon its algal environ- ment, 562 F. W. GAMBLE. In former papers on the subject, the relation existing between the pigments of Hippolyte and the coloration of its surroundings was explained as due to light effects, as if the weed backgrounds in virtue of their disposition, of their luminous character, and colour, acted as stimuli to the chroma- tophores of the prawn. However, before we accept that explanation, the influence of two other factors must be con- sidered: First, the effect of darkness on pigment-formation, and second, the source of these pigments, whether derivative or not. The first factor—darkness—is discussed on pp. 577— 579, and it is there shown that the red (vermilion) pigment does not require the stimulus of light for its development, and that it increases in amount when the Hippolyte are kept in darkness. ‘The yellow pigment, however, is more dependent on light for its formation and increase, diminishing in amount in specimens kept in darkness, especially if little or no food is supplied to them. ‘The crimson pigment and the diffuse blue colouring matter are not at present investigated from this point of view. There is evidence, however, that light is essential to the production of all varieties of Hippolyte, except the reddish-brown ones. ‘The other factor—the source of pigment itself—is less known than are the conditions which determine each particular tint. The colouring matter of the food is one possible source, and this has to be briefly con- sidered, since, if proved, it would simplify the problem of sympathetic coloration. ‘That the sub-hypodermal colours of caterpillars and beetle larvee are due to diffusion of fatty pigments from the food-contents of the gut is a conclusion reached both by Poulton and Towers, though the physio- logical details of this remarkable process have never been ascertained. But the hypodermal colours of these animals are of an entirely different nature from those of Hippolyte, and appear to be determined by enzymes, elaborated by this layer acting upon the ‘primary ” cuticle or retained within the hypodermal cells. In Hippolyte and in Crustacea generally (as in the insect larve), the first formed pigments are developed independently of the plant-food present in the THE RELATION BETWEEN LIGHT AND PIGMEN'T-FORMATION. 563 mother, and it would be of great interest to know how they were formed. In order to test the influence of food-pigments on the development of pigment in Hippolyte, the following experi- ment was carried out at Millport, N.B. A series of double olass vessels were prepared, the Hippolyte being placed in the inner chamber and a mass.of weed in the outer one. ‘I'wo series of pressure-bottles, one in diffused light, the other in darkness, were set up for isolated specimens. The food employed was chosen from the following: The natural alga chopped up into fine pieces so as not to act as a massive back- ground ; etiolated Laminaria, also subdivided ; the muscle of Hyas, the colourless ovary of Hyas, and the scarlet, mature ovary of the same crab. The specimens of Hippolyte em- ployed were 5-7 mm. in length, colourless, and tending on a black background to assume a faint brown-lined colour pattern. Taste A.—Feeding Experiment. Colour of Hippolyte after exposure to contrasted colours in food and surroundings, Millport, 1909. Colourless foods employed are crabs’ muscle, etiolated Laminaria, and colourless ovary of Hyas. The Hippolyte used were from 6-8 mm. long and colourless to the naked eye. Experiment lasted seven to ten days. Colours of foods employed. Colour and nature of surroundings. 1 a Red (scarlet a Colourless. Biases Brown (alge). | Green (alge). | ; | aah Darkness . - | Pale brown-| Reddish- — Reddish or lined brown colourless, | 1 green. Green alge - | Green Green 2 green, =: f 1 brown Red alge . 1 = uals — — — ; Greyish Brown alge = .5 | ea — -- — | 1 grey Parti-coloured | . A ere J 5, 4 . yellowish oilcloth |B lack-lined | Brown-lined | Brown-lined | Brown-lined. | \ 564 F. W. GAMBLE. ‘he results are shown on Table A, and at once bring out the fact that colourless muscle, white or red, ova are greedily taken up, but that the background is the dominating factor in the resultant coloration in daylight. Thus against a back- ground of green weed Hippolyte fed with colourless food, with red ovary, and with fine brown weed became green. In darkness, however, the amount of pigment in the food has a rough relation to the resulting colouring that will need further experimental testing, but there is no good evidence that the colour of the food determines that of the prawn. V. ANALYSIS OF THE CoLoureD LicHr EXpeRIMEN’s. (1) In green light, and amongst red weed, Hippolyte develops crimson and deep, not superficial, colouring. ‘The presence in the experimental vessels of a fair quantity of finely branched red weed (Ceramium) would, under the action of diffused, strong green light act as a black back- ground, and this, as we know, in the presence of white light, encourages the formation of vermilion and yellow pigments, and these are most notably absent. ‘he crimson effect in green light cannot, therefore, be merely due to dim light acting on a dark background. It must be due to a distinctive factor not present in the other experiments, and that factor can only be the green rays. In the presence of these rays, not only is the crimson pigment developed, but the vermilion and yellow pigments are dis- missed. Whether a similar result would follow if a colour- less food were used is of course a subject for further research. The most striking feature of this crimson colouring obtained during exposure to green light, is the fact that it is comple- mentary in colour to that of the incident hght. This relation may have a considerable significance. In an earlier paper (1905) it was pointed out that strongly insolated Hippolyte showed mobile fat in their chromatophores, and as this fat disappeared in specimens transferred to darkness there was some ground for the inference that the production of this THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 565 fat was associated with the presence of light. If that were so the assumption of a complementary colouring would be obviously the best means of absorbing the maximum amount of coloured light, and of obtaining any other benefit which light might confer upon metabolism. Under the conditions of deep water, where the green or green-blue rays have filtered down from the surface, such a colouring would be the most efficient absorbing pigmentation, and it is well known that in hauls made from the deeper water of the English Channel the Hippolyte are uniformly of a crimson colour. The facts as to these crimson Hippolyte produced in green light would be most comprehensively explained by saying that the red Ceramium acted merely as an excitement to coloration, but that the carmine pigmentation is produced under the direct stimulus of the green light employed. 2) Green weed . Green coloration . Superficial and deep. ep Red weed . Yellow or brownish- chromatophores. Red light yellow The action of red light is less easily analysed. The constant effect associated with it, is the production of yellow pigment and the maximal expansion of that pigment into networks producing a grand colour. ‘Then, according to the absorbent or reflecting nature of the background (i.e. green weed or red weed), we have a green or a brownish tint, in the latter case accompanied by a development of scarlet chromato- phores both at the surface and along the lines of the alimentary tract and of the nerve-cord. In the case of red light, therefore, it would seem that the direct action of the rays lies in the production of yellow pigment, and that the nature of the background, indirectly modified by the further action of the red lheht, modifies this yellow coloration less or more. If the background be red, the action of the rays is intensified, and a red background is thus instituted. Probably this is the factor that gives the scarlet chromatophores, for, as will be seen subsequently, that is the effect of a red background in white light; the resultant colour is then brownish-yellow ; but where, as in VOL. 55, PART 3.—NEW SERIES. 37 566 TY. W. GAMBLE. the case of green weed, the background is of a less Iuminous character, the red colour contracts in the Hippolyte and the resultant coloration is then green, owing, in some cases, to the presence of diffuse blue mingling with the yellow network, and in the longest experiment to an apparent change in the pigment from yellow to green. ‘The most im- portant and most clear influence of red light, however, is the spread of the yellow pigment. These results are so strikingly dissonant from those obtained by subjecting Hippolyte to green or to red backgrounds that an explanation is clearly called for. They differ not only in being totally opposed to the sympathetic colouring so charac- teristic of the latter, but also in being slowly acquired. It may fairly be asked, if red light reflected from red surround- ings gives red Hippolyte, why does red light diffused give green or yellow ones? The same contradictory relation obtains between the action of green surroundings and diffused green light. In answer to this objection attention may be drawn to the double nature of the light affecting Hippolyte under natural conditions. here is the light reflected from the background and there is also the general diffuse light. The rapid sympathetic background colour-relations obtained experimentally have been made in strong daylight, and as the depth of water is increased or as the red end of the spectrum is cut off the conditions of the experiment are materially altered. A strongly colonred background becomes black in every light except that of its own colour, and in the presence of it we should expect the usual black background effect (brown, i.e. red and yellow pigments) to be produced in Hippolyte in any light except that with which it agreed in colour. But whilst this background effect is an undoubted factor, its potency is determined by another factor, namely, the definite action of diffused monochromatic light. The action of many rays has yet to be determined, but from the foregoing account a case has been made out for the action of green and of red light. This action, though slow, is very THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 567 precise, and it would certainly help to account for the crimson and yellow colouring found in deep-waterand shallow- water Hippolyte respectively. The results, then, of these two factors, the action of diffused coloured light and that of backgrounds in white or mono- chromatic light, are not contradictory. They are the two factors which, so far as we yet know, are associated in the production of pigmentation in Hippolyte. The green speci- mens on Zostera are green, not only because they are ona green background in bright or fairly bright light, but because at or near the surface of the sea the red rays are miost potent, and their action is to produce that network of yellow pigment in Hippolyte, which is the basis not only of green tints but of those yeilowish tints that this animal assumes on the etiolated parts of Zostera, and of the brown specimens on various brown weeds so characteristic of the Plymouth littoral flora. The diffuse red hght, on penetrating to more densely absorb- ing backgrounds, such as coarser brown weeds, is checked in its action upon Hippolyte by the tendency for such back- grounds to produce red pigment in them. Hence the absence in such cases of that more precise colour-relation to the inci- dent light. The brown pigmentation contains many red and yellow chromatophores, but the red is scarlet and not the crimson of the deeper zones. Passing out of the range of the action of the red rays, the characteristic zone of the Floridez is encountered, and it is in this zone that the green rays are more potent. Their effect in producing crimson pigmentation is seen in parti-coloured specimens of the red-lined variety and in occasional pink specimens of the Laminarian zone, but it is not until a fair depth is encountered that their action is made clear by the dominance of this peculiar carmine pigment, which has hitherto been confused with the vermilion or scarlet one under the confusing term “ red.” No doubt there are similar effects of yellow, orange, and blue rays to be analysed before a full analysis of the coloration of Hippolyte can be given. The main conclusion derived from these experiments is that 568 i. W. GAMBLK. Influence of Light on the Colours of Lepidopterous Pupe. (After Poulton, Petersen, etc.)! Light. None Red (pure) Red (red and some yellow) Yellow (red to green) Green (pure), green glass Green (some red, yellow, and green) Blue. (General absorption least | in blue ; some red, yellow green and blue rays are transmitted) Colour of background. Red Orange Yellow Light green Dark green Blue White Light wood Orange White | Light wood Dark _ Light | | ; Dark Plain wood White Dark Green White Red Orange Blue Light | Dark Spectrum of back- ground, | | Red Red to yellow Red to green Red to green (red | to yellow largely absorbed in some | experiments) General absorp- | tion least in green | General absorp- tion least in blue ee Resulting colouring. Vanessalo.|Pierisrape. Irregular (dark and | light) Darkest) - | l| @ (green) Very light (green) Light green ) Dark (brown) Dark Light (green) Ditto Light (green) Unknown Green Green Light (green) Darkish (V.urtice)| Darkish Ditto 2 light, 3 dark Dark Ditto Ditto Dark (Poulton), Light Green Light (green) eterson) 1 The references to these papers are given fully in Bachmetjew’s quoted on p. 582. See especially ‘ Trans. Entomol. Soc. London,’ 1892. P.brassice. iI Dark | green. Green. Darker. work THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 9069 the pigments developed in Hippolyte, when kept in diffused monochromatic light, are not the same as those which appear in specimens kept in daylight on a background reflecting these rays. Ona red background in white light, Hippolyte becomes reddish-orange; in pure red light it becomes yellowish or green. On a green background in white light Hippolyte becomes pale green. In pure green light it becomes crimson. On backgrounds of weeds, young colour- less specimens speedily acquire the corresponding tint. Mono- chromatic light, then, when saturated, has an entirely different effect from the same light diluted with daylight. As we pass from the surface to the deeper waters of the sea this dilution becomes less marked. The “ background effect,” so potent in producing the more littoral colour varieties, becomes less overwhelming as the red and yellow rays are absorbed by the surface waters. Further down, in British coastal waters, the blue end of the spectrum is said to be absorbed, so that at eight fathoms the dominant light rays are greenish or bluish-green (Oltmanns!). Consequently the effect of satu- rated monochromatic light is most probably felt in the region below the eight-fathom line. If this distinction between the effects of coloured back- grounds in white light and of diffused monochromatic light on pigment production is well founded, it should be supported by analogous results in other animals. Fortunately the work by Poulton and others upon Lepidopterous pup give a closely comparative result. As will be seen from the appended table extracted from their papers, the effects of monochromatic light are very different according as to whether the dominant rays are or are not diluted by white light. Although these experiments have not been made with a view to excluding background results so completely as those given in this paper, yet the distinction between the effect of red light, for example, when concentrated and when diluted, is quite analogous in the case of larval pigmentation in insects to its effect on pigment-production in Crustacea. As a pure 1 * Jahrbuch. Wiss. Botanik.,’ 1892, p. 420. SA) F; W. GAMBLE. concentrated light, both red and green rays act like orange- yellow ones in suppressing pigment. When diluted, however, with white light,red rays produce pigment and pure green rays do likewise. As a background in daylight, therefore, the monochromatic rays act in one way ; as a pure incident light they act in an‘opposite fashion. ‘This apparently contradictory result is therefore supported by the evidence from experiments on two widely different groups of animals, Crustacea and Insecta. What exactly, then, are the factors that determine the extraordinary close sympathetic colour-rendering of the environment in the pigmentation of these animals? First of all in both groups, light is not essential to the production of pigment. Poultou’s results, as well as my own, show that dark-kept animals become dark coloured, though somewhat irregularly. In the case of Hippolyte darkness does not induce the formation of all the pigments. Red (vermilion), the dominant one, and yellow to a less extent (giving a brown coloration), are the only colours formed in the absence of light. In the insect larve, brown pigment is likewise formed in darkness, and develops as a sheath upon the green sub- epidermal layer. The action of light, then, in both groups is ‘ather directive or inhibitory than effective. In the case of insects, the orange-yellow rays are apparently those which, when reflected from backgrounds, inhibit this brown pigment and allow the subjacent green pigment to confer its full value on the colour of the larva or pupa. In Crustacea the case is different; the action of these rays upon them is at present quite unknown. ‘The colours are pigmentary, con- tained in chromatophores and not “hypodermal” as in insects, but the production of the well-known green, brown, and reddish varieties of Hippolyte is due mainly to manipu- lations of a reddish-yellow coloration which is formed in the absence of definite stimulation. The light reflected from natural aigal backgrounds is of a mixed character, but with some yellow, some green, and varying amounts of red in it. All we have to imagine is THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 571 that in the production of a green Hippolyte on Ulva the yellow and blue pigments are encouraged, the red discouraged. We do know that this effect occurs in pure red light, but in this case few red rays are reflected. We are driven to the conclusion that in daylight the yellow of Ulva directs the expansion and development of yellow pigment, and the green the expansion and development of blue pigment. In other words, we have here Wiener’s effect or conclusion confirmed. But when the water deepens, the red (vermilion) pigment, no longer inhibited by hght rays, develops more strongly, and yellow and brown, and even blackish, Hippolyte occur in response to the diffused background of brown weeds, the hght from which contains chiefly red and yellow-green rays. At this depth the incident light has lost some of its red and yellow rays, and is of a more bluish-green colour. From this depth onwards the action of diffused light becomes more and more apparent, that of the background less so. In the dominantly green water the crimson and diffuse blue pigments of Hippolyte develop to the exclusion and repression of the red and yellow ones, thus giving the various shades of carmine, purple and violet, that characterise Hippolyte taken in deeper water and in deep, shady crevices near the shore. Ina greater depth than that to which light extends, Hippolyte varians is not found. Indeed, it does not appear to extend beyond the range of some ten fathoms. In deeper water the genus is represented by Spirontocaris, the colour problems of which have not yet been investigated. If we accept this conclusion, that carmine, purple, violet, are colour effects, related directly to the diffuse green light in which many animals of deeper water live, an explana- tion may be found for the prevalence of these colours in many other groups. For example, carmine is a tint acquired by some fish, Crustacea, many echinids, starfish, and corals. Violet or purple is an even more characteristic pigment ot the deep-sea fauna. This purplish tint is complementary to green, and the relation has given rise to much speculation, but, so far as I am aware, the above experiments with or2 F., W. GAMBLE. Hippolyte give the first indication that the purplish colour is actually developed in a few weeks when the animal is exposed to green light. The significance of the scarlet colouring, so characteristic of abyssal Crustacea and of certain more shallow-water forms, e.g. Hemimysis lamorng, is still obscure, but the obser- vations made above as to the development of red (vermilion) pigment in young specimens kept in darkness may throw some light upon the subject. With regard to Hippolyte varians, the facts so far ascertained are these : The red pigment is the first to appear. It arises in the larva, even if this is reared in darkness, and the amount at the time of hatching is roughly proportional to that in the mother. In adolescent specimens subjected to darkness the scarlet pigment increases in amount. VI. Summary or REsutts. Crenilabrus melops. (L) The colouring of young specimens is due in part to the blue endo-skeleton and in part to chromatophores. (2) On backgrounds of weeds these fish assume varied coloration. On brown weed they become brown, on green weeds green, on red weed green. (3) In light transmitted through weeds, Crenilabrus assumes a colour, the complement of that which is most strongly represented in the incident light. ‘Thus, in light mainly green, a brownish red colour (due largely to red pigment) develops. In light mainly red, a green colour (due largelv to yellow pigment) develops. Hippolyte varians. (1) In any brood the amount of larval pigment (which is always red) is constant, and is correlated with the amount of red pigment present in the female parent in all colour- varieties except green. (2) A given green Hippolyte throws one of three kinds THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 573 of young; red, colourless, or a mixed brood, containing red and colourless individuals in the proportion of nearly 3: 1. (3) This result suggests what is probable on other grounds —that green Hippolyte are of two, and possibly of three kinds: (1) Brown forms that have become green; (2) green forms that have undergone no change of colour; and (3) a cross