i SL ■Bmllilii ffilii »■■ ^^^™ ^^^^■M _D t^M^_ ^s^^s r-^ LO ^ l_) i.— S^SSS L_i ^ r-q aaaa ^s D ' m — : m THE VOYAGE OF H.M.S. CHALLENGER. ZOOLOGY-VOL. XIX. REPORT ON THE ^ £& SCIENTIFIC RESULTS OF THE VOYAGE OF H.M.S. CHALLENGER DURING THE YEARS i 8 7 3-7 6 UNDER THE COMMAND OF Captain GEORGE S. NARES, R.N., F.R.S. AND THE LATE ^<-r.>.CAPTAiN FRANK TOURLE THOMSON, R.N. %m Sir C. WYVILLE Knt, F.R.S., &c. PREPARED UNDER THE SUPERINTENDENCE OF THE LATE MARINE BIOLOGICAL LABORATORY LIBRARY WOODS HOLE, MASS. | W. H. 0. I, REGIUS PROFESSOR OF NATURAL HISTORY IN THE UNIVERSITY OF EDINBURGH DIRECTOR OF THE CIVILIAN SCIENTIFIC STAFF ON BOARD AND NOW OF JOHN MURRAY ONE OF THE NATURALISTS OF THE EXPEDITION Zoology— Vol. XIX. IPuIjlt'sljeD bj> ©flier of i£>er Jfflajestp's ®obeniment PRINTED FOR HER MAJESTY'S STATIONERY OFFICE AND SOLD BY LONDON :— EYRE & SPOTTISWOODE, EAST HARDING STREET, FETTER LANE EDINBURGH :— ADAM & CHARLES BLACK DUBLIN :— HODGES, FIGGIS, & CO. 1887 Price Twenty-five Shillings. PRINTED BY NEILL AND COMPANY, EDINBURGH, FOR HER MAJESTY'S STATIONERY OFFICE. xr q o THE Editor of the Challenger Reports will be greatly obliged to Authors sending him copies of separate papers, or references to works in which the Challenger discoveries are referred to, or the observations of the Expedition are discussed. This will greatly facilitate the compilation of a complete Bibliography, and the discussion of the results of the Expedition, in the final Volume of the Series. Letters and Papers should be addressed — JOHN MURRA Y, Challenger Office, j 2 Queen Street, EDINBURGH. (The Manuscript ivas received in Instalments between \lth October and 15th December 1886.) III. — Report on the Phyllocarida collected by H.M.S. Challenger during the years 1873-1876. By Professor G. 0. Sars, of the University of Christiania. (The Manuscript ivas received 15th January 1887.) IV. — Report on the Pteropoda collected by H.M.S. Challenger during the years 1873-1876.— First Part. By Paul Pelseneer, D.Sc. (Brussels). ( The Manuscript ivas received 15th November 1886.) it 7 o CONTENTS. I. — Report on the Nemertea collected by H.M.S. Challenger during the years 1873-1876. By Dr. A. A. W. Hubrecht, LL.D., C.M.Z.S., Professor of Zoology in the University of Utrecht. (The Manuscript ivas received in Instalments between I2t/i October and 18th November 1886.) II. — Report on the Cumacea collected by H.M.S. Challenger during the years 1873-1876. By Professor G. 0. Sars, of the University of Christiania. (The Manuscript was received in Instalments between Wth October and 1 5th December 1886.) III.— Report on the Phyllocarida collected by H.M.S. Challenger during the years 1873-1876. By Professor G. 0. Sars, of the University of Christiania. (The Manuscript was received 15th January 1887.) IV.— Report on the Pteropoda collected by H.M.S. Challenger during the years 1873-1876.— First Part. By Padl Pelseneer, D.Sc (Brussels). (The Manuscript ivas received 15th November 1886.) EDITOKIAL NOTES. This Volume contains Parts LIV., LV., LVI., and LVIII. of the Zoological Series of Reports on the Scientific Results of the Expedition. Part LIV. — On the return of the Expedition the specimens of Nemertea, along with the Annelida, were placed in the hands of Professor W. C. M'Intosh, F.R.S., for description. In the year 1884, however, Professor M'Intosh's time being fully occupied with the Annelida, Professor A. A. W. Hubrecht, of the University of Utrecht, was requested to undertake the investigation and description of the Nemertea, and the results of his labours are presented in this interesting and valuable Report. Parts LV. and LVI. are respectively Reports on the Cumacea and Phyllocarida, by Professor G. 0. Sars, of the University of Christiania, who has already contributed a most excellent Memoir on the Schizopoda collected by the Expedition, which forms Part XXIX. of the Zoological Series of Reports. Part LVIII. — This Memoir, which is the first instalment of the Report on the Pteropoda, by Dr. Paul Pelseneer of Brussels, treats of the Gymnosomata, and embraces a critical examination of all the genera and species of this little known order of pelagic aimals. The Report will be completed by subsequent Parts dealing with the Thecosomata and with the Anatomy of the whole class. John Murray. Challenger Office, 32 Queen Street, Edinburgh, 9th February 1887. THE VOYAGE OF H.M.S. CHALLENGES. ZOOLOGY. KEPORT on the Nemertea collected by H.M.S. Challenger during the Years 1873-76. By Dr. A. A. W. Hubrecht, LL.D., C.M.Z.S., Professor of Zoology and Comparative Anatomy in the University of Utrecht. INTRODUCTION. It was in September 1884 that the collection of the Challenger Nemertea was handed over to me for investigation, on the express condition that the whole of the MS. and plates were to be ready within fifteen months. I mention this, not only with a view of finding an excuse for omissions and curtailings which cannot fail to have been occasioned in this Report by its rather rapid elaboration, but at the same time in order to be able at the earliest moment to express my thanks to Professor W. C. MTntosh of St. Andrews, in whose hands the Nemertea were placed on the return of the Expedition, for his suggestion that the working up of this group should be entrusted to me, his own time being fully occupied with the description of the Annelida of the Expedition. He moreover placed at my disposal such notes as he had already found occasion to make upon the contents of the collection, some of them referring to species which he recognised as new to science. When in the following Report these notes are made use of, such passages will be specially indicated. The material, as it was put into my hands, was contained in about thirty small bottles, and was without exception preserved in spirit. It looked far from promising from a (ZOOL. CHALL. ESP. — PART LIV. — 1886.) Hllh 1 2 THE VOYAGE OF H.M.S. CHALLENGER. systematist's point of view. Hardly any traces of coloration were visible, no external appendages distinguished the different species, no definite shape, marked out by hard portions of the integument, which facilitate the recognition of representatives of so many other divisions of the animal kingdom, were anywhere noticeable. The majority of the specimens in the collection were cylindrical or flattened fragments, generally truncated, and rarely so intact that it was possible at first sight to distinguish between head and tail. The cephalic slits of those specimens which belong to the group of the Schizonemertea, and the mouths of many of the fragments, were the only external marks that could serve both for the discernment of what was posterior and anterior, and for a rough and provisional arrangement of the forms as they were probably related to each other. The various figures on PI. I. will give a general impression of the appearance of the fragments just alluded to. I more especially insist upon the poor aspect presented by the preserved material of the Nemertea in order to impress the reader with the fallacy of pronouncing an unfavourable judgment on a collection of marine invertebrates by relying merely upon the external appearance of the specimens. For I can hardly sufficiently emphasise the exceedingly good state of preservation of the large majority of the Challenger Nemertea. They were, indeed, in so perfect a state that the internal anatomy of all the fragments could be determined, and in very many cases delicate histological details could be revealed with as much facility as if the specimens had been captured a few days instead of ten years ago. I feel the more called upon to make this statement, and to express my admiration for the extreme care which the scientific staff have so evidently given even to unattractive and small-sized fragments, as it has been occasionally stated (even in certain of these Eeports) that the Challenger material was sometimes unfit for the minute investigation of internal anatomical struc- tures. So far as the collection of Nemertea is concerned, this statement is absolutely unfounded. The study of the fragmeuts and complete worms constituting this collection was only possible by aid of the microtome. This instrument was very freely made use of, and the most important improvement to which it has been subjected by Caldwell, whose automatic microtome was available, has enabled me to go into many more details than would otherwise have been possible. The total number of sections through different specimens of the Challenger material which have been prepared in the drawing up of this Report amounts to 19,560. They were all of them stained with Ranvier's picrocarmine. As already mentioned in the Narrative of the Voyage,1 the number of Stations from which the Challenger obtained Nemertea is more than twenty. It cannot be said that any of the three large subdivisions of the group is limited to any special region of the globe, although representatives of the very lowest and most primitive genera of Palseo- 1 Narr. Chall. Exp., vol. i. pt. ii. p. 831, 1885. REPORT ON THE NEMERTEA. 3 nemertea are as yet only on record from the Atlantic and the Mediterranean. It is, however, very probable that these genera (the Carinellidas) are cosmopolitan, and have as yet only escaped detection because even in the region from which they are known they count among the rare forms. I have here only to add that in drawing up the list of the Challenger Nemertea, I will follow the subdivision into larger groups that was proposed by myself several years ago (VII, p. 204),1 and will successively treat of the Palseonemertea, Hoplonemertea, and Scbizonemertea. It may be remarked that in the first named group, which contains the most primitive and least differentiated representatives, the genera Valencinia and Ewpolia ( = Polio) were also provisionally placed. These two may be looked upon as, to a certain degree, transitional forms. New light has been thrown by the Challenger material upon at least one of these genera, and it appears advisable to let them stand in that subdivision, however far they may differ in certain respects from the typical Palseonemertea, such as Carinella, Carinoma, Carinina, &c, and however strong their affinities may be in other respects either to the Scbizonemertea or to the Hoplonemertea. 1 The heavy numerals refer to the Bibliography at the end of the Report. DESCRIPTION OF THE GENERA AND SPECIES. NEMERTEA. A. PALjEONEMERTEA. Family Caeinellida Carinina, n. gen. Closely allied to Carinella, from which it differs in the presence of a distinct posterior brain lobe, situated with the rest of the brain and nerve-stems in the integu- ment, outside of the body musculature. A ciliated canal penetrates into this posterior brain lobe. Carinina grata, n. sp. (PI. I. figs. 1-3; Pis. II., III., IV.; PI. VI. figs. 1-3; PI. XL figs. 1, 2). Two specimens of this new genus and species were obtained in the dredge, both from considerable depths, and from the same part of the Atlantic Ocean, namely, to the east of the United States (Stations 45, 47). The bottom is recorded to be blue mud, and the depth 1240 and 1340 fathoms respectively. This is the greatest depth from which Nemertea have been brought to light, and it is worthy of notice that this deep-sea form, which is at the same time the representative of a new and distinct genus, should be characterised by peculiarities of structure, hereafter to be more fully recorded, which are diametrically opposite to certain of the most striking features of the pelagic genus Pelagonemertes and of other forms that generally occur close to the surface. The most striking of these characters is the exceptionally strong development of muscular tissue in the body-wall, and coincident with this, the considerable reduction of the inter- muscular connective tissue, which in the surface forms becomes a gelatinous matrix in which both the internal organs and the musculature are embedded. It can hardly be doubted that this opposite line of development is to a large extent influenced by the much more considerable resistance to be overcome by an animal that has to move about at so great a depth of water. Of the external appearance of the fragments of this new species little need be said. b THE VOYAGE OF H.M.S. CHALLENGER. They are reproduced both in the Narrative of the Cruise (vol. i. pt. ii. p. 831) and on PL I. figs. 1-3. There is no trace of longitudinal lateral slits — so characteristic of the Schizonemertea — but a terminal crescentic groove (marked out by darker pigment and by more profuse cfliation in one of the two specimens) was present in both. The pro- boscidian opening could be easily detected in both specimens, the subterminal mouth in one of them. Incomplete as was the information that could be gathered from superficial examina- tion, very interesting data came to light after the two specimens had been transferred to the microscopic slides. From specimen a, which was dredged at Station 45, a con- tinuous series of transverse sections was made, whereas specimen b, from Station 47, was cut transversely along the region behind the brain, nearly horizontally through the brain and tip of the snout, and longitudinally through that region of the body where the nephridia are found. It is very striking that in all these sections the cellular integument is of a consider- able thickness when compared with the muscular (PI. II.; PI. XL figs. 1, 2). In it we may distinguish several strata successively characterised (PL IV. fig. 1) by an accumu- lation of nuclei, by profuse integumentary gland-cells, &c, which will be more fully described in another chapter of this Eeport. Curiously enough the contents of the deeper glands have a well-marked green colour in the anterior portion of one of the body fragments of specimen b, whereas they are brownish-red in the posterior portion, the whole fragment having first been stained with picrocarmine. A homogeneous basement-membrane separates the integument from the subjacent muscles. This membrane is more deeply stained than other portions of the intercellular substance and thus stands out very clearly. In the outermost cellular layer, distinct unicellular, flask-shaped glands are present (PL IV. fig. 1), although they are not so numerous as in many Schizonemertea, These facts authorise us to look upon the integument of Carinina as similar in all its essential elements to that of other Nemertea. We will further insist upon this similarity when describing Eupolia and Cerebratulus. Before we pass from the integument to the muscular investment of the body we have to mention the central nervous system, which is found outside the homogeneous basement layer just referred to. Where the tissue of brain and nerve-stems takes its course in the deeper layers of the integument, it is directly applied against the subjacent muscles, the basement-membrane being indistinct if not interrupted beneath these central parts of the nervous system (PL III. fig. 7). In addition to the two lateral stems, each transverse section reveals the presence of a dorso-median thinner nerve-stem, corresponding to what I have formerly termed the proboscidian sheath-nerve. REPORT ON THE NEMERTEA. 7 There is, moreover, positive evidence as to the presence of a delicate nervous plexus, situated just outside of the basement membrane already alluded to, and which, judging from the available fragments, is present throughout the whole length of the animal. The transverse sections also show that a second nervous plexus may be presumed to be present just outside the inner layer of circular muscles. The details of this will be discussed hereafter when considering the nervous system. Passing forwards along the lateral nerve-stems we find them in every section fixed to the subjacent muscles by semicircular fibres indicated in figs. 7 and 8 on PI. III. Anteriorly the stems pass insensibly into the brain, which, as a mere thickening of the lateral stems, has as yet only attained to a very low degree of differentiation. In this respect the brain much resembles that of Carinella, from which, however, it differs in a very important point, viz., the presence of a posterior lobe into which penetrates a ciliated duct ending blindly and communicating with the exterior. This posterior lobe is situated, as is the anterior one, outside the muscles of the body- wall (PL VI. figs. 1-3), the inner channel is coated by a ciliated epithelium, differing in texture from the surrounding nerve-cells. The latter, however, can hardly be sharply distinguished from adjacent cells of a more indifferent character, and belonging to the lower strata of the integument. The muscular elements partaking in the formation of the body- wall are kept distinctly apart from the tissue, which we have described as the integument, by the homogeneous membrane above mentioned. Below this membrane we find a thin, circular, muscular layer (PL XL figs. 1, 2, (3), then follows the much thicker layer of longitudinal fibres (a), and finally an inner layer, thinner again, of circular fibres (8). The comparative thickness of these two latter layers throughout the oesophageal region may be gathered from PL II. fig. 5. In the two circular layers the fibres appear to be more closely set than in the longitudinal. In the outer circular layer the direction of all the fibres is, however, not perpendicular to the body-axis, a very regular network of other fibres which have their direction at an angle of 45° both with the longitudinal and the transverse axis, being closely interwoven with this layer. These are, however, not massive enough to form a distinct layer by themselves. The homogeneous intercellular substance, which is also present here between the bundles of muscular fibres, and which stains very distinctly with picro- carmine, is of course best visible when the bundles are widest apart. Such a portion is figured in PL III. fig. 6. This intercellular substance is also seen to be again traversed by radial fibres passing between the two circular layers ; nuclei are, moreover, present both in the intercellular substance and enclosed along with the bundles of fibres. Within the muscular body-wall are lodged — (l) the proboscis and its sheath ; (2) the intestine ; (3) the blood-spaces ; (4) the nephridia ; and (5) the generative sacs. The space not occupied by any of these is entirely filled up by a tissue, which I will call the gelatinous tissue, and regarding which more ample details will be given in the chapter 8 THE VOYAGE OF H.M.S. CHALLENGER. dealing with the anatomy of the group. It is wholly continuous, and a body-cavity in which the above-named organs may be said to float, or to be suspended, is totally absent. A rapid survey of the peculiarities which these five systems present must complete this descriptive account of the new genus. The aperture for the proboscis, which is situated terminally, leads into an anterior cylindrical compartment, which remains passive when the proboscis is protruded or retracted. It is coated by ciliated cells, and at the posterior end the anterior insertion of the proboscis into the body musculature takes place (PI. II. fig. 8 ; PI. III. fig. 5). Although the name is etymologically not wholly adapted for the purpose, I still am inclined to adopt for this compartment, which is present in all Nemertea, the name of rhynchodseum. This name as clearly separates it from the cavity of the proboscis or its sheath, as that of stomodseum and proctodseum distinguishes certain portions of the intestine of other invertebrates from the mid-gut. The rhynchodseum of Carinina has a great resemblance to that of Carinella, more especially because of the wide and much distended blood-space which wholly surrounds it, and in which it is kept in place by numerous strings of tissue starting from the muscular body- wall and inserting themselves on the muscular investment of the rhynchodseum (PL III. fig. 5 ; cf. IX., pi. i. fig. 2). From the same figure it may be gathered that the internal cellular coating of the rhynchodseum is more than one row of cells thick, and that these cells have a clear and distended aspect, with a comparatively small nucleus. The proboscis itself is inserted in a very simple way in the muscular tissue of the body-wall. The muscidar investment of the proboscis curves round at an angle of 90°, and becomes continuous with the longitudinal muscular layer of the body-wall. The details of this arrangement may be gathered from PL III. fig. 5, and it will there also be seen how the protruded proboscis remains fixed to the body all along this annular point of attachment. Thence it extends backwards as far as the proboscidian sheath permits, which, in the forms allied to Carinella, is only the anterior portion of the body. It is drawn back again by its retractor. How far backwards the proboscidian sheath reaches in Carinina could not be made out, as I only possessed two small anterior fragments, in neither of which the proboscidian sheath terminated. The proboscis itself could be examined with detail in the single specimen which was cut longitudinally. An anterior and a posterior portion of different textures are exceedingly distinct. They are separated from each other by a constriction. In the posterior portion the cells are eminently glandular, high and flask-shaped ; in the anterior portion they are less high and appa- rently less glandular (PL III. figs. 1, 2). Great differences in aspect, but not in actual texture, are of course occasioned by the different stages of contraction in which the various parts of the proboscis happen to be. The mouth, situated ventrally close to the anterior extremity, was very small in both specimens. The cellular coating of the oesophagus is very distinct, and the direct REPORT ON THE NEMERTEA. 9 application of the oesophageal epithelium against the muscular tissue most marked (PI. IV. fig. 3), not even a basement membrane separating the two, whereas a suspension of the intestine by means of the gelatinous tissue is of common occurrence in other forms. The longitudinal sections prove that further back the intestine does not con- tinue as a straight tube as it sometimes does in Palasoneniertea, but is constricted (PI. IV. fig. 2), the constrictions and resulting caeca being, however, much less marked and prominent than in the Schizonemertea and Hoplonemertea. Right and left of the intestine are situated the two longitudinal blood-spaces, which are direct continuations of the blood-spaces already noticed in the head, and which com- municate with these by passages that are encircled, together with the proboscis and its sheath, by the annular nerve ring formed out of the right and left halves of the brain and their superior and inferior commissures. The arrangement of these spaces has been fully described for Carinella by Oudemans,1 and I may refer to that description, the arrangement being on the whole very much the same. There is no median dorsal blood-vessel in Carinina, and there is a very distinct internal epithelium to the longitudinal blood-spaces, two of which are figured on PI. IV. figs. 2, 5, 6. Transverse vessels of communication are not present in these two forms. I may perhaps remark that my researches (XIV.) on the development of one of the Schizo- nemertea have rendered it probable that also in the Palseonemertea we shall have to look upon the blood-spaces in the same light as upon the cavity of the proboscidian sheath, viz., as a direct derivative or continuation of the blastoccele, for which cavities (in the adult state) I have proposed the name of archiccele. The nephridia are situated partly in the anterior portion of the blood-spaces, another portion traversing the muscular body-wall and leading to the exterior. In the portion of the paired nephridia exclusive of this excretory duct we may distinguish two distinct parts, one a continuous tube of varying dimensions, formed out of very regularly arranged cells with large nuclei, but not in any way forming a series of perforated cells such as are known in the nephridia of both Turbellaria and Discophora. These cells are distinctly ciliated and figured on PL IV. figs. 4-6, Nc. The structure of the second part of the nephridium is not so easily unravelled, and my preparations of the two specimens do not suffice to reveal all the details. I find it to consist of a cellular mass of spongy appearance protruding along a certain distance into the blood- space, here and there giving evidence of a tubuliform structure, no internal funnels being, however, anywhere recognisable (PI. IV. figs. 4-6, N. sp.). I must here remark that the researches of Oudemans, who described in detail the arrangement of the nephridia of Carinella and Carinoma,2 render it very possible that in 1 Quart. Journ. Micr. Sci., Suppl. volume, 1885. 2 hoc. cit., p. 71, pi. i. figs. 4, 5; pi. iii. figs. 56, 57. (ZOOL. CHALL. EXP. — FART LIV. — 1886.) Hhh 2 10 THE VOYAGE OF H.M.S. CHALLENGER. Carinina internal communications between the blood-spaces and the nephridian channels exist, communications by which the cavity of the blood-spaces is thus directly connected with the exterior. As I have, however, remarked, I could not detect the presence of similar communications in my two specimens. The aspect of the spongy portion of the nepbridium and its connection with the canalicular portion, as well as of this with the exterior, is represented in PI. IV. figs. 1, 2, 4. As to the generative apparatus of Carinina, 1 can only observe that one of the Challenger specimens is a male, that the fragment contains only two sperm-sacs in its posterior portion, and that these communicate with the exterior, each by a separate pore. Whether in Carinina.. the sperm-sacs are disposed metamerically as in most Nemertea, or irregularly distributed beneath the dorsal integument as in Carinella, could not be made out from this specimen. The general distribution of integumentary and muscular tissue, as well as of the cavities of the intestine (D), the proboscidian sheath (Ps), and the blood-space (bl) in the body of Carinina is indicated by the various figures of PL II. The proboscis itself is here indicated by Pr, the rhynchockeurn by aPr. Family Eupoliida Eupolia, n. gen. Polia, delle Chiaje. Integument generally thick in comparison with the body musculature, the two layers of contractile fibres of the integument never coalescing with the outer larger one of longi- tudinal body muscles as in certain Cerebratxdi. Proboscis and proboscidian sheath thin and inconspicuous. Brain-lobes compact, posterior lobe long, wedged in between the superior and inferior ones. Often a commisssure of the longitudinal nerve-stems beloiv the anus. No longitudinal cephalic slits but transverse grooves as in many Hoplo- nemertea. The necessity for creating a new generic name for the species of Palseonemertea I am now about to discuss is evident from the following considerations. The generic name Polia, when it was applied by delle Chiaje to a genus of Nemertea which he introduced into science {Polia delineata being the typical species of this genus) had already been preoccupied in Zoology by Ochsenheimer, who in 1826 so designated a genus of Lepidoptera. This reason alone suffices to reject it henceforth from Neniertean nomen- clature, and this rejection is also facilitated by the fact that the same generic name has been used by other naturalists, such as Quatrefages (XXVIII.), Schneider (XXXI.), &c, for Nemertea widely different from delle Chiaje's type. It was an error of judgment on REPORT ON THE NEMERTEA. 11 my own part, when giving my critical revision of Nemertean genera and species (VII.), to retain the name Polia in that list, although I was aware of its inapplicability according to the accepted rules of nomenclature. I retain delle Chiaje's Polia delineata as the type species for Eupolia. Eupolia delineata, (delle Chiaje) Hubrecht (PL VII. figs. 9, 10). This species, which is very common in the Mediterranean, more especially at Naples, is represented in the Challenger collection by one fragmentary specimen, captured at St. Vincent, one of the Cape Verde Islands. The fragments showed no head, but the characteristic colouring of the specimen by thin brown stripes on a lighter brown back- ground, was still so distinct in the spirit specimen, that even in the absence of a head, I do not hesitate to identify this form with the above-mentioned species, especially after com- paring the sections through the fragments with those through Mediterranean specimens. In one of the fragments, which is the tail, a terminal commissure between the two nerve-trunks can be demonstrated ; by a curious twisting of the fragment the sections seem to prove this commissure to lie above the intestine. Such a commissure is found in other species of Nemertea (e.g. Amphiporus, Drepanophorus), but a close inspection soon reveals that here, and also in other Eupolise, the posterior commissure is indeed found below the posterior portion of the intestine instead of above it. In MTntosh's preliminary notes on the Challenger Nemertea, I find a notice made of this specimen to the following effect : — " The specimen is incomplete, neither head nor tail being present. The body is firm and rounded, measuring about 45 mm. in length and about 8 mm. in diameter at the wider part anteriorly The body is closely striated longitudinally, dorsally, and ventrally by alternate white and brownish belts, the pigment constituting the latter being situated on the inner side of the basement layer of the cutis, which forms a simple stratum." Eupolia giardii, (M'Intosh) Hubrecht (PI. I. figs. 7-9; PL V.; PL VI. figs. 4-11 ; PL VII. figs. 4, 5, 8 ; PL X. fig. 6 ; PL XL fig. 12). Euborlcuiia giardii, M'Intosh, in lift. This new species, which shows interesting peculiarities, is represented by one specimen, which was cut up into different portions when I first examined it, so that I can only give a sketch of the head but no figure of the animal in toto. A well-marked peculiarity of this species of Eupolia is its shortness, which even surpasses that of Eupolia curta from Naples. The head shows (see PL I. figs. 7-9) a faint annular constriction, not continuous in the median ventral line. In this constriction the right and left external openings leading into the posterior brain-lobe are situated. Judging from what we find in other 12 THE VOYAGE OF H.M.S. CHALLENGER. Eupoliie, I suppose that this constriction is not so marked in the living animal, but that here, as in the Mediterranean species, two very shallow, strongly ciliated grooves in the integument, curving laterally round both sides of the head, were present, and that, during the process of preservation in spirit, the fold just mentioned made its appearance in the region where normally these transverse grooves are situated. What immediately distinguishes Eupolia giardii from its congeners is the thickness of the circular muscular layer in the oesophageal region. In M'Intosh's preliminary MS. notes on the Challenger Nemertea, I find the following remarks upon this specimen, which he perfectly recognised as a new sjnecies (the specific name giardii is taken from M'Intosh's provisional label), without, however, at that time referring it to delle Chiaje's Mediterranean genus. " A comparatively large form, measuring about 40 mm. in length, with a diameter in its widest part of 6-5 mm. This specimen is colourless, bluntly rounded at each end and somewhat fusiform in outline. " The anterior end is almost truncate, with a dimple in the middle, caused by the proboscidian aperture, and there are traces of a transverse and a vertical groove, thus forming a cross at the tip of the snout. " The latter is separated from the body by a well-marked fold which probably indicates a furrow, and which on each side does not quite reach the mouth. The mouth occurs on the ventral surface somewhat behind the foregoing furrow and in the form of a triangular dimple. A small aperture (anus) is situated at the dorsal margin of the blunt posterior end The small size of the proboscidian sheath is remarkable. It has an external layer of circular fibres and an inner layer of longitudinal muscular fibres. Both are thin. It is continued to the posterior third In the middle and towards the posterior third the body-wTall presents a decided change from the foregoing — the alimentary cavity forming a large central space and the solid wall is considerably diminished in proportion." In studying the sections of this species several additional points of interest came to light. Those concerning the brain-lobes will be discussed in the paragraph treating of the nervous system in general; the general aspect of the brain as it may be gathered from a reconstruction of the sections is figured on PL Y. The outline of the whole of the lobes and that of the internal fibrous core are here figured side by side in order to show the relations of the parts and the actual position of the ciliated canal that penetrates into a separate part of the braindobes more clearly. There is a terminal commissure between the longitudinal nerve-stems below the anus (PI. VII. fig. 8). The right and left longitudinal nephridial ducts (PI. VI. fig. 9, Nep.) communicate by deferent ductules (PI. VII. fig. 5, Nep.) with the exterior. Of the latter there are several; in the available transverse series through the head and oesophageal region I count five to the left and seven to the right, some of these (sections 298-325 left and REPORT ON THE NE.MERTEA. 13 303-308 right, as well as 448-485 left and 450-485 right) being unmistakably opposite. The duplicity of these deferent ductules, as figured on PI. VII. fig. 5, is the exception ; it was only noticed in this one case, all the other ductules being single. As to the generative caeca I find in this specimen (which is a male) that they arc very full, and that dorsally and ventrally they assume a conspicuously lobed and arbor- escent appearance. The integument offers many points of interest which will not be detailed here as they will be more fully described in the paragraph devoted to it. Eupolla australis, n. sp. (PI. I. fig. 6 ; PI. VII. figs. 1-3, 7). From M'Intosh's notes on this specimen I copy the following : — ■ "Another type of a whitish colour, measuring about 19 mm. in length and about 2 mm. in its widest part in front. The body is tapered from the wide anterior region to the fractured posterior end. It is rounded in front, somewhat flattened towards the pos- terior region. The head having been retracted forms a short blunt cone projecting from the folds of the wider nuchal region. No trace of furrows exists, but the mouth seems to be at the bottom of the transverse dimple at the base ventrally The inner longitudinal (muscular) layer is peculiar, for its fibres are somewhat regularly arranged, in long, parallel, and occasionally pennate fasciculi, which in transverse section run inward from the former coat. There is a slight hiatus in the dorsal middle line above the proboscidian sheath The proboscidian canal is somewhat thin The specimen is a male and the sperm-cells form large masses." The sections showed that the species was distinct from Eupolia giardii, which comes from the same locality, as also from the Japanese Ewpolia nipponensis, which will be described below. They furnish the following data wThich it may be of use to recapitulate, in order to facditate identification of the species wdien it is again captured in the same waters, and may then be described with its external coloration, of which no indication can here be given. The primary difference between every section of Ewpolia australis and all the other species of Eupolia here described is found in its integument. That portion of the integu- ment which lies outside of the secondary basement membrane, B (PI. VII. figs. 1-3), is by far the most prominent and the thickest portion, whereas in the other Eupolise it attains only half or even less of the thickness of the whole integumentary layers that are found outside the primary basement layer, Bet (PL VII. figs. 5, 9). In correspondence with this the secondary basement membrane is much thinner in these latter species than it is in Eupolia australis. The regular distribution of the blood-spaces round the oesophagus, and just behind it, is such that in addition to the medio-dorsal and the two ventral blood-vessels (br) it 14 THE VOYAGE OF H.M.S. CHALLENGER. would appear as if there were two other longitudinal ones, situated right and left of the proboscidian sheath (PI. VII. fig. 1). This may perhaps also turn out to be a special feature of this species. Certain other peculiarities observed concerning the intestinal system must for the present be passed over in silence, for want of material to verify them. It may, however, be added that the nervous plexus and the dorso-median nerve are much less conspicuous (though present) in Eupolia australis than in Eupolia nipponensis, where the plexus is in some places very thick (PL VII. fig. 11). In this respect Eupolia australis more resembles Eupolia giardii, where the plexus is not so very prominent, although the dorso-median nerve (PI. VII. fig. 4) is distinct though not massive (cf., PI. XL fig. 12). Eupolia nipponensis, n. sp. (PL I. figs. 4, 5, 10; PL VII. figs. 6, 11, 12). By this name I wish henceforth to designate a species of which fragments, partly heads, partly posterior body regions, which obviously belonged to different specimens, were collected by the Challenger in the Japanese waters. The series of sections reveal enough of common characters to deter one from assigning the fragments to different species. The species is characterised by certain features ah-eady alluded to in the foregoing- description of Eupolia australis. If it resembles Eupolia giardii in the disposition of the different layers of its integument, it differs from this species in the absence of the unusually thick circular muscular layer (PL VI. fig. 9) found in the oesophageal region of the latter species. The deeper layers of the integument are most conspicuously developed and vacuolated. That its nervous plexus is more conspicuous than that of the other Eupoliw was noted before, and I may add that in the available sections a very good horizontal aspect was obtained of the brain-lobes, which showed these to differ in certain minor but still easily verifiable points from those of Eupolia giardii. The upper lobe appears to be much more cylindrical ; so does the inner fibrous core. There is no superior additional gyrus to the superior braindobe with special fibrous core corresponding to what is described and figured for Eupolia giardii (PL V. figs. 1, 5, 7-9 ; PL VI. fig. 8). The connection between the posterior braindobe carrying the ciliated canal and the rest of the brain is, however, very intimate ; they are soldered together along a very extensive surface. As to the proboscidian sheath, one of the sections clearly demonstrates how exceed- ingly thin and delicate it is, and how the separation of its cavity from that of the blood lacuna is even difficult to observe. In this as in other species of Eupolia the distinction in the oesophageal epithelium REPORT ON THE NEMERTEA. 15 between the layer of ciliated cells immediately surrounding the lumen and the deeper layer of granular gland-cells is very marked. There is no muscular layer in the oesophagus as in Eupolia giardii (PI. VI. fig. 9) or in Cerebratulus corrugatus (PI. XIII. fig. 6) ; the thick layer of glandular cells just alluded to may here and there show a longitudinal fibre in addition to the radial ones that serve to suspend it in the circum-oesophageal lacuna ; for the greater part these cells project freely into this cavity and are bathed by the fluid it contains. B. HOPLONEMERTEA. Amongst the Hoplonemertea collected by the Challenger none are so different from those that are at present known as to necessitate the establishment of a new genus. Stdl several of them present certain notable points of interest by which our knowledge of this order of worms is extended, and which at the same time offer valuable material for more general speculations. The three genera Drepanophorus, Amphiporus and Tetrastemma, to which all the Challenger Hoplonemertea belong (when we except Pelagonemertes) appear to be very cosmopolitan ; the same remark, however, applying to the Schizonemertean Cere- bratulus, as will be shown in the sequel. Family Amphipoeida Drepanophorus, Hubrecht. Tlie mouth and the aperture of the proboscis are separate openings. The exceedingly muscidar proboscidian sheath communicates with lateral spaces that are metamerically placed, and have thin cellular or membranous walls. Tlie armature of the 'proboscis often conforms to an aberrant type. Drepanophorus rubrostriatus, Hubrecht. This species, although not represented by complete specimens, was dredged by the Challenger off St. Vincent, Cape Verde Islands, in July 1873. MTntosh identified the fragments before they came into my hands, and remarks upon them in his preliminary notes as follows : — " Two fragments of a form apparently closely related to Drepanophorus rubrostnatus, the two measuring about 14 mm. by about 3 mm. in breadth. The colour of the animal is reddish-brown on the dorsum, with longitudinal pale stripes " The ganglia and cephalic sacs are remarkably distinct and the nerve-cords have a cellular investment. A very remarkable feature is the presence of large granular tubes which communicate with the cavity of the proboscis 16 THE VOYAGE OF H.M.S. CHALLENGER. " The structure of the proboscidian sheath is peculiar, since the longitudinal fibres are clasped in spaces made of the circular coat and in transverse section the lining is papillose." Not only the remains of the external coloration but also the internal anatomy 4- fd'^' Z^WF\^^ Fig. 1. — Armature of the proboscis of an adult Drepanophorus serraticollis. H, curved handle of stylets with muscle- fibres attached to it ; st, nail-shaped stylets in action ; si', nail-shaped stylets in accessory reservoirs, the number of these reservoirs more or less corresponding to that of the active stylets, and also increasing with age. convinced me of the identity of this fragmentary specimen with the Mediterranean species, which has, moreover, already been found by Langerhans at Madeira. Although REPORT ON THE NEMERTEA. 17 the Challenger specimen contained no proboscis (which had apparently been expelled), I cannot refrain from giving a woodcut of the curious and divergent armature of the proboscis as it was observed by me both in young and in older specimens of Drepano- phorus from the Mediterranean. In young specimens the number of pointed stylets and of reserve sacs is less considerable. This proboscidian armature is certainly one of the most marked and distinctive features of the genus, although, as we shall presently see, I feel justified in assigning other species to it even when the presence of a similar armature has not been definitely demonstrated. The specimen was a female ; the generative products are, however, yet very far from ripe. Drepanophorus serraticollis, Hubrecht (PI. IX. figs. 5, 6 ; PI. X. fig. 5 ; PI. XI. fig. 8 ; PL XII. fig. 6 ; PI. XV. fig. 17). Drepanophorus serraticollis, Hubrecht, Aanteekeningeu over Anat. van eenige Nemertinen, Utrecht, 1874. Concerning the specimens here referred to this species, I find the following notice in M'Intosh's preliminary MS. : — " Two specimens were dredged at Station 162 (off East Moncceur Island, Bass Strait), 38-40 fathoms, sand, length about 30 mm., with a diameter of about 7 mm., but both are broken. " The ventral surface is marked by a median and two lateral longitudinal grooves. " Externally the dorsum is tinted of a pale madder-brown without stripes. A darker patch runs in the centre of the head in front of the cephalic furrows. The under surface is pale. " The head is wider than the neck and seems to have been somewhat bluntly conical. The aperture for the proboscis is slightly inferior. It is marked by a prominent ridge indicating the cephalic furrows, which slope slightly forward on each side to the middle fine, where they are separated by a short interval. Inferiorly they slope more distinctly forwards and inwards, and are separated by a wide interval, from which a median ridge goes forward to the proboscidian aperture. In front of this furrow, both dorsally and ventrally, there are a series of secondary furrows about thirteen or fourteen in number, running forward from the main groove " In regard to the structure of the proboscis it agrees with the others of the genus, presenting no stylets. " The proboscidian sheath presents a regularly interwoven or basket-like pattern of circular and longitudinal fibres, and the inner surface is papillose in transverse section. No diverticula seem to be present Many ova are found partially projecting through apertures a little external to the nerve-cord and corresponding to the very evident raised line on the ventral surface." (ZOOL. CHALL. EXP. PAST LIV. — 1886.) Hhh 3 18 THE VOYAGE OF H.M.S. CHALLENGER. It needs no comment that it is at the least rather hazardous to identify with the Mediterranean species (which seems also to have been examined and figured by Quatre- fages when he gave the description of his Cerebratulus crassus), a specimen in which the proboscis, as well as its armature, is absent. Still the transverse sections offer such a very close resemblance to those of actual specimens of Drepanophorus serraticollis, that it would be again hazardous to establish a new species for the fragments, of which the colora- tion affords a less decisive clue than in the case of the foregoing Drepanophorus rubro- striatus — the madder-brown hue referred to by M'Intosh being all that is preserved of the uniform though bright coloration which the specimen must have had when alive, if it agreed in this respect with the Mediterranean Drepanophorus serraticollis. I have, moreover, hazarded the identification with the foregoing specimens of a third fragment collected in the Kerguelen waters, of which not only the proboscis but also the head was absent. Here, too, the internal characters enabled me to refer the specimen to the genus Drepanophorus (the transverse cseca of the proboscidian sheath being in this case the guiding feature). The systematic position of this specimen thus only rests upon the similarity of the transverse sections and on the general yellow hue of the fragment, darker on the dorsal than on the ventral surface. The very thick- walled proboscidian sheath with its delicate lateral sacs, different in certain respects from that of a new species of Drepanophorus hereafter to be described, is figured on PI. X. fig. 5. Drepanophorus lankesteri, n. sp. (PI. I. fig. 22; PI. IX. figs. 1, 2, 10; PL X. figs. 2, 4 ; PL XII. figs. 5 ; PL XIV. figs. 9, 10 ; PL XV. fig. 13). Of the three species of Drepanophorus contained in the Challenger collection, this is without doubt in several respects the most remarkable. One specimen measuring 30 mm. in length and 3^ mm. in breadth was obtained ; it was dredged at Station 49, in the waters of Nova Scotia. As to its colour when alive, the spirit specimen allows of no other certain conclusion than that the dorsal surface is darker than the ventral, which may have been whitish. No special markings are now traceable on the dorsal integument, and we may thus surmise that its natural colour, which has been only partly preserved in spirit, was in life brown or red. If I nevertheless feel justified in creating it a new species, it is because certain internal characters are so well marked as to allow of no confusion with the species of Drepano- phorus hitherto known. The two characteristic features which immediately attract attention in studying a series of sections through this species are, first, the presence of a series of transverse commissures (PL IX. fig. 10) metamerically placed at intervals of about 0"2 to 0'15 mm., and connecting the two longitudinal nerve-stems all along their course below the intes- EEPOET ON THE NEMEETEA. 