between these two. In the absence of knowledge ot the male parentage of the broods, the last suggestion needs confirmation. (4) Light is not essential to the production of red pigment in the larva. Darkness does not prevent the continued pro- duction of red pigment in young forms. (5) The action of monochromatic light upon the pigment- formation of Hippolyte is entirely different from that of a monochromatic background in white hght. (6) In pure red light, yellow pigment develops. In some cases this leads to a green coloration: in others the colour remains yellow. (7) In green light a carmine pigment is produced, and any red or yeliow pigment existing in the experimental batch is either destroyed or disappears almost completely. (8) On a red background in white light, Hippolyte becomes reddish-orange. (9) Ona green background in white light, Hippolyte be comes green, but the colour is not retained if the batch is transferred to an absorbing dark background. (10) Continued exposure to daylight and a white back- ground produces hypertrophy of the red pigment along the nerve-cord and a disappearance of the red and yellow pigment elsewhere. (11) The production of sympathetic colouring in the shallower zones of the coast is explained as a background effect, in which the incident diffused lhght plays little part. The influence of background is predominant. The prod uction of crimson colouring in deeper water is explained as due to diffused green light. (12) There is no evidence that the pigments of the food (algze) are the sources of the pigments of Hippolyte. AMBLE. ri W.G F. ) ~ e Suol | -xe Wo 90U0 Ye SuTysNy | ‘UM FZ puv ‘TZ | kq dn poaoys quq “qureyz ‘ST ‘poppe suowto | 4say ge spurgq ‘uooad | | Y90% ‘aSuryo on | -ods popuvq-oped g | jo oSury yLH YstuMoag | -- — — ysneny oovjans [erjzuoA WO YST WOOL TWMOTG Curt TI- | JIM popueq |OT) poppe suo UMOIG | TAAIH JFYSraq spleq poyreut-[pomM | [Joa pormoyoo | -to0ds popuuq | YET -ystppot spre, Smee) “qjgep ouo ATUG yaIM TM OA - YSU) : -yunp Aaa [py | UMorg ATPUIETE) GsLoy ‘(qeT wo poppe ody potieqg Apurey peusn jo F) “pappe poo (qa¢T (IST Jsu.sny wo poy [ep UMOAG-Ystp | Jsusny uo peppy | peppy atom F) “Woot -por YJIM ouo TTeUs T | aOM F) “pomreq -YystUMoAq Zi ‘Woot WO9LG YUOL Url | ‘uoato §=6ATQOULSIP [LV | -UMOIG T “Wood Z JO oSury YIM UMOIG Z — UMOLET -vdsirty o[vg | qsnony *poarireq A[IUTBT [[B SSOSSRIM F ‘up ApouIeayxo !190LT 07 pat | “past iter | ‘snommoads patrmq APUIEy F | {aurey ie esos M & vuSeee aban ar ae suommods g *37uR4) gnd strouroads be “116 atrp ATA “AOI WULTLA “aur ere aetna es od Nita Maeda pee ga on “Spey OL! i qs fenimo"ebes a ae Ree BeOS Supe HOM YES oO Tbe pate an a ae e 4 cakant etek ap pee Lass |\—"pogooaTsod Fy sty | B YALM poprtaoad o10M S[assoA TTIIM ILRI OY RIS He ie Sea Sh Ee Se iu SUT “paom UMOAG 1OF : : : : yO wo puvavl| ssvpo avo fel | guow | sR ISUBIIY “BASU pue “poomM UMOAG OJ VRTYI OF TIO “PAM ICM pos LY awel-[paq V—‘SSouy1ecy Bnitewenne b aaa aia ee ails ¢ydo4arn) incite 9 ALOIS JUOUTOS URL WU POSMOUIUAT SUM SLIT puR (Lod I ce ta out U1 ante ae hearts ae oY TN) POM PON | _s(eat A) poom uaoty) ‘aul tmosnur ve ut paord aon = a wer ; i x Pa are ow | | SOSSBIM OIL —"(RUTIBTIOUS “IMAI aia }R ECO PINUS | BILBVULMLBT) POOM WMO | “STOO NT pue spodiydmy yt pez otam Ao, “punog oy} Ul purjsy seyxVaq punosl opty MOT Fv Uday a.10M sosswim ounod asol{y, ‘“alayMosfa URY} SUOLSaA OSat UL S1oquINU JayeaIs UL payesoisse soroyd -OFLWMOLYI Youlq puv pos OF onp 9.1B savq asaaAsuRly UMOIG OYJ, ‘So1oydoqemowyo MO[[Ox FO ¥.1OM -jaM VB HNO] Laas Sttiaq (Sapeos yuadsozedo of} 07 os[e Ayqavd pure) uoyo[aysopus onjq of) 07 pur skul-ug eq} Suisnyns quowsid onpq oqy 0} onp udots yuotedsurcy-1iuos “pup B st Inofod punoss ey, ‘(FFG ‘d uo uorjditosop pur Z sy) uoretofoo jo ody paasreg Ajurey [ensn oy Fo a0 Aoy, “Ypousl seuladjxe Ul UL ZT PUB “WU OQ] SULSRADAR ‘SAZIS OA\Y JO OAOM OSSVAIM SLY} FO SMauTtOads at I, OGL “YQnowdé|q ‘spunoas -youg SNOIIVA WOAT pOZOOTFor yoy :]T guowmodxy ‘sdOTAN SAUAVIINGNAD “OSsVI,A—'|[ ATAVY, N'T'-FORMATION. 575 7. By sHT AND PIGM EEN LI¢ THE RELATION BETW “quout -std mopped Jo WoryvUt -10F oonput sAvt pat Aygueurumtop ey 4eqy MOYS 0} sutees 4[NSet aed Appoqyoodxoun sty], “ATV LO qos Quoutsid poy ‘popuvdxe pur padoy -oAop TOM Apes {tout -o1d MOTTOX “spurred qUIvE ATA Woado MOT[OA SpuUNoALS-taats uo potreq-WMo0.aq Ayquiey Zz “puno.s -yoeq yovyq wo yng ‘ot “‘spueq UMOIG ITM “USTIIdLy ‘popuvdxe Ajtaerodute} saxtoyd -OJVULOAYD por ‘ spueq YStuUMorg, pamoys ‘toumoaq ATJOUTAST(T SoUO DATPOTO ayy Ayquered -de sAvi o5ue10 -pat oy, punoas -yoeq pot Fo yyy ayIyUN jou 4[Nsey spurq WMO TTA MoTTAA Z “YAO -jou MOT[AA poos pomoys worneurut -exo dIdOdsO.0Or TAT “MOT[OA - YSTtto90.a.5 Z punoroyovg yoelq wo Aang potreg-WMo.aq Zz ‘UMOIG ST -“Md0ls T “Waal T /UMOIG G “WooLs T pepueq -UMOIG & "WAIL T asuvyo ON pUNoASyoV WERT ITT | ov 07 SMlees Poo UMOA poo Waa.ts 10 pat YP yuorttod -xo UT UeYy yuoutoId pot AMO “PpUNoOLS YSTILIALS UO UMOIG TYITM pa.red Tk “TOTSURAX: | -TT9AA worsuvdxa oinstea OF punodrs -oVd YOVL uo Jug aor) ON YSTWoa.1.5 Zz ‘anopoo-Apoq wWee.to (uo spuvq jury ygIm Z IMopoo puno.xs (TST -ue0ds) roped rayyet Z ‘papurq-[Tom pure yarep g —aouURYD ON ayAjoddryy jo aseyd-anojoo [eumyoou | OF apqvaeduoo | ‘satoydozvut OLD FO WOTPORAY “OD OUT. XA | | | | UMOIG | Jo a0ury YIM | ONL “UsTTg quomdsuety, (q48zZ JSnony / wo worsuvdxea Aaetoduta J, ) ‘soroydoq |-BUTOATLD JO WOTy -OBAIPTOD OUTA.TY “XW “ony gue /-edsuviy Ato A (yQ9ZISony Ut qnd ouoy ‘urut | O&-0Z suout |-toods MMO. Q) | ysy [ey uourLtedxe UTP ILE gk BIBS) -yarq “puno.s | -youq yortq Fo qooyfo [Busn ot], SIV ALeF FTA onTq juored -SUB.Iy “peo out oUt 0491 popurq [Tom pure yap Agoa [Ty -edsuvay opeg | ysusny | pund.azcyorq OFIYM JO qooyo Teusn oy, | | | | | | | (peom | weeds) JuoUt |-1todxo Toyjour, | OF patty -SUBIY MOTLOTOM asoy, “OUT Wao FUuat soouo -LofUuy 106 4ysn.ony ILE 4ysn.ony W4S6 PUS qsnony ISTE 576 F. W. GAMBLE. Taste II.—Experiment II. Creniiasrus metors. Light transmitted through Weed. A. Green weed transmitting | 3 soee orange to green and a little | er a Le 00d te red hight.—T wo rectangular aaaition a % aa “bere f vessels fitted one inside the». Brown weed trans- red licht, Bite Date.— | other, with a space of lem. miutting orange and a Tae ‘taclon een D 2 a August |between the two. This space trace of redand green, nd Mito ‘i ites 23rd, was filled with Ulva (3-4 light of low intensity.— an facainy a M9 a 1908. fronds). ‘The inner con- Similar receptacles = te rs a 2 ieee tained the fish and was = with Laminaria. Sreenieh; 4 See Without weed. 3 specimens, 17 mm., faint barred; 1 ask, ? | 8-10 mm., greenish ; 2 speci-| ae 10 ee brown mens, 8-10 mm., brownish. it | August 2green,3 brown-barred. 5 green specimens 2 greenish; 3 faint- 26th (1 small green exa- | barred. mined: — pale blue pigment diffused roundthe gut. Yellow | pigment well ex- panded) August 1 green, 3 pale grey- | All green = 27th brown barred August All brown, consider- Ditto 3 green (2 with tinge 31st able amount of red of brown). pigment in 3 Infer-| Conversion of green The amount of Under the infiu- ence colour to brown by light was pro- ence of red light development of red! bablyinsufficient the complementary pigment. The effect) todo more than colour is retained of orange-green light | act as dimness, or developed. favours this change which produces a greenish color- ation TWEEN LIGHT AND PIGMENT-FORMATION, 577 1 vy THE RELATION BE “SSoTINOTOO Z@ iO [ ‘swqoy peult + uwMOAq-YsTp -pod IIB} IO poosgg | (yggTgsusny wo. 7) “U48SE qysnony wo praq | ‘peutt-poet q{urey |z ‘“peuly-pet poos z ‘4uotedsursy Ata A | | | SOTRT |B sypeqs oAn (ua8T qsnony wo peqyirys | OZ jo Jo[T «aoygo | |-Uy) ‘peur, pet [Ty inchicys) queustd pat oy punore peyreut [pea onjq “jues -qe quourctd Molla x ‘UOSUULIO «=PIATA ATOA eee -tadns puv deep yy40q | pedoyaaop-T[eM “yueur -std wosmi1 aand B Jo JIB satoydozeut qysty por UI osoyy URYy pea aSMzUL 910 — “yoo yo “pod ALey T ‘pat poos G ‘pert [NF ‘soroydoqzeut -oayo daop pure [eIoy _-aeadns ut mopped pur pet youyT “qs. ZU OT-O7FIp OF VpTUITS SUILLOF Paty por ystmozjoad @ Spurd |p ‘pour-pex Atoa [TY Z ‘pour, por peordAq g Ystusea.6 Y40q “Jol Z soTey aN] ALA purw poqzorary -0N—pey ‘pesva.ro “ul youtm ‘ynzryue,d puv pepurdxe TOA —'MOT[OX ‘(per outos | pue Moyjed Yonur) yst | OID ‘OUTUTIV ATESIAG T | SULMOF pouUT[-uMOAq 9 | -MOT[OA T *(ULLOF poury | -WeeI8 [ “Weds IVT T WdeIs POoOs Z) Wda.L0 F popurd -xo pur podorasop yon yuourstd Moyo ‘moo. - yStMoTToA ‘sumqOJ PoUT[-Ystppea 9 ULIOF PoUl][-pod JUIVF T “ULAOF pouUry -Ystumorq [ ‘Sotoydozyeutorys ocuv10 ddvzns aoe] [T “poptoood ota atoyy Worsuedxe WNUWTXeUL UIL{GO 07 FYSTL WIP Ul punotoyoug youyTq 0} otnsodxe 41oys vB to4yyy ‘A[UO pot T ‘mopfe owos pur pet ITM z ‘satoydozeutorya MOT[OA puV pat TZIM CG “W409 FOvIq YIM palaaoy "poof oF [TINTUERIIQ “pezyRpNoaTd TOBA TOL T.SnoOIy4 ‘rel tanesnur SurmL1eyu09 YURy ayv[s y—"ssouyrvqd *yUR9 UL SUIpURIg *posuyip pus yo11q@—" I svyt ‘aqRuoaryd “4od JO 9dB.14 TIE OL) Jo (ory “WLd G.[) WOTNTOS "411990 zed 09 B Aq poyeiedas saaxzvaq OMY, —'YIeT usny —"4y.sTT W8eLH ‘O o€ ZT-LT YUV] UT surpuryg 10 JOO —"IYHvT “pooy roy WUtIMVvAIY “You wd G.T ulsoryyAsd *0°d QOF AG poyeredoas siaxyRoq 9.1] OM J, —TYIET ISNSNY— "FSI poy “qSI[UNS pesnytp “qILorpuns pesuyiq—Iysvy “suo, ‘uu 8-9 suauttoads pauiy-ared OT 1190z OY UO poppe BY daz0m -OLOIUM OFT Y “pooyx IOJFUINTMBAIID “Lassa IOUUT UT UOLYBTNIATO 194BM -yURyT, ‘WoAMJaqY WISsOItyy Ao jo LOART “UO ZB YIM soystp SSBLS ODAIVl Z—"4UST poy | SUMOF poulT-UMOIG | -op-]JomM = sjzuoursrd ames [AIM soroyd -oyeutoayo = daa ‘govduavo m0 sa.coyd -ogeutoayo «= MOTTO pur port acaey Ato A podoyoa | | | | (ese INN. ‘939T § | yutd oped T ‘SuUIIO} poul[-pot cG A | *q.SI[UNS posnuyip puByDotIp + 4SvT *.5U0T | “mu QT-J Suatateds pouty | -aywd § “pooy 1oOf poam UMOIG JUG SuMsypR pur UIN[T} Bp«royH ‘yuey a1 YO YStuseds [[Mp Wo sur -pUuUBIS LOYROC SSBLS TBATO 0°09 008 UWY— IAT] ONT MA 206] ‘SNVINVA HLATOddIWR ‘JUOWMIIedxyY 4qoly poinojog—]]] Wavy, WP ‘qdag W966 qsnony 1406 4qsnony YI6L 4gsnony | WALT qsnony “78 qsnsony —"ayu sAMBLE. EY yea xD “pt0o -aAIOU PUL JS punos sesseut = Aysnq = ut MOTLOA PUL WOLPTULAL A. ‘sotoydozeutoayo oovVyz “INS MAT “WAOF pouly -umoaq doaqy “Ava SUIALS PIS Surzpod “SULLOJ POULL- Por JFULLT lO aez [TV “SULLOF POUL[-Pot PULLF i “SULLOF POUT] 0} SUIMO 4JoT T ALO '4yeL ZI ‘quoredsuet oped Ano 4 “pot poytvul-]poa [Ly *poqiwgs cz InNoqy Pat) WULUER19:) “AagVAL JO VOTPRI DOL JUBISUOL) “YURI OBIS UL Pesopoua PUB YS Suypoq WLM pataAoo wef umasnyy ouNE peWweEIg—ssouyaug “pooy a0y (pao “pag pAod-OAOU PUL JG oy} punot quourord OUTULIWO JO SosSvUT qUooSatogau — osMA(T soto dozvutoay [eIoy -radns oyg Jo seovry popurdxe A[png “OTRULOATLOIG, “god yo sontg + 8 Ou Jo ATU “UOSUILLD FILS & aaXBl “ud | BAG poguandes Soy BOG OMT —"IUSIL Uae) JF MoToq pure Apo eyy Jo aoujAns oy to YOR pedopoaop [pon soaoyd -OFWULO.LYD —- PaO Oo UOTPLULIO A YOM JOU Bw Suro; quoursid MOTTO “yurd T ‘pouty -pot T ‘mop yst “UMOLY [ SPOUL-UMOAG, T (4y9LI—pue oun?) JULOIT OPTI Ut Spooa pod WO poul[-pot AULD = PRL YOTIpAL poy ‘SULIOF pout[-por g stoUtLoeds YALM popteyy Salrguir eat) UMIYIVT = “FULorL pay PUOULS TL WOT[LUOA poepuvdxe — pony ‘quounstd = Moppoa popyepoorjot — You, “SMOT[OAP *YSTUOOLS | SSvT smmopoo T ‘ysppurd z ‘sour mopped YF -YSTUMOUd ZT. LOMB NOA1D-ALy “LoyBoq LAU punosre OULULIBO LUNIA JO MOAR “UL ZW “4ULoI] poy potoqyle oouEgsq ns OULPOT “SULLOF oped ZT -up MOT[oA Ol YOU] poutp-pot Oy “eZ “[q uo uMOTS | ad.44 og Jo Suouttdeds 67 qQnoqy “poeM UMOAG OUT] “UOLPR[NOALO ALY “Layne 9°02 OOSL V—"FU51T OU M. QOL Ayne YLT oun U9ST oune WIOL oun “6061 ‘W948 ouny —'97Bq WULULB IIL) ‘pooy toy (paaa. paul) "606 ‘SNVIUVA ULATOddI}R—'(penutqyuod) [J] Alavy, ‘(ecg ‘d aas) oAOqe WOT SU [JAM SB MO[OG MOAT poloqzUo FYSIT SpEssoA osoyy [[v UT Taste IV (1908-9).—Summary of Influence of Light on the Starting point for these experiments: Small specimens (43-54 transparent colourless (on black background), and red | THE RELATION BETWEEN LIGH T AND PIGMENT-FORMATION. 579 Development of Pigments in HipponyrE varians. mm. long), faintly red-lined (on white background). chromatophores present, but not in sufficient quantity to give rise to a definite colour (Pl. 25, fig. 4). Weed (for Length of uniform Yellow chromatophores Light. Background. food). pepe ‘Resulting colour. eae ee & | | Re ( 1% ‘Reddish-brown Both yellow and red : edand Z None (Darkness) | },own 2 Red-lined increased, espe- 5 6 | Brown-lined cially in the deeper | layers. Surface pig- ments disappear. Avoided by Ditto 13-8 That of weed | Surface and deep | uniform | pigments well de- ee iation | | veloped. | f 1 | Almost colour-| Surface pigments White | Ditto |- less absent. aka | 2 | Deep crimson | Deep crimson on gut below | and nerve-cord. | Green weed | Green 2 days Green in 66 | Evenly distributed. | | per cent. Red weed Red 2days | Red in 80 per Deeppigments better | | cent. | developed. Glass resting Red and) ( 2 | Reddish | Red disappeared, | | on slate then |; 4 |Greenandyel-| yellow developed, Red | green \ _ lowish green | blue developed. ee Avoided by Red | 4d Yellow and | Red present, yellow | uniform | greenish developed, blue in illumination | one. Glass resting, Red 3 Crimson Carmine and blue on ivory- | | only. “Green . glazed brick ee | : Avoided by | Ditto As Ditto | Carmine. Surface | | illumination have almost dis- appeared. 580 Kf. Wie GAMBLE, Taste V.—Summary of Results showing the Colouring obtained by subjecting Young, almost Colour- less, Hippotyre varirans to Diffused Transmitted Monochromatic Light. Red light. Period in weeks. rae So... 1908. 1 Reddish 2 Reddish and brown 3 Green | and yellow 4. Greenish Food Ceramium and (latterly) a little fine green weed Final pig- | Much yellow. ‘mentation of) Trace of red chromato- | (vermilion). | phores. Much blue or | (Figs. 8-9) green See figs. 5-7. Green light. 1909. Brownish- yellow. Brownish-red Yellow and greenish Ceramium Much yellow. Some red (ver- milion). Blue in one (greenish) specimen 1908, Red Carmine (pure ) Ceramium Carmine abun- Blue dant. (fair). No yellow. No vermilion Carmine. Ceramium | Carmine. Trace yellow. | THE RELATION BETWEEN LIGHT AND PIGMENT-FORMATION. 581 Taste VI.—Showing Hffect o f Background on the Development of Pigments in young HippoLyre VARIANS. 1908. 1909. | White porcelain vessels with an air- circulation. Fine red weed used for food. One vessel covered with a sheet of green glass and two layers of Baker’s green gelatine giving pure green light. T:me in White light. ee weeks. | White background. eround. it Allremained very faint red-lined parent faint forms. red-lined | | forms; trans- ferred to white light 2 Superficial chro- — matophores had disappeared. Deep’ carmine ones clus- tered round the | gut and nerve-cord \(see Pl. 23, fig.11). | Specimens | appeared trans- | parent witha narrow crimson line down the | centre Infer- | Remarkably pro- | On reflecting | ences | tective develop- | backgrounds, ment of crimson green hight pigment in bright inhibits forma. | white light tion of pig- ments when employed for | a short time. VOL, 09, PART 3,—NE W SERIES. 20. Pale trans- Colourless (7), Colourless (2), 2 fect. It merely _ Large museum jars painted with | several coats of flatting, a clear) space being left in the front. The red flatting reflected red light only, the green flatting reflected green lightand atrace of bme Finered weed was used for food. Water- circulation employed. 25specimens. — — | White light. Green back- ground. White light. Red background. reddish (4), red (4), ereenish (4) ereenish (4). Colourless or faint ereenish (8) Orange (7). | deft; all faint| Bright ereen, but re- reddish-orange | verting to ned (2). lined forms on exposure to dark back- ground | Green light | Red light suf- suffused with | fused with bright white white light has) light has no adefinite effect, distinctive encouraging the develop- ment of red and yellow pigments. acts like dim white heht 582 1894. 1904. 1905. 1906. 1907. 1909. 1909. F. W, GAMBLE. LITERATURE, Noé, J., and Dissard, A.—‘‘ Déterminisme de ’homochromie chez les poissons,” ‘C.R., Soc. de Biol. de Paris,’ sér. 9, vol. v, pp- 100-101. _ Wiener, O.—‘ Wiedemann’s Annalen,’ vol. lv, p. 225. _ Holt, E. W. L.—‘‘ An Observation of the Colour-changes of a Wrasse, Labrus maculatus, Donovan,” * Journ. Marine Biol. Assoc.,’ N.S., vol. iv, pp. 193-194. . Nagel, W. A.“ Ueber fliissige Strahlenfilter,” ‘ Biol. Centralbl.,’ 1898, p. 654. . Gamble, F. W.,and Keeble, F.— Hippolyte varians: a Study in Colour-change,” ‘ Quart. Journ. Micr. Soc.,’ vol. 43, pp. 589-698. . Bachmetjew, P.—‘ Experimentelle Entomologische Studien,” Leipzig. (Gives full list of literature on experiments with lepi- dopterous larvee and pupx, and also a summary of these results and of the bearing of Wiener’s conclusions upon them.) _ Gourret P.— Monograph on the Labride Annales Museum, Marseilles,’ 1893, plate iv. Gamble, F. W., and Keeble, F.—* Colour-physiology of Higher Crustacea,” ‘Phil. Trans. Roy. Soc.,’ B., vol. elxxxix, pp. 195-388. “ Colour-physiology,” Part III, ‘ Phil. Trans.,’ B., vol. exeviii, pp. 1-16. Van Rynberk.—* Ueber den durch Chromatophoren bedingten Farbenwechsel,” ‘ Ergebnisse der Physiologie,’ v, pp. 347-571. Minckiewiez.—“ On the Range of Coloration acquired by Hippo- lyte,” ‘ Arch. de Zool. Expér. et générale.’ Tate Regan.— Observations on the Colour-changes of Certain Fish,” ‘ Proc. Zool. Soe. London,’ p. 130. Townsend, C. H.—‘ Observations on Instantaneous Changes in Colour among Tropical Fishes,” ‘Thirteenth Annual Report New York Zoological Society.’ THE RELATION BE'TWEHEN LIGHT AND PIGMENT-FORMATION. 583 EXPLANATION OF PLATE 23, Illustrating Professor Gamble’s paper on “ The Relation between Light and Pigment-formation in Crenilabrus and Hippolyte.” Fig. 1—Young Crenilabrus melops (x 5) in the dark-banded phase induced by exposure to dark backgrounds. Fig. 2.—The green phase in the same fish induced by exposure to red light transmitted by red weed, and also by exposure to backgrounds of red weed for one week. Fig. 3—The reddish brown banded phase assumed by exposure for a week to light transmitted through green weed. The red colour is a shade too pronounced in the figure. Fig. 4.—Young Hippolyte varians in the almost colourless con- dition in which it is taken among weeds when 4-5 mm. long. (xX 24.) These colourless Hippolyte formed the starting-point for the experi- ments recorded in this paper. Fig. 5.—The brilliant carmine colouring induced in Hippolyte by exposure to pure green light for three to four weeks. (x 22.) Food- plant, Ceramium. Fig. 6.—The green colouring induced in Hippolyte exposed to red light for four weeks. Food-plant, fine green weed. Fig. 7.—The yellow colouring induced in some Hippolyte exposed to red light for four weeks. Food-plant, Ceramium. Fig. 8.—Chromatophores from fig. 7, highly magnified. (x 390.) Fig. 9.—Chromatophores from fig. 6, highly magnified. (x 390.) Fig. 10.—Chromatophores from fig. 5, Fig. 11.—Chromatophores from Hippolyte exposed to white reflected light for one month. mah a «ue : ee = Over brown weed, u Under red weed Quant.ourn Mien Sev. Wl, b5 NS&ZE. Red light Original Stock ‘Huth, Lith? London IS THE TROPHOBLAST OF HYPOBLASTIC ORIGIN ? 585 Is the Trophoblast of Hypoblastic Origin as Assheton will have it ? By A. A. W. Hubrecht. With 7 Text-figures. In the ‘ Quarterly Journal of Microscopical Science’ there has lately appeared (vol. 54, part 2) an article by my friend Assheton, in which he points out certain objections which he feels inclined to raise against some of the views that, were developed by me in a contribution to the fifty-third volume of this Journal, entitled ‘ Karly Ontogenetic Phenomena in Mammals.” Although I regret that he has not seen his way to comply with the invitation which I addressed to my fellow-embry- ologists in October, 1901 (it was published on p. 5 of my article on ‘“ Tarsius” in the ‘Verh. Kon. Akad. v. Weten- schappen te Amsterdam,’ vol. viii, No. 6, 1902), and which was intended to minimise printed disputes, where personal inspection of the preparations might bring about consensus of opinion, still, I accept his challenge (loc. cit., p. 221), and will now “ discuss more fully the difficulties which have arisen in the minds of some who are unable to accept (my) theoretical conceptions.” In doing so I wish to remind my readers that I am not going to treat all the objections raised by Assheton one by oue. Many of them will remain sub lite until new facts have been discovered, settling the point in dispute either one way or the other. I will on this occasion 586 A. A. W. HUBRECHT'. restrict myself to a point of very fundamental importance on which Assheton’s and my own views are diametrically opposed to each other, ever since 1898. If the new facts which I bring forward in this paper should be convincing enough to change the minds of those who feel inclined— following Assheton’s example—to look upon the trophoblast as hypoblastic, I have no doubt that my proposal to exclude from the phylogeny of Eutherian mammals any ancestor who deposited megalecithal eggs, like the Sauropsids and the Ornithodelphia, will find a more easy acceptance on their part. Assheton’s reasons for considering the trophoblast as an essentially entodernial foetal envelope were first developed in 1898, in his article on ‘‘The Segmentation of the Ovum of the Sheep” (‘Quart. Journ. Micr. Sci.,’ vol. 41). Plate 18 of that article presents us with a series of diagrams most delicately shaded in red and blue, which were meant to explain the mutual relations of trophoblast, epiblast, and hypoblast in ten different genera of mammals, and to compare them with the Sauropsidan arrangement. These diagrams have not found favour with later authors on this subject, and have been taken no notice of in Hertwig’s extensive ‘ Entwickelungsgeschichte,’ in three volumes. At that time I refrained from entering into any polemical dis- cussion, considering that later observations would show the untenability of Assheton’s ingenious but unsatisfactory generalisation. In writing his latest article Assheton has, however, allowed himself to come too strongly under the influence of his own hypothesis of twelve years’ standing. I see no necessity for entering upon any detailed discussion concerning the numerous and different arguments which have led other embryologists as well as myself to reject that hypothesis of Assheton’s now that new facts have come to light concerning the very earliest segmentation stages of Galeopithecus. This very archaic genus may be looked upon as a derelict representative of a group that in earlier geological epochs gave rise to the modern bats. There are —_— IS THE TROPHOBLAST OF HYPOBLASTIG ORIGIN? 