19 tinal cseca. Close to the posterior end of the body I cannot vouch for their presence ; their extreme tenuity, and a folding of the sections, preventing the transverse commissures, if present, from being seen. Nor could I make out with certainty in the one specimen available, whether the longitudinal stems themselves coalesce above the anus as they do in the other Hoplonemertea, but on a priori grounds, I can hardly doubt their doing so. Anteriorly, the transverse commissures were present even in the immediate neigh- bourhood of the brain, up to the point where the so-called vagus nerve branches off and stretches forwards towards the oesophagus. Although, on the whole, they have a very regular course, and are situated at equal intervals, still a few irregularities in these commissures must be noticed ; some of them branching into two, others being connected with the preceding or the succeeding commis- sure by a small bundle of nerve-fibres. The significance of this nervous arrangement will be discussed further on. The second characteristic to which I wish to draw attention is the presence of trans- verse cseca belonging to the proboscidian sheath. Although they are present in other species of Drepanophorus, so that we are justified in looking upon their presence as one of the typical generic characters, still I never found their walls so markedly developed as in Drepanophorus lankesteri. Generally the walls are exceedingly thin and membranaceous (e.g., Drepanophorus rubrostriatus) ; here, however, they attract attention by the thick cellular coating which immediately reveals its presence both in longitudinal and in trans- verse sections. On PI. X. fig. 4, the nature of this arrangement is clearly shown. Another peculiar feature of these cgeca of the proboscidian sheath is that I have found a few of them coalescing peripherally with the one preceding or following them by means of a short longitudinal extension, which allows these few successive cseca to intercommunicate not only by means of the proboscidian cavity, but also by means of this distal connection. The muscular body-wall of this, as of most other Hoplonemertea, may be shown to contain, in addition to the two layers a and /3 (cf. PI. XL fig. 8), certain cross fibres not forming a definite layer, but arranged at angles of 45°, and visible in sections parallel to the surface. The openings of the longitudinal canals of the nephridia to the exterior are situated ventrally, posteriorly, and at the same time terminally; this constitutes another difference as compared with Drepanophorus rubrostriatus which has been already described and figured by Oudemans.1 On one side of the specimen investigated two openings of the nephridial duct piercing the integument are at all events observable, although somewhat more internally, before these deferent ducts have pierced the muscular body-wall, they coalesce. There is a very close proximity between the anterior nephridial ramifications and the lateral longitudinal blood-vessels. They do not, however, intercommunicate, nor do, as was supposed by MTntosh, the proboscidian sheath-ceeca and the blood vascular system. 1 Loc. cit., pi. i. fig. 7. 20 THE VOYAGE OF H.M.S. CHALLENGER. Eyes are present in Drepanophorus lanhesteri. In the (detached) posterior lobes of the brain there is a double canal as in most other Hoplonemertea, the one branch taking- its course along the glandular cells, the other in the ganglionic part of this brain-lobe. As to the sexual elements, they are in this species enclosed in sacs that are ventrally situated, and although I have only one specimen at my disposal, which is just beginning to ripen, I still believe I may lay it down as a rule for the species that the genital caeca are arranged in pairs in the vicinity of each nerve-stem, so that four are very often simultaneously met with in one section. This is, as we will presently see, a more compli- cated arrangement than that which obtains in the other species of Drepanophorus, but it is a simplification of the more profuse and less regular distribution of the genital sacs, as it occurs in Amphiporus moseleyi. Amphiporus, Ehrenberg. Stylet in the proboscis of the normal shape. Oral and proboscidian aperture con- fluent. No lateral cseca to the proboscidian sheath. Numerous longitudinal nerves in the proboscis as in Drepanophorus. Amphiporus moseleyi, n. sp. (PI. I. figs. 20, 21; PL IX. figs. 4, 7-9, 11; PI. X. fig. 3; PL XV. figs. 11, 12, 20). Professor MTntosh has drafted the following notes on a provisional examination of the specimens before they were handed over to me, which I may here be allowed to insert : " A large flattened species, the largest specimens about 68 mm. in length, and about 12 mm. at the broadest part. The body in those best preserved is somewhat flattened and with an acute edge along both sides of the tip of the tail. Anteriorly the body is thick and rounded both dorsally and ventrally, but posteriorly it is much flattened. In the preparation the anterior end is more pointed than the posterior. The ventral surface throughout is flatter than the dorsal. ' The larger specimens are deprived of much of their cutaneous tissue so that they are comparatively pallid ; in one (the smaller) the dorsum is of a dull blackish-grey ; while the ventral surface is either whitish or pale greyish. The lip of the snout is pale, and from this a pale line runs backwards to the tail on each side. This is not altoo-ether due to pigment, for in those devoid of cutis a very evident whitish band is found along the anterior third, but it becomes indistinct posteriorly. ' The head is somewhat truncated anteriorly and marked by a series of eyes which are rather deeply seated. In the large specimens these form a marginal band on each side along the antero-lateral margin of the snout. In the smaller a series begins on each side of the median line of the snout and runs in a tolerably straight line backwards to the slight narrowing of the furrow, whfle a somewhat triangular area superiorly is covered REPORT ON THE NEMERTEA. 21 with similar eyes. The posterior boundary especially being so distinct as almost to make a special row. " Behind these a band of similar eyes runs upwards and slightly forwards, a consider- able interval on the summit of the dorsum separating those of each side. " The cephalic furrows slope outwards and backwards on each side to the margins dorsally, and from the latter point are continued ventrally outwards and forwards. " So far as can be observed in this form only a single aperture exists for the proboscis and mouth. This forms a well-marked sbt in the ventral surface, a little behind the tip of the snout " In minute structure the proboscis corresponds to that of the typical form .... the stylet is simple and normal." To this description of the external characters I have nothing to add, but may proceed to remark that the examination of the internal structure by means of sections has revealed the significance of the white lateral stripe, noticed by Professor M'Intosh as not being due to pigment. It is, indeed, a pecubar feature by which this species is characterised, and which I have hitherto not observed in other Nemertea. All along the extent of this lateral and longitudinal whitish line (PI. IX. fig. 8) the sections show the presence of numerous glandular (or sensory ?) cavities, opening to the exterior by very numerous pores piercing the integument, and both accumulated at, and bmited to, the region where the dorsal musculature merges into the ventral (PL XV. fig. 11), and where, as in so many Hoplonemertea, the muscularity of the body-wall is reduced to a minimum, i.e., in the right and left lateral line. In the posterior portion of the body these organs were no longer present ; anteriorly, however, they could be traced even in the head (PI. X. fig. 3, gls.). Further details about their structure will be given in the anatomical part of this Report. The other chief peculiarities of the species which are revealed by a study of the sections, and which must be briefly enumerated in this summarising description, are : — the situation of the longitudinal nerve-stems, in the portion of the body where the intestinal caeca are clearly developed, above these cseca about one-third or halfway between the lateral margin and the proboscidian sheath. It should be remarked that this arrangement is the opposite of what is observed in Drepanophorus, where the longitudinal nerve-stems, as elsewhere described, have their course below the intestine, or below the lateral cseca. The significance of this different arrangement will be elsewhere discussed ; in itself it is a feature very much facilitating the discrimination of Amphi- porus moseleyi from other Amphipori, where the nerve-stems are found much more strictly laterally, at least in those hitherto known. There is a very distinct commissure between the longitudinal stems above the anus. Another most characteristic feature which may generally be distinguished in every transverse section of the animal, especially when it is ripe for reproduction, is the situation 22 THE VOYAGE OF H.M.S. CHALLENGER. of the numerous reproductive receptacles. These are not situated alternately between each pair of intestinal caeca right and left, as we find in the majority both of Hoplo- nemertea and Schizonemertea. In Amphiporus moseleyi, the distribution of the genital receptacles appears to follow a more primitive arrangement, and offers many points of similarity with what obtains in the Palseonemertean Carinella, where there is not yet a regular metamerical arrangement of the genital sacs, but where there are short independent cavities, irregularly distributed under the dorsal body- wall, which they pierce by means of short ducts. The outer openings of these ducts are seen on the dark dorsal surface of the animal as so many fine white dots irregularly spread between the transverse and longi- tudinal white lines that form such well-defined external markings in the species in cpiestion. Amphiporus moseleyi, as will be seen on comparison of figs. 4 and 7 of PL IX., has its generative sacs distributed very much in the same way, with this difference, however, that generative pores are situated not only on the dorsal but also on the ventral surface of the animal. When the animal is very ripe and the generative sacs are overfilled, it is manifest that this specific character may be more easily detected in every transverse section than in young or unripe specimens. As many as seven separate sacs in one section were noticed. Both the male and female sex were found to agree in this respect. Another character peculiar to nearly all Amphipori — the coalescence of the oral and the proboscidian aperture into a common wider opening, situated just below the tip of the snout — is also met with in Amphiporus moseleyi. Amphiporus marioni, n. sp. (PI. IX. fig. 3; PI. X. fig. 1; PI. XV. figs. 14, 15). A second species of Amphiporus is represented in the Challenger collections by two specimens, the larger coming from Marion Island, and having been collected on December 26, 1873 ; the other from Christmas Harbour, Kerguelen, at a depth of 120 fathoms. The place at which the first specimen was obtained was an inducement to dedicate this species, in preference to any other of the novelties of the Challenger, to the indefatig- able naturalist of Marsedles, so well known by his numerous researches in the field of invertebrate morphology. Amphiporus marioni, was one of the larger sized specimens, measuring 5^ mm. in diameter anteriorly in its widest region. The body musculature may be said to be stronger than in most of the other Hoplonemertea (cf. PL IX. figs. 1-6) ; the longi- tudinal muscular layer shows a very marked pennate arrangement of the bundles (PL X. fig. 1) between which the gelatinous tissue penetrates, carrying with it massive nerve- stems which assume a more or less flattened, plexus-like arrangement, just between this longitudinal layer a and the circular layer /3 (PL X. fig. 1, ne). The proboscidian sheath is also very muscular ; the proboscis has the stylet of the normal Amphiporean shape. The nephridial system is comparatively short and has one pair of deferent ducts REPORT ON THE NEMERTEA. 93 situated at the posterior extremity ; the longitudinal canal is anteriorly very copiously branched (PI. X. fig. 1, Nep). There is a very thick basement membrane (B) to the integument, and very strongly de- veloped gelatinous tissue (Gt) inside the muscular body- wall. Curious granular enclosures (inc) occurring in this tissue, both in the head and in the body, will be elsewhere described. The longitudinal nerve-trunks are not wholly lateral but nearly so; there are no ventral commissures between them. The generative caeca assume the ordinary character of paired dorsal receptacles meta- merically distributed between the intestinal casca. The generative pores are dorsal and situated above the nerve-trunks. The ova, present in both specimens, are in both of them characterised by a curious refractive body constantly present in addition to the nucleus, and staining deeply with picrocarmine. This " paranucleus " can be seen to be present at the very earliest stages of the development of the eggs which came under observation ; stages at which the eggs could stdl hardly be distinguished from the surrounding cellular elements in the wall of the generative caeca (PI. XV. figs. 14, 15). Family Teteastemmid.i, Tetrastemma, Ehrenberg. Eyes four ; arranged so as to indicate a square or oblotig. Specimens generally small. Tetrastemma agricola, Willemoes Suhm. Of this species, collected by Suhm in Bermuda (Mangrove swamps, Hungry Bay) and which is the only Land Nemertean procured during the voyage, no specimens have been preserved, although Suhm tells us that he collected a good many of them. So I must content myself with reproducing the chief points of its anatomy as they were made out by him in the Ann. and Mag. Nat. Hist, for June 1874. At the same time I have repro- duced one of his figures in woodcut. Suhm writes (loc. cit., p. 409): — " The largest of these worms have a length of 35 mm. by 2 mm. in width. They are of a milky- white colour. Their movements are slow and sometimes catterpillar-like ; they shoot out their long proboscis, fix it at some distant point to which it adheres by means of its papillse, and draw their body after them. Their skin is filled with rod-like bodies as described by Max Schultze and others, and is covered on the outside all over with cdia. In the front we find two pairs of eyes, one of them near the entrance of the proboscis, the other smaller one further out ; they consist of a fine granulated pigment, imbedded in a colourless substance, which holds these granules together, in which, how- ever, a regular lens could not be observed ; underneath these eyes is seen the prominent centre of the nervous system (fig. 1, g); it consists of two lobes and a ring which connects 24 THE VOYAGE OF H.M.S. CHALLENGER. them and encircles the proboscis. From the lobes depart the two lateral nerves (n) and some other cephalic nerves, which were not quite clearly visible The cephalic fissures or ciliated sacs .... are either very small or wanting entirely. Sometimes a folding of the skin seemed to indicate their presence ; but in the contractile bodies of these worms it is very difficult to say whether you have a small cephalic fissure or a folding of the skin before you. " Underneath the ganglion, on the under surface, is the semicircular opening of the mouth (o) leading into an intestinal tube (i), which runs through the whole length of the animal, without showing anything particular, and is terminated by an anus (a) The proboscis is divided into two portions — the papilligerous part and the glandular part. At the bottom of the former we find a peculiar spine .... this spine is remarkable because it differs in form according to the sex of its owner. In the male it has a rounded base and is pear-shaped (fig. 2, pr3),1 whde in the female the base has sharpened angles (fig. 3, pr*).1 I do not think that such sexual differences have hitherto been observed in Nemerteans. " The ovaries and testes are, as usual, situ- ated between the intestine and the walls of the body I . . . . establish for it the specific name of agricola, as there is probably no described marine species of Tetrastemma with which it could be identified. " I, however, do not attach much import- ance to this point, as the object of these lines is only to show that in America also land Nemerteans exist. Hitherto they were only known from the Pelew Islands, where Semper has found another, to which he has given the name of Geonemertes palseensis. I think it is highly probable that land Nemerteans exist to a greater extent in tropical countries than has hitherto been supposed, and that from their hidden life, and the impossibility of preserving them, they have hitherto escaped the attention of travelling naturalists. Especially in such islands as the Bermudas, where the earth of the lower grounds contains a great deal of salt, it may easily be imagined how marine animals have taken to terrestrial habits ; and it was interesting for me to see that one Tetrastemma 1 These figures referred to by Suhui have not been here reproduced. Fjg. 2. — Tetrastemma agricola, Willemoes Suhm. o, mouth ; g, brain; n, lateral nerve; i, intestine; Pt 1, rhyn- cnodteum ; Pt 2, papuliferous ; Pt I, glandular part of the proboscis ; Pt 3 and ca, region of the stylet and reservoir. REPORT ON THE NEMERTEA. ZO when put into salt water would live there for twenty -four hours, but when put into fresh water died after a few hours time. Fresh water, however, poured over the earth which contained them, did not damage them in the least." Tetrastemma fuscum, Willemoes Suhm (nee OErsted). A second species of Tetrastemma was noticed by Suhm during the first year of the Challenger cruise. In the paper just referred to, in which he describes Tetrastemma agricola, it is noticed in the following words: — " I may here also add that on our cruise from the Bermudas to the Azores I found parasitical Nemerteans on Nautilograpsus minutus, one of the gulf-weed crabs. They were small brownish animals, and occupied especially the underside of the crab, under the abdomen of which I found most of them. They did not exceed the length of 2 mm. and in none of them could I see genital organs. In fig. 4 I have figured one of these small parasites, which probably also belong to the genus Tetrastemma, though the second pair of eyes is only punctiform, situated on both sides of the proboscis. Nervous system and digestive apparatus do not present anything particular, and the proboscis (fig. 4, jyr1, pr3), is remarkable for its shortness. " I do not think that these worms attain their full size on the crab, but believe them to be young parasitical stages of some Nemertean which possibly lives on the gulf-weed." In his journal, of which an abstract is given in the Narrative of the Cruise, vol. i. p. 169, Suhm further remarks about this species : — " The worm presents no modification induced by parasitism ; it appears to be a new species, and from its colour may be called Tetrastemma fuscum. . . . The ganglia are especially large and conspicuous. . . . The proboscis is very short, and distinguished from all other species I know of by having the stylet-sac placed close behind the ganglion and just above the mouth. . . . Length 0"75 mm., breadth 0"25 mm." Suhm was apparently not aware that the specific name which he gives to his specimen was preoccupied for another Tetrastemma, as early as 1844, by CErsted for the species that is now known as Tetrastemma dorsalis. No specimen being preserved, the special features enumerated by Suhm do not justify me in proposing a new specific name. For completeness' sake it was, however, necessary to mention his observations. Family Pelagonemektidj;. Pelagonemertes, Moseley. Pelagonemertes rollestoni, Moseley. It was indeed a novel and startling fact when detailed news ' appeared regarding the capture by the Challenger naturalists of a pelagic Nemertean, which, in addition to other 1 Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 165, 1875 ; vol. xvi. p. 377. (zool. en ALL. EXP. — part liv. — 188G.) Hhli 4 26 THE VOYAGE OF H.M.S. CHALLENGER. characteristic differences, was recognisable by a change in the constitution of its tissues, similar to that which is noticed in pelagic animals belonging to other groups, when compared with their non-pelagic allies, viz., the hyaline transparency of the body, with undiminished, or even with rather increased bulk. The first specimen captured was a ripe female, the second a very young female. No further specimens were met with. The first specimen, though somewhat lacerated, was preserved ; the second was observed alive and figured, but was destroyed. Before mentioning the results which have been obtained by a careful microtomy of the available specimen, I will here insert in full the interesting descriptions with which Professor Moseley furnished us as early as 1875,1 soon after the specimens were captured, and when he had been able to observe them in the fresh state. His first article 2 On Pelagonemertes rollestoni runs as follows : — " This remarkable form was found in the trawl, together with a number of deep-sea animals, from 1800 fathoms, near the southern verge of the South Australian current, lat. 50° 1' S., long. 123° 4' E., March 7, 1874. Its appearance at once pronounced it a pelagic animal, the body being gelatinous and transparent as in Salpa, with the exception of the alimentary canal, which stood out in relief, being of a deep burnt-sienna colour (as is the nucleus in many Salpse), and the region of the sheath of the proboscis, which was less transparent than the remainder of the body. The animal was living when obtained, and when placed in fresh sea-water gave evidence of life by a feeble, irregular, peristaltic contraction of the external muscular tunic, which increased on irritation ; the proboscis was also protruded and retracted several times. " The animal was about 4 cm. long and 2 broad, and 5 mm. in thickness. Hence its dimensions, and especially its thickness, render it unfavourable for a perfect examination of its structure under the microscope whilst in the entire condition. As only one specimen was procured, and as this was believed to be unique, no dissection was resorted to, excepting the removal of a small portion of the epidermis and external muscular tunic for microscopic examination. Hence the investigation of the structure of this Nemertine necessarily remained an imperfect one, and the affinities of the animal amongst other Nemertines could not be determined. "The animal is leaf-like in shape, narrowing to a blunt point at the posterior extremity, and commencing abruptly at the anterior. The proboscis is protruded from the summit of a protuberance occupying the middle region of the anterior extremity. The mouth is situate on the ventral surface of the body, just posterior to the aperture for the proboscis. It is a simple aperture with a plaited margin composed of five or six folds. It is the commencement of a short muscular tube, the oesophagus, which was seen to pass behind the most anterior prolongation of the main mesial digestive canal, but the communication of which with the latter was not traced. The digestive system stands 1 Ann. and Mag. Nat. Hist, 1875, No. 87 and No. 96. 