587 certain points of agreement between its early development and that of Pteropus, whilst Leche’s anatomical work (‘ Kgl. Svenska Vet. Akad. Handl.,’ Bd. xxi, 1886) upon Galeopithecus points in the same direction. Of this genus I have now in my possession several series of sections made through segmen- tation phases, some of which I have here figured. These sections leave no doubt that the trophoblast of Galeopithecus originates by delamination at as early an age as the two- and four-cell segmentation stage, and render it utterly futile to try and explain the Galeopithecus tropho- blast as “due to an overflow of the yolk or hypoblast cells over the epiblastic rudiment ”’ (Assheton, |.c., p. 228). If we look more closely at the three stages of Galeopi- thecus here figured and begin with the one that is the furthest developed (Text-fig. 1), we find full coincidence with a similar stage described by Assheton for the sheep (I.c., 1898, PI. 16, figs. 14, 15), by Keibel for the stag (‘ Arch. f. Anat. and Phys. Anat. Abt.,’ 1902, p. 292), by Weysse (‘ Proc. Amer. Acad.,’ vol. xxx, p. 283) for the pig, by van Beneden for the rabbit and bat (‘Archives de Biologie,’ vol. i), by myself for the hedgehog, for the shrew (‘Quart. Journ, Mier. Sci.,’ vol. 30, Pl. 17; vol. 31, Pls. 36, 37), for Tupaja, for Tarsius (‘ Verh. Akad. Wetensch. Amsterdam,’ vol. iv, 1895, Pls. 1, 2; vol. viii, 1902, Pls. 1, 2), and for Nycti- cebus (‘ Keibel’s Normentafeln,’ 1907), as well as by other embryologists for various other mammals. This is the common starting-point in which there is a trophoblast and an embryonic knob with a cavity below it, and in which a hypoblast is not as yet distinctly developed, although just beginning to make its first appearance. It should be borne in mind that this very stage is thus characteristic for genera of mammals so diverse as those mentioned above. The way in which Assheton attempts to prove from yet earlier stages of the sheep that the outer trophoblastic layer is in reality a derivate of the hypoblast appears to me to be so pre-eminently artificial (c.f. l.c. his figures 9-14) and the argumentation so weak, that I must ascribe to a similar 588 A. Av W. HUBRECH'. incredulity on van Beneden’s part that this latter author in the important article which appeared one year later than Assheton’s (‘ Anat. Anzeiger,’ 1899, p. 305), does not take the slightest notice of the English author’s view that the trophoblast (van Beneden’s “couche enveloppate”) should be looked upon as an entodermal derivate. If we now return to T'ext-fig. 1 of this paper and inquire how this stage in the ontogeny of Guleopithecus has been reached, Trxt-Fic. 1. Section of a blastocyst of Galeopithecus with embryonic knob and enveloping trophoblast, just before the establishment of the continuous hypoblast. we see that it has been preceded by the stages of which Text-figs. 2 and 3 are the representatives. In Text-fig. 2 the centre of the different sections is occupied by comparatively large nuclei, evidently belonging to a central group of cells—the mother cells of the embryonic knob. Outside this embryonic knob and forming the peripheral layer in these sections is protoplasm in which distinct cell- boundaries are not visible, but in which a certain number of nuclei (smaller than those of the embryonic knob) clearly indieate that in the live blastocyst a peripheral cell-layer was differentiated in addition to the embryonic knob. IS THE TROPHOBLAST OF HYPOBLASTLG ORIGIN ? 589 Going back yet one stage earlier, in which the two first cleavage-cells are just on the point of splitting up into four (as is distinctly indicated by the karyokinetic figures of fig. 3C), we notice, besides the two cleavage-cells, three polar bodies of comparative large size, as they are known for mammals in general. Moreover, at the periphery of the two cleavage- cells we find separate nuclei, indicating the very first origin by an early delamination process of the cells which in Text-figs.2 and 3 constitute the continuous layer of trophoblast. TEXT-FIG. 2. Fig? Galeopithecus. A series of five sections through a stage of cleavage preceding the blastocyst of Text-fig. 1. Embryonic knob with bigger nuclei contrasting with the smaller peripheral trophoblast-nuciei. No central space as yet developed. In the mammalian genera Iitherto examined with respect to the origin of the trophoblastic layer (Tarsius, Tupaja, rabbit, sheep, pig, stag, dog, mouse, guinea-pig, etc.), the tropho- blast undoubtedly makes its appearance at a somewhat later stage of cleavage, or rather the distinction in the morula stage between the mother-cells of the embryonic knob and those of the trophoblast is not so soon evident as it is in Galeopithecus. However, the karyokinetic processes by which in this latter genus the trophoblastic nuclei separate from the segmentation nucleus (which in its turn owes its 590 A. A. W. HUBRECHT. origin to the union of the male and female pronucleus) at so early a moment are not revealed by my preparations, and we cannot for the present come to any sound conclusion as to which of the two modes of formation of the trophoblast is the more archaic one. Recognising that the definite answer to this question can only be given when a number of new observations will be at our disposal, I may still be allowed to call attention to the fact that in Galeopithecus the spot where the polar bodies are TEXT FIG. 3. Galeopithecus. A series of six sections through a cleavage stage just preceding the formation of the second pair of cleavage- cells. In C karyokinetic figures indicate this. In B—E the polar bodies are visible. Apparent trophoblast nuclei are situated peripherally. applied against the egg (see lext-fig. 3 B—H) remains without trophoblast nuclei somewhat longer than other parts of the egg’s surface. ‘The question presents itself—supposing the process 18 more primitive in Galeopithecus—whether this particularity might have led (in such mammalian genera that should be considered as phylogenetically younger) to the arrangement which has induced van Beneden, Duval, and Assheton (in his later publications) to consider the cleavage- process of those mammals as revealing epibolic characteristics. In case this question will later have to be answered in the IS THE TROPHOBLAST OF HYPOBLASTIC ORIGIN P 591 affirmative, the so-called blastopore which van Beneden (1875) described in the rabbit?s morula-stage might correspond to the spot referred to in the three figures (Text-fig. 3 B—D), where the polar bodies lie. Having thus shown that Assheton’s hypothesis of the hypoblastic nature of the trophoblast is irreconcilable with the phenomena in Galeopithecus, I emphatically repeat my conclusion that we are not justified in accepting it for- any other vertebrate. He himself will admit that, such being the case, the comparison of the trophoblast of mammals with the ‘‘ deckschicht ” of fishes comes to the foreground with increased validity. I have already stated above that it is not my intention in this paper to follow Assheton’s criticism step by step. A more extensive article on the ontogeny of Galeopithecus will appear in the course of this year. I shall there find occasion to reply more fully to other parts of Assheton’s criticism. There is, however, one point on which I feel bound to apologise, viz. that I have not allowed enough space for the recognition of the fact that my kephalo- and notogenesis had already been partly forestalled in several of Assheton’s papers, and had by him been termed proto- and deuterogenesis. I ought to have particularly mentioned these names in my paper of 1908. Still, I must maintain my terminology now that Assheton himself states (l.c., p. 240) that his and my names “ signify a different interpretation,” and now that he maintains that mine ‘‘does not represent the actual facts.” As matters stand I feel that the important issue which is at the base of the whole question of gastrulation in vertebrates (very fully treated in Keibel’s contribution to vol. x of the ‘Ergebnisse der Anatomie und Entwickelungsgeschichte,’ but since then looked upon in a somewhat different hght after his and my own short papers in the ‘Quart. Journ. Mier. Sci.’ [vol. 49] and in the ‘Anat. Anzeiger’ [vol. xxvi] had appeared) renders any polemics about the nomenclature that should be adhered to untimely. Very numerous investigations 592 A. A. W. HUBRECHT. are yet necessary, and will undoubtedly soon be undertaken, before we dispose of the comparative material which is necessary for settling this important point in Vertebrate ontogeny, and for finally deciding which nomenclature ought to be adhered to. I gladly leave the latter decision to others, but would not let this paper see the light without recognising that until lately I have not sufficiently been aware that Assheton already in 1894 expressed opinions to which Keibel and myself have come along other roads, and which, though far from identical, still overlap each other in many respects. APPENDIX. While this paper was in the press, attempts were made b , y TEXT-FIG. 4. Part of a section through the blastocyst of Manis. The ectoderm (ec.) and endoderm (en.) of the embryonic knob are transversely cut. The trophoblast cells appear darker in this figure. me to ascertain whether other genera of mammals might perhaps exist which furnish evidence concerning the early phases of the trophoblast that might further corroborate the facts such as they are presented by Galeopithecus. I was all the more anxious to obtain information concerning the earliest stages of the scaled ant-eater (Manis), as, by a regrettable lapsus calami, which disfigures both the English and the German version of my “early ontogenetic phenomena in mammals, ete.,” a gastrula stage of Manis is erroneously attributed to Galeopithecus. Tt is fig. 18 on Pl. C, in vol. 53. of the ‘Quart. Journ, of Mier. Sci.,’ and fig. 46 in the German publication. I here —— IS THE TROPHOBLAST OF HYPOBLASTIC ORIGIN P 593 reproduce the misnamed figure of this early Manis, and have since had the good fortune of obtaining sections of yet earlier cleavage stages of the same animal. Sections of early blastocysts of two specimens of Manis are TExT-FIGS. 54 AND B. Fig.5a.- Two consecutive sections of very early blastocysts of Manis, which show what is presumably the earliest trophoblastic covering of the mother-cells of the embryonic knob. reproduced here in 'l'ext-figs. 5 a,b, and 6. The stage of 'ext- fig.5 is presumnably a two-cell, the other (as far as I can follow it up in the consecutive sections of the series) a four-cell cleavage stage (purposely but incorrectly not counting the TEXT-FIG. 6. Fig.6. Another section through another blastocyst of the same genus in the same stage. trophoblast cellsas such). In both the differentiation between the mother cells of the embryonic knob on the one hand, and the already so much more numerous trophoblast cells, leads to the inevitable conclusion that the phenomenon of the separation of the larval trophoblast from the remaining 594, A. A. W. HUBRECHT, embryonic cells takes place at quite as early a moment as we have above described it for Galeopithecus, and that’ also in Manis it is perfectly excluded to look upon the tropho- blast cells as hypoblastic. And so the early Manis may be joined to the early Galeopithecus as fatal to Assheton’s interpretation of the trophoblast. I cannot yet say with certainty, but I have reason to believe that also in the very young hedgehog similar peculiarities occur, At the same time it is very suggestive that the quaint and aberrant mode in which the trophoblast cells of Galeopithecus and Manis arise offers so many points of mutual resemblance between these two genera, and differ not inconsiderably from what we find in Primates, Rodents and Carnivores. Later investigations will have to decide whether the phenomenon, as it presents itself in Galeopithecus and Manis, is one of precocious segregation, FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 595 The Origin and Formation of Fibrous Tissue Produced as a Reaction to Injury in Pecten Maximus, as a type of the Lamellibranchiata. By G H. Drew, B.A., Beit Memorial Research Fellow; and W. De Morgan, F.Z.S. With Plate 24. ConrENTs. PAGE Introduction ; : : ; F . O95 Methods : . 596 Description of the Tissues of Pee ber maximus inv oly ed in the Experiments, and the Normal Process of the “ Clotting” of the Blood. . 598 The Formation of Fibrous eau at fie site of the Truplantatinn of a Mass of Gill-Tissue : . 600 The Formation of Fibrous Tissue at the site of the Implantation of Digestive Gland Cells : . 604 The Reaction of the Tissues to the Taplantation of a Mass of Sterile Agar Jelly : : ‘ , . 606 Summary of Results : ; ; : . 608 References ; : : ; . 609 INTRODUCTION. ‘HE experiments described in this paper were performed on Pecten maximus at the Laboratory of the Marine Biological Association at Plymouth. The object of our work was to investigate the histology of the reaction of the tissues to the presence of a foreign body, and to determine the origin and method of formation of the fibrous tissue formed around it. 596 G. H. DREW AND W. DE MORGAN. As one type of foreign body we chose sterile agar jelly, which has little or no irritative or toxic action on the tissues, and is not removed by phagocytosis. As another type we chose masses of gill-tissue and of the tissue of the digestive gland, taken from an animal of the same species. Neither of these could be injected under aseptic conditions, and both were capable of removal by phagocytosis. Considerable irritation was set up by the implantation of these tissues, especially in the case of the digestive gland. This produced marked degeneration of the neighbouring tissues, possibly owing to the liberation of ferments and consequent digestive action. Pecten maximus was selected for these experiments on account of the large size of its adductor muscle, which presents a homogeneous mass of tissue particularly suitable as a site for implantation of foreign bodies. Before making this choice, experiments were tried on several other animals, but it was found that in most cases the technical difficulties encountered in endeavouring to make implantations into small masses of tissue, and in determining the exact relation of the underlying organs to the superficial anatomy, were too great to render these animals suitable subjects for experiment. Such experiments were tried on Carcinus menas, Pagurus bernhardus, and others of the smaller species of crabs, on Palemon serratus, Ligea oceanica, Aphro- dite aculeata, Patella vulgata, Aplysia punctata, Archidoris tuberculata, and many Lamellibranchs, but none Offered such promise of success as Pecten maximus. Meruops. Pecten maximus can be readily obtained in the Saleombe Estuary. It was found necessary to allow these animals to become acclimatised to living in the laboratory tanks before proceeding to the experimental work. When first placed in the tanks the mortality was heavy, often amounting to 30 FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 597 per cent. in the first three days, but after the lapse of about a week the survivors appeared to be fully acclimatised to the changed conditions, and often remained healthy tor some mouths. Experiments on animals whose health was doubtful were of no value, both because the shock consequent on the injection of the foreign body frequently caused death, and also because the reaction of the tissues was not normal in unhealthy specimens. When a Pecten is healthy it lies with the valves of the shell shghtly apart, the tentacles are expanded, and it responds rapidly to any stimulus by closing the shell; when held up in the air, the water which drains away is Clear and contains no slime. An unhealthy specimen lies with the valves of the shell wide open, there is little or no response to stimuli, and the valves only close under pressure. ‘lhe tentacles are retracted, and the gonads, gulls, and tissues generally, look flabby and unhealthy. ‘lhe water which flows out between the valves is shmy and viscid, and this is generally the first sign of deterioration. All instruments used in the experiments were carefully sterilised in boiling water. The transplanting needle resembles a large hypodermic needle about 1 mm. in diameter and 6 cm. long. Into the hollow needle a somewhat longer stylet fits closely and works like a piston. Any material taken up in the point of the needle is sucked in by drawing the stylet back, and again ejected by pushing it forward. For injecting into the muscle, a solution of agar in sea- water, coloured by a little hamatein, was used. The agar jelly was liquefied by heating in boiling water, and was drawn up into the transplantation needle. On cooling it forms a cylinder, of the diameter of the needle, which is easily intro- duced into the muscle. The adductor muscle of Pecten maximus 1s so large that there is no difficuly in selecting a spot at which to bore the shell. The apex of an equilateral triangle, having for its base the line of junction of the posterior auricula with the VOL. 55, PART 3.—NEW SERIES. 39 598 G. H. DREW AND W. DE MORGAN. right valve, marks roughly on the surface a point at which the shell may be bored without damage to any organ. But as the animal gapes when removed from its tank, it is easy to slip a cork between the valves and select a spot by inspection. The holes were drilled in the convex or right valve by an ordinary dentist’s drill, the head of which was prevented from penetrating too deep by a lapping of thread. The spot selected for drilling was sterilised with a saturated solution of corrosive sublimate, washed off with a solution of hydrogen peroxide (30 vols.) or distilled water, care being taken not to allow any of the sublimate to run between the valves. ‘he transplanting needle was then introduced to the required depth, slightly withdrawn, and its charge projected into the channel. The hole was then thoroughly dried, and stopped with sealing-wax. If the drying is thorough the wax will adhere after the animal has been returned to the tank. It would, of course, have been possible to implant directly into the muscle through the opening of the valves, but the risks of sepsis would have been greater. When required for examination, the shell was opened by cutting the adductor muscle at its attachment to the right or convex valve, and a portion of the muscle containing the implanted material removed. ‘This was fixed by three or four hours’ immersion in Gilson’s fluid, then thoroughly washed, passed through the alcohols, cleaned in xylol, and embedded in paraffin wax. It was then cut into serial sections eight pe thick. Delafield’s hematoxylin, followed by Van Gieson’s stain, or Benda’s iron mordant and hematoxylin were used for staining. DescrieTioN OF THE TISSUES OF PECTEN MAXIMUS INVOLVED IN THE EXPERIMENTS, AND THE NorMAL PROCESS OF THE “‘CLoTTING”’? OF THE BLOoD. The adductor muscle of Pecten maximus consists of two portions, bound together by the same sheath of connec- FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 599 tive tissue, but differing in structure. The larger, semi-trans- parent and whitish, consists of striated fibres. The fibres of the smaller, which is opaque and dead white, and lies against the posterior surface of the larger mass, are non-striated. It was into the larger mass that all material in our experiments was introduced. There is a large blood supply to the muscle from the adductor artery (Dakin, 2), and it contains numerous lacunar spaces. Scattered through if are numerous strands of con- nective tissue. These contain fibroblasts with deep staining nuclei and long fibrillar processes. The digestive gland has a tubular structure and com- pletely surrounds the stomach, mto which its ducts open. The ducts break up into numerous alveoh, which ramify and ultimately form cca. The ducts are lined with ciliated epithelium, and the alveoli with secreting cells. These secre- ting cells are said to degenerate and become filled with a eranular pigment, and are ultimately shed into the lumen of the ducts (Dakin, 2). Thus in their younger stages they appear to have a secretory, and in their later stages an excre- tory function. In addition to these glandular cells, fibrous connective tissue and unstriated muscle-fibre are present. The ducts contain particles of food material, alge, diatoms, and bacteria, and consequently as a rule septic conditions prevail in the experiments. The blood of Pecten maximus is a slightly cloudy, colourless fluid. It does not coagulate, but when shaken a number of small, white, floccular masses appear, which soon fall to the bottom of the tube, leaving the supernatant fluid clear and transparent. ‘hese masses consist of blood-cor- puscles agglutinated to form plasmodia. The corpuscles, although varying in size, appear to be only of one kind. ‘They are amceboid bodies, which when expanded protrude a number of slender pseudopodia. When contracted, they are ovoid or spherical. There is a single compact nucleus, staining readily with methylene-blue. The cyto- plasm is finely granular, and stains with eosin, but there are 39 § 600 G. H. DREW AND W. DE MORGAN. no large eosinophile granules. According to Cuénot (1), they originate in a ‘ glande lymphatique ” situated at the base of the gills. One of us (Drew, 4) has shown in the case of Cardinm norvegicum that when the corpuscles come in contact with a rough foreign body, or with injured tissue, they possess the power of agelutinating and forming a compact plasmodial mass. In this way bleeding from a small wound is stopped. When the edges of a wound are covered with this mass of agolutinated corpuscles, protoplasmic strands are formed across the wound, connecting the plasmodia; these strands thicken and contract and so approximate the edges of the wound. So far as our observations go, there is no reason to suppose that the blood of Pecten maximus differs in any of these particulars from that of Cardium norvegicum. That Lamellibranch blood-corpuscles are capable of a phagocytic action towards degenerated cells has been shown by De Bruyne (8) in the case of Mytilus edulis, Ostrea edulis, Unio pictorum, and Anodonta cygnea. Sir Ray Lankester (5 and 6) has shown that certain corpuscles of Ostrea edulis have a phagocytic action on diatoms and minute green alge, and it has been shown by Drew (4) that the corpuscles of Cardium norvegicum have a phagocytic action on bacteria, and are attracted towards extracts of dead tissues. Tae Formation or Frprous Tisstk IN THE SITE OF THE IMPLANTATION OF A Mass or GiLL-TIssuE. As bacteria are normally present on the gill-filaments, the conditions when gill-tissue is implanted differ totally from those obtaining when sterile agar is used. The implantation soon produces an intense inflammatory reaction on the part of the animal. The blood-spaces in the immediate neighbonrhood of the implanted tissue become distended and crowded with corpuscles, which escape from the lacunar spaces and migrate towards the source of irrita- FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 601 tion, travelling in all directions between the muscle-fibres. On reaching the gill-tissue the corpuscles come to rest, and form a dense, agglutinated, plasmodial mass, completely sur- rounding and shutting off the gill-tissue from the neighbour- ing muscle (fig. 1). They soon appear as if they had under- gone some degree of pressure and the nuclei are slightly flattened, probably owing to the contraction of the plasmodial mass as it tightens round the implanted gill-tissue (Drew, 4). In time the corpuscles show signs of degeneration; the nuclei become irregular in outline, and the chromatin is represented by numerous granules staining darkly with hematoxylin. The degenerated mass of corpuscles is then invaded by fresh blood-cells, and is more or less completely removed, apparently partly by a process of phagocytosis and partly by autolysis. While this is going on, the cells of the gill-filaments have degenerated, their outlines are ill-defined, and the nuclei no longer discernible ; the bacteria present multiply consider- ably. The degenerated gill-tissue is then invaded by blood-cor- puscles which have penetrated through the surrounding mass of agglutinated cells, and in most cases the bacteria and epithelial débris are removed by phagocytosis, leaving only the chitinous supporting-rods of the gills. In the course of this process many of the invading cells also are destroyed, and appear in their turn to be removed by other phagocytes. In time the whole space originally occupied by the gill-tissue becomes filled with a loosely packed mass of blood-cells, among which the chitinous supporting bars are the only relics of the original implanted mass. In many of our experiments bacteria multiplied so rapidly that the phago- cytes were unable to cope with them. Consequently the bacteria invaded the neighbouring tissues, entered the blood- spaces, and rapidly caused death. In preparations from obviously unhealthy animals, it was commonly found that the bacteria had penetrated beyond the protecting mass of agglutinated cells and had invaded the 602 G. H. DREW AND W. DE MORGAN. muscular tissue, which showed signs of degeneration in its somewhat swollen fibres and faint striation. When a blood-space had been entered, bacteria were often seen ingested by the blood-corpuscles, but in later stages it was obvious that the number of bacteria was so out of pro- portion to the number of corpuscles that they could not all be removed by phagocytosis, and were of necessity distributed all over the body in the blood-stream. During these processes the fibroblasts in the walls of the blood-spaces, and in the intermuscular connective tissue in the neighbourhcod of the implanted mass, undergo rapid division. This rapid division, resulting from the reaction of the tissues to the irritation caused by implantation, appears to be always amitotic. Mitotic division was only observed in much later stages, when the source of irritation had been removed by phagocytosis, and the rate of division of the fibroblasts was much slower. Before amitotic division the fibroblasts lose their spindle shape and become oval; a split then appears at one end, and progresses in the plane of the long axis of the nucleus until two daughter nuclei are formed, attached to each other at one extremity, and inclined at an acute angle to one another. These gradually straighten out until they form an hour-glass- shaped mass of nuclear material. Finally the two nuclei are separated at the constriction and become almost circular in shape. As a result of this active multiplication of the fibroblasts, the strands of connective tissue bounding the blood-spaces and forming the intermuscular connective tissue become crowded with nuclei. The bodies of the fibroblast cells become very indistinct, and little beyond rows of elongated nuclei is discermble. As the multiplication becomes more ‘apid the typical spindle shape of the nuclei is lost, and they become first oval and finally circular. There appears to be a constant migration of these cells, with round and oval nuclei, towards the site of implantation. They have very little cytoplasm, and from this, and their FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 603 smaller size, are easily distinguished from the blood-corpuscles (figs. 2and 3). These fibroblasts largely follow the course of the strands of fibrous tissue bounding the blood-spaces, and they appear to travel along in the spaces, being most plentiful near the walls. At the same time, when they multiply very rapidly, many migrate in ali directions between the muscular fibres towards the implanted tissue, and are not confined to travelling only in the proximity of pre-existing connective- tissue strands. Onreaching the degenerating layer of agglutinated corpuscles surrounding the implanted tissue, they arrange themselves in rows, and their nuclei elongate in such a direction that their long axes form arcs of a circle surrounding the implanted tissue. Some fibroblasts penetrate among the degenerating cells of the gill-tissue, which are being removed by phago- cytes, and in this position start the formation of fibrous tissue. The surrounding layer of fibroblasts gradually thickens, and presents a somewhat stratified appearance. At first this layer contains a number of blood-corpuscles, but these even- tually are removed, probably by autolysis, leaving only the fibroblasts, which can now be seen to be connected with each other by numerous fine processes of the cytoplasm, the whole presenting a somewhat reticulated appearance. In time this tissue becomes more compact, and the reticulation vanishes. It would appear that this has been caused by the contraction of the processes of the fibroblasts, with consequent approxi- mation of the cells. Finally, the nuclei become long and spindle-shaped, the amount of cytoplasm slightly increases, and a layer resembling normal fibrous tissue results. In our experiments the great variation in the rapidity with which the various changes described took place was very noticeable. The health of the animal after the experiment seems an Important factor m accounting for this, for the slow rate of fibrous tissue formation in unhealthy, as compared with healthy animals, was very marked. Unfortunately none of the animals into which gill-tissue 604. G. H. DREW AND W. DE MORGAN. was implanted lived long enough for all the elements of the gill-tissue to be completely replaced by fibrous tissue, but in healthy specimens most of the signs of inflammation had vanished, and the implanted tissue was surrounded by a wall of apparently healthy fibrous tissue, in four or five days. Formation or Frprous Tissuz AROUND HE SITE OF IMPLANTED Digestive GLAND CELLs. After the implantation of portious of the digestive gland, a marked degeneration of the muscular fibres in its neigh- bourhood is noticeable. ‘l'hey swell slightly, all trace of striation is soon lost, and they stain less intensely. The area of degeneration gradually extends, and the muscular fibres in the immediate neighbourhood of the gland tissue are slowly dissolved. ‘This action is presumably due to the presence of ferments in the digestive gland, which digest and render soluble all tissues in the immediate neighbourhood. At the same time the cells of the gland itself degenerate aud appear to undergo auto-digestion, so that eventually only the brown pigment-granules originally contained within the secreting cells remain. Under these conditions bacteria do not seem to multiply, though they must have access to the ceeca of the digestive gland, as these are in direct communication with the alimentary canal. In none of our sections have we been able to find bacteria, though it is quite common to find the siliceous skeletons of diatoms in the ceca. It seems, there- fore, probable that the presence of digestive ferments inhibits the multiplication of bacteria. Asa result of the implantation of this tissue a condition of intense inflammation is set up, and all the blood-spaces in the neighbourhood become distended with blood-corpuscles. ‘here appears to be an endeavour on the part of the orga- nism to shut off all the implanted gland, together with the area of muscular tissue which has undergone degeneration, from the general blood-stream. ‘This is effected by the formation of a layer of agglutinated blood-corpuscles around FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 605 the whole of the affected area (fig. 4). It was very notice- able in our preparations that the degenerated area was always larger in specimens that had been implanted with the digestive gland for some time (up to six days), than in those implanted for shorter periods, and thus it would seem that the range of action of the digestive ferments gradually increases. The degenerated area was always found sur- rounded by a layer of agglutinated corpuscles, though in different specimens this layer varied considerably in thickness. It would seem that while the degenerative process is spread- ing the layers of corpuscles must be continually dissolved, and others formed a little further back by the spread of the digestive ferments. During this process the fibroblasts undergo division as in the case of the gill-tissue, but while the inflammation is much more acute, the multiplication of fibroblasts is not so rapid, and they are not nearly so notice- able a feature in the sections. In the form of rounded cells, with oval or spherical nuclei, they migrate in small numbers towards the layer of agglutinated blood-corpuscles. Here they share the fate of the corpuscles, being dissolved by the digestive ferments, and accordingly there is no formation of fibrous tissue. We were never able to keep the animals alive for more than six days. At the end of this time all that remained of the digestive gland was the brown pigment-granules and a little epithelial débris. This was surrounded by a space from which most of the muscular tissue had been dissolved, aud this again by a relatively large area of degenerated muscle-fibres. Finally, the whole was surrounded by a layer of agglutinated blood-corpuscles, into which a few fibro- blasts were making their way. These experiments show that the protective layer of cor- puscles must very completely shut off the space it encloses from the neighbouring tissue. If this were not the case the digestive ferments, once they had gained access to the blood, would rapidly become disseminated over the whole body. Instead of this, we have distinct evidence that there is a slow and 606 G. H. DREW AND W. DE MORGAN. steady invasion of the tissues by the ferments, and that the area of their action is always contained within a protective layer of agglutinated blood-corpuscles. It seems probable that the digestive gland, when implanted, contains little or no free enzyme, and quickly becomes surrounded by the protective layer of corpuscles, and that later the enzymes are slowly evolved from the zymogens contained within the cell. The vitality of these cells has been impaired by removal from their normal connections and by implantation into the muscle tissue, and accordingly they are dissolved by the enzymes they have themselves evolved. Tae Reaction or tHe Tissurs to ImMeLANTED AGAR JELLY. Sterile agar jelly has no irritative action on the muscle, and so differs from the tissues previously described. Agar jelly may be regarded as a physiologically inert substance, and as in these experiments it was made from sea- water in which the Pecten were living, it was approximately of the same salinity as their blood (Dakin, 2), and so was of the same osmotic concentration. Further, the cylindrical rods of agar are remarkably smooth, and if unbroken present no rough surface, except possibly at the extremities. One of us (Drew, 4) has shown that in the case of Cardium norvegicum, the agglutination of the blood- corpuscles (in vitro) is much influenced by the nature of the substance on which they impinge, and that it occurs very much more readily when they come in contact with a rough surface from which a large number of small points may be imagined to project, than when they impinge on a smooth, polished body. It seems probable that similar conditions obtain in the case of the blood of Pecten maximus. In accordance with these properties of the agar jelly, it was found that absolutely no inflammation resulted from its implantation in the muscle. No layer of agglutinated corpuscles was formed round it, and there was no sign of the collection of unusual numbers of the corpuscles in the FIBROUS TISSUE PRODUCED AS A REACTION TO INJURY. 607 vicinity, nor of any distension of the blood-spaces. ‘The fact that the rod of jelly was always implanted as far as possible parallel to the long axes of the muscle-fibres, and that they were usually rather separated from each other, than cut by the insertion of the transplanting needle, probably contributed towards this result. After a period of about seven to eight days there were signs of division of the fibroblasts in the neighbourhood of the implanted mass, and a slow migration of the new-formed cells towards the agar took place. By about the tenth day these cells had arranged themselves so as to form a thin and delicate ensheathing layer. The process presents marked differences from that which occurs after the implantation of a substance which causes an inflammatory reaction, with the consequent development of a protecting layer of agglutinated corpuscles. he division of the fibroblasts, instead of being rapid and amitotic, is comparatively slow, and frequently, though not always, mitotic. The nuclei of the young fibro- blasts retain their elongated shape, and though the nuclei of the dividing cells lose their typical spindle-like appearance and become oval, they do not become round, as in the case of rapid division after inflammation. The layer of fibrous tissue formed is thinner and less compact, the proportion of cyto- plasm to nucleoplasm is greater, and the nuclei assume their typical spindle shape more rapidly. ‘The process seems to be complete by the tenth day, and the appearance is almost identical with that shown in fig. 5, which represents the condition after seventeen days. In some of our experiments the sealing-wax with which the drill holes were closed became detached in the tank. ‘The holes were re-sealed as soon as this was noticed, but the animals seldom survived long. On sectioning, an area of inflammation was usually found surrounding the agar, and rapid division of the fibroblasts in the vicinity was in pro- gress. In specimens that survived longer a complete sheath of fibrous tissue had formed round the agar, and the con- dition resembled that resulting from implantation of gill- 608 G. H. DREW AND W. DIS MORGAN. tissue. It seems that in these cases bacteria must have entered through the drill-hole, and, travelling between the agar and muscle, have caused an inflammatory reaction. In one other case, in which the hole had not come unsealed, in- flammation and formation of fibrous tissue occurred, but as this ouly took place once out of twenty-six implantations made with sterile agar, it is probable either that the sealing-wax plug leaked at the edges or that bacteria found their way in when the agar was introduced. SumMMARY or Resutts. Our experiments show that the implantation of a tissue, such as that forming the gills, accompanied by the bacteria which adhere to it, produces an intense inflammatory reaction. This is characterised by the active migration of blood- corpuscles, which form a plasmodial mass around the im- planted tissue, shutting it off from the general circulation. This protective layer is gradually removed by phagocytosis and autolysis, and at the same time the gill-tissue is invaded and removed by phagocytes. - While this is going on, rapid amitotic division of the fibroblasts in the neighbourhood occurs ; they lose the typical spindle-shape of the nuclei, and the new-formed cells consist of rounded or oval nuclei, with a scarcely perceptible amount of cytoplasm. ‘These rounded cells migrate towards the implanted tissue, aud arrange themselves in layers around it, the nuclei become elongated, and the proportion of cytoplasm increases. Finally, a layer of typical “scar ” fibrous tissue is formed, enclosing the chitinous skeletons of the gill-bars. In the case of the implantation of digestive gland tissue a similar protective layer of agglutinated corpuscles is formed, but this is continually dissolved up and reformed, as the sphere of action of the enzymes in the cells of the digestive gland extends, All the muscle-fibres within this protective layer soon lose their striation, swell, and are partially dis- solved, presumably by the digestive enzymes. The fact that FIBROUS TISSUE PRODUCED AS A REACTION 'TO INJURY. 609 there is a progressive extension of this digestive action shows that the layer of agglutinated corpuscles performs its pro- tective function very completely, as otherwise the enzymes would escape into the general circulation. Simultaneously the fibroblasts in the vicinity multiply and migrate, as in the case of implanted gill-tissue, but the multiplication does not seem to be so rapid. No permanent layer of fibrous tissue is formed, as the migrated fibroblasts are dissolved in the course of the extension of the sphere of action of the digestive ferments. In the case of the implantation of sterile agar jelly, made with sea-water, no inflammation results, and for some time there is no sign of any reaction of the tissues 1f absolute asepsis has been ensured. After seven or eight days there is a Slow and often mitotic division of the neighbouring fibro- blasts; they migrate and rearrange themselves to form a thin layer of fibrous tissue around the agar. It is noteworthy that though the tissues and the blood, especially in its manner of forming a “ clot,” present marked differences from those in Vertebrates, yet the formation of fibrous tissue, as a reaction to injury, does not differ in any essentials from the process which takes place in the higher types. REFERENCES. 1. Cuénot, L.—* Etudes sur le Sang et les Glandes Lymphatiques,” ‘Arch. de Zool. Expér. et Gen.,’ Deuxieme serie, tome ix, Paris, 1891. 2. Dakin, W. J.—* Pecten,” ‘ Liverpool Marine Biological Comunittee Memoirs,’ xvii, London, 1909. 3. De Bruyne, C.—* Contribution A l'étude de la Phagocytose (1),” ‘Arch. de Biol.,’ tome xiv, Paris, 1896, p. 161. 4. Drew, G. H.—*Some Points in the Physiology of Lamellibranch Blood-Corpuscles,” ‘Quart. Journ. Mier. Sci., vol. 54, part 4, February, 1910, p. 605. 5. Lankester, Sir E. Ray.—* On Green Oysters,” ‘Quart. Journ. Micr. Sci.,’ vol. 26, 1886, p. 71. 6. ——— “Phagocytes of Green Oysters,” ‘Nature,’ vol. xlviii, 1893, Ley) VOL. 59, PART 3.—NEW SERIES. 4.0 610 ‘ H. DREW AND W. DE MORGAN. DESCRIPTION OF PLATE 24, Illustrating the paper by Messrs. G. H. Drew and W. de Morgan on “ ‘The Origin and Formation of Fibrous ‘Tissue produced as a Reaction to Injury in Pecten maximus, as a type of the Lamellib ‘anchiata.” REFERENCE LETTERS. ag. Agar. agg.lyr. Agglutinated layer of blood-corpuseles. — .c. Blood-corpuscles. deg. gill. Degenerated gill-tissue. deg. msl. Degene- rated muscle. dig. gl. Digestive gland-tissue. div. fbl. Dividing fibro- blasts. fbl.lyr. Fibroblast layer. mig. fl. Migrating fibroblasts. msl.-fbr. Muscle-fibres. [N.B.—In the figures the bundles of muscle-fibres are shown as a whole: the individual fibrils and their striations are not differentiated. The size of the muscle-hundles differs considerably in different parts of the adductor muscle. | Fig. 1—x 400, Gill-tissue which has been implanted for sixteen hours. A layer of agglutinated corpuscles divides the degenerated gill- tissue on the left from the muscular tissue on the right. Corpuscles ure making their way between the muscle-fibres to join the agglutinated layer. Fig. 2.— x 300. A later stage of fig. 1, tuken seventy-two hours after implantation. A definite layer of fibrous tissue has been formed round the gill-tissue, which is completely degenerated and invaded by phago- cytes. The fibroblasts are dividing and migrating towards the lesion. Fig. 3—x 700, A more highly magnified portion of one of the blood-spaces drawn from the same section as fig. 2. The fibroblasts are undergoing amitotic division, and migrating towards the gill-tissue, where they arrange themselves to form a layer of fibrous tissue. Fig. 4—x 450. Digestive gland-cells (on the left) which have been implanted for ninety-six hours. External to them is a region of degenerated and partially dissolved muscle-fibres, which is divided from the normal muscle by a thin layer of agglutinated corpuscles. These are also rapidly degenerating, but are reinforced by the continued arrival of fresh corpuscles. The cellular structure of the alveoli of the digestive gland has been lost, leaving little beyond traces of the original cell walls and the brown pigment-granules. Fig. 5—x 450. Agar jelly (to the left) which has been implanted for seventeen days. It is divided from the muscle-tissue by a delicate layer of fibroblasts. DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 611 The Division of the Collar-Cells of Clathrina coriacea (Montagu): A Contribution to the Theory of the Centrosome and Blepharo- plast. By Muriel Robertson, M.A., and E. A. Minchin, M.A. With Plates 25 and 26. INTRODUCTORY. Ar the present time there is a great deal of confusion in the use of the words “ blepharoplast” and “‘ centrosome.” Two distinct questions arise with regard to the significance of these bodies; the first is the question of the homology of blepharo- plasts and centrosomes; the second is that of the nature of the centrosome, and more particularly whether or not it is to be regarded as equivalent primarily to a nucleus. With regard to the first of these questions, it is now generally admitted that blepharoplasts and centrosomes are essentially bodies of the same nature, for reasons that will presently be considered at greater length. The difference between a centrosome and a blepharoplast, on this view, is entirely a matter of divergence of function. A centrosome may be briefly characterised, in a general way, as a body which exerts or governs kinetic functions in relation to the division of the nucleus; a blepharoplast may be defined as a centrosome which governs the movements of motile organs, such as flagella, which arise from it and are in direct or indirect connection with it. With regard to the second of these questions, namely, the nature of the centrosome, two opposite views are current, VOL. 55, PART 4.