2 The figures which accompanied the article have all been reproduced on PL I. figs. 24-27. EEPORT ON THE NEMERTEA. 27 out very conspicuously in the fresh condition of the animal, from being of the deep burnt-sienna colour already mentioned. It consists of a broad, flattened mesial canal, somewhat broadest in the middle region of the body, anteriorly ending in a bluntly terminated caecal prolongation, and posteriorly narrowing gradually. As the posterior part of the animal was somewhat injured it could not be determined whether the canal terminates in an anus or not. " The mesial canal receives on either side lateral tributaries in pairs, which tributaries remain simple for some distance of their horizontal course, and then break up into ramifications. The most anterior pair of lateral canals is split up into by far the most ramifications. The ramifications become less and less in each pair towards the posterior extremity of the body, some of the most posterior lateral canals being simply bifurcate, and one merely enlarged at the extremity. There are thirteen pairs of lateral canals in all. " The nervous system was plainly seen in part. A pair of rounded ganglia lie on the ventral and lateral surface of the sheath of the proboscis, being a little posterior in position to the mouth. A commissure passes above the oesophagus and between it and the proboscis-sheath. From the ganglia a pair of fine simple nerve-cords pass in a curved course down to the posterior extremity, where their termination could not be ascertained. " The cords cross ventrally the lateral digestive canals about the point where ramification commences. Further connections of the ganglia could not be ascertained. " The specimen obtained was a female. A series of ovaries, consisting of pear-shaped masses of minute ova, were present, situate between each of the pairs of lateral digestive tubes immediately external to the nerve-cord on each side. The masses of ova are contained in small cavities in the gelatinous internal body-tissue. When pressure was exerted the ova issued from small corresponding apertures on the ventral surface, and the small empty cavities remained. The ova were spherical, about "28 mm. in diameter, and appeared composed of fat-globules and granular matter. " The proboscis-sheath, which is wide and capacious, is very plainly seen on the dorsal aspect of the body, and dimly through the thickness of the body from the ventral aspect. It has a firm muscular attachment at its orifice, and bundles of muscular fibres (apparently retractor) are attached to it here on either side (pi. xv. fig. B, 1 = PI. I. fig. 24). The proboscis itself is, when retracted, coiled up in the usual manner within its sheath, as seen in fig. D ( = P1. I. fig. 25). It could, unfortunately, not be ascertained whether the proboscis is armed or not. It was never entirely retracted, but a small portion of it always remained exserted. " The outer surface of the body of the Nemertine is covered with a hyaline, very thin integument, which is thrown into numerous folds and wrinkles, which are so arranged along certain lines around small spaces nearly free from them as to produce on the surface of the body an appearance of a series of small polygonal areas, separated by fine reticular 28 THE VOYAGE OF H.M.S. CHALLENGER. network (fig. D = fig. 27). This condition of the surface was most conspicuous about the anterior part of the body, but the body was much lacerated by the meshes of the trawl, and, therefore, I cannot say whether the whole integument is in this condition in the fresh state or not. The folds and plaits in the integument are so sharp that they give the appearance under the microscope of somewhat spindle-shaped bodies with sharply pointed extremities (fig. c, 1, 2, 3 = fig. 2G). At first I supposed that these bodies were urticating organs, resembling those of Bipalium, but on carefully teasing up a portion of the integu- ment with fine needles, and being unable to isolate a single one, I concluded that they were mere folds. They are, however, of remarkable appearance, from their extreme abundance and the manner in which they cross each other at all angles. They are well preserved in glycerine preparations of the skin hardened in picric acid. " Beneath the integument is some granular glandular matter. Immediately beneath the integument, and in close adherence to it, is the muscular tunic, evidently the homologue of the cutaneous muscular system of Bipalium and other Planarians. As in these, the outermost fibres are circular in direction, the inner longitudinal. " The muscular tunic encloses the entire body. It is thin, and in the fresh condition of the animal transparent and inconspicuous, but becomes opacpie when the animal is hardened in picric acid. The inner longitudinal layer consists of stout bands of fibres running parallel to one another. The outer circular fibres are far less developed, and are not gathered into bundles, but cross one another slightly obliquely in their transverse course, forming a slight meshwork over the longitudinal fibres. " Beneath the muscular tunic and between its meshes the body mass is filled up with a gelatinous, hyaline, structureless matter, imbedded in which lie the viscera and the muscles attached about the orifice of the sheath of the proboscis. Internal muscles, except those referred to, were not observed. " No eyes or other sense-organs were found, and ciliated sacs were not seen. " From the circumstance of the only specimen of Pelagonemertes having been much lacerated, and from the animal not having been dissected, it will of course require further examination. In the specimen as procured there was a deep constriction of the body at about the junction of the first with the second fourth of its length. This, it appeared pretty evident, had been caused by the meshes of the net. The posterior extremity was somewhat injured, and its form may not be quite correctly given. Ciliated sacs may be present, and the structure of the proboscis might throw light on the affinities of the animal. " The form of the digestive system is the most remarkable feature about Pelago- nemertes in its close resemblance to that of Dendroccela. In other respects Pelago- nemertes is thoroughly Nemertine in structure, being merely modified for pelagic existence. It is remarkable that the gelatinous hyaline mass of the body is not tegumental in character, but apparently homogeneous with internal structures. REPORT ON THE NEMERTEA. 29 "The occurrence of a peculiar burnt-sienna colour in many very different pelagic animals is remarkable. With many the colouring may be explained as protective resemblance to the oceanic seaweeds. For its occurrence in others, such as Salpa and Pelagonemertes, in an otherwise hyaline body, there may be some common cause, possibly also protective. " Diagnosis of the genus Pelagonemertes, H. N. M. : — Body leaf-shaped, gelatinous, hyaline. The anterior extremity of the body broad and abrupt, the posterior narrowed to a 'point. TJie digestive canal with thirteen pairs of lateral ramifications, as in Dendrocoda. Integument thin and hyaline, with a thin muscular tunic immediately beneath it, consisting of external circular and internal longitudinal fibres. The animal free-swimming, oceanic." Moseley's second article, which appeared nine months later, was again accompanied by figures, which will be found reproduced on PL I. figs. 23, 28-31. Tbe contents were as follows : — "On June 5, 1875, in lat. 34° 58' N., long. 139° 30' E., about halfway between Vries Island, Oosima, and Cape Sagami, the trawl was used by H.M.S. 'Challenger' in from 755 to 420 fathoms. A young specimen of a peculiar pelagic Nemertean, which has been described by me (Ann. and Mag. Nat. Hist., ser. 4, vol. xv. p. 1G5, March 1S75) under the name of Pelagonemertes Rollestoni, in honour of my friend and instructor Professor Rolleston, was found by Dr. von Willemoes Suhni adhering to the net, and by him handed over to me for examination. The adult specimen before procured and described was in a similar manner found adhering to the trawl-net after a deep-sea dredging by Dr. von Willemoes Suhm. " The animal was very much smaller than the one obtained before, measuring only 13 mm. in extreme length, and 11 mm. in extreme breadth, and about 1 mm. in extreme thickness. It was in good preservation when found, and living, and, being extremely transparent, much more of its structure could be observed than in the case of the more full-grown specimen. Unfortunately, an attempt to preserve the specimen by treatment with perosmic acid and subsequent action of glycerine failed, and the specimen perished. The trawl came up late in the evening, when only an hour of daylight remained ; the examination made was thus a hasty one. " The animal showed the same feeble pulsating movements which had been shown by the adult. " The external gelatinous investment of the body was perfectly transparent, and none of the peculiar corrugations of a thin superficial epidermic layer were visible as in the adult specimen. The contours of the body were well preserved, including those of the hinder portion, which was broken in the specimen before obtained. "The forepart of the body is wide, with rounded margins; the posterior narrowed, with a series of indentations on its margin corresponding to the successive pairs of 30 THE VOYAGE OF H.M.S. CHALLENGER. diverticula of the digestive tract. At the extreme hinder termination of the body is a shallow notch, at the bottom of which is the anus. " The mouth, which is a simple opening at the apex of a small, short, conical protuberance, was situate just in front of the nerve-ganglia on the ventral surface of the body (it is not shown in the figure, which represents the animal from the dorsal aspect). "The central canal of the digestive tract terminated in front in a wide, rounded, blind end, and tapered gradually to the anus at the posterior end of the body. "The lateral diverticula in this young specimen were evidently in an immature condition, and the successive pairs showed successive stages of development, the most anterior being the most fully formed. This most anterior pair is the only one which shows a commencement of ramification at the peripheral extremities. The ramifications, so ample and well marked in the adult worm, are seen here to be developed as csecal buds from the outer ends of the long diverticula. The diverticula themselves, of which five pairs were present in the young specimen here figured, arise, as can be seen from the figure (pi. xi. fig. 1 = PL I. fig. 23), as simple lateral buds from the central digestive tube. These buds gradually increase in length, their peripheral eaecal ends being always larger than the tubes connecting these with the central digestive tract, and eventually these caecal ends give off buds and form ramifications. A slight enlargement in the rectum, situate just anteriorly to the anus, and shown in the figure, probably represents the spot where a sixth pair of diverticula were about to bud off from the digestive tube. The diverticula, with the exception of the first pair, were not placed exactly opposite one another, the right diverticulum in each pair being situate anteriorly to the corresponding left one. " The digestive tract was filled with a dark reddish-brown matter, consisting of large granules and oil-globules. The contents of the diverticula were darker and more opaque, and contained numerous clear oily globules of a bright yellow and bright red colour, mingled with similar opaque globules (fig. 4, a = fig. 28). The brightly coloured globules exactly resembled those of the main tract. Similar coloured globules occur in larval Nemertines, and I have observed them also in a marine Planarian larva, possibly that of a Tliysanozoon, or the Planarian larva described by Johannes Miiller from the Mediterranean, supposed to be that of Eurylepta (Claus, : Grundztige cler Zoologie,' p. 286). " The sac of the proboscis corresponded exactly with that described in the adult. It was here found to terminate posteriorly in a blunt point at a short distance from the hinder end of the body. The fluid contained in it appeared transparent and without corpuscles. The proboscis itself could be carefully examined in the present transparent specimen. It was without stylets and quite simple, invaginated in the usual manner ; it was not seen fully protruded, but when so protruded must be slightly longer than the animal's body ; it has an outer pellucid gelatinous investment, and an inner muscular layer (fig. 5 = fig. 29). No retractor muscle was observed to be connected with it. " The nervous system consists of two pairs of ganglia, of which the upper are by far REPORT ON THE NEMERTEA. 31 the larger and give off the stout nerve-cords. The cords stretching backwards on either side unite with one another above the rectum at the hinder end of the body. The nerve- ganglia .are shown enlarged in fig. 3 ( =fig. 30). No sense-organs of any kind were detected. On the outer margin of the large superior ganglion (fig. 3) a series of elongate pellucid cells were arranged side by side perpendicularly to the curved surface which they form. Abundant fine nerves were given off from the entire length of the nerve-cords to the surface of the body, the muscles, &c, arising both from the inner and outer margins of the cords. At the origins of these nerves from the cords there are very slight swellings on the margin of the cord, but these do not contain any nerve-cells. The nerves are very fine, hyaline, with a nearly rectilinear course, and they generally divide into two near their points of distribution ; they are never tortuous or much ramified. Terminal organs on the surface of the body in connection with the nerves were carefully sought for but could not be found. " A pair of vascular trunks follow the course of the nerve-cords through the body, lying internally to them and beneath them. The vessels unite with one another posteriorly, as do the nerve-cords ; their course is undulating. Just behind the nerve-ganglia the vascular trunks are enlarged into wide reservoirs. No branches of these vessels were seen, and though the animal was living when examined, no pulsation iu them was observed. The vessels had a pellucid wall, in which were imbedded elongate oval nuclei (fig. 4, b = 6g. 28), but which otherwise appeared structureless. No motion of any fluid within the vessels was seen. " Although the specimen under description was evidently so immature, well-developed ovaries were present, the specimen being a female, as was the adult one before obtained. The ovaries follow in their disposition the vascular trunks so closely as to appear as if connected with them. The ovaries are simple ovoid sacs with a distinct wall (fig. 2 = fig. 31), filled with ova (in various stages of development) and granular matter. A dark irregular fissure appeared on the centre of each ovary as viewed from the dorsal surface, which I believe to be an opening by which the cavity of the organ communicates with the exterior, thus dorsally. The ovaries were not quite regular in disposition, an extra anterior one being developed on the right side of the body. In the interspace between the most anterior and larger pair of intestinal diverticula and the next posterior pair were four pairs of ovaries, whereas in the succeeding corresponding interspaces were only single pairs of these organs. In the adult specimen described in the 'Annals' (March 1875) a single ovarian sac only was present in each interspace between the diverticula of the digestive tract. It would therefore seem probable that on further development three pairs of diverticula would have budded out between the first and second pairs in the present specimen. " The muscular system consists of a series of excessively fine transversely or circularly disposed fibres, which are external in position to a series of broad band-like longitudinal 32 THE VOYAGE OF H.M.S. CHALLENGER. muscles. The longitudinal muscular bands are in close relation with the proboscis-sac. Their exact disposition was not made out, and their arrangement, as shown in the figure, will possibly need correction. " On the whole, Pelagonemertes is a form of considerable zoological importance. In the flattened form of its body, and in its dendroccelous digestive tract the animal resembles Planarians. Amongst the Rhabdocceles the Prostomeas possess an exsertile proboscis like that of Nemertines, but such an organ is present in no Dendroccele. In all particulars — in being unisexual, in the simplicity of the generative organs, in the form of the nervous and vascular systems and of the proboscis, in the position of the mouth and presence of an anus — in all essential structures Pelagonemertes is most distinctively a Nemertine. Only in its remarkable dendroccele intestine does it differ from all other Nemertines, and (but this is of far less importance) in the modification of its tissue into the peculiar hyaline gelatinous condition which is characteristic of so many otherwise most widely differing pelagic animals. " The development of the dendroccele intestine is very remarkable, in that the lateral ramifications are apparently to be regarded as a series of buds occurring successively from before backwards from a previously straight digestive tract, such as exists in other Nemertines. In this the digestive tract differs entirely from that of dendroccelous Planarians, such as Leptoplana tremcllaris, in which, as we know from the observations of Keferstein (' Beitriige zur Anatomie und Entwickclungsgeschichte einiger Seeplanarien von St. Malo,' Abhandl. der k. Gesellschaft der Wiss. zu Gottingen, 4ter Band, Gottingen, 1868, Taf. iii. figs. 19, 20, 21, text p. 34), 'the great yelkballs arrange them- selves in the embryo with regularity and map out the form of the future digestive tract,' the peripheral ramified part of the tract being formed at the same time as the central portion. " The peculiar form of the front of the body of Pelagonemertes may be regarded as an instance of the excessive formation of the head lappets of many Nemertines. In having no ciliated sacs and an unarmed proboscis, Pelagonemertes resembles Cephalothrix, but the animal must evidently be placed in a new family of Nemertines, for which I propose the term Pelagonemertidse, thus characterised : — "Animal pelagic in habit. Body gelatinous, hyaline, broad and flattened. Proboscis unarmed. Ciliated sacs absent. Special sense-organs absent. Digestive tract dendro- ccelous. " The occurrence of a second specimen of Pelagonemertes off Japan shows that the animal has a wide distribution. It was found on both occasions adhering to the trawl- net, and is, from its very slight consistency, easily overlooked. Hence it may have been often missed by us, and probably is as widely distributed as other oceanic forms. Since it has never been taken by former observers of pelagic animals nor by us in the tow-net, it is very probable that it occurs only in deep water, and does not come to the surface ; it is, however, most evidently not an inhabitant of the sea-bottom. REPORT ON THE NEMERTEA. 33 " Postscript. — Since the above was written, my attention has been directed by Dr. von Willemoes Suhm to Lesson's original figure of Pterosoma in the ' Zoology of the Voyage of the Coquille ' (which work we have been able to consult, with a splendid series of similar publications, in the Hawaian Government Library at Honolulu), and to the many points of resemblance between Pterosoma and Pelagonemertes. Fio. 2.— Pelagonemertes rollestoni, Moseley, enlarged, viewed from the dorsal surface ; the proboscis is partly extruded. P, proboscis ; PrS, sac of proboscis ; IP, invaginated portion of proboscis within the proboscis-sac ; G, superior nerve- ganglion ; NC, nerve-cords ; V, vascular trunk (the upper V points to an enlargement of the vessel lying just posteriorly to the superior nerve-ganglion) ; I, intestine ; D, diverticula of intestine ; 0,0, ovaries ; CM, circular muscles ; LM, longitudinal muscles. " Pterosoma plana is described by M. Lesson, ' Voyage de la Coquille, Zoologie,' Paris, 1830, p. 254, and figured, pi. iii. figs. 3 and 3 Us. Pterosoma was obtained in great abundance by Lesson between the Moluccas and New Guinea, August 31, 1828. "The animals measured 3 inches and some lines in length, 18 lines in breadth, and 3 to 4 lines in thickness. In general form and gelatinous structure Pterosoma resembles (ZOOL. CHALL. EXP. — PAET LIV. — 1886.) Hhh 5 31 THE VOYAGE OF H.M.S. CHALLENGER. Pelagonemertes further, in that a series of polygonal areas are marked out on its surface. The spirally wound organ, described as a tube, which is indicated in the figure of Pterosoma, can scarcely be anything else than the proboscis of a Nemertine, the mouth, at the extreme end of the body, being probably the aperture of the proboscis-sac, and the fusiform nucleus the sac itself. On the other hand, it is difficult to conceive that Lesson, with a number of sj)ecimens available for examination, could have missed seeing the very conspicuously burnt-sienna-coloured ramified intestine of Pelagonemertes had such been present in his Pterosoma. Further, in Pterosoma, a pair of elongate, closely opposed eyes are described and figured, having transparent coloured cornese. " On the whole, now that a pelagic Nemertine is known to exist, there seems little doubt that the animal seen and figured by Lesson was a Nemertine and not a Mollusc, but it seems to have been a distinct form, with a pair of eyes and an unbranched digestive tract." For the sake of facilitating reference to Lesson's two figures of Pterosoma, in which Moseley is so strongly inclined to see a second species of pelagic Nemertine from the tropical seas — a supposition in which I entirely concur — I have reproduced these on PL VIII. figs. 1, 2. The single available specimen of Pelagonemertes, when it came into my hands in September 18S4, was no longer entire, but consisted of two fragments of the body, and of three small fragments of the proboscis. Both body-fragments were slit open on the ventral side, the internal surface of the digestive tract, and partly that of the proboscidian sheath, having thus been laid bare. The terminal portion of the body was missing (vide Moseley, supra). My friend Professor MTntosh, who made these incisions and examined the specimen before it came to me to be sectionised, has made the following notes: — " The structure of the cutis corresponds with that of other fonns. In the preparation the transparent gelatinous basis-substance for the most part alone remained, the granular cells and clear mucous or gelatinous contents of these spaces having escaped. Beneath this layer is a remarkable deep investment of basement tissue, which forms an elastic investment all round the body. Numerous ducts, often having a zig-zag appearance, pass through this coat. "The muscular layers of the body-wall, as Professor Moseley observes, are compara- tively thin. The external circular fibres are hardly to be distinguished in ordinary transverse sections, and form a thin layer outside the longitudinal. Such a condition contrasts strongly with the well-marked circular coat in Amphiporus lactijioreus. The longitudinal muscular coat is likewise comparatively thin, though it is better developed than the circular. " Proboscis. — In the preparation the proboscis is partly extruded, and issues in the same manner as in the ordinary form. It is streaked by longitudinal lines externally. REPORT ON THE NEMERTEA. 