—NEW SERIES. 4] 612 MURIEL ROBERTSON AND E. A. MINCHIN. which may be summarised as follows: (1) The centrosome is to be regarded as primarily a body of achromatic! nature, elaborated and evolved, in all probability, in the nucleus or in connection with it, but not itself equivalent to a nucleus ; (2) the centrosome is regarded as the equivalent of a nucleus, and as representing primarily a nucleus which has become modified and specialised both in function and structure. These two theories may be termed conveniently the achromatic and the nuclear theory of the centrosome respectively. According to the second of these views, which has recently been revived and advocated by Hartmann and Prowazek (6), every cell is to be regarded as primarily and essentially binucleate; the two nuclei, at first, doubtless, equivalent and similar in all respects, became modified in two directions respectively, the one becoming specialised for trophic, the other for kinetic functions, with corresponding differentiation of structure. In the metazoan cell, aecording to this theory, the nucleus represents the original trophic nucleus deprived of all kinetic structures, while the centro- some represents the kinetic nucleus deprived of all “ vegeta- tive” functions and of its chromatic apparatus. On this interpretation of the centrosome, the minute granules which are the centre of kinetic functions are termed ‘‘centrioles,”’ in order to distinguish them from the centrosomes as a whole. In fact, from the point of view of the nuclear theory of the centrosome, the centriole requires to be defined in exactly the same way as the centrosome itself on the achromatic theory. The confusion produced by these two theories of the cen- trosome reaches its height in the nomenclature of the different parts of the body of a trypanosome. In these organisms, and in allied genera of flagellates, there are three distincts parts of the nuclear apparatus to be reckoned with. First, there ! Meaning by the term “achromatic ” something which is not com- posed of chromatin, not necessarily something which is not coloured by stains. All is not chromatin that stains, even with a so-called nuclear stain. In our opinion a great deal of error and misconception has arisen by identifying as “ chromatin ” all bodies in the cell that are coloured black, for instance, by the iron-hematoxylin method. DIVISION OF COLLAR-CELLS OF CLATHRINA CORTIACEA. 613 is achromatic body, which may be denoted temporarily by the symbol N, situated usually in or near the middle of the cell-body, and in no special connection with the flagellar apparatus. Secondly, there is a second chromatic body, which may be denoted by the symbol m, distinctly con- nected with the flagellum or flagella, when they are present, and apparently kinetic in function. In the genera Try- panosoma, Herpetomonas, Leishmania, and Cri- thidia, N is always much larger than n, but in Trypano- plasma the reverse may be the case. Finally, the flagella arise, probably in all cases, from basal granules, which are often very minute and exhibit staining reactions quite different from either N or 1. According to the nuclear theory of the centrosome advo- cated by Hartmann and Prowazek, these three parts of the trypanosome body are to be interpreted and named as follows: N is the trophic nucleus, while n represents the second nucleus of kinetic function, in other words, the cen- trosome, which, since it controls the activities of the flagellar apparatus, is to be termed a blepharoplast. The basal granule is a mere thickening of the proximal end of the flagellum, of no special significance, or at most representing a centriole. Thus a trypanosome would represent the ideal binucleate cell of Hartmann and Prowazek in a very primitive state. An interpretation of the trypanosome body, quite different to that of Hartmann and Prowazek, has been advocated by one of us (12), which may be briefly stated as follows: N is a trophic nucleus, which contains its own centrosome or division-centre in itself; 2 is a distinct kinetic nucleus, a specialisation of the nuclear apparatus for a particular func- tion ; it has nothing to do with a true centrosome, though it may, like the trophic nucleus, contain a body of this kind, nor is it to be regarded as a blepharoplast, a body which is repre- sented by the basal granule of the flagellum.! ' It is not our purpose here to summarise the various views that have been put forward with regard to the morphological interpretation of the trypanosome-body, but only to select two which show in sharp 614 MURIEL ROBERTSON AND E. A. MINCHIN. In consequence of these divergent theories and interpreta- tions, a great confusion in nomenclature has arisen, especially with repard to n, which is always termed the blepharoplast in German works, the centrosome in French works, aud in this country is sometimes named the micronucleus, but more usually the kinetonucleus, In Flagellata other than the trypanosomes and their allies there is usually only one structural element other than the principal nucleus (N) to be reckoned with in the nuclear apparatus, namely, a deeply staining grain or set of grains, from which the flagellum or flagella take origin, and to which the name ‘ blepharoplast,” or the synonymous term ‘“‘diplosome,”’! is commonly applied. ‘The question at once arises, How is the arrangement seen in a trypanosome to be compared to that of other flagellates, and to which element in the nuclear complex of a trypanosome should the blepharo- plast of an ordinary simple flagellate be compared? Does it represent the basal granule (true blepharoplast, on our view) or the kineto-nucleus (x)? In our opinion, the bodies in question are true blepharoplasts, comparable to the basal granules of the flagella of trypanosomes, and the kineto- nucleus or German blepharoplast of the trypanosomes and their allies isa nuclear body peculiar to them, and not found in ordinary flagellates. To this extent, at least, we are in agreement with the idea expressed by Hartmann, who has placed the trypanosomes and forms regarded as contrast opposed views with regard to the nature of the blepharoplast and the proper application of this word. Thus Layeran and Mesnil in their well-known work on trypanosomes use the term * centrosome” for n; so also Moore and Breinl, who contrast the extra-nuclear centrosome (n) with the intra-nuclear centrosome (karyosome of 1). 1 “The term “diplosome,’ meaning literally and etymologically a double body, is commonly applied, by an abuse of language, to the single grain from which a flagellum arises. It should, of course, be used only for those cases where twin granules give origin to two or more flagella, that is to say it should not be regarded as synonymous with blepharoplast or basal granule, but as implying a condition in which such bodies are doubled. DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 616 allied to them in a separate order of Flagellata termed the Binucleata. (The question as to whether or not the Heemo- sporidia should be included in the Binucleata is one which, in the present memoir, we do not wish to raise or discuss.) A trypanosome is, in our opinion, a_ binucleate organism, possessing a trophic nucleus (N), a kinetic nucleus (nz), and a blepharoplast (basal granule). In order to settle these disputed points, more knowledge is required regarding nuclear and other structures connected with the locomotor apparatus in other organisms, and in the hope of throwing some light on these questions we have studied the division of the collar-cells of a caleareous sponge, of which preserved material was in the possession of one of us. I greatly regret that in my account. of the Sponges in Lankester’s ‘Treatise on Zoology’ (Part II, 1900, p. 56) I gave an entirely erroneous account of the division of the collar-cells of Clathrina coriacea, stating that after division of the nucleus the cell divides transversely to its long axis, and then the basal portion forms a new collar and flagellum. I have re-examined the figures and preparations on which these statements were founded, and see that I was misled by sections passing obliquely through the epithelium, so that the top part of a dividing cell, with the nucleus at the apex, appeared superposed on the base of an ordinary cell, with its nucleus in the usual position. The account given in the present memoir will show clearly the error of my former statements.—H. A. M. 620 MURIEL ROBERTSON AND E. A. MINCHIN. Most of the sponges were preserved in osmic acid followed by picrocarmine, a good method for showing clearly the cyto- plasmic structures, especially the collar and flagellum, but not suitable for demonstrating the finer details of the nuclear apparatus. Some of the material, however, had been pre- served in Hermann’s fluid, and it is on this that we base the results set forth in this memoir. Sections cut from sponges preserved in this way were stained with various stains, more particularly by Heidenhain’s iron-hematoxylin method, and counter-stained with eosin or Lichtgriin, the latter being found to be of great assistance in making out the details of the collar and flagellum, since these parts are tinged by it. (1) The Resting Collar-cell._—In Clathrina coriacea, as in all sponges of the family Clathrinide, the nucleus lies invariably, in the ordinary ‘‘ vegetative” or resting condition, at the base of the columnar collar-cell, that is to say, at the end which is furthest from the collar and flagellum. At the apex of the cell, in the centre of the area enclosed by the base of the collar, lies a minute granule—the blepharoplast—from which the flagellum takes origin. These structures, no less than the general form of the collar-cell and its position in the epithelium, of which it forms a part, give a definite orienta- tion to the cell; any direction parallel to an imaginary axis continuing that of the flagellum and passing through the blepharoplast and nucleus may be termed vertical, while any plane at right angles to the vertical axis may be termed horizontal. Fa The form of the collar-cell and the dimensions of their different parts vary considerably with the condition of the sponge, whether expanded or contracted, and may be different also in different parts of the same sponge. In specimens in which the pores are fully open, and in which all appearances indicate that the collar-cells are in full functional activity, the bodies of the cells are fairly broad, and about 12-13 4 in height by 5-6 w in breadth ; the collar reaches a length of 10-11 uw, and the flagellum some 25-27 uw. When, on the other hand, the pores are closed up and the sponge is partially con- ON DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 621 tracted, the collar-cells become taller and narrower and the collar much shorter. In each cell the basal three fourths of the body is broader and more or less cylindrical in shape ; this part of the cell is in contact with the neighbouring cells, and constitutes the main body of the cell. We have not found processes connecting the bodies of the cells with one another. It has been shown by Minchin and Reid (14) that when the collar-cells are carefully brushed away and the wall of the sponge is stained with picro-nigrosin, a delicate blue-stained network is visible in surface view, representing a honeycomb- like structure, the spaces in which were originally occupied by the bodies of the collar-cells. Hence in life the bodies of the collar-cells are probably not in actual contact, but are separated by a delicate extension of the gelatinous ground- substance of the body-wall of the sponge. If, as would seem probable on theoretical grounds, the bodies of the collar-cells are connected across this intervening substance by proto- plasmic fibrils, such connections have escaped our notice, possibly on account of their being of extreme tenuity and requiring, perhaps, other methods of technique, in order to demonstrate their existence, than those employed by us for the study of the mitosis. It is well known that in other sponges the collar-cells may be connected by protoplasmic processes, as, for instance, in Hexactinellids, where such processes are extremely obvious, forming the so-called membrana reticularis. The cylindrical basal portion of the cell ends in a distinct rim or flange, and from this level arises a narrower portion, which may be termed the ‘neck,’ and which is quite free from any contact with neighbouring cells. The summit of the neck is rounded off, forming a convex lens-like area enclosed by the base of the collar, and giving off centrally the flagellum. The so-called collar has more the form of a cuff or sleeve when fully expanded. It is distinctly thicker and more rigid in its basal portion, becoming very delicate at its distal end, which is usually more or less shrunk and dis- torted in preparations. The uppermost limit of the collar is 622°, MURIEL ROBERTSON AND E. A. MINCHIN. often very difficult to make out. It is best preserved in the osmic-picrocarmine preparations ; after Hermann’s fluid it appears collapsed and shrunk or frayed out. A short way above its origin the collar usually shows a distinct thickening, visible as a horizontal hoop-like structure, especially when the collar is a little contracted ; when it is expanded to its fullest extent the hoop is difficult to make out as a horizontal line, but its presence is marked by the fact that all the part of the collar below it stands out stiff and firm, and is not creased and folded like the part above. It is evident from the appearances seen both in the resting and the dividing cell that the collar for about 2 4 from its origin is thickened and strengthened as compared with its distal portion. The nucleus of the collar-cell is about 5 « in diameter and more or less spherical in form, sometimes slightly flattened in the vertical direction. ‘he most conspicuous element in its structural composition is a large grain, which stains deeply with iron-hematoxylin, and appears to be of the nature of a karyosome. This structure is always present, and sometimes double (figs. le, 3,7). The karysome is sometimes lodged in a clear space (fig. 36, d, e, f) ; its position in the nucleus varies. The remainder of the nuclear contents appear granular, but in thin sections of the nucleus a fine network can be made out (figs, 36, e, f, g), in the nodes of which the granules of chromatin are lodged. These granules vary very much in different nuclei in the same preparation, being sometimes so fine as to be scarcely visible individually, while in other nuclei they are coarse and irregular in size and shape (figs. 30 and 36, a, b,c). All transitions can be found between the finely and the coarsely granular condition, but the two extremes form two: well-marked types, which may be characterised as the light and the dark type respectively. It is worthy of note that nuclei of cells about to divide are always of the light type, as will be pointed out in the next section. The above description of the nuclei refers to preparations stained with iron-hematoxylin. In material preserved and i ,, DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 623 stained by the osmic-picrocarmine method the nuclear structure is not shown at all as a rule, but the nucleus simply stains evenly pink. Sometimes the karyosome can just be made out, sometimes not. A peculiar feature of the prepara- tions is that the red stain often does not extend up to the nuclear membrane; the stained portion forms a mass lying in the centre of the nucleus, and between this stained mass and the nuclear membrane a clear space remains, which can often be seen to be traversed by delicate radiating lines, as if fine filaments started from the membrane to support the central stained mass. Comparison with nuclei stained with iron- hematoxylin shows in many of the latter a distinct alveolar border to the linin-framework; sometimes the alveolar border is relatively very broad (fig. 36), and shows the radiating partitions of the alveoli very distinctly. It would appear as if the action of the osmic-picrocarmine method was to cause a shrinkage within the alveolar border, with the result that this inner portion of the nuclear framework contracts and appears as a homogeneous mass, which contains all the chromatin and stains deeply, leaving the alveolar border unstained. It should be noted that by no means all the nuclei of the collar-cells show the clear border within the membrane ; many of them stain evenly up to the membrane, and this is always so in those cells which are about to divide, The blepharoplast and flagellum stain black with iron- hematoxylin, but by the osmic-picrocarmine method they are not stained. The cytoplasm of the collar-cells is finely granular and usually very vacuolated. The neck is free from vacuoles as a rule, but in many cases a round vacuole-like structure, which differs in appearance from the other vacuoles, can be seen in the neck region. ‘lhe ordinary vacuoles in the body of the cell are clear and appear as empty spaces, doubtless representing drops of fluid in the living condition, but in the direct line between the nucleus and blepharoplast there is generally to be seen a vacuole, which has finely granular 624. MURIEL ROBERTSON AND E. A. MINCHIN. contents and sometimes a minute central granule (fig. 30, cell on the extreme right). This body is sometimes nearer the blepharoplast, sometimes nearer the nucleus, but usually it lies at a level midway between the neck and the main body of the cell or in the neck itself; its significance is doubtful. In addition to the vacuoles, the cytoplasm almost always contains one or more coarse refringent granules of irregular, angular form and yellowish-brown colour. They are lodged in any part of the cell and are often present in the vicinity of the blepharoplast. They probably represent excretion-grains. After the iron-hematoxylin stain they become darker, but still retain their characteristic yellowish-brown tint, and can be easily distinguished from chromatin grains. No other enclosures, as a rule, are to be found in the collar-cells, but occasionally they contain large rounded bodies (figs. 31-35 and 50, 51), which stain deeply with iron-heematoxylin and appear to be of the nature of organisms, though whether they represent parasites or food ingested by the cells is difficult to say. In some parts of the sponge they are found more com- monly than in others, and in one case (fig. 34) no nucleus could be made out in the cell; it may, however, have been cut off in the section. (2) Preparations for Division.—Before the nucleus begins to show any of the changes in its minute structure which initiate mitosis certain events take place in the cell, namely, the migration of the nucleus bodily from the base to the summit of the cell, the disappearance of the flagellum, and the division of the blepharoplast. As a general rule these three events take place in the order named, but not invariably, so that a number of different combinations arise in different cases. The migration of the nucleus is always the first sign that a collar-cell is about to divide, and this peculiarity is a great aid to the study of the division, since in a section of the sponge which shows the collar-cells cut vertically those that are dividing or preparing to divide arrest the attention at once, even with a comparatively low power of the microscope, DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 625 owing to the fact that the nucleus is no longer in its usual position at the base of the cell, but has either migrated to the apex or has been preserved in the act of doing so, and is found in some position between the base and the apex (figs. 1-5, etc.). Such cells are also characterised by being much broader and stouter than the ordinary resting cells, but they do not increase in height to an appreciable extent. By this process of migration the nucleus comes to lie immediately under the blepharoplast, and at this stage a curious appearance has been observed in two instances (figs. 4 and 5); the nucleus is seen to be flattened on the side nearest to the blepharoplast, and from the blepharoplast itself two streaks appear to radiate to the two ends of the flattened side of the nucleus. Careful examination of each of these preparations gives the impression that these two streaks are in reality the optical section of a cone-shaped mass of proto- plasmic substance, the base of which rests on the flattened side of the nucleus, and which is, perhaps, the cause of the flattening. A comparison with the resting cell suggests that this conical mass is derived from the peculiar vacuole with granular contents, which was described in the last section as situated in the direct line between nucleus and blepharoplast, and that by the upward migration of the nucleus the vacuole in question is pushed up until it is caught, so to speak, between nucleus and blepharoplast, when, coming under the influence of the forces of attraction or repulsion exerted by the blepharoplast, it assumes the conical form seen. If this is a correct interpretation of the phenomena, the vacuole should, perhaps, be regarded as an archoplasmic vesicle, such as has been described in other cases, and which supplies some part of the material of the achromatic spindle in the mitosis. In fig. 4 it is seen that the flagellum is still present, though short, while in fig. 5 the flagellum has entirely disappeared and the blepharoplast has divided. The disappearance of the flagellum and the division of the blepharoplast are two events which take place independently so far as their relative sequence in time is concerned, that is 626 MURIEL ROBERTSON AND E. A. MINCHIN. to say, the flagellum may disappear completely before the blepharoplast divides or may persist until after this has taken place. In either case the two daughter-blepharoplasts migrate inwards and place themselves on opposite sides of the nucleus in order to become, as will be seen, the two centro- somes in the mitosis. If the flagellum persists during this process of events it remains attached to one of the two blepharoplasts (figs. 6 and 10), and becomes drawn into the body of the cell, as seen in figs. 7-9; in each of these three specimens the flagellum, though greatly shortened, is still persistent, and can be seen passing into the body of the collar-cell and terminating in one of the two blepharoplasts, while the other blepharoplast can be seen on the other side of the nucleus quite independent of the flagellum. On the other hand, figs. 5, 11, and 12 show the two blepharoplasts very close together at the apex of the cell and apparently very recently separated from one another, with no trace of a flagellum. The exact method in which the flagellum disappears is difficult to determine simply by comparison of different stages in sections; it could only be made out satisfactorily by watching the process in the living cell. In collar-cells in which the upward migration of the nucleus is taking place, the flagellum almost always appears much shorter than in the surrounding cells, an appearance too constant in occur- rence to be explained simply as due to artificial curtailment of the flagellum in the process of section-cutting, especially when the collar is intact and the flagellum does not project beyond it (figs. 9 and 10). But a remarkable feature of this stage is the frequent occurrence of a protoplasmic projection, like a small pseudopodium, from the apex of the cell round the base of the flagellum (figs. 7, 9, 39, 40) ; this process persists for atime after the flagellum has completely disappeared (figs. 14, 41). The appearances suggest that the cell throws out a pseudopodial process, by the help of which the flagellum is retracted and absorbed at its base ; DIVISION OF COLLAR-CELLS OF CLATHRINA CORLACEA. 627 in all cases the protoplasmic process in question is very short in proportion to the length of the original flagellum. The division of the blepharoplast takes place with forma- tion of a distinct centrodesmose connecting the two daughter- blepharoplasts (figs. 6, 7, 18). During these changes the collar remains practically unal- tered, except that it begins to show more or less clearly the appearance of shrinkage and degeneration characteristic of the succeeding stages of the division. (5) The Mitosis.—The general course of the mitosis in the collar-cell is similar to that known to occur in the cells of other Metazoa generally, and described for sponges by Maas _(10, 11) and Jorgensen (9). It is unnecessary, therefore, to do more than describe its most characteristic features. As already stated in a previous section, the nucleus of a collar-cell about to divide, but before any changes prepara- tory to division have begun in the chromatin contents, is of a pale type—that is to say, the granules of chromatin dis- tributed over the general framework are very fine and scattered evenly, so as to give the nucleus an almost homo- geneous appearance relieved only by the karyosome, stained a deep black, after iron-hematoxylin, in contrast with the pale grey tint of the remainder of the nucleus (figs. 1, 10,11). The nucleus at this stage is also distinctly larger than the average nucleus of a resting cell. The first changes to be observed in the chromatin contents of the nucleus are that they stain darker and become more blotchy and uneven in appearance, apparently as the result. of the minute granules of chromatin being clumped together to form coarse grains or masses. Figs. 6 and 7 show early stages in this process; the masses of chromatin still stain faintly, appearing to be loose in texture and ill-defined in ontline, and the karyosome stands out sharply. In later stages (figs. 8, 12) the chromatin masses become more definite in outline and somewhat smaller, and the deep stain they take gives the impression that they are more closely knit and of denser texture ; the karyosome, however, is still distinct. VOL. 55, PART 4,—NEW SERIES. 4.2 628 MURIEL ROBERTSON AND E. A. MINCHIN, Finally, the chromatin masses become very definite and stain very deeply, and no distinct karyosome can be made out ; this body seems to break up and to contribute by doing so to the general store of chromatin, At first the chromatin masses, or chromosomes, as they may now be termed, appear to be con- nected together by delicate filamentous junctions (fig. 9) ; this stage corresponds apparently to the spireme stage. Next, the connections between the chromosome disappear, and they are seen lying separately from one another as irregular rounded masses, showing more or less distinctly indications of division, each into two (tig. 15). In spite of much searching we have not been able to find any stages other than those described, and, in particular, nothing more nearly resembling an ordinary spireme stage than the specimen shown in fig. 9. These changes in the interior of the nucleus also go on quite independently of the changes in the flagellum and blepharo- plast described in the previous section. Thus the flagellum may have vanished, and the two daughter-blepharoplasts may have taken up their definitive position when the nuclear con- tents are at the beginning of their changes (fig. 14); or the nucleus may be comparatively far advanced when the blepharoplast has only just divided (fig. 12), or betore the flagellum is absorbed (figs. 8,9). Finally, however, a stage is reached when the nucleus has resolved itself into a mass of separate chromosomes, and the two blepharoplasts, or, as they may now be termed, the centrosomes, are placed on opposite sides of it, indicating the two poles of the future nuclear spindle (fig. 15); when this stage is reached the nuclear membrane is absorbed and cannot be discerned. The formation of the nuclear spindle is seen in the two stages drawn in figs. 16 and 17. After the absorption of the nuclear membrane the chromosomes arrange themselves to form an equatorial plate, to which delicate rays can be seen to pass from the centrosomes, forming the characteristic achromatic spindle. The two centrosomes appear to be pushed further apart by the formation of the spindle, so that they DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 629 come to lie at the extreme surface of the cell. The spindle is lodged in that portion of the cell which we have termed the neck in a previous section, and the centrosomes are situated about midway between the origin of the still per- sistent collar and the flange. The chromosomes appear massed together, and are difficult to distinguish individu- ally when the equatorial plate is seen in side view (figs. 17, 18), but can be seen better in cells cut parallel to the plane of the equatorial plate (fig. 22). The number of chromosomes appears to be about sixteen. At this period, while the equatorial plate is still simple and undivided, an important event takes place. From the cen- trosomes at the two poles of the spindle the two daughter- flagella grow out, appearing as two minute hair-like projec- tions from the surface of the cell (figs. 18-21). This stage is a very common one, and it is, in fact, rare to find a collar- cell with a mitotic spindle without the two daughter-flagella projecting from the two centrosomes ; this indicates that the first formation of the flagella must be an extremely rapid one. Sometimes only one daughter-flagellum is to be seen, but in such cases the cell is usually shghtly oblique, and the missing flagellum has probably been cut off by the knife in cutting the section. ‘The two new flagella are formed entirely outside the original collar, which is still persistent. The condition of the collar is best studied in osmic-picrocarmine preparations (figs. 42—4.5) in which it is seen that the formation of the nuclear spindle causes the cell to become much broader, with the result that the base of the collar is greatly stretched. The thicker portion of the collar, below the hoop, retains its form more or less, but the portion above the hoop tends to collapse and fall together. From the stage with the single equatorial plate the diaster- stage arises in the usual way (figs. 23, 24). It is remarkable that we have succeeded in finding but few specimens of the diaster-stage, and, unfortunately, most of those have been cut obliquely or horizontally, and hence do not show well the relation of this stage to the cell as a whole. Figs. 23 and 24 630 MURIEL ROBERTSON AND E. A. MINCHIN. show the two best diaster-stages we have found. Fig. 23 shows the spindle well, but the cell is cut almost horizontally, and the collar and one daughter-flagellaum are sliced off; in fig. 24 the cell is cut more vertically, and shows the collar, but the plane of the spindle lies obliquely, and only one cen- trosome and daughter-flagellum can be made out. The scarcity of the diaster-stage indicates that it is passed over very rapidly, and this conclusion receives further support from the fact that in the subsequent stages, when the daughter- nuclei are being reconstituted, the daughter-flagella are scarcely longer than they were in the stage with the undivided equatorial plate. After the diaster-stage, and with the reconstitution of the daughter-nuclei, the cell-body begins to divide (figs. 25-28a). Between the two daughter-nuclei there are seen for a time streaks, the remains of the achromatic spindle, stretching across from one to the other (figs. 25-27); these streaks per- sist until the division of the cell-body is far advanced. The details of the reconstitution of the nuclei are difficult to make out clearly ; the chromosomes appear to fuse together into a compact mass in which their individuality is masked, if not lost. The division of the cell is effected by means of a con- striction in the vertical plane, producing a cleavage which is much more marked at the upper than at the lower end of the cell. The cleavage goes right through the old collar, and leads to its destruction and disappearance ; it appears to break down into a granular mass which disintegrates and vanishes. When division of the cell-body is complete the new collars of the daughter-cells grow out round the short but growing flagella. At their first origin the new flagella projected in an oblique direction from the dividing cell, as figs. 18-28 show clearly; they took origin from that portion of the surface of the parent-cell which lies between the flange and the base of the collar. When the division is nearly complete (figs. 28a and 47), the point of origin of the flagella becomes slightly shifted so as to be placed at the uppermost level of DIVISION OF COLLAR-CELLS OF CLATHRINA CORIACEA. 631 the cell, with the result that the young flagella come to point vertically upwards. After complete division the form of the two daughter-collar-cells undergoes a change, becoming elongated in the vertical direction, so that the cell as a whole acquires a slender columnar form, with a shallow collar sur- rounding the short flagellum at the upper end (figs. 29, 30, 48). hal Quart. JounN. Micr. Scr., Vou. 55, N.S. Pu. 38. — | . | | ae eee 5 7 ate, Se ——— oe y " Quart. Journ. Mrcr. Scr, Von. 55, ns. PL, 39. Quart. JouRN. Mier. Sctr., Vou. 55, n.s. Pr. 40. _—s ithe STINGING-CELLS IN CRASPEDOTE MBEDUS#. 763 On the Origin and Migration of the Stinging- Cells in Craspedote Medusz. By Charles L. Boulenger, M.A.(Camb.), Lecturer on Zoology in the University of Birmingham. With Plates 42 and 43 and 5 Text-figures. ConreNTS. PAGE 1. Introduction. ; : ; . 763 2. The Stinging-cells of the Adult Medusa of Merisia . 765 3. Migrating Stinging-cells in other Meduse . a eau 4. The Development of the Medusa of Mcrisia 771 5. General Conclusions : : : BA7ic) 6. Bibliography . : ; Bigs! 1. IyrropucTion. Lewis Murbach (1), in 1894, definitely established the fact that the stinging-cells of the Hydromeduse have the power of active movement in the tissues by the formation of pseudo- podial processes from the cnidoblasts. These observations were confirmed by K. C. Schneider! (2) who published a detailed account of the development of nematocysts in A galmopsis and other Siphonophora, and in his paper stated emphatically that: “Alle Nesselzellen der Siphonophoren enstehen an locali- ' Schneider, as early as 1890, pointed out the fact that developing nematocysts were excessively rare in the tentacles of Hydra, and suggested that they might be formed on the body of the animal. He did not, however, pursue the subject any further (véde “‘ Bibliography,” 8.) VOL. 55, PART 4,—NEW SERIES. D1 764 CHARLES L. BOULENGER. sierten Bildungsherden, von denen aus sie in einem bestimmten Entwickelungsstadium als Wanderzellen auf die Verbrauchs- stiitten tberwandern.” The subject has been recently revived by Jovan Hadzi (4) in a remarkable paper in which he records his observations on the thread-cells of marine hydroids. Hadzi’s results are of the greatest interest, as he was able to examine the living tissues as well as preserved material. His main conclusions are as follows : (1) The thread-cells of hydroids are not formed “in situ” but in the ectoderm of the ccenosarcal branches, where, on account of the thick perisarcal investment, they can obviously not become functional. (2) When completely developed, except for accessory structures such as the cnidocils and the stalks, they migrate to the important nematocyst batteries on the tentacles. This migration can take place in two different manners. In simple forms, e.g. Campanularia, the thread-cells move actively by means of their pseudopodia, making their way between the ectodermal cells of the colony. In Tubularia however, they adopt a quite different method of locomotion : from the ectoderm of the coenosare they force a way through structureless lamella and endoderm into the cavity of the hollow stem, whence they are carried by the current caused by the flagella of the endoderm cells to the hydranths. Here the thread-cells re-enter the tissues and migrate actively by their own movements to the ectoderm of the tentacles. In a recent paper, whilst describing the structure of the Egyptian lacustrine medusa, Mcerisia lyonsi (5), I called attention to the fact that large nematocysts were to be found in abundance among the endoderm cells of the manubrium. Being at a loss to account for their presence in this position I cut sections of a large number of specimens, careful examination of which convinced me that I was dealing with a case similar to that investigated by Hadzi in Hydroids. As this phenomenon has not been described previously in ——ES STINGING-CELLS IN CRASPEDOTE MEDUSA. 765 Medusze, I have endeavoured in this paper to give as complete an account as possible of the origin and distribution of the nematocysts of this form. The material used for this investigation was collected by Dr. Cunnington and myself in Lake Qurun, and was carefully fixed either with osmic acid or with hot corrosive sublimate. Sections were cut by the ordinary paraffin method and a number of stains were tried, the best results being obtained with hematoxylin followed by eosin; this produced an excellent double-stained effect, the eosin bringing out the nematocysts and rendering them most conspicuous. Borax carmine followed by picro-indigo-carmine was another good differential stain and iron-hematoxylin was useful when examining sections of the developing Medusee. The work in connection with this paper was carried out partly in the Morphological Laboratory at Cambridge and partly in the Zoological Laboratory of Birmingham University. I wish to express here my sincere thanks to Professor F. W. Gamble, who very kindly read through my manuscript and made many valuable suggestions. 2. Tur STINGING-CELLS OF THE ADULT MeEpusA or Marista. As mentioned above, a striking feature of the anatomy of this medusa is the presence of numerous thread-cells! in the endoderm at the base of the manubrium. At first it seemed possible to account for their occurrence in this unusual position by assuming that these stinging-capsules were used ones taken in by the jelly-fish together with its food. On careful consideration this view was found to be quite un- tenable, for— 1 The nomenclature of the different parts of the stinging-cells is somewhat cumbrous and complicated; moreover, the various names have been used very loosely. In this paper I have employed the terms thread-cell or stinging-cell for the whole structure comprising the nematocyst (the actual stinging capsule), and nematoblast (the cell in which the former is embedded, and of which the enidocil and the stalk are parts). 766 CHARLES LL. BOULENGER. (a) The nematocysts found in the endoderm are always undischarged. (b) Favourable sections show them to be accompanied by their nematoblasts. (c) The nematocysts are never to be found near the free margins of the endoderm cells, but, for the most part, between the more basal portions of these cells near the structureless lamella. These thread-cells can, obviously, not become functional in this position, and the only possible explanation of their occurrence here is that they are making their way from their place of origin to some battery where they can be of use. At this point it may be well to review the distribution of stinging-cells in the ectoderm of the manubrium. The chief battery is situated around the mouth-opening; here the thickened ectoderm formsa circular lip crowded with nemato- cysts, and constitutes a powerful organ of offence (Pl. 42, figs. 3and 4). ‘lhe ectoderm of the remainder of the manu- brium proper consists of a single layer of low, closely fitting epithelial cells with occasional isolated nematocysts ; it is to be noticed that here, as well as on the oral lip, interstitial cells are completely absent. At the base of the manubrium is the broad stomach, the ectoderm of which is considerably thickened and forms the conspicuous gonad. Interstitial cells and developing thread-cells being absent from the more distal parts of the manubrium, the question arises— Where are the nematocysts of the oral battery formed, and how did they attain their position in this region? An answer is, I think, afforded by the study of the distribution and arrangement of the nematocysts in the manubrial endo- derm. ‘The greatest number of these are to be found just below the region of the gonad, where, in most specimens, numerous thread-cells are to be met with among the large digestive cells of the endoderm. In this position one can usually find a number of dark-staining interstitial cells, some of which contain rudiments of stinging-capsules, and are obviously nematoblasts (Pl. 42, figs. 1 and 2). STINGING-CELLS IN CRASPEDOTE MEDUSA. 767 In the more distal parts of the manubrium we find nemato- eysts to occur less abundantly, and their position in the endoderm is very regular, the longer axes of the capsules being parallel with the structureless lamella and their broader ends directed towards the mouth of the medusa (PI. 42, fic. 3). Previous authors have shown this orientation to be characteristic of migrating thread-cells, and we must come to a similar conclusion; namely, that they are making their way from the base of the manubrium to the oral battery. This view is confirmed by an examination of the tissues of the mouth region, where one can often find thread-cells actually forcing their way through the structureless lamella to the oral battery. Here they take up their definitive position and develop accessory structures, e.g. cnidocil and stalk, from the nematoblast. A stinging-cell occasionally turns aside before reaching the oral region (Pl. 42, fig. 4), and passing through the lamella, forms one of the isolated nematocysts to be met with in the more proximal parts of the manubrial ectoderm. The route followed by the thread-cells of the medusa is readily explained. These structures, when the nematocysts are completely developed, are of considerable size, whereas the ectoderm of the manubrium is very low, and, moreover, forms a very definite epithelium of closely fitting cells, between which the large stinging-cells could scarcely force a passage. We need, therefore, not be surprised that they adopt the much easier way between the large and _ loosely packed cells of the endoderm. From the above account it appears, therefore, that in Meerisia the nematocysts of the oral battery of the medusa are developed in the endoderm at the base of the manubrium ; this does not necessarily imply that the nematoblasts are themselves endodermal in origin, as will be explained in the section of this paper which deals with the development of the medusa-bud. In addition to that surrounding the mouth opening, the main nematocyst batteries of the medusa are situated on the 768 CHARLES LL. BOULENGER. four perradial tentacles suspended from the umbrella edge. These tentacles are slender and of great length when fully extended; at their bases they are swollen to form the very conspicuous ocellar bulbs, each of which bears on its ex- umbrellar surface a bright red eye-spot. The tentacles are hollow, their cavities being continuous with that of the circular canal; the ectoderm is thickened at regular intervals to form conspicuous transverse rings crowded with nemato- cysts, and becoming very noticeable and almost bead-shaped when the tentacles are fully extended. On examination of sections and maceration preparations of these organs, one is again struck by the almost complete absence of nematoblasts or other interstitial cells, and we are driven to the only possible conclusion, namely, that the stinging-cells have developed elsewhere and have migrated to the batteries on the tentacles. The large, eye-bearing bulbs at the bases of the tentacles immediately suggest themselves as possible nematocyst “factories,” and sections of these structures show that such a function must be assigned to them (Text-fig. 1). An ocellar bulb consists of a mass of thickened ectoderm crowded with small, irregularly shaped cells and nematocysts in various stages of development. The fully formed thread- cells are devoid of enidocils or other accessory structures, and the capsules are never orientated so as to lie at right angles to the surface; we must, therefore, conclude that they do not become functional in this region. In the centre of the bulb the nematocysts lie in all directions, but near the base of the tentacle we find a distinct tendency for these organs to be arranged with their longer axes parallel with the structureless lamella, a position, as mentioned above, characteristic of migrating thread-cells. The above-mentioned facts lead us to the conclusion that the stinging-cells of the tentacles, like those of the oral battery, are not developed “in situ,” but migrate into these organs from “ factories” situated ina more central position on the medusa, in this case from the ocellar bulbs, whence a STINGING-CELLS IN CRASPEDOTE MEDUSA. 