35 " In minute structure the anterior region quite corresponds with the typical form, presenting externally a layer of elastic fibrous tissue, which in transverse section presents a series of circular fibres. The external longitudinal muscular coat is well developed. The reticulated layer cuts the foregoing layer into sections, as usual in most other forms, and the connecting bands seem to be broad, the longitudinal belts rounded in transverse section. The condition of the specimen, which is imperfectly preserved, however, probably causes this layer to be more prominent, and the separation between the two longitudinal coats wider. The inner longitudinal muscular coat has the usual thickness and appearance, and the same may be said of the inner circular muscular coat. The basement layer is very largely developed, and fills up a considerable part of the central area. It is somewhat regularly streaked by a radiating series of granular channels. Occasionally tortuous and numerous granular cells- occur in it. The central granular glandular tissue has mostly disappeared, but in minute structure it corresponds with that in other forms. " The middle region of the organ presented no visible stylets, but otherwise it agrees in general configuration with the typical form, though less definitely formed. The glands of the posterior region are largely developed. "The posterior part of the proboscis sheath had a quantity of flocculent material microscopically presenting numerous granular, and often nucleated cells, mixed with granular material ; such might be connected with the proboscidian gland. " The mouth, as Professor Moseley observes, forms a well-marked aperture with a frilled margin. It is proportionally the most distinct aperture yet observed. The minute structure of the oesophageal region is similar to that of other forms, and it terminates similarly. " The dendritic arrangement of the digestive system, Professor Moseley states, is the most remarkable feature about Pelagonemertes, indeed it differs from all other Nemerteans. The condition of the organ in the other forms, however, renders the arrangement less striking. In Nemertes gracilis, for instance, it is considerably divided. Microscopically it agrees with other forms in cells, &c. Professor Moseley describes a distinct anus in the young form, but the specimen may have been incomplete posteriorly. It is certainly unusually distinct. The vascular system offers no particular feature of interest so far as observed, and seems to follow the ordinary arrangement. The lateral vessels appear to vary a little from the ordinary relation to the nerve-cords, being often internal rather than inferior, but probably such is due to the yielding nature of the tissues. The inner lining of the vessels is granular, and a few granular cells appear in the centre, but the nature of such is problematical " The nerve-trunks are enveloped in a coating of nerve-cells. The branches from tbe ganglia and nerve-cords are remarkably distinct, presenting a clear sheath and granular axis cylinder. The gelatinous interstitial substance extends between the muscular 36 THE VOYAGE OF H.M.S. CHALLENGER. substance and fills up the central area of the body in transverse section to an unusual extent, as pointed out by Professor Moseley. " Besides the organs mentioned by this author within the body-cavity, a large number of granular nerve-cords run outwards to the surface (muscular wall), besides fine fibres, which are very apt to assume a coiled or zig-zag appearance. Some of these much resemble blood-vessels, but they are probably nerves, and they divide into fine branches towards the muscular coat. " The proboscidian sheath or chamber is very large, and its anterior aperture would seem to be unusually distinct. The structure of the parts in unrolling of the organ is the same as in the others, and the wall of the chamber presents continual circular internal longitudinal fibres; the posterior region of the proboscis is firmly fixed as usual to the internal wall of the proboscidian sheath at its narrow posterior end. The fibres of attachment are short, so that the cul-de-sac of the posterior chamber is brought close to the surface of the wall of the sheath. From the appearance of the parts posteriorly, it is possible that the region is in process of repair after laceration. Indeed, it is not unlikely that both anterior and posterior ends can easily be repaired after rupture, and that might account for the absence of eyes (see Lesson's Pterosoma, 'Voyage of the Cocpuille'). Such gelatinous forms are especially prone to rupture, and sufficient is known of the recuperative power of the Nemertean to render repair rather than permanent injury the rule. The proportionate size and firmness of the proboscidian chamber, with its glistening internal surface, are certainly remarkable, and, on the whole, my impression is that the form is incomplete posteriorly." All the fragments of Pelagonernertes that came into my hands were carefully treated with staining reagents, hardened, imbedded and sectionised. The sections were all trans- verse. PI. VIII. fig. 3 represents one of the average sections with parts of all the important organs imbedded in the common gelatinous basis. No sufficient sections of the brain were available, nor were distinct traces found of a nephridial system. The integument was in most parts of the surface wholly deficient ; in a few others its general correspond- ence in transverse section with what is typical for the Hoplonemertea could be verified. Certain other striking particulars, also visible in this section, and on the whole only confirmatory of what we already knew by Moseley's and MTntosh's observations, are (l) the absence of a dorsal median blood-vessel; (2) the ventral openings of the generative sacs; (3) the numerous transverse fibres, both contractUe and nervous, running in different directions through the body -jelly. Further, it may be observed, even with a rather low power, that each of the ripe ova is surrounded by numerous small follicle cells (PL VIII. figs. 3, 11), and that the nucleolus is often only represented by very numerous small chromatin spheres. By the aid of stronger powers the details also represented on PL VIII. by figs. 4-13, can be more fully studied, and of these mention will successively be made in REPORT ON THE NEMERTEA. 37 the different paragraphs, which are more especially devoted to the internal anatomy of the different representatives of the species, that are here described in their systematic arrangement. C SCHIZONEMERTEA. Family Lineidj;. Cerebratulus, Ren. To this genus I wish to refer all the Schizonemertea collected by the Challenger. I have elsewhere (VIIl) insisted on the difficulty of distinguishing the genera Cerebratulus, Lineus, Micrura, &c, of which perhaps the two first may be distinguished by an ontogenic difference (Pilidium or Desor-laxva). And even this distinction is not definitely established. It is simply impossible to refer spirit, specimens to any one of these genera rather than to any other, and having formerly included Micrura as a synonym amongst Cerebratulus, I now even feel inclined to do the same with Lineus. What value has a generic distinction when it can never be of any use to a taxonomist ? And why should a developmental difference, such as that which obtains between a Pilidium and a Desor larva, not be sufficiently honoured by a specific distinction ? As to describing new species in this genus, it is even more difficult than in any other, because of the number already existing, which are partly solely distinguished by their coloration in life, a character not available in determining the Challenger specimens. Hence, only in seven cases can I hold myself justified in referring the fragments to separate species, six of which are new. The remaining fragments and heads, which clearly show the Cerebratulus type (e.g., PI. XV. fig. 8), I will not specifically distinguish; what remarks I have to make about them will appear when treating of the anatomy, and will then be noticed as pertaining to the genus Cerebratulus in general. Following the order of succession of the stations at which the Schizonemertea were successively dredged during the expedition, we will now proceed to describe them. Cerebratulus truncatus, n. sp. (PI. I. figs. 11, 12). This species, collected on the coast of Nova Scotia, and, as we refer a specimen from Bermuda to it, also frequenting the ocean in the vicinity of those islands, was captured May 8, 1873 (Le Have Bank), and once more, on May 20 of that year, when it came from a depth of 85 fathoms at Station 49. The anterior part of these two specimens is figured on PI. I. figs. 11, 12, the first being a side view, the second a ventral view, the latter more considerably, the former about four times enlarged. From both figures it is seen that the mouth is small and close 38 THE VOYAGE OF H.M.S. CHALLENGER. to the tip of the head, which is more or less truncated, that the cephalic slits are compara- tively short, extending on to the anterior surface of the head, where they do not, however, coalesce, but leave a small interval in which the proboscidian aperture is situated. This is all that can be said of the external appearance. The internal structure further confirms the supposition which the external characters point to, viz., that these two specimens belong to the same species. Thus, certain chief points in the series of transverse sections made through both the specimens coincide to a degree that may be judged of by the following table : — The superior brain commissure is situated in section No. The inferior brain commissure is situated in sections No. Canal of the posterior brain lobe is situated in section No. The mouth begins in section No. The mouth ends in section No. First appearance of nephridia in section No. . In Specimen a. In Specimen h. 30 34 35-39 38-44 43 51 55 60 84 103 97 112 In specimen a (May 8, 1873) there are noticed four transverse deferent ducts to the longitudinal nephriclial duct on the left side (in sections 138, 175, 217 and 278), and four on the right side (sections 147, 191, 217 and 278). In specimen b (May 20, 1873) there is a less regular arrangement, some of the deferent ducts being double, i.e., two at the same level or in the same section. Once in this specimen this is so arranged that in the one section (No. 197) there are four deferent ducts cut nearly throughout their whole length. In the portion of the oesophageal region sectionised, which, however, does not embrace the whole nephridial region, I count in this specimen on the left side six deferent ductules (sections 112, 144, 156, 159, 197, 207), and to the right also six, which are only partly opposite to the left ductules (sections 142, 168, 178, 180, 197, 208). In judging of this apparent discrepancy it should not be lost sight of that the short distance separating the two ductules 156 and 159 on the left is still symmetrically repeated on the right side in 178 and 180, though somewhat further backwards. A commissure uniting the two vagus nerves after they have left the brain and before they have yet reached the oesophagus was distinctly noticed. The histological details of the integument fully correspond to the type of Cerebra- tulus corrugatus, which will hereafter be more fully described and figured (cf. PI. XIII. fig. 6), and which is diagrammatically represented in fig. 9 of PI. XL as differing from figs. 10 and 11. One of the specimens showed a very curious pathological degeneration, which I will however, only touch upon very briefly. The muscular tissue on a restricted but ring- shaped region not far behind the head, and only as far as the circular and longitudinal muscular coats are concerned, is replaced by (or gradually passes into) a homogeneous REPORT ON THE NEMERTEA. 39 gelatinous mass, in which granular spherical bodies with large and distinct nuclei (about one-fifth of the diameter of the body-wall on this spot) seem to be suspended. Is this a local injury? Are these bodies enclosed gregarines or other parasitical unicellular organisms ? These questions can as yet only be formulated but not answered, and I would call the attention of future observers dealing with fresh material to similar cases, which no doubt must be accompanied by some externally visible distinction between the tissues. Of the Bermuda specimen, which I refer to this species only with doubt, and which measured 33 mm. in length by 1\ mm. in diameter, the integument was very imperfectly preserved ; still it corresponds in the remaining characters with the other specimens of Cerebratulus truncatus. There are also points of agreement in the structure of the proboscidian sheath. The transverse blood-vessels are very thick-walled in this specimen, a feature more particularly due to the basement-tissue below the inner epithelial lining. Cerebratulus medullatus, n. sp. (PL XL fig. 10 ; PL XII. figs. 9, 10). It is hardly consistent with the most lenient rules of zoological nomenclature, and it is certainly not consistent with those which I have myself advocated in my introduction, to establish a new species on a fragment which has neither head nor tail ! Still, as I have been rather careful on other occasions, and that especially with regard to genera, I may be trusted to be anxious to guard against superfluous additions to an already cumbersome synonymy. The reasons by which I am guided in putting up this mutilated spirit fragment as the type of a new species are purely morphological, and, as will be seen in the paragraph treating of the nervous system, the peculiarities offered by this species are of sufficient morphological interest to give it a place by itself, and (in at the same time naming it) to direct the attention of collecting naturalists to this form. It came up in the same haul of the dredge as did one of the specimens of Cerebratulus truncatus just described, viz., at Station 49, off Nova Scotia, from a depth of 85 fathoms. In sections it is, however, immediately seen to be distinguished from the foregoing species by several features, the first of which concerns the integument. Instead of the integument being separated from the longitudinal outer muscular layer by a more or less massive expanse of basement-tissue (PL XIII. fig. 6), the integument itself being divided into a superficial and a deeper glandular layer (which are separated by a secondary basement membrane and an expanse of fibres, such as is found in Cerebratulus corrugatus, Cerebratulus truncatus, &c, also in the Challenger collection), our present species has the superficial layer of the integument immediately applied to the outer longitudinal muscles, from which it is thus only separated by the outer secondary basement layer just alluded to (PL XII. fig. 10). This fact causes the integument in transverse 40 THE VOYAGE OF H.M.S. CHALLENGER. sections to appear very thin in comparison to that of other species. However, closer inspection reveals (as will be more fully discussed in the paragraph devoted to the integument) that the deeper glandular layer is not wholly deficient, that it has only become intercalated amidst the outer longitudinal muscular layer, which in this species is not very massive, and that its cellulo-glandular elements thus reach amongst these muscle fibres, even as far down as the nervous layer just outside the circular muscular coat (cf., PI. XII. figs. 2, 10). The strong affinity which these gland-cells possess for staining reagents brings this out in the sections very clearly. Nor is this the only distinctive feature of Cerebratulus rnedullatus; the second, and none the less important, is that in the nervous plexus just alluded to, the clorso-median longitudinal thickening, which I shall presently, in the anatomical part of this Report, designate as the medullary nerve, is exceptionally massive (cf., PI. XI. fig. 10 ; PI. XII. fig. 9), being about one-third to one-fourth of the thickness of the fibrous core of the lateral longitudinal nerve-trunks. Amidst the fibres of this dorso-median stem a few nuclei, more faintly coloured and marking the presence of nerve-cells, are also seen. In longitudinal sections through the same specimen, the unusual size, distinctness, and marked individuality of this nerve-stem was also very obvious. No other features of this species will for the present be enumerated. The specimen on which it is founded was of the female sex. I sincerely hope that the points here enumerated may enable American naturalists, when recapturing specimens of the species, to recognise it and to give us indications of its colour and other peculiarities in life, of the shape of its head and cephalic slits, and of its brain lobes and proboscis. Cerebratulus longifissus, n. sp. (PL I. fig. 16 ; PI. XV. figs. 1, 9, 10). No other Schizonemertea were collected between the last-mentioned station and Marion Island (Station 144a). A Cerebratulus was here brought to light, which it will in future be easy to recognise by the fact of its having uncommonly long cephalic slits. The specimen, which was perfect, and is also characterised by its comparative shortness, is fio-ured on PI. I. fig. 16. On inspecting it with the naked eye the mouth was found to be small and to be situated anteriorly; the proboscidian aperture occupied the place indicated in the figure, whereas on the dorsum of the animal, just behind the end of the slits, two rows of sublateral, very small pores were noticed, being visible as extremely small white punctures. These rows were continued very far backwards. Although, on account of their reaching so very far forwards, they might at first sight be taken for the exterior openings of the nephridial system, the sections showed that they are indeed the generative pores. Moreover, that the cephalic slits, though long, are comparatively shallow, especially in their posterior portion, and that the canals leading into the posterior REPORT ON THE NEMERTEA. 41 brain-lobes do not open out in the lower angle of the fissure but far forwards, in agree- ment with the situation of those lobes. It will perhaps best serve the purpose of conveying an idea of the relative situation of the more important organs in the head and trunk to give, as was done for the fore- going species, the number of the section of the whole transverse series (into which head and anterior oesophageal region were cut up) in which these organs occur. We then find: — Section 30. First appearance of the lateral blood- lacunae in the head. Sections 64-66. Superior commissure of the brain- lobes. Sections 71-78. Inferior commissure of the brain- lobes. Section 80. Left canal from cephalic fissure into posterior brain-lobe. Section 87. Right canal from cephalic fissure into posterior brain-lobe. Sections 84-104. Left posterior brain-lobe. Sections 89-105. Right posterior brain-lobe. Section 71. The median blood-vessel enters the proboscidian sheath. Sections 350-360. It again leaves it. Sections 129-159. Mouth. Section 129. First appearance of the nephridial canals just above the level of the longitudinal nerve-trunk. Sections 355-400. Left deferent ducts of nephri- dial system. Sections 329-335. dial system. Right deferent ducts of nephri- These two deferent ducts (the only pair) are seen in the same section which still shows the cephalic fissures. The ducts do not, however, open out into these fissures, but just above them. This also proves how far the fissures reach, all along the oesophageal region. In the 410th section the end of the cephalic fissures may be said to be reached, the whole of the nephridial system, excretory ducts and all, being thus situated within the region of the fissures. Another specimen from the same locality was without a head, and though the principal points of comparison are thus deficient, I feel confident, by the internal characters which are shown by the sections, that we may look upon this specimen as belonging to the same species. It was exceedingly well preserved and showed certain interesting points with regard to the generative organs, which will be more fully discussed in the paragraph devoted to those parts. Another feature by which the species is — if not distinguished from its neighbours — at least characterised as far as the fragments allow us to judge, is the great thickness of the transverse connecting vessels between the medio-dorsal vessel and the two lateral ones. Cerebratulus corrugat us (M'Intosh), Hubrecht (PI. I. fig. 17 ; PI. XI. fig. 9 ; PI. XI h figs. 3, 4 ; PI. XIII. figs. 1-6 ; PI. XIV. figs. 2-4). This species, which was first described by M'Intosh in the Transactions of the Royal Society for 1879 (extra vol., p. 262), has again been recognised by this author as occurring (ZOOL. CHALL. EXP. — PART LIV. 1886.) Hhh 6 42 THE VOYAGE OF H.M.S. CHALLENGER. amongst the Challenger Nemertea. I have already indicated (p. 37) why I wish to bring it under the genus Cerebratidus, rather than under Linens, where M'Tntosh placed it. The following notes were made by him concerning the Challenger specimens. " Dredged in considerable abundance in Royal Sound, Kerguelen, January 27, 1874, in 25 fathoms. "Stat. 149, off Christmas Harbour, Kerguelen, 120 fathoms, small specimens. " Stat. 151, off Heard Island, 7th Febr. 1874, 75 fathoms. Some reach the length of 200 mm., and much contracted forms have a diameter of about 15 mm. In the oesophagus of one was a blackish mass of sand-grains, sponge-spicules, Diatoms, and mud." In M'Intosh's. previous publications the description of Lineus (Cerebratulus) corru- gatus contains the following characteristic features : — "Body rather abruptly pointed anteriorly, and more gradually posteriorly. The oesophageal region is marked externally by a series of prominent and somewhat regular rugae, which sweep from the mouth dorsally and ventrally. " Colour dark olive throughout, with the exception of a white band, which crosses the anterior border of the snout, and passes backward to the posterior third of the lateral fissure, where it bends dorsally and terminates." " The special characters are the very large mouth, with the prominent rugse, which show that the ammal probably possesses unusual powers of oesophageal protrusion — a supposition borne out by the great development of the external circular muscular fibres, the dorsal longitudinal coat, and the other fibres of the organ. The internal glandular lining is also very firm." A couple of figures are added, one a section of the proboscis, the other a section of the ventral body-wall. The latter should be compared with our fig. 6 on PI. XIII. It will then be seen that M'Intosh's " pigmentary layer, divided by a definite black band," is our superficial and deeper layers of the integument (Isg and Idg). The " black band" between them is no other than our external secondary basement membrane. What MTntosh designates as the " curious translucent stratum cut into somewhat regular spaces" is our basement layer proper (B), comparable to that of Eupolia and Garinina, and radially traversed by bundles of contractile and nervous fibres, which bring about the "regular spaces" alluded to. On the whole, our two figures will be seen to correspond very well, only M'Intosh omits the nervous layer and the innervation of the oesophagus. The large mouth and folded lips were very conspicuous in the Challenger specimens, the head of one, seen from the ventral surface to show the mouth, being figured on PI. XIII. fig. 5. The different series of sections which I have made through four specimens of Cerebra- tulus corrugatus were very instructive in several respects, although they all conform to the well-known Schizonemertean type. It is especially the considerable development of basement tissue of the integument (PI. XIII. fig. 6, B), which not only brings out the REPORT ON THE NEMERTEA. 