769 TEXT-FiG. 1. A longitudinal section through the ocellar bulb and the base of a tentacle of Mcerisia lyonsi. tent.b. Ocellar bulb. nem. Nematocyst migrating into the tentacle. TEXT-FIG. 2. Section of the umbrella edge of Merisia lyonsi showing the velum (vel) and part of an ocellar bulb (tent.b.) nem. Nematocyst migrating towards the edge of the velum, 770 CHARLES LL. BOULENGER. continual stream of thread-cells are being poured forth. ‘The majority of these are obviously on the way to their tentacular batteries, although occasionally one may wander into the velum, as shown in T'ext-fie. 2. 3. MIGRATING STINGING-CELLS IN OrHeR Mepusm. In the preceding paragraph I have attempted to prove that the conspicuous bulbous swellings which occur so constantly as the bases of the tentacles of craspedote Medusz have an important function besides that of bearing the ocellar sense- organs. In such craspedote Medusz as are devoid of tentacle- bulbs, e.g. the Trachomedusez and Narcomeduse, we find that the edge of the umbrella is provided with a special thickened 1ing of ectoderm, containing stinging-cells, some- times known as the “ nettle-ring.” Further, those forms in which the tentacles take their origin some distance from the margin of the bell on the exumbrellar surface are provided with special bands of nematocysts, called peronia, which connect the above-mentioned uettle-ring with the bases of the tentacles. These facts make it very tempting to assume that the marginal ring of nematoblasts replaces the ocellar bulbs in function, and reference to the figures of this organ, given by various authors, seems to show that this assumption is probable correct. It is a point which requires special investigation, and I will at present merely refer to the evidence which is at my disposal. The Hertwigs’ most accurate figure of the umbrella edge of Carmarina (6, Pl. iv, fig. 5)! shows the nettle-ring to be packed with thread-cells without definite orientation ; at the base of the tentacle, however, a number of nematocysts are drawn arranged in such a manner that there can be little doubt that they are migrating from the marginal ring to the batteries on the tentacle. I have examined sections through the tentacles of a medusa of the same genus, and 'T should like to express my indebtedness to Dr. 8. F. Harmer, F.R.S.. for calling my attention to this figure. STINGING-CELLS IN CRASPEDOTE MEDUS. rival these showed the same orientation of nematocysts as in the specimen figured by the Hertwigs. I have figured one of these sections (‘Text-fig. 3) chosen from a series in the Cam- bridge Morphological Laboratory ; comparison with that of Mecerisia (Text-fig. 1) is very instructive. TEXT-FIG. 3. A longitudinal section through the base of a tentacle of Car- marina sp. vel. Velum. nem. Nematocyst migrating through the ectoderm of the tentacle. Giinther’s figure of Limnocnida (7, fig. 6) shows that a similar migration of thread-cells must occur in that medusa. 4. Tue DEVELOPMENT or THE MeEpuUsA oF Marista. As shown above, the nematocysts of the main stinging batteries of Mcerisia are formed in two quite distinct positions in the medusa: (a) The manubrial endoderm, (>) the ecto- derm of the ocellar bulbs. In order to properly understand the origin of these VOL. 05, PART 4.—NEW SERIES. 52 Tae CHARLES L. BOULENGER. different situations of the stinging-cell factories it is necessary to examine the development of the medusa in some detail. Until recently the accepted view of the development of the gonophores of the Hydromedusz was based essentially on L. Agassiz’s observations on Syncoryne mirabilis, published in 1862 (8). His account of the process was confirmed by Hertwig (9), Weismann (10), and almost all later workers on the same subject, and is essentially that to be found in the majority of modern text-books. ‘The following description of the development of the medusa of Bougainvillea is taken from one of the latter (18), and represents the prevailing ideas on the subject : The medusa-bud makes its first appearance as a simple hollow bud formed by the evagination of the two layers of the mother-polyp. Multiplication of the ectodermal cells at the apex results in the production of a lens-shaped mass of small cells which sinks below the level of the superficial ectoderm, pressing the endodermal wall in front of it into the shape of a cup. This mass of ectoderm is called the entocodon (Glockenkern), and a cavity which appears in its interior is the rudiment of the subumbrella cavity. It is followed by an invagination of the superficial ectoderm, the wall between the new cavity thus formed and the subumbrella cavity being the future velum. Growth of this subumbrella cavity results in an approximation of the endodermal walls of the coelenteron, and these ultimately fuse into an endoderm lamella except where the circular and radial canals are to lie. The upgrowth of the manubrium from the floor of the subumbrella cavity, the formation of the tentacles and the perforation of the velum and manubrium complete the formation of the medusa. A. Goette (11) has recently made a thorough examination of the development of the gonophores of Podocoryne carnea and a large number of other hydroids, and has published a long and elaborate paper on the subject. As the result of his investigations this author concludes that the current views on the origin of these structures are quite STINGING-CELLS IN CRASPEDOTE MEDUSA. 773 erroneous, and states that carefully cut series of sections of developing medusa-buds show that a double-walled eup of endoderm is not present at any stage; moreover, the four radial canals arise from four unconnected pouches of endo- derm which grow out separately, although simultaneously, from the ccelenteron of the bud, and are completely indepen- dent of the entocodon. The endoderm lamella is formed later by the lateral extensions of the solid edges of these pouches, which finally fuse with one another. Again, an invagination of the superficial ectoderm does not take place and the forecast of the velum is present at a quite early stage, and is then represented by the flattened apex of the bud, where the superficial ectoderm and the distal wall of the entocodon come into contact with one another. Goette’s paper has not received (at any rate in this country) the attention which so important a communication deserved, and the only confirmation of his results is that of his pupil, Walter Richter (18), who, acting on his professor’s advice, worked cut the development of the gonophores in Rhizo- phora, Physalia, and other Siphonophora and described a similar origin for these structures in this division of the Hydromedusz. In my account of the anatomy of the hydroid stage of Meerisia I did not go into this subject with any detail, but merely stated that the development of the medusa-buds seemed quite typical, the growth of the cavity in the ento- codon causing the approximation of the endodermal walls of the bud, The examination of a large series of sections during my investigation of the origin of the nematocysts has shown me that this statement was erroneous, and that the development of the medusa of this form agrees very closely with that of Podocoryne carneaas desribed by Goette. My error, like that of other writers on the same subject before Goette, was due to the use of optical sections, and partly to the examination of single sections of the buds instead of complete series. 7174 CHARLES L. BOULENGER. In Meerisia lyonsi the medusa-buds are to be found scattered irregularly on the broadest region of the hydranth between the bases of the tentacles (Text-fig. 4), thus differing in position from the asexual lateral buds, which are restricted in the majority of cases to the more proximal parts of the body. TrxtT-FIG. 4. Outline sketch of a hydranth of Merisia lyonsi to show developing gonophores and a small asexual bud. x 30. tent. Tentacle. m.b. Gonophore with conspicuous ocellar bulbs. as.b. Asexual lateral bud. The ectoderm of this region is somewhat deeper than in other parts of the hydroid, the boundaries of the large musculo-epithelial cells are difficult to detect, and the whole tissue is crowded with interstitial cells, for the most part nematoblasts, containig nematocysts i various stages of development. The endoderm consists of large vacuolated a EE STINGING-CELLS IN CRASPEDOTE MEDUSA, 715 digestive cells, between which are numerous characteristic gland-cells with coarse granular contents which stain deeply. The first indication of a developing medusa-bud is to be traced in the ectoderm, an accumulation of interstitial cells causing this layer to project slightly outwards. The endo- derm soon begins to take part in this bulging out of the tissues, and owes its increase in area chiefly to the prolifera- tion of the large cells, but partly also to the accumulation of interstitial cells, which are to be found in the endodermal tissue in the region of a developing bud. These cells I believe to be ectodermal in origin, for favourable sections show occasional interstitial cells to migrate from the ectoderm through the structureless lamella into the endoderm. In this way a hollow, double-layered bud is formed (Pl. 43, fig. 5) by a process which cannot be called one of simple evagination, but in some respects resembles that of the formation of the early stages of the lateral buds in Hydra, as recently des- cribed by J. Hadzi (14). As long ago as 1891, W. B. Hardy (15) showed that in the early development of the gonophores of Myriothela phrygia there was a certain mixing up of endodermal and ectodermal cells to form a kind of blastema, and it seems probable that further investigations will prove that the pro- duction of a bud from the body of a hydroid is by no means so simple a process as has been made out by some authors. The entocodon is next formed by the proliferation of the ectoderm at the apex of the bud, and consists of a small-celled plug of tissue between ectoderm and endoderm. Four pouches of endoderm are arising simultaneously from the ccelenteron ; from them the radial canals of the adult are to be derived. Reference to fig. 6 will show that there is nothing of the nature of a double-walled endodermal cup in the bud, one side of the obliquely cut section showing a radial pouch, the other the contact of the entocodon with the superficial ectoderm. It is to be noticed that this superficial ectoderm has not 776 CHARLES L. BOULENGER. changed in character and is identical in structure with that covering the hydranth, consisting of large epithelial cells, interstitial nells, and nematoblasts, with occasional nemato- cysts. The independent origin of the four radial pouches of endo- derm is still more obvious in figs. 7, 8, and 9, which are three sections in different planes of a slightly later stage. In the transverse section (fig. 7) the entocodon is seen to be roughly square in section, being in contact with the super- ficial ectoderm at the four corners (interradii); the four perradial pouches are thus completely separated from one another. A median longitudinal section (fig. 8) through the perradii at this stage shows, of course, two of the endodermal pouches separated by the hollow entocodon. As pointed out by Goette, it is from the examination of such a section, independently of others of the series, that the idea arose that a double-walled cup of endoderm was formed by the growth of the entocodon. A tangential section taken a short way on either side of this median section will naturally show a single pouch only, as illustrated in fig. 9. In this stage the forecast of the manubrium is already conspicuous, and is, of course, clothed externally by the proximal wall of the entocodon. The four endodermic pouches continue their growth out- wards to the very tip of the bud, and at their terminations push out the ectoderm, causing the formation of four perradial buibous projections, which are the forecasts of the ocellar bulbs. A section, therefore, taken through a perradius gives rise to the false idea of an invagination of ectoderm towards the entocodon (Pl. 48, fig. 10). The four bulbs are very conspicuous features of the external anatomy of the medusa, even at this relatively early stage of development (Text- fig. 4). The formation of the endoderm lamella is exactly as described by Goette for Podocoryne carnea; the central part of each endodermal pouch becomes a radial canal, the large cells at the edges growing out to form two solid wings of endoderm, which meet similar projections from the other oa. STINGING-CELLS IN CRASPEDOTE MEDUSZ&. ray pouches at the interradiu (Pl. 43, fig. 7, r.p.e.!). The ring- canal is formed by the fusion of the distal ends of the radial pouches at the bases of the bulbous swellings referred to above. Up to this point the histology of the two layers has been quite constant; the superficial ectoderm has retained its original character and remains crowded with interstitial cells of all kinds, in striking contrast with the small-celled regularly arranged tissues derived from the entocodon. The endoderm lining both the ccelenteron and the radial pouches consists of large clear cells, with somewhat indefinite outlines and containing numerous large nutritive spheres, which stain deeply with iron-hematoxylin; a few irregularly shaped interstitial cells are to be found, most numerous between the endoderm cells hning the manubrium. In the last stage of the development described above we found all the organs of the adult medusa already well defined, with the exception of the tentacles. From this point onwards the more important changes are to be found in the structure of the umbrella, which now grows rapidly, especially in the region between the ocellar bulbs and the base of the manu- brium, so that the superficial ectoderm loses its characteristic features, as noticed above, and gives rise to a low, small- celled epithelium covering the external surface of the bell. The endoderm behaves in a somewhat similar fashion. The ocellar bulbs, however, remain unaltered; the endoderm still consists of large irregular cells with nutritive spheres; the ectoderm is still crowded with interstitial cells, thread-cells, and nematoblasts, the latter increasing rapidly and forming new nematocysts, both large and small (Text-fig. 5). The ocellar bulbs give rise to the tentacles, their main function being obviously that, already mentioned on p. 768, of supplying these organs with stinging-cells. In the preceding paragraphs I have tried to emphasise the fact that beyond an increase in the actual number of cells, the ectoderm of this region has remained practically un- changed throughout the development of the gonophore. The 778 CHARLES L. BOULENGER. tentacular nematocysts of the fully-formed medusa thus arise in the interstitial cells derived from the ectoderm of the parent hydroid, The ocellar bulbs are, of course, retained throughout the life of the medusa, and, as mentioned above, keep on supply- ing the tentacles with stinging-cells; they are no doubt especially active during the regeneration of these organs. This explains the constant presence of such swellings at the THXT-FIG. 0. A longitudinal section of an ocellar bulb of Mceerisia lyonsi just before the development of a tentacle. vel. Velum. nem. Small nematocyst. nem.’ Large nematocyst. bases of the tentacles of the Hydromeduse, as well as their early appearance and relatively enormous size in the develop- ing medusa-buds. The function of bearing the ocellar sense- organs must bea secondary one, for such bulbs are conspicuous in the formation of medusze which do not possess ocelli, e. g. Podocoryne carnea, concerning which Goette (11, p. 19) remarks: “ Bald nach der Fertigstellung des Velum verdickt sich das Ectoderm jedes Randwulstes dicht iiber dem Velum zu | = STINGING-CELLS IN CRASPEDOTE MEDUSA. 779 einem vorspringenden Polster, das den Ocellarbiidungen andrer Medusen entspricht, aber, wie schon die Alteren Beobachter (Allman, 16; Grobben 17) feststellten, keine Ocellen entwickelt.” Ina young medusa of Mcerisia a short time before its liberation the manubrium is still without a mouth opening, and is clothed externally by a single layer of low ectodermal cells (Pl. 43, fig. 11), the internal lining consisting of large clear endoderm cells containing nutritive spheres and occa- sional irregularly shaped interstitial cells. The latter become more numerous as development proceeds, and some can be clearly recognised by their enlarged nuclei to be sex-cells. These at a later stage, no doubt, become transferred to the ectoderm of the stomach region, and by their further division form the gonad. Owing to the absence of individuals of the right age, [am unable to state how the transference of sex- cells from one layer to another takes place. I have never met with them migrating through the structureless lamella, and it is quite possible that the transference is a passive one, similar to that described by Goette in the male gonophores of Hydractinia (11, p. 70). In the youngest free-swimming medusz examined by me the endoderm of the slightly swollen stomach had lost its small cells, and was separated by a very thin lamella from the ectoderm, which contained a few rows of developing sex-cells. The endoderm slightly distal to this region had retained a number of interstitial cells, some of which prove to be obvious nematoblasts and contained developing nematocysts. These are, of course, the rudiments of the fully formed stinging-cells, which, as described in the first part of this paper, are to be found in the endoderm, just below the stomach of the adult medusa, and which later migrate to the battery at the oral extremity of the manubrium. From this we must infer that the nematoblasts of the manubrium arise in the endoderm of the developing gonophore in exactly the same way as do the sex-cells; like the latter they are able to migrate through the tissues of the medusa. 780 CHARLES L. BOULENGER. When we remember the similar origin of the two kinds of cells from undifferentiated interstitial cells, we need not be surprised that they both possess the same powers of active movement. That the thread-cells are identical in origin with the sex- cells is further emphasised by the fact that in exceptional cases part of the testis of Moerisia can give rise to a nemato- cyst battery instead of producing sperm-cells, as shown in Pl. 43, fig. 12. Both kinds of cells are first to be recognised in the endo- derm of the medusa-bud; this does not necessarily imply that they originate in that layer ; in my account of the early development of the gonophore, I showed that interstitial cells of the ectoderm occasionally migrate through the structureless lamella of the hydranth and become incorpor- ated among the proliferating cells of the endoderm. It is probable that these cells or their derivatives give rise to the sex-cells and nematoblasts. In my description of the anatomy of Mcerisia lyonsi (8), I mentioned that exactly the same types of nematocysts were to be found in the medusa as in the hydroid; in this paper I hope to have proved that they are not only identical in structure, but actually originate from the same cells. This fact is one which might be of use in systematic work on the Hydromeduse, where the assignment of Medusz to hydroids is often only a matter of inference ; a careful com- parison of the nematocysts of the two stages should be of great value in this connection. 5. GENERAL CONCLUSIONS. (1) The stinging-cells of the medusa of Mcerisia lyonsi are not developed “in situ” on the principal batteries, but migrate to their final positions on the oral lip, or on the tentacles. (2) The stinging-cells of the oral battery are formed in the endoderm of the manubrium, just below the stomach ; those ts STINGING-CELLS IN CRASPEDOTE MEDUSA. 781 of the tentacles in the ectoderm of the conspicuous ocellar bulbs at the terminations of the radial canals. (3) There is reason to believe that the bulbous swellings at the bases of the tentacles have this function throughout the craspedote Medusz. In the sub-divisions Trachomeduse and Narcomedusex, they are probably replaced by the thickened ring of thread-cells on the margin of the bell. (4) The development of the gonophores of Mcerisia takes place in the manner described by Goette for other hydroids, There is no double-walled cup of endoderm at any stage, the radial canals and the endoderm lamella being derived from four separate pouches of endoderm, which grow out simul- taneously from the ccelenteron of the simple bud. (5) The stinging-cells of this medusa are developed from cells, which, like the sex-cells, arise directly or indirectly from the ectoderm of the parent hydranth. BIRMINGHAM, June 19th, 1910. BIBLIOGRAPHY. This bibliography includes only those works actually mentioned in the text of my paper; for a more complete list of literature I must refer the reader to the memoirs of Hadzi (4) and Goette (11). 1. Murbach, L.—* Beitrage zur Kenntnis der Anatomie und Entwicke- lung der Nesselorgane der Hydroiden,” ‘ Arch. f. Naturg.,’ Jahrg. 60, Bd. i. 2. Schneider, K. C.