43 distinction between the outer longitudinal muscular layer 7 and the integument much more clearly than is so often the case in other Schizonemertea, where these two have be- come blended, but which also enables us to trace the course of radial nerve-fibres coming from the plexus or the longitudinal stems, and innervating (after having traversed the muscular layer y and this basement tissue) the glands and sense-cells of the integument. Cerebratulus corrugatus is, moreover, the species in which, for this reason, I was able to determine the direct part which the longitudinal nerve-trunks take in the innervation of the skin (PL XIV. fig. 2), whereas, even in this very favourable specimen, I never noticed the faintest trace of a similar participation of nerve-branches directly springing from these stems in the innervation of the subjacent musculature. In that case such branches would have to take an opposite course, and would have to traverse in the first instance the circular muscular layer /3. This was never observed. It will be seen in the paragraph more especially treating of the nervous system, as well as in the chapter devoted to general considerations, that this fact is, in my opinion, not without morphological importance. The specimen of Cerebratulus corrugatus is also of great importance in demonstrating the relation of the medio-dorsal medullary nerve and the branches springing from it at right angles as so many thickenings of the plexus (PL XIII. fig. 2) . Cerebratulus parkeri, n. sp. (PL XIV. fig. 5 ; PL XV. figs. 5, 16). A well-preserved specimen of a Schizonemertean was collected in the New Zealand waters, which, I think, may safely be looked upon as belonging to a distinct species. I have dedicated this species to the naturalist who of late years has done so much for our knowledge of the New Zealand fauna, and whose anatomical preparations and zoologi- cal collections from those regions have excited the admiration of visitors to the Colonial and Indian Exhibition of 1886. The head and anterior body fragments of Cerebratulus parkeri are thus described in M'Intosh's preliminary notes : — "A fragment of the anterior portion, measuring about 34 mm. in length, and about 7 mm. in diameter at its flattened and widest region. The cephalic furrows and mouth conform to the ordinary type. The body is somewhat rounded anteriorly ; flattened towards the posterior end of the fragment. " Transverse sections of the rounded anterior region show that the muscular walls of the body are greatly thickened .... the longitudinal muscular fibres form a very thick coat all round, especially, as usual, opposite the nerve-cords The circular muscular coat is uniformly thick, the longitudinal layer within it being excavated superiorly by the large channel for the proboscis, and laterally by the very large and very muscular vascular canals — a little below each nerve-trunk. The thickened part of 44 THE VOYAGE OF H.M.S. CHALLENGER. this coat is just beyond the proboscidian canal, while the thinnest part is in the median line above it. " The proboscidian sheath has a remarkably thick wall, which is chiefly composed of an interwoven circular coat, presenting an irregular looped appearance in transverse section. The fibres in thin sections seem to anastomose. On the inner surface of this coat are some longitudinal fibres, with a glandular epithelial coat internally. " Posteriorly the body-cavity is greatly dilated by the presence of a vast series of ova, so that all the muscidar coats are much thinned, and the digestive canal contracted. The inner longitudinal coat is especially affected, three thickenings only being left, viz., one on each side of the proboscidian sheath, and one in the middle line ventrally. The ova are compressed into various angular forms, between vertical partitions which occur at short intervals in the region, and which fill up the entire area, except the small space for the proboscidian and digestive canals superiorly." For some farther particulars concerning this species, I refer to the description of Cere- bratulus macroren that is still to follow, and where the points of agreement and of differ- ence between these two evidently closely related species will be more fully entered upon. A second specimen, also from the New Zealand waters, shows that the conspicuous development of the secondary basement layer of the integument, and the considerable thickness of the medullary nerve, are further characteristic features of this species. Cerebratulus angusticeps, n. sp. (PI. I. fig. 15 ; PI. XIV. figs. 1, G ; PI. XV. fig. 4). This is another Schizonemertean from the New Zealand waters (Station 167a, Queen Charlotte Sound, June 27, 1874 ; 10 fathoms). Its head is figured on PI. XV. fig. 4. In MTntosh's preliminary notes this fragment is referred to in the following words : — " A fragmentary form, resembling in shape C. angulatus, and with an acutely pointed snout. The diameter of the widest flattened region is about 4 "5 mm. The colour is dull yellow, with a darker stripe down the middle of the dorsum The vascular trunks lie opposite the nerve-cords, and this in the angular lateral region. The external longitudinal muscular coat is largely developed, and the fibres are uniformly fine. The very much produced lateral angle of the body is mainly composed of this coat. In such forms the lateral nerve-cord seems to be much flattened from above downwards." The sections which I made of this fragment, horizontally through the head and trans- versely through the trunk, were in many respects very instructive. To the flattened and pointed shape of the head they added an internal distinctive feature of the species in the unusual size of the posterior brain-lobes, which, although in intimate connection with the superior brain-lobes in the same way as in the Schizonemertea, in general equalled or even surpassed the latter lobes in length. PL XIV. fig. 6, gives a representation of this, illustrating, at the same time, how in a horizontal section the blood-space (c.o.l.) REPORT ON THE NEMERTEA. 45 surrounding the oesophagus may very evidently be seen to be in direct connection with that in the head — a fact known to Blanchard, and for the first time fully described by Oudemans (loc. cit.), but never adequately figured. This figure, at the same time, demonstrates the projection of the posterior brain-lobe into this blood-space, by the con- tents of which it is thus bathed. Another figure reveals how the sections of Cerebratulus angusticeps were important in another respect, viz., that of the medio-dorsal medullary nerve and the transverse branches springing from it in metamerically arranged pairs (PI. XIV. fig. 1). Nor was this regularity exceptional in the region figured ; it was characteristic wherever these trans- verse nerve-tracts (which will be discussed more fully in their relation to the plexus in the special paragraph) were met with, and was thus equally distinct ventrally and dorsally. Ventrally, however, there is no median longitudinal stem. On the other hand, the transverse tracts in Cerebratulus angusticeps may be traced as high up as the lower brain-lobes, which they connect till just behind the strong ventral commissure of these lobes. These transverse commissural trunks are distinctly separate from those by which the two> stems of the vagus nerve are united close to their origin and in front of the mouth. The latter (cf. PL XIV. fig. 5) are also present as distinct commissures in Cerebratulus angusticeps. With respect to the nephridia, which often offer certain points of comparison for the different species, I must state that the anterior part of the longitudinal duct is very distinctly seen in the horizontal sections through the anterior extremity of the body, and occupies its usual position in the circumoesophageal blood-space, whereas its size or its histological details showed no special features. The deferent ducts of the nephridial system were not contained in this section, nor in those made through the remaining fragments of the further portions of the body. They were very probably situated in the intervening part, which was sacrificed by MTntosh in drawing up his preliminary notes, and I cannot thus state with certainty whether these ducts were single or more numerous, terminal or not. To Cerebratulus angusticep>s I must refer another specimen which came up in the dredge at Station 168 from a depth of 1100 fathoms. It was much torn and lacerated (PI. I. fig. 15), a phenomenon which it would, however, be rash to attribute to the depth from which it was brought up, although the possibility of that being the cause cannot be wholly excluded. The reasons for identifying the specimens with Cerebratulus angusticeps are the following, and are deduced from the comparison of the sections : — (1) The length of the superior brain-lobes stands to that of the posterior in the same (uncommon) rela- tions as in the specimen above described ; (2) the ventral and parallel commissures of the inferior brain-lobes are similarly very conspicuous ; (3) certain histological details of the superior brain-lobes and ciliated canal, and also of the nerve plexus, are identical ; (4) the aspect of the rhynchodseum is very similar. 46 THE VOYAGE OF H.M.S. CHALLENGER. X. figs. 8, 7-9; PI. XIV. fis;s. 9 ; PI. XL 7, 8, 11; Cerebratulus macroren, n. sp. (PI. I. figs. 13, 14, 18, 19; PI. fig. 11 ; PL XII. figs. 1, 2, 7, 8 ; PL XIII. fi PL XV. figs. 2, 3, 19). From Japan, at Station 232a, a Schizonemertean has been brought home by the Challenger Expedition, which I was first inclined to regard as identical with Cerebratulus parkeri, but which I have reluctantly been obliged to distinguish specifically — reluctantly, because there being no possibility of giving outwardly visible distinctive features of colour or shape, I must indicate an anatomical difference as the principal point of distinction between the two ; and all the more reluctantly because of a second specimen, which, dredged at 700 fathoms off New Zealand (Station 169), was identical with the Japan specimen, and differed from its much closer geographical neighbour in this very same point. I could not, however, evade assigning the latter specimen to the species now about to be described, knowing by experience that if the confusion created by the unnecessary multiplication of specimens is often very troublesome, the premature com- bination under one name of forms that afterwards may be shown to be different is often quite as apt to lead to inextricable confusion, when characters which, as a matter of fact, only belong to one of them, are attributed to both. About this second specimen M'Intosh's rough notes contains the following passage : — " A small Linens, presenting the ordinary external characteristics, and of a dull yellowish hue, somewhat brownish on the dorsal surface in front. It was tapered from the snout backwards. The cutis has a simple areolar structure ; the gelatinous contents were slowly extruded in the sections after mounting, as cylindrical or clavate, translucent gela- tinous processes. The basement tissue is largely developed as a translucent belt all round." Of the larger specimen from Japan, the head is figured (PL I. figs. 18, 19). The cephalic slits are perhaps comparatively a little longer than are those of Cerebratulus parkeri (PL XV. fig. 5), although the New Zealand representative (PL I. figs. 13, 14) is again very similar, even in this respect, to Cerebratulus parkeri. The internal anatomical character, to which I alluded just now as being a distinctive feature, by which the species differs from Cerebratulus parkeri, is the size of the longitudinal nephridial ducts, which are exceptionally conspicuous (both in the Japanese and in the New Zealand specimens), and which have their deferent ducts leading to the exterior situated at the very hindmost end, whereas in Cerebratulus parkeri there are two deferent ducts, the one situated very closely behind the other, and placed about the middle region of the longitudinal canal. The accompanying wood- cut diagramrnatically illustrates this difference between the two species. Oudemans (loc. cit.) has already demonstrated that the number of deferent ducts to the nephridial Fig. 4. — Nephridial ducts and their communication with the exterior, a, for Cerebratulus parkeri ; b, for Cerebraltdus macroren. REPORT ON THE NEMERTEA. 47 system increases with age ; there thus might remain one chance that the New Zealand specimen of Cerebratulus macroren could still be assigned to Cerebratulus parkeri, if we assume that increase in growth can have brought about a further extension backwards of the principal nephridial duct, and at the same time the appearance of a second deferent duct immediately behind the first. For the present this assumption appears to me to be more strained than my own, which unites the New Zealand and the Japanese specimens by laying more stress upon the large size of the longitudinal tube, combined with the terminal situation of the deferent duct. The further peculiarities that reveal themselves on studying the microscopic sections, certainly show that the two species, Cerebratulus macroren and Cerebratulus -parkeri, cannot be very far apart. Both have in common the very thick and homogeneous secondary basement layer beneath the outer glandular layer of the integument (PI. XL fig. 11). It would also be difficult to point out salient points of disagreement in the muscular body-wall, the proboscidian sheath, and the proboscis which would hold good when respectively comparing the head, the oesophageal, or the posterior body region. Cerebratulus, sp. inc. (PI. X. fig. 7; PI. XV. figs. 6-8, 18.) At the close of our systematic description of the Schizonemertea I must mention certain fragmentary specimens, which have all the aspect of belonging to distinct species, but which I cannot venture definitely to unite with any of the species here described, or with such as have been pubbshed elsewhere. The fragments here alluded to are mostly without a head, and some of them of not inconsiderable size. I will discuss them in the order of the stations at which they were obtained. The first was procured in the Kerguelen waters. It is important, in consequence of peculiarities in its integument, which will be more fully discussed in the paragraph devoted to this system. A part of a section was figured on PI. X. fig. 7, and from that section it may also be gathered that the dorso-median medullary nerve is comparatively very massive. This might eventually prove that it was related to Cerebratulus meclullatus ; the differ- ence in the integument, though important from a morphological point of view, hardly justifying the establishment of a different species, supposing all the other characters, external and internal, might prove to be identical. That difference might then be con- sidered as indicative of a variety. The second Cerebratulus, about which I must remain in doubt, was obtained among the Philippine Islands (off Zebu). MTntosh has made the following notes about these fragments : — " Fragments of a large species. The fragments in all measure over 100 mm., with a diameter of 12 mm. at the widest part The carrying of the vascular trunks far inwards towards the ventral middle line seems to be a feature . . . in this form." 48 THE VOYAGE OF H.M.S. CHALLENGER. There were further no particulars, nor any features deviating from the general type of Cerebratulus to be gathered from these fragments. Numerous other fragments, of which one was a head, were obtained at Kobe (Japan). Three of these are figured on PI. XV. figs. 6, 7, and 8. Sections were duly made of these fragments, but do not give much additional light beyond the general result that we have a Cerebratulus before us, which cannot be definitely identified with any of the species hitherto described. The vicinity of the Japanese waters to scientific centres from wrhich accurate descriptions of the Japanese marine invertebrates may probably be expected ere long, makes it all the more advisable to refrain for the present from creating or identifying species from these regions of which sufficient data cannot yet be obtained from the available fragments. Still the specimen figured on PI. XV. figs. 6 and 7 deserves some attention. The fragment, on which longitudinal and transverse white stripes were visible, as indicated in the figure, was also distinguished by a peculiar rigidity. Trans- verse sections (PL XV. fig. 7) showed that this phenomenon was occasioned, or at all events accompanied by, an extraordinary development of intermuscular gelatinous tissue. The section figured, when compared with that of Pelagonemertes (PI. VIII. fig. 3), will demonstrate this, and at the same time show it to possess the arrangement of muscular layers and other peculiarities that are typical for Schizonemertea. These muscular layers are, however, exceedingly reduced in thickness, and occupy a very inconsiderable fraction of the vertical or horizontal diameter. The proboscidian sheath is, in the fragments investigated, thin and unimportant ; in more posterior sections there are indications of its place being taken by more irregularly shaped, cellular material, without a lumen. I am, however, not satisfied with the details that could be gathered from these fragments concerning these important morphological points, and must refrain from more particularly insisting upon them. The transverse blood-vessels are exceedingly numerous and tor- tuous— the latter phenomenon causing their lumen to be transversely cut a large number of times in every section ; these apparent perforations of the gelatinous tissue giving a very peculiar appearance to most of the sections. Another reason why I do not venture to establish a distinct species upon these peculiar fragments, is the fact that I found them to be very considerably infested by a large unicellular parasite (probably a Gregarine) which was not (as the Gregarines that infest Nemertea generally are) found in the intestine, but which was present in consider- able numbers in all the different tissues, both without and within the muscular layers. When surrounded by gelatinous tissue, there was always a well-marked space round the parasite in which it was contained. There was no distinct capsule, and the free space may perhaps not have existed during the life of both host and parasite. When first noticed, these unicellular parasites, with very distinct nuclei and granular protoplasm, might have been mistaken for ova ; not only their distribution throughout the animal as isolated individuals, but also the REPORT ON THE NEMERTEA. 49 presence of true ova in the fragments, which proved to have belonged to a female individual, definitely excluded the possibility of any confusion on this head. The question arises, whether the peculiar appearance of the different tissues recorded above might be pathological, and somehow in causal relation to the infesting parasites. It is this which necessitates extreme caution in the identification of these fragments. I may here remind the reader that another case, presumably of a parasitic organism (also unicellular, but of much smaller size) infesting the different tissues, was met with in Amplii-porus marioni. The last fragmentary specimen which I wish to record may, for all I know, have belonged to the common Cerebratulus margvnatus. It was collected in the Atlantic Ocean, at Station 321, off the Brazflian coast. It was a very large and flattened speci- men, but without head or tail. In MTntosh's notes I find these fragments referred to as follows : — "Two fragments, respectively 70 and 108 mm. in length, and with a transverse diameter ranging from 21 to 24 mm., the vertical (in the centre) being only 5 or 6 mm. The dorsal surface was dull olive, with a dark median band, the greater part of the breadth being marked by fine transverse striae, leaving only the borders untouched. Various transverse lines, passing quite across the body, also occurred ventrally. The median line and the borders had each a smooth belt, the rest being marked by the closely arranged and tran verse lines. A median ridge occurred along the ventral band." The ova of this species, polygonal by reciprocal pressure, and surrounded by a gela- tinous outer layer (capsule), are figured on PI. XV. fig. 18. (ZOOL. CHALL. ESP. — PART LIV. 1887.) Hllll 50 THE VOYAGE OF H.M.S. CHALLENGER. List or the Stations at which the different Species of Nemertea COLLECTED BY THE CHALLENGER WERE OBTAINED. Distin- guishing No. of Date. Latitude and Longitude. Depth in Fathoms Nature of Bottom. Species collected. Station. 1873. 45 May 3 38° 34' 0" N., 72' 10' 0" \V. 1240 Blue mud Carinina grata. 47 7 41" 14' 0"N., 65° 45' 0" W. 1340 Blue mud )) 51 ,, 8 Le Have Bank, N. Scotia 75 /', i-rhwhilii.x truncatus. 49 ,, 20 43° 3' 0"N., 63°39'0" W. 85 Gravel, stones Drepanophorus lankesteri, Ccrebratulus truncatus, Ccrebratulus medullatus. ... June Bermuda Ccrebratulus truncatus, Tctrastemma agricola. Bermuda to Azores Gulf weed Tct rastemma fuscum. July St Vincent, Cape Verde Islands Eupolia delincata, Drepanophorus rubro- striatus. 144a Dec. 26 1874. 46°48'0"S., 37° 49' 30" W. 69 Volcanic sand Ccrebratulus longifissus , Amphiporus marioni. 149 Jan. 9 49° 8' 0" S., 70° 12' 0" W. (Acces- sible Bay) 20 Volcanic sand Amphiporus moseleyi. 149E „ 21 49° 37' 0"S., 70° 16' 0" "VV. Royal Sound, Kerguelen 30 Volcanic sand Cerebratulus corrugatus. Amphiporus moseleyi, Drepanophorus serra- ticollis, Ccrebratulus corrugatus. j» j > Christmas Harbour Cerebratulus sp. inc. (medullatus?), Amphi- porus moseleyi. 151 Feb. 7 52° 59' 30" S., 73° 33' 30" E. 75 Volcanic sand Cerebratulus corrugatus. 158 Mar. 7 50° 1' 0" S., 123° 4' 0" E. Pelagoiiemertes rollcstoni. 162 Apr. 2 39° 10' 30" S., 146°37'0"E. "38 Sand and shells Drepanophorus scrratieollis. 167a June 27 41° 4' 0" S., 174° 19' 0" E. 10 Mud Cerebratulus parkcri, Cerebratulus angusli- 168 July 8 40° 28' 0" S., 177° 43' 0" E. 1100 Blue mud Cerebratulus angusticeps. 169 „ 10 1875. 37° 34' 0"S., 179° 22' 0" E. 700 Blue mud Eupolia giardii, Cerebratulus macrorcn, Eupolia australis. 209 Jan. 22 10°14'0"N., 123°54'0"E. 95 Blue mud Cerebratulus sp. inc. 232 May 12 35° 11' 0"N., 139°28'0"E. 345 Green mud Ccrebratulus macrorcn, Ccrebratulus sp. inc., Eupolia nipponensis. June 5 34° 58' 0" N., 139° 30' 0" E. off Japan Pelagoiiemertes rollcstoni. 1876. 