—‘* Mittheilungen itber Siphonophoren: V, Nessel- zellen,” ‘ Arb. Zool. Inst. Wien,’ Tom. xii, 1900. “Histologie von Hydra fusca,” ‘Arch. f. Micr. Anat.,’ Bd. xxxv, 1890. 4. Hadzi, J—‘Ueber die Nesselzellwanderung bei den Hydroid- polypen,” ‘ Arb. Zool. Inst. Wien,’ Tom. xvii, 1907. 5. Boulenger, C. L.—“ On Merisia lyonsi, a new Hydromedusan from Lake Qurun,” ‘ Quart. Journ. Mier. Sci.,’ vol. 52, 1908, 6. Hertwig, O., and R.—‘ Das Nervensystem und die Sinnesorgane der Medusen,’ Leipzig, 1878. 7. Ginther, R. T.—‘A further Contribution to the Anatomy of Limnocnida tanganyice,” ‘Quart. Journ. Micr. Sci.,’ vol. 36, 1594. 782 CHARLES L. BOULENGER. 8. Agassiz, L.—‘ Contributions to the Natural History of the United States of America,’ vol. iv, Boston, 1862. 9. Hertwig, O., and R.—‘ Der Organismus der Medusen und seine Stellung zur Keimblattertheorie,’ Jena, 1878. 10. Weismann, A.—‘ Die Enstehung der Sexualzellen bei den Hydro- medusen,’ 1883. 11. Goette, A.— Vergleichende Entwicklungsgeschichte der Gesch- lechtsindividuen der Hydropolypen,” * Zeit. f. wiss. Zool., Bd. Ixxxvii, 1907. “Ueber die Entwicklung der Hydromedusen,” ‘ Zool. Anz.,’ Bd, xxvii, 1904. 13. Richter, W.—‘ Die Entwicklung der Gonophoren einiger Sipho- phoren,” ‘ Zeit. f. wiss. Zool.,’ Bd. lxxxvi, 1907. 14. Hadzi, J—‘ Arb. Zool. Inst. Wien,’ Tom. xviii, 1909. 15. Hardy, W. B.—* The Histology and Development of Myriothela phrygia,” ‘Quart. Journ. Mier. Sci.,’ vol. 32, 1891. 16. Allman, G. J.—‘ Monograph of the Gymnoblastic or Tubularian Hydroids,” ‘ Ray Soc.,’ 1871-2. 17. Grobben, C.—‘ Ueber Podocoryne carnea,” ‘Arb. Zool. Inst. Wien,’ Tom. ii, 1875. 18. Fowler, G. H.—‘ Hydromedusee” in Lankester’s ‘Treatise on Zoology,’ vol. ii, 1900. EXPLANATION OF PLATES 42 anp 43, Illustrating Mr. C. L. Boulenger’s memoir ‘‘On the Origin and Migration of the Stinging-cells in Craspedote Meduse.” PLATE 42. EXPLANATION OF LETTERING. ect. Ectoderm of the manubrium. end. Endoderm of the manubrium. gl.c. Gland-cell. nem. Endodermal nematocyst. nem.' Nematocyst of the oral battery. nem.2 and nem.s Migrating nematocysts. s.l. Structureless lamella. test. Testis. Fig. 1—A longitudinal section of the proximal part of the manubrium of the medusa, Merisia lyonsi, to show the nematocysts in the endoderm. STINGING-CELLS IN CRASPEDOTE MEDUSA. 783 Fig. 2.—A transverse section through the same region. Fig. 3.—A longitudinal section of the distal part of the manubrium showing the oral battery and a stinging-cell (nem.’) migrating through the endoderm towards it. Fig. 4.—A similar section showing a stinging-cell (nem.*) making its way through the structureless lamella to the ectoderm. PLATE 48. EXPLANATION OF LETTERING. c.b, Cavity of the medusa-bud. c.e. Cavity of the entocodon, i.e. subumbrella cavity. c.m. Cavity of manubrium. ect. Superficial ecto- derm of the developing gonophore. end. Endoderm of the same. ent. Ectoderm of the entocodon. g.c. Gland-cell. 7.c.e. Interstitial cell of the endoderm. xem. Small nematocyst. nem.' Large nematocyst. n.s. Nutritive sphere of the endoderm. 7.p.e. Radial pouch of endo- derm. r.p.e.! Lateral solid entension of the same, which later forms the endoderm lamella. s.c. Sex-cell. tent.b. Tentacle-bulb. fest. Testis. v. Velum. Fig. 5.—Longitudinal section of an early stage in the formation of the gonophore of Merisia lyonsi (cf. text, p. 775). Fig. 6.—Tangential longitudinai section of a young bud showing the entocodon and a single radial endoderm pouch. Fig. 7.—Transverse section of an older gonophore to illustrate the complete independence of the four radial pouches. The entocodon already has a large cavity (subumbrella cavity), and at 7.p.e.' can be seen the solid extension of the edge of a pouch which later forms the endo- derm lamella. Fig. 8.—Radial longitudinal section through a similar (slightly younger) bud, showing two radial pouches of endoderm separated by the entocodon. Fig. 9.—Tangential longitudinal section of the same medusa-bud ; only asingle pouch is shown. Fig. 10.—Longitudinal section of an almost completely developed medusa to show the bulbous swellings at the termination of the radial pouches. Fig. 11.—Manubrium of the same medusa under a higher magnifi- cation. Fig. 12.—Section through the testis of an adult medusa, part of which has given rise to a stinging-cell battery. IJuant.Lourn. Mier Se. tb, 55 NS Cee. 59 Sie alo) 2 AANA, eRe Hoth Lith? London. C.L.Boulenger, del. a « a ee oe Se a ee =a ae ek aw > er nn vat a i ee 8 a i Si C.L.Boulenger, del fi 2 rey, ae &, SEOCOos (m) ren Oe es en ay als D ce) aD A O09 is » ES Quart. Sourn Micr:Ser.VA.b5 NSM. Fhith Lith? London Ay. MUTATIONS IN CRUSTACEA OF THE FAMILY ATYIDA. 785 The Researches of Bouvier and Bordage on Mutations in Crustacea of the Family Atyide. By Ww. TT. Calman, D.Sc., of the British Museum (Natural History). With 4 Text-figures. Some six years ago Professor H. L. Bouvier (?04, 705)! called attention to the remarkable dimorphism of certain tropical river-prawns of the family Atyide, which he compared with the phenomenon of mutation described by de Vries in the vegetable kingdom. He pointed out that the case was espe- cially noteworthy, not only because of the marked discon- tinuity and constant occurrence of the variations, but also because they affected characters regarded as distinctive of genera; and he drew the conclusion that these genera had originated by a process of mutation. M. EH. Bordage has recently published (’08, 09a, 7098) the results of some obser- vations and experiments on the living animals which seem to support Bouvier’s views, and to indicate, at all events, a promising field for further investigations. At the suggestion of Sir Ray Lankester the following account has been pre- pared in the hope that it may induce some naturalists, who have the opportunity of studying the animals under natural conditions, to give attention to the matter. The Atyidz (see Text-fig. 1) are a family of Decapod Crus- tacea belonging to the tribe Caridea (whichincludes most of our common prawns and shrimps), and are widely distributed in fresh waters in the warmer regions of the globe (see Ortmann 1 The numbers refer to the list of papers on p. 796. 786 W. T. CALMAN. 94 and Bouvier 705). Some.of the members of the family show very primitive characters, having, for instance, swim- ming branches or exopodites on all the thoracic limbs, as im the so-called “Schizopods.” In this and in other features they resemble the deep-sea Hoplophoride, from which, or from some allied forms, most authorities are agreed in con- sidering them to have been derived. Other members of the family, however, are considerably specialised. In some characters this specialisation has pro- ceeded along lines parallel to those followed in other series of the Caridea—for example, in the progressive disappearance Trxt-FiG. 1. Atya bisuleata. Ovigerous female of the Atya-form. x3. From a specimen in the ‘* Challenger” collection from Honolulu. of the exopodites and, later, of the epipodites of the legs, and a diminution in the number of the branchiz. In other characters specialisation has followed lines peculiar to the family, and thisis especially the case with the modifications of the chelate first and second pairs of legs. In nearly all Atyidee these limbs are comparatively small, not dissimilar in size, and have the fingers each tipped with a brush of long hairs (Text-fig. 1). Fritz Miiller (’92) has described how these brushes are used in collecting pellets of mud on which the animal feeds.! Among the more specialised members of the 1 T do not understand Bordage’s statement that the chele are used for excavating burrows in the mud, for which their structure would appear to be ill-adapted. MUTATIONS IN CRUSTACEA OF THE FAMILY ATYIDE. 787 family the characters used as distinctive of the genera are chiefly drawn from the modifications of the chelipeds, and some of these may now be considered in fuller detail. In the very numerous species of the genus Caridina (Text- fig. 2) the chele themselves do not differ greatly, except in carrying brushes of set, from the typical form found in many other Decapods. The dactylus (d.) or terminal segment of the limb, forming the “ movable finger,” is opposed to a thumb-like process (‘ immovable finger ’’) of the penultimate TEXT-FIG. 2. Caridina nilotica var. 1, 2, first and second chelipeds. ¢., carpus; d., dactylus ; p.. palmar portion of propodus. x 40. - From a specimen collected by Dr. W. A. Cunnington in the Victoria Nyanza. segment or propodus. ‘he proximal part of the propodus, expanded to contain the muscles moving the dactylus, forms what is known as the “ palm” (p.) of the chela. In Caridina the two pairs of chelipeds differ in the form of the segment which supports the propodus, the “ wrist” or carpus (c.). In the second pair it is more or less elongated and slender, and the propodus articulates with its distal end; in the first pair, on the other hand, it is short and broad, its distal margin is more or less concave (cf. Text-fig. 2,1, and Text-fig..4, 4’), and the propodus articulates with its lower corner. VOL. 55, PART 4.—NEW SERIES. 53 788 W. T. CALMAN. The species of the genus Ortmannia (formerly known as Atyoida) differ from those of Caridina chiefly in the fact that the carpus of the second pair resembles that of the first pair (Text-fig. 3, B’, B”’), being short and broad, with its distal margin excavated and articulating with the propodus at its lower corner. It is to be noted that these characters are not equally well marked in all the species referred to Ortmannia; in some the second carpus is still, as in Caridina, somewhat longer than the first, and the excavation of its distal margin is shallow (asin Text-fig. 4, B”) ; in other species the carpus is nearly similar in the two pairs and so deeply excavated as to assume an almost crescentic form (as in Text-fig. 3, B’, BY’). Associated with this excavation of the carpus is a shifting (already begun in Caridina) of the carpo-propodal articulation from the proximal end to the lower border of the propodus. Further, while in some species the chele themselves are quite is much “ce ? similar to those of Caridina, in others the “ palm’ shortened, or, in other words, the articulation of the movable finger is carried backwards towards the base of the propodus. These modifications lead towards the conditions found in the genus Atya, which includes the largest and most highly specialised members of the family. In these the two pairs of chelipeds (‘l'ext-fig. 3, A’, A”) are quite similar, and the carpus is reduced by the excavation of its distal border to a narrow crescent, with the lower limb of which the propodus articulates. The propodus itself assumes a form unlike that of any other Decapod; the backward shifting of the articulation of the dactylus has been carried so far that the palm has entirely disappeared, and the chela is composed of two similar parts, hinged together at one end, like the legs of a pair of compasses. Although, within each of the genera, there is some varia- tion in the degree to which these characters are developed, this variation is so far discontinuous that all the known species could, prior to Bouvier’s researches, be referred without much difficulty to one or other of the genera. If it be objected that such apparently trivial differences should MUTATIONS IN CRUSTACEA OF THE FAMILY ATYIDM. 789 not be regarded as of generic value, it may be pointed out that, as a rule, though not in every case, they are coincident with other features which help to characterise, although they do not define, the generic groups; and further, there is no criterion by which the generic value of a character may be estimated, except that of its constancy throughout a croup of species. Bouvier’s discovery may be shortly expressed by saying that certain species were found to be dimorphic and to oscillate, as it were, in a state of unstable equilibrium between one generic group and the next. Thus, Miss Rathbun (’01) THXT-FIG, 3. Atya bisuleata. ” ” GOVERNORS (Life Members of Council) 500 0 0 Members have the following rights and privileges :—They elect annually the Officers and Council; they receive the Journal free by post; they are admitted to view the Laboratory at any time, and may introduce friends with them ; they have the first claim to rent a table in the Laboratory for research, with use of tanks, boats, ete. ; and have access to the Library at Plymouth. Special privileges ure granted to Governors, Founders, and Life Members. Persons desirous of becoming members, or of obtaining any information with regard to the Association, should communicate with— The DIRECTOR, The Laboratory, Plymouth. ees ya > mA With Ten Plates, Royal Ato, 5s. CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA AND AMPHIOXUS. By:E. RAY LANKESTER, M.A., LL.D., F.R.S. London: J. & A. CeCe, & Great See eoren Street. Quarterly . Journal of Microscopical Science. The SUBSCRIPTION is £2 for the Volume of Four Numbers ; for this sum (prepaid) the JouRNAL is sent Post Free to any part of the world. BACK NUMBERS of the Journat, which remain in Bret are now sold at an uniform price of 10/- net. The issue of SuppremMenr Noumpers being found inconvenient, and there being often in the Kditor’s hands an accumulation of valuable material, it has been decided to publish this Journal at such intervals as may seem desirable, rather than delay the appear- ance of Memoirs for a regular quarterly publication. The title remains unaltered, though more than Four Numbers may be published in the course of a year. Kach Number is sold at 10/- net, and Four Numbers make up a Volume. London: J. & A. CHURCHILL, 7 Great Marlborough Street. TO CORRESPONDENTS. Authors of original papers published in the Quarterly Journal of Microscopical Science receive fifty copies of their communica- tion gratis. All expenses of publication and illustration are paid by the publishers. Lithographic plates and text-figures are used in illustration. Shaded drawings intended for photographic reproduction as half- tone blocks should be executed in “‘ Process Black” diluted with water as required. MHalf-tone reproduction is recommended for uncoloured drawings of sections and of Protozoa. Drawings for text-figures should nor be inserted in the MS., but sent in a separate envelope to the Editor. Contributors to this Journal requiring eatra copies of their communications at their own expense can have them by applying to the Printers, Messrs. ADLARD & Son, 22}, Bartholomew Close, E.C., on the following terms : For every four pages or less— 25 copies : ‘ ; i d/- aOR S35 , : 6/- 718 ae ‘ ; : é 6/6 100. ,, 2/- Plates, 2/- per 25 if uncoloured ; if colour ed, at the same rate for every colour. Prepayment by P.O. Order is requested. ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRESSED TO THE CARE or Messrs. J. & A. Courcaiti, 7 Great Marizoroucs Srreet, Lonpon, W. THE MARINE BIOLOCIGAL ASSOCIATION OF THE UNITED KINGDOM, Patron—HIS MAJESTY THE KING. President—Sir RAY LANKESTER, K.C.B., LL.D., F.R.S. 505 THE ASSOCIATION WAS FOUNDED ‘“ 10 ESTABLISH AND MAINTAIN LABORATORIES ON THE COAST OF THE UNITED KINGDOM, WHERE ACCURATE RESEARCHES MAY BE CARRIED ON, LEADING TO THE IMPROVEMENT OF ZOOLOGICAL AND BOTANICAL SCIENCE, AND TO AN INOREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS OF BRITISH FOOD-FISHES AND MOLLUSCS.” The Laboratory at Plymouth was opened in 1888. Since that time investigations, practical and scientific, have been constantly pursued by naturalists appainted by the Association, as well as by those from England and abroad who have carried on independent researches. Naturalists desiring to work at the Laboratory should communicate with the Director, who will supply all information as to terms, etc. Works published by the Association include the following :—‘ A 'l'reatise on the Common Sole,’ J.'I'. Cunningham, M.A., 4to, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British Islands, J. TI. Cunningham, M.A., 7/6 net (published for the Association by Messrs. Macmillan & Co.). The Journal of the Marine Biological Association is issued half-yearly, price 3/6 each number. In addition to these publications, the results of work done in the Laboratory are recorded in the ‘Quarterly Journal of Microscopical Science,’ and in other scientific journals, British and foreign. Specimens of Marine Animals and Plants, both living and preserved, according to the best methods, are supplied to the principal British Laboratories and’-Museums. Detailed price lists will be forwarded on application. TERMS OF MEMBERSHIP. ANNUAL MEMBERS . : . £1 1 Operannum. LIFE MremBErs . : : ; . 15-15 0 Composition Fee. FOUNDERS . 100 0 O a Governors (Life Members of Council) 500 O 0 Members have the following rights and privileges:—They elect annually the Officers and Council; they receive the Journal free by post; they are admitted to view the Laboratory at any time, and may introduce friends with them; they have the first claim to rent a table in the Laboratory for research, with use of tanks, boats, ete. ; and have access to the Library at Plymouth. Special privileges ure granted to Governors, Founders, and Life Members. Persons desirous of becoming members, or of obtaining any information with regard to the Association, should communicate with— The DIRECTOR, The Laboratory, Plymouth. With Ten Plates, Royal 4to, 5s. CONTRIBUTIONS TO THE KNOWLEDGE OF RHABDOPLEURA AND AMPHIOXUS. By E. RAY LANKESTER, M.A., LL.D., F.R.S. London: J. & A. CHa, a Great Sean Street. Quarterly Journal of Microscopical Science. The SUBSCRIPTION is £2 for the Volume of Four Numbers ; for this sum (prepaid) the JournaL is sent Post Free to any part of the world. BACK NUMBERS of the Journat, which remain in print, are now sold at an uniform price of 10/- net. The issue of Suppnement Nomsers being found inconvenient, and there being often in the Hditor’s hands an accumulation of valuable material, it has been decided to publish this Journal at such intervals as may seem desirable, rather than delay the appear- ance ot Memoirs for a regular quarterly publication. The title remains unaltered, though more than Four Numbers may be published in the course of a year. Kach Number is sold at 10/- net, and Four Numbers make up a Volume. en: J: & A. CHURCHILL, 7 Great eee street. TO CORRESPONDENTS. t Authors of original papers published in the Quarterly Journal of Microscopical Science receive fifty copies of their communica- tion gratis. All expenses of publication and illustration are paid by the publishers. Lithographic plates and text-figures are used in illustration. Shaded drawings intended for photographic reproduction as half- tone blocks should be executed in ‘‘ Process Black” diluted with water as required. MHalf-tone reproduction is recommended for uncoloured drawings of sections and of Protozoa. Drawings for text-figures should nor be inserted in the MS., but sent in a separate envelope to the Hditor. Contributors to this Journal requiring evtra copies of their communications at their own expense can have them by applying to the Printers, Messrs. ApiarD & Son, 224, Bartholomew Close, E.C., on the following terms : For every four pages or less— 25 copies : ; f j 5/- 5). ey eye ; , : 6/- hie. 5 : ; ; : 6/6 100 We Plates, 2/- per 25 if uncoloured; if coloured, at the same rate for every ‘colour. Prepayment by P.O. Order is requested. ALL COMMUNICATIONS FOR THE EDITORS TO BE ADDRUSSED TO THE CARE or Messrs. J. & A. Cuurcuint, 7 Great Mariporouca Srreet, Lonpon, W. THE MARINE BIOLOGICAL ASSOCIATION OF THE UNITED KINGDOM. Patron—HIS MAJESTY THE KING. President—Sir RAY LANKESTER, K.C.B., LEZD: 7k R.S: 20; THE ASSOCIATION WAS FOUNDED “ 10 ESTABLISH AND MAINTAIN LABORATORIES ON THE COAST OF THE UNITED KINGDOM, WHERE ACCURATE RESEARCHES MAY BE CARRIED ON, LEADING 10 THE IMPROVEMENT OF ZOOLOGICAL AND BOvraNIcaAL SCIENCE, AND TO AN INCREASE OF OUR KNOWLEDGE AS REGARDS THE FOOD, LIFE CONDITIONS, AND HABITS OF BRI’ISH FOOD-FISHES AND MOLLUSCS.” The Laboratory at Plymouth Was opened in 1888. Since that time investigations, practical and scientific, have been constantly pursued by naturalists appointed by the Association, as well as by those from England and abroad who have carried on independent researches, Naturalists desiring to work at the Laboratory should communicate with the Director, who will supply all information as to terins, etc, Works published by the Association include the following :—‘ A ‘Treatise on the Common Sole,’ J.T. Cunningham, M.A., 4to, 25/-. ‘The Natural History of the Marketable Marine Fishes of the British Islands,’ J. T. Cunningham, M.A., 7/6 net (published for the Association by Messrs. Macmillan & Co.). The Journal of the Marine Biological Association is issued half-yearly, price 3/6 each number. In addition to these publications, the results of work done in the Laboratory are recorded in the ‘Quarterly Journal of Microscopical Science,’ and in other scientific journals, British and foreign. Specimens of Marine Animals and Plants, both living and preserved, according to the best methods, are supplied to the principal British Laboratories and Museums. Detailed price lists will be forwarded on application. TERMS OF MEMBERSHIP. ANNUAL MeMBEns : - £1 1,0 per annum. LIFE Mempers . : : : - 15 15 O Composition Fee. FOUNDERS . - —100°40%:0 = 2 Governors (Life Members of Council) 500 0 0 Members have the following rights and privileges:—They elect annually the Officers and Council; they receive the Journal free by post; they are admitted to view the Laboratory at any time, and may introduce friends with them; they have the first claim to rent a table in the Laboratory for research, with use of tanks, boats, ete. ; and have access to the Library at Plymouth. Special privileges are granted to Governors, Founders, and Life Members. Persons desirous of becoming members, or of obtaining any information with regard to the Association, should communicate with— The DIRECTOR, The Laboratory. ¢) Plymouth. ee . = ’ pee tapes ors Journal of micy | N.S. vol 55. 910, | 1O- 46556 Date Loaned | R Borrower’s N ame 4 ? Ve —— Tp ) aC) e) JUN 1 | b, [JE Si) | i ya stra University Library : irae a Uma J i) a% t “a Pm ‘ 4 ‘ ‘ ’ wy , ‘ ; r , 5 “ ‘ ‘ { | ’ f fy 7 ‘ , wig li ( , H \ ' : ' A ¥ ‘ A HO de , y " WENA ? 4 r ‘i y whey LY r . a & : wk . “ « . fan. “ ‘ ‘ . 4 ay ne + > + ® ‘ * * ; ® ‘ wie + . al , ae 7) . . ’ * A * * * « . am