321 Feb. 25 35°2'0"S., 55° 15'0"W. 13 Mud Ccrebratulus sp. inc. A glance at the foregoing table shows that the very large majority of Nemertea were captured in shallow water, and that they may, as a rule, be said to belong to the littoral fauna. The most aberrant new types are the pelagic Pelagoiiemertes, and the genus that has come from the greatest depth. Carinina grata. The interesting peculiarity of a strongly developed medullary nerve, which is much less conspicuous in the European species hitherto investigated, occurs in species so widely apart as the waters of Nova Scotia {Cerebratulus medullatus), of Kerguelen {Cerebra- tulus sp. inc. (medullatus?), of New Zealand {Cerebratulus parkeri), and of Japan {Cere- bratulus macroren). If we group the species according to geographical distribution, we find as novelties in the American part of the Atlantic — Carinina grata, Drepanophorus lankesteri, Cerebratulus medullatus, and Cere- bratidus truncatus. EEPOET ON THE NEMEETEA. 51 Off Marion Island : Amphiporus marioni, Cerebratulus longifissus. Off Kerguelen : Amphiporus tnoseleyi, Cerebratulus corrugatus. Off New Zealand : l Eupolia giardii, Eupolia australis, Cerebratulus parTceri, and Cerebratulus angusticeps. Off Japan : Eupolia nipponensis and Cerebratulus macroren. It might Lave been expected that a collection of Nemertea from different parts of the globe, preserved in spirit, would prove to be less valuable for our intimate knowledge of genera and species, and for our elaboration of the systematic arrangement of this group, than it would be for anatomical and histological purposes. Where external specific characters have come to be totally effaced, the details of the internal framework remain most per- fectly preserved for microscopic investigation. In the preceding pages we have encoun- tered difficulty in attempting specifically to determine the fragments forming this collection, and we have at the same time seen that only in four cases (Eupolia delineata, Drepanophorus rubrostriatus, Drepanop>horus serraticollis and Cerebratidus corrugatus) coincident with extraordinarily favourable circumstances, was identification possible with species already known. That nevertheless new species could be established with certainty must merely be ascribed to the fact that the distinguishing internal specific characters, gathered by means of microtomy, were so marked and so divergent. Still we cannot picture to ourselves the appearance of these new species when alive and in the fresh state. As just noted, we have in this collection a much more reliable basis for the study of the anatomical and histological details. We must, indeed, recognise that we have to thank the exceedingly well-preserved collection of Challenger Nemertea for several new facts and suggestive results, to the description of which we will now devote a new chapter. 1 Mention is made in M'Intosh's Monograph (XIX., p. 96) of a New Zealand Nemertean, distinguished by Baird as Lineus novx-zelandix. I have nowhere been able to find any description of this species by that author. After personal inspection of Dr. Baird's specimen, which is preserved under that name in the British Museum, I do not, however, feel justified in identifying with it any of the Challenger specimens from New Zealand. Externally it much more resembles Cerebratulus corrugatus, and if this latter species is really encountered in New Zealand, Baird's name will have to be dropped altogether. ANATOMICAL INVESTIGATIONS. INTEGUMENT. By "integument" I wish to designate the cellular layers as they are found outside the more or less stratified connective tissue, which is known as the basement membrane. The latter is of very varying thickness, and, for reasons to be given subsequently, will be treated in another paragraph. Glandular structures of the integument, whether enclosed within the cellular layers just alluded to, or stretching inwards between the muscles and piercing a secondary basement membrane, hereafter to be described, only by means of their communicating ducts with the exterior, will also be treated of here. Commencing with a description of the integument of the Palseonemertea, it will be well to take the more important genus of which representatives are found amongst the Chal- lenger Nemertea, viz., Carinina, as a type. This is all the more desirable as we shall here find the central nervous system still clearly belonging to the integument, its constituents imperceptibly merging into those of the deeper cellular layers of the skin, and also lying outside of the basement membrane. The two specimens of Carinina at my disposal revealed the same features with respect to their integument, although in one of them the granular secretion in the glands that form part of it is much more copious. When we leave out of consideration the basement membrane, that can be easily detected in the sections by the uniform and deep red tint it acquires by the staining reagent (picrocarmine), we can roughly distinguish four constituent strata in the integument not in any way separated by sharp boundary lines, but characterised by the presence of different histological elements which we will now proceed to describe more fully. It wdl be understood that the absence of fresh material and the scanty supply of spirit specimens has necessarily limited the exact discrimination of these histological elements. Of the four strata alluded to, the one adjoining the basement membrane is extremely important, being the seat of those cellular modifications which must be looked upon as the differentiation of the central nervous system within the domain of the integu- ment. This position — it was already noticed in the description of the species given 54 THE VOYAGE OF H.M.S. CHALLENGER. above — remains the same in the adult, and we have in this more superficial situation of the central nervous system one of the surest indications of the more primitive position that ought to be assigned to Carinina and the allied genera, when compared with the other Nemertea. The intermixture of integumentary and nervous tissue is none the less evident in the medio-dorsal longitudinal nerve than in the lateral stems. The first named nerve, which in former publications (IX, X) I have, not wholly adequately [cf. p. 132), desig- nated as the proboscidian-sheath-nerve, can readily be distinguished in my transverse sections of Carinina as a delicate stem. It is not situated, as in the Schizonemertea, just outside the circular and below the outer longitudinal muscular coat, nor, as in Drepano- phorus, Amphiporus marioni, &c, below the basement membrane just outside the same circular muscular coat, but it lies in this case actually outside the basement membrane, and forms part of the deepest layer of the cellular integument. Further, a plexus-like distribution of nervous tissue between dorso-median and lateral nerve-stems obtains in this species, as was more fully described elsewhere for the Schizonemertea, connecting the three longitudinal stems and spreading round the body as a cylindrical investment. It must be remarked that this plexus-like arrangement is thus necessarily situated in the very layer of the integument with which we are occupied, and it may be added, that in numerous transverse sections of Carinina the presence of fibrillar nerve-tracts in this layer can be easily demonstrated, and these exactly resemble those that are met with in the nervous plexus of Schizonemertea. This plexus has, since my first notice of its presence (IX, X) been again observed by Dewoletzky (II) and other naturalists, and will be more fully discussed in a succeeding paragraph. The jnexus here alluded to merges into the lateral nerve-stems. I may here once more emphasize the fact that the whole system lies outside the basement membrane. At the same time, however, the lateral stems would seem to be separated, though very incompletely, from the integument by bundles of fibres which bind them down to the underlying layer of circular fibres (PI. III. figs. 7, 8.) In the vicinity of the brain-lobes it is impossible to distinguish between the cells of the deepest integumentary layer and the nerve-cells. With respect to this layer I have further to state that its nuclei are less conspicuous, its cells jxaler, and the boundary lines of the latter less easily distinguished than in the other layers. Extremely delicate fibrillar tracts were already noticed as occurring in it. The next layer to this, when we pass outwards, is the glandular layer of the integu- ment. It is the thickest of the four, beings alone often as massive as the three others taken together. The large, flask-shaped, and tubular glands it encloses contain a thick granular secretion, which is partly stained brownish-yellow, partly dark red in my pre- parations (PL III. figs. 3, 7, 8, gi ; PI. IV. fig. 1 ; PI. VI. figs. 1-3). The ducts leading to the exterior penetrate the two outermost layers. The presence in the most peripherally REPORT ON THE NEMERTEA. 55 situated of the two latter layers of similar though much shorter glandular cells (PL IV. ficr. 1 ; PL VI. fig. 2, E), immediately contiguous with others of a larger size, makes it a subject for inquiry whether these are connected by transitional forms to those larger ones which form the layer we are now describing. This involves the not unimportant question, whether the glands constituting this layer are or are not unicellular. While I am inclined to accept the latter proposition, I feel that the question can only be solved by the aid of a careful inquiry into fresh specimens. The layer external to the glandular is one in which very numerous and deeply stained nuclei are heaped together. These nuclei occupy several rows in one section. The same may be said to apply, as will be indicated further on, to the corresponding layer of deeply staining nuclei in the integument of the Hoplonemertea. The very outermost layer bearing the cilia clearly contains, in Carinina, the same elements as will be described more in detad (vide infra, pp. 58, 61) for the Schizonemertea, i.e., nervous end-cells, alternating with supporting cells, " Stiitzzellen." This layer, containing fewer nuclei and less granular protoplasm, is more transparent in transverse sections, and distinguished by fine radial striae, indicating the boundary lines between the contiguous cells. The four layers here described are not equally distinct in all sections, nor are they equally well marked in all the sections figured. Thus, for example, in fig. 4 of PL III. the glandular layer is feebly developed, and the two external layers are so indifferently preserved that their distinctive character, just described, fails to attract attention. The chief points enumerated can, however, even there, be easdy ascertained. The partial absence, or, at any rate, temporary indistinctness of these gland-cells in certain portions of the integument can also be observed in tangential sections, such as the one of PL III. fig. 8. There, too, the groups of glandular cells to the right and the left are separated by a band of integument, in which they are decidedly absent. The same figure shows the different ways in which the contents of these gland-cells react on staining agents ; those to the right in this section have decidedly yellowish contents, whereas the contents of the group on the left had a deep carmine tint. This description of the integument of Carinina must now be followed by that of Eupolia, the only other genus of Palaaonemertea contained in the Challenger collection. The interesting genera Carinella, Cephalothiix and Carinoma, are not represented in those collections and certain intermediate characters displayed by these genera, which serve to justify the identification of the different layers, which as I am going to propose can only be touched upon as far as they elucidate the phenomena. This I will defer till after the detailed description of the integument of Eupolia. Instead of four layers it would not be difficult to distinguish eight in the integument of this genus, not all of them separately and clearly represented in every section, but sometimes (e.g. PL VII. fig. 5, 9) sufficiently distinct. This arrangement may be looked upon as a further differentiation of an earlier phase, corresponding to that of Carinina, 56 THE VOYAGE OF H.M.S. CHALLENGEE. the chief difference being that the basement membrane upon which the whole integument rests, is not here, at least not in all specimens, so clearly defined, nor comparatively so structureless, as it is in Carinina, and as we shall again find it in the Hoplonemertea. Moreover, identification is somewhat obscured by the appearance of a second, homo- geneous, very thin basement membrane, which has also a strong affinity for the staining reagent, and which we must be careful separately to distinguish, if we wish to establish an adequate comparison between the parts in the different genera. This second basement membrane (B, in PL VII. figs. 2, 3, 5,6,9; b, in PI. XII. fig. 2 ; PL XIII. fig. 6) divides the integument into two strata — an external one, comprising the peripheral sense-cells and ciliated cells, the unicellular glands, and the layer of deeply-stained nuclei ; and an internal one, containing the longer and more tortuous glands, the deepest integumentary tissue, and, moreover, at least two very thin layers of fibres. If we count the thin basement membrane alluded to for one layer, this makes three strata externally and four internally to it — eight altogether. The different aspects of these strata may be gathered from the figures on PL VII. Fig. 6 shows the three layers outside the membrane B, and though the histological elements were not isolated, it was very obvious that large unicellular glands were here pouring their secretion to the exterior. The very outermost layer was here, as in all Nemertea, formed of strictly radially arranged cells, with far less distinct nuclei, whereas between and just below the secreting cells strongly stained nuclei give to this part of the integu- ment the peculiar radially striped appearance which it has when viewed with lower powers (figs. 2, 3, 5, 9). This same peculiarity is only less visible in fig. 6 because of the very copious discharge of secretion in the gland-cells. From figs. 5 and 9 it is sufficiently obvious, however, that these glands are not the only ones, but that in the layer indicated by Gi the darkly-stained secretion of more deeply lying, larger, and more irregularly-shaped glands is unmistakable, and is also seen to communicate with the exterior by fine tortuous tubes piercing the superposed membranous and cellular strata, about ten of these ducts being specially indicated in fig. 5. Their direct passage into the respective glands is not always visible in one section, the course of the tubes being tortuous. The same glands, though also present in the sections shown in figs. 2 and 3, are there less marked, because the secretion has not yet so distinctly accumulated. Here, too, the reference letters Gi, point to the stratum in which we find them imbedded. A second constituent of this stratum is seen in cells similar to the secreting cells in the unripe stage (figs. 2, 3), but having afterwards a very distinctly vacuolated character, and then forming the surrounding and sustaining tissue for the functional glands. They might, then, best be designated as vesicular connective tissue (" blasenformiges Bindegewebe "), with hardly any intercellular substance. The comparative thickness to which this part of the integument may attain is best understood from PL VII. figs. 5, 9, Gi. That there is a sharp line of demarcation between it and the gelatinous or REPORT ON THE NEMERTEA. 57 lamellar connective tissue of the basement layer Bet, upon which the whole of the integu- ment rests, may here be specially insisted upon, and is indicated in the same figures. Finally, I have to mention the two layers of fibres which, though very delicate, form in Eupolia very constant and characteristic parts of the integument. They are imme- diately applied against the inner surface of the supernumerary basement layer B — an outer layer of circular and an inner one of longitudinal fibres. Fig. 9 {Eupolia delineate/,, long, sec.) shows them to be more conspicuous than figs. 2, 3 {Eupolia australis). Fig. 9 moreover, serves to demonstrate that the pigment, to which in this species the peculiar longitudinal brown stripes are due, is accumulated in the same stratum of the integument, where these fibrous layers are found ; the section represented shows an unpigmented zone between two pigmented ones. The pigment is granular, and appears to be limited to this stratum. It was not met with in the other species of Eupolia. That I am justified in looking upon the integumentary arrangement of Eupolia as a higher differentiation of a lower type, which in general resembles the integument of Cari- nina, must now be shown by a short account of the condition of things in Cephalothrix and Carinoma, two other Palasonemertea, both of them inhabitants of the European seas. Cephalothrix shows an advance in differentiation upon Carinina, by the deeper situation of the lateral nerve-stems (PI. XL fig. 15), imbedded in the outer longitudinal muscular layer. Still it deserves special attention, that in this genus the medio-dorsal nerve is still situated in the deepest layer of the integument (PI. XL fig. 5) outside of the basement membrane, and that, at the same time, the integument wholly answers to the description that we have given of the integument of Carinina. Carinoma, whilst generally agreeing with Cephalothrix in the situation of the lateral nerve-stems, no longer retains the longitudinal medio-dorsal nerve as part of the integu- ment, but in a somewhat deeper situation, enclosed in the basement-membrane. It also shows very decided complications in the structure of the integument. Whereas the basement layer is most closely similar to that of Carinina and Carinella, the outermost integumentary layer is much more distinctly cellular, provided with unicellular glands, and separated from the deeper glandular layer by the development of two layers of muscular fibres that were first noticed by MTntosh (XXIV), and by him interpreted as two accessory muscular layers of the body-wall. They are such, in fact, although I feel confident that we may look upon them as forming an integral part of the integument, and as being, together with it, wholly of epiblastic origin. Not wishing to give a figure of Carinoma, which does not form part of the Challenger collection (the diagrams on PI. XL may, however, be consulted) , I must needs appeal to the confidence of the reader that a glance at a transverse section of the integument of Carinoma carries with it the conviction that it is, in this respect, truly the most conclusive intermediate form between Carinina and Eupolia, so that we are amply justified (l) in declaring the basement membrane (B of PL III. figs. 3, 4, 7 ; PI. IV. fig. 1) to be not homologous with the one (ZOOL. CHALL. EXP. — PAST LIV. — 1887.) Hllh 8 58 THE VOYAGE OF H.M.S. CHALLENGER. also lettered B on PI. VII., but to be so with the deeper layer, Bet of the latter plate ; and (2) in comparing the glandular structures of Carinina more especially with the deeper layer of glands of Eupolia, and in looking upon the unicellular glands of the latter as more particularly developed in the outermost layer. These latter are then comparable to the smaller and superficial unicellular glands which are met with in Carinina (PI. IV. fig. 1). Similar considerations concerning the homology of the differ- ent portions of the integument can be applied, as we shall see in the sequel, to the Schizonemertea, their integument in so many points resembling that of Eupolia, and being thus also linked to that of Carinina by intermediate forms, such as Carinoma. Amongst the Hoplonemertea it is most difficult to obtain specimens in which the integument is in a fair state of preservation. Even in the specimens that have been treated with special care, it is a rare occurrence to find a portion of the integument in which its different layers can be clearly made out. From what I have noticed in the Challenger specimens, there is a certain amount of uniformity which must first be noticed and compared with what obtains in Palseonemertea and Schizonemertea. Sharply distinct from the basement-membrane, which, after removal of the integument, would even show a more or less honeycombed surface, are the deeper cell-layers of the integument, the rounded bases of these cells fitting into the honeycomb- like pits in the basement-membrane just alluded to (PI. X. fig. 2). These cell-layers, with very distinct nuclei, are in thin sections many rows thick. The nuclei are, however, never so close together as they are in the subsecpient layer, which is generally situated about halfway between the basement-membrane and the ciliated surface (PL X. fig. 1). On teasing out the elements of this layer, we find long spindle-shaped cells, considerably thinning out at the ends, and only bulging at the spot where the deeply stained nucleus is situated. These nuclei, again arranged in several rows even in the thinnest transverse sections, are there disposed so as to fit close between each other, the thin ends of the cells being directed one towards the outer surface and the other towards the deeper layers before mentioned. It cannot be doubted that sense-cells, which very generally have a similar shape and position, are among these (cf. Dewoletzky, II). Towards the outer surface, the remaining stratum of the integument has the peculiar radially streaked appearance already described for the Palfeonemertea ; nuclei being rarer in this layer than in the two foregoing. Finally, the cilia are implanted upon the outer margin of this region, and teased preparations reveal the presence of special cells with a nucleus of a very much paler hue, and very faintly coloured. Only in a few preparations have the cilia been well preserved ; a cuticula upon which they are implanted, as elsewhere (XIV) described in embryonic stages, was also only noticed in certain favourable sections. These are the principal features marking the integumentary system of the Hoplone- mertea. A few further details, however, may still be added. The deepest layer (PI. X. fig. 2) contains fibrous nerve elements, not distinctly indicated in the figure. The absolute REPORT ON THE NEMERTEA. 59 thickness of the layers, above described as integument, does not vary very much in smaller and larger-sized individuals. Also, in the tip of the tail, where growth is continually going on, and where the newly formed parts are thus, in the first instance, found, the integument is very much in advance of the underlying tissues in obtaining its definite size. For example, in Drepanophorus lankesteri I find the cellular integumentary layer (the basement membrane not included) to be on the thickest part of the body, 0-15 mm. (i.e., one-twentieth the horizontal diameter in this region), and close to the tail end O'l mm. (i.e., one-tenth the horizontal diameter in this region). In the largest specimens of Amphiporus moseleyi, which have more than twice the length and more than two and a half times the transverse diameter of Drepanophorus lankesteri, I find the integument to measure 0"12 to 0"15 mm. Both in specimens of Amp)hiporus moseleyi and in one of Amphiporus marioni, as well as in one of Drepanophorus, it is easy to demonstrate glands in the integument \>y the evident presence of the secretion, partly extruding towards the exterior, partly still enclosed between the deeper layers (PI. X. fig. 2). These glands are, however, neither so massive, nor do they form such a conspicuous layer as in Carinina, above described. Whether they may be regarded as unicellular, and as comparable to the superficial ones already mentioned in the integument of Eupolia, and also present in the Schizonemertea, or whether they are more directly homologous to the deeper glands of that division, is for the present difficult to decide by means of the preserved material in the Challenger collection. Histological investigation of fresh specimens will be necessary to clear up this point. Drepanophorus serraticollis would certainly tend to a direct comparison with the flask glands of Cerebratulus, &c. In a general way, however, the Hoplone- mertean integument offers more analogy to the more primitive arrangement than to the higher differentiation of the layers in Eupolia and the Schizonemertea. One further detail deserving mention is the presence of a granular deposit (pigment?) in the lower cell-strata of the integument of the last mentioned specimen of Drepano- phorus. This deposit is different from the one hereafter to be mentioned in the tissues of Amphiporus marioni. A deposit comparable to the latter was absent in the specimen of Drepanophorus here alluded to. The granules have more resemblance to the pigment granules of Eupolia delineata, described above, and may probably be looked upon as such. While we see that the Hoplonemertean integument is directly connected — at least in general outlines — with that of the Pak-eonemertean genera Carinina and its allies, the Schizonemertea are linked to the primitive stock by the intervention of Eupolia and Carinoma. Certain Schizonemertea (e.g., Cerebratulus corrugatus) in some portions of the integument reveal a complete uniformity with what we have described for Eupolia ; a stratified basement layer separating the outer longitudinal muscle-bundles of the body- wall from the layer of vesicular tissue sustaining the deeper skin glands, these glands 60 THE VOYAGE OF H.M.S. CHALLENGER. communicating by tortuous tubes with the exterior, and being separated from the outer- most cell-layers with unicellular glands by a special secondary and continuous, though thin basement layer (PI. XIII. fig. 6). Below the latter, longitudinal and circular fibres proper to the integument are also present, corresponding, even in their more massive development, to those of Carinoma. There appears to me to be no doubt that this same arrangement holds good for the great majority of Linei and Cerebratuli, and the only reason why the separation of the parts is often less marked is the stronger development of the outer longitudinal muscles of the body-wall, concordant with the disappearance of the connective tissue separating the integument and body musculature, and also the fusion to a smaller or larger extent of the longitudinal muscle-fibres proper to the integument with those of the body-wall. From this it inevitably results that the line of separation between the body-wall and the integument seems to be formed by the external, secondary basement membrane, the deeper glands having the appearance of being imbedded within the longi- tudinal muscles of the body (PI. XI. figs. 10, 11 ; PL XII. figs. 2 and 10). That this is a secondary arrangement, and that the real and original line of separation was another one, has been demonstrated in the foregoing pages. It is not necessary, after the detailed description of Eupolia given above, once more fully to discuss the same details for those Schizonemertea that wholly correspond to the same type of integument. In those species in which the more developed longitudinal muscular layer more or less effaces the boundary line between integument and muscles (Cerebratulus macroren, Cerebratulus meduttatus, &c), the characteristic and sometimes massive layer of vacuolated cells surrounding the deeper glands is considerably reduced. The other con- stituent parts have retained their original character, with the exception of the thin muscu- lar strata of the integument, which are no longer separately recognisable (PL XII. fig. 10). The integument is generally very completely preserved in the cephalic fissures; it may here be noted that there, too, the deeper gland-structures of the integument may be noticed, although they are much more sparingly set. In a few cases it would appear as if they are wholly absent, and as if only the outer integumentary cell-layer is preserved in the cephalic fissures; others, again (PL XIV. fig. 11), offer special differentiations in the region of the cephalic fissures of the glands, which may there be united in paired accumu- lations. I must also mention a somewhat aberrant type of integument, as we find it represented in a Cerebratulus sp. inc. (meduttatus?), from Kerguelen Island. The integu- mentary layers offer more general resemblance to what obtains in the more primitive Pakeonemertea (Carinina, Cephalothrix, &c.) than to Eupolia. Eventually it might be said to retain a more primitive embryonic condition. I have at least described a develop- mental phase of the integument very similar to what I am now about to describe for adult forms, as occurring in the ontogeny of Lineus obscurus (XIV). The integument in question may, however, also be looked upon in another light, i.e., as in no way REPORT ON THE NEMERTEA 61 more primitive, but rather as a special differentiation of the normal type of the Schizo- nemertean integument, the result being an apparent simplification. This simplification (PI. X. fig. 7) consists in the absence of the deeper layer of gland-cells below the superficial secondary basement membrane b. The whole integument of this species would thus only seem to correspond to the very outermost layer (the layer of the unicellular flask glands) of the other Schizonemertea. I have, indeed, the conviction, that far from being more primitive, this condition may be linked with what obtains in other Schizonemertea and in Eupolia by such transitional forms as Cerebratulus medullatus, &c. (PI. XII. figs. 10 and 2 ; PI. XL figs. 10, 11). These species have evidently well- developed glands belonging to the deeper layer ; and whilst the glands are on all sides surrounded by the longitudinal muscles, they at the same time penetrate more deeply into this layer, even at a few points touching the nervous stratum, which is superposed upon the layer of circular muscular fibres (PI. XII. fig. 10). The special character of the integument of Cerebratulus sp. inc. (medullatus ?), and the deceptive reminiscences it evokes of the more primitive stages of the integument may well be said to be due to the strong secondary basement membrane. Thus in this species the layer of the deeper glands seems to have altogether disappeared. Before passing to another paragraph, I cannot refrain from pointing out the many points of resemblance that may be noticed between the integument of the Polyclada, now so well known, thanks to A. Lang's beautiful monograph, and of certain Nemertea, viz., those in which the integument is secondarily simplified as in those last discussed. Our external layer of unicellular glands is evidently comparable to what Lang and Graff call the " Schleirnstabchenzellen " or " pseudorhabdites," and these in their turn are compared by Lang, on very plausible grounds, with the " Rhabditeuzellen," in which the peculiar rod-like enclosures of the integument are found. The highly refractive, uniform contents of what I have called the unicellular glands, their general shape and properties, wholly coincide in their semi-viscous nature with what are described by Lang as the " Schleimkorper," and looked upon by him as merging into true glandular structures. The Nemertean layer of deep glands is also found in the Polyclada, below the (also secondary ?) basement membrane. So is the layer of nuclei regarded by Lang as belonging to a continuous stroma, by which both sense-cells and gland-cells are sustained and which was recognised by me in all the subdivisions of Nemertea. The figure given by Lang (XVIII ; PL XL fig. 11) would fit very well for different genera of Nemertea ; only in Nemertea the rod-like viscous bodies are not subdivided into superposed blocks. More- over, the tactile and sensory cells in the integument, as Lang describes and figures them, more especially for the tentacular integument, offer without doubt a close analogy to that outermost layer of the Nemertean integument, with its triangular cells (tip down- wards), which is also found in the vast majority of species where the skin is uninjured, which is uniformly distributed over the body, and which, if indeed sensory, as appears 62 THE VOYAGE OF H.M.S. CHALLENGER. to me most probable, would go a long way to explain the high degree of sensibility of every portion of the Nemertean body-wall. MUSCULAR SYSTEM AND CONNECTIVE TISSUE (GELATINOUS TISSUE, BASEMENT MEMBRANE, &c). In describing in the foregoing paragraphs the integument and its varied constituents, glands, sense-cells, ciliated cells, &c, the tacit assumption has been made that the structures there described might be looked upon as so many derivatives of the epiblast. Although reliable embryological data are as yet very scanty, my own experience on this head (XIV, XV) appeared to me to afford justification for this assumption. However, I agree that the question, whether the thin layers of longitudinal or circular fibres, that, more especially in Eupolia and Cerebratulus corrugatus (PI. VII. figs. 5, 9 ; PI. XIII. fig. 6), form so intrinsic and conspicuous a part of the integument, are also epiblastic derivatives, or whether they are due to mesoblastic elements, is open to dispute, and cannot be solved for the present on any other than the a priori arguments just alluded to. Hence, if I look upon the tissues that are treated of in the present section as essentially meso- blastic structures, I wish it to be well understood that this distinction may after all not be a final one. I have purposely omitted discussing the basement membrane of the integument under the head of the integument, because it appears to find its more natural place amongst what we are now going to describe : the tissues between the outer cell layers and the intestinal epithelium, i.e., the muscular body- wall and the connective tissue (better, gelatinous tissue, " Fiill-Gewebe"). The latter is not only present in the space between the body-wall and the intestine (so far as it is not encroached upon by the generative, blood-vascular, or nephridial systems), but also between the individual muscle-bundles, when these are not very closely applied against each other, and outside of these, between the muscles and the integument, as the so-called basement membrane above mentioned. The question as to the exact nature of this tissue is, in my opinion, a very important one. It represents the tissue which in Ccelenterata fills the space between epiblast and hypoblast, the "jelly" of Medusas and Ctenophora, with its multifarious inclusions of muscular, fibrous, and eventually nervous nature. This jelly is the more im- portant since its distribution, in the way above defined furnishes a strong argument for the view, also held by me, that the Nemertea are devoid of a body-cavity comparable to that of Arthropods, Annelids, and of Vertebrates. The only body-cavity proper to the Nemertea is the modified segmentation cavity, the archiccelome, as I have elsewhere pro- posed to call it (XIII, XIV). Of the cavities of the generative sacs and of the nephridia mention will be made in the respective paragraphs. REPORT ON THE NEMERTEA. 63 It will be well to consider this connective tissue more closely before we pass to the description of the muscular layers of the body-wall. From the foregoing it may already be inferred that there is a direct continuity between the different parts of this gela- tinous tissue, be it situated close to the intestinal epithelium or to the integument, and that this continuity is more or less completely interrupted by the muscular layers. We may thus conclude that it will be most prominent in those species that have the muscular body-wall reduced to a minimum, whereas it will be hardly visible in species that have a very strongly and massively developed musculature. This is indeed the case, Pelagonemertes offering a very striking example of the first category, Carinina of the second. Although the latter species is by far the most primitive, I would hesitate very much in at the same time regarding the relation of the gelatinous tissue to the body-muscu- lature of this deep-sea form as typically representing the original arrangement. Both species mentioned represent an extreme ; the normal starting point may be more easily derived -from what we find in Eupolia and in most Hoplonomertea. It then becomes obvious that our gelatinous tissue, though uniform and continuous, still appears in three principal modifications, which, however, are often connected by transitional phases having the characteristic features of more than one of these modifications. The first of these modifications is found between the muscles and the integument, the second in the midst of the muscular bundles, the third between the muscular invest- ment and the internal organs. The first modification just alluded to appears in Carinina (as also in Carinella and other Palseonomertea not collected by the Challenger) as a wholly homogeneous base- ment layer, on which the deeper cell layers of the integument are implanted, partially honeycombing it in the way above noticed. It is strongly stained by picrocarmine, and in Carinella traversed along circular and longitudinal tracts by nervous tissue. In Carinina the corresponding tracts are still situated in the deeper layers of the integument itself. Nuclei are very rare. What is a distinct basement layer in the more primitive Palseo- nemertea just named, retains this character with but little change in the Hoplonemertea. One change which is revealed at first sight is a distinct though exceedingly fine stratifica- tion, that becomes apparent in the basement membrane of nearly all Hoplonemertea. Along with this we very often find included in the Hoplonemertean basement layer distinct and sometimes numerous nuclei. The other inclusions in it, as they are figured on PL X. fig. 1, B, appear to be parasitic unicellular organisms infesting this particular specimen (see p. 49). Besides the exceedingly fine stratification which is parallel to the surface of the body, and which is often thrown into wavy folds, there are numerous radial tracts that would seem to transverse this basement layer, but are often only due to slight differences in texture and coloration, or to hardly susceptible folding or contraction (PL VIII. fig. 13). Where actual radiating fibres can be demonstrated, they 64 THE VOYAGE OF H.M.S. CHALLENGER. sometimes are found to be extremely thin nerve-stems connecting the cellular integu- ment with the central nervous system, whereas in other cases their spiral coiling and their affinity for staining solutions permits us to define them as contractile or elastic fibrils. It is, however, not to these radial fibrils that the extreme pliability and continual change in thickness of the basement membrane of the Hoplonemertea can be . ascribed. This phenomenon must be an inherent quality of the tissue itself, and may be studied in every transverse section, where the outer boundary line of the basement membrane is only very rarely parallel to the inner one. Generally it is strongly undulated, in accordance with the folds and wrinkles into which the integument may be thrown, not only during life, but also when the animal is preserved in spirit. The consequence of this undulation is, that in several places the integument much more closely approaches the muscular body-wall than in others, where it is kept very widely apart, the basement membrane being in the first case compressed ; in the second, extended to its utmost limit. Together with this extension and contraction, the fine parallel stratification changes its aspect, becoming more coarse, and sometimes so coarse that it might be difficult not to look upon the basement layer as composed of fibres. A confusion with subjacent mus- cular layers would in some cases be pardonable. A comparison of longitudinal and transverse sections reveals, however, the fact that it is indeed no fibrillar, but a stratified condition. Another change accompanying these phenomena of extreme plasticity is the change in colour, the staining appearing far more intense when the strata are in the contracted than when they are in the expanded condition. The nuclei remain visible in both cases. How the change of shape and the successive expansions and contractions are actually brought about in this homogeneous though laminated tissue, which has more the appearance of being intercellular ground substance than anything else, must here remain an open question, which we shall again meet when treating of the contractions of the muscular body -wall. In PL VIII. fig. 13, and PI. X. figs. 1, 2, different aspects of the Hoplonemertean basement membrane are given. I will now pass to those of the Palseonemertean genus, Enpolia, that in so many respects leads over to the Schizonemertea. In certain speci- mens of this genus an arrangement, more or less corresponding to what has just been described, was in a few cases met with, the basement membrane being of about the same thickness all round, finely striated, with imbedded nuclei, and sharply separated from the integument (PI. VII. figs. 3, 9, Bet). In other cases the membrane separating the muscles and the integument is much more folded, more irregularly striated, and less characterised as a separate band of distinct tissue (PI. VII. figs. 2, 5, Bet; PL X. fig. 6, B). Especially in these latter cases it is quite clear that this band of tissue, to which the name of basement membrane can only be applied with particular restrictions, imperceptibly passes into the connective tissue found between the longitudinal muscular bundles of the outer layer, where it is moulded REPORT ON THE NEMERTEA. 65 between these bundles, and also carries in the nephridial regions the terminal deferent portions of the nephridial ducts (PI. VII. fig. 5, Nep). The originally cellular nature of this connective tissue is retained in the head and in the anterior portion of the trunk, where the outer longitudinal layer of muscles is not less thick, but certainly contains a very much smaller number of fibres, because of the permanence of the intermuscular cellular stroma just alluded to. A. comparison of figs. 2 and 3, PL VII., the one taken through an anterior, the other through a more posterior region of the trunk of a Eupolia, will elucidate this, as will also the comparison of fig. 5, PI. VII., with the more enlarged fig. 6 of PL X. (taken from the same specimen), which represents a section through the region marked Bet in the former figure situated further forwards, and thus decidedly cellular as far as concerns the intermuscular tissue. One important fact is clearly indicated in this latter figure, viz., that the cells situated between the muscle bundles of the outer longi- tudinal layer (y.vl), although their general aspect, vacuolation and arrangement very much resemble that of the similarly vacuolated cells of the deepest layers of the integu- ment {Jdrf), may nevertheless be immediately distinguished from these by their much larger nucleus. The same fact follows quite as unmistakably, though somewhat less clearly, from PL VII. fig. 2. It gives some support to the hypothesis, that the whole of the deeper cell-rows of the integument, vacuolated or otherwise, being substantially different from the subjacent mesoblast cells, may be looked upon as epiblastic. However, this question, which pertains more to an ontogenetic than to an anatomical investigation, may safely be left out of further consideration. This basement tissue of Eupolia, much less regularly arranged than in Hoplonemertea and in the Carinellidse, is thus still directly homologous with that of the latter. A secondary external homogeneous basement layer is found immediately below the outer stratum of unicellular glands of the integument ; in the paragraph devoted to the integument the comparison with Carinina, Carinella and Carinoma has been already instituted, and it was at the same time shown in what way these different arrangements may be identified with one another. In the anterior portion of the body the stratified basement layer Bet (PL VII.) fuses with the sparse intercellular tissue that is present round the vacuolated cells, and appears to be a direct continuation of it. In the posterior portion, however, where the muscular bundles are more strongly developed, this stratified tissue appears more limited to the region between the muscles and the integument, principally because here the character of the intermuscular tissue is also changed and becomes more homogeneous, although it is here and there traversed by radial fibres, is also provided with nuclei, and contains numerous nerve-tracts. The general aspect, and the effect of the staining reagents, show this intermuscular tissue to be identical with the homogeneous, more or less gelatinous tissue, that is observed between the outer longitudinal bundles of Carinina (PL III. fig. 6). The cells, of which traces are found (ZOOL. CHALL. EXP. — PART LIV. 1887.) Hhll 9 66 THE VOYAGE OF H.M.S. CHALLENGER. around the nuclei contained in this homogeneous intercellular substance, show a very- delicate granulation, and it is ' often very difficult to decide whether a given one belongs to the nervous network, or whether it is a more indifferent cell, appertaining to the gelatinous ground-substance. In the former case the connection with the nerve- plexus is of great advantage in the decision, and but for this such a decision would often be wholly impossible. The fibres generally offer less difficulty, the delicate nervous fibres being sufficiently distinct from the elastic or contractile fibrils. . The intermuscular homogeneous tissue of Eupolia, and its inclusions in the region somewhat behind the head, are figured in PL VII. figs. 4, 5. In its deepest part, immediately surrounding the circular muscular layer, we find the nervous stratum, that will be more fully discussed further on. We must mention this, because in certain of the Schizonemertea (PL XII. fig. 10), to whose basement membrane and intermuscular tissue we have now to direct our attention, glandular structures belonging to the integument reach as far down as this layer ; a factor which we have to keep well in view when discussing the tissues to which this paragraph is devoted. This is all the more necessary, because in that case the other deeper cellular components of the integument are reduced in number, whereas the outer longitudinal muscular layer having become more compact and dense, the inter- vening region between these two, the region Bet of PL VII., has vanished from view. The secondary basement membrane (B) is then the sole representative of such a structure, and might easily, but as I hope I have demonstrated, injudiciously, be looked upon as homologous with the basement membrane of Carinina, Carinella, &c. (of. PL XL). An arrangement of the basement membrane, wholly comparable to what we have described in Eupolia, is found in such Schizonemertea as Cerebratulus corrugatus (PL XIII. fig. 6, B,b). In most of the others the strongly developed and massive outer longitudinal muscular coat so much encroaches upon the deeper layers of the integument in the way just noticed, that it is no longer possible clearly to distinguish between the two integumentary muscular strata {Bern) and the subjacent one constituting the body-wall (7). The extreme representatives of this development are figured on PL X. fig. 7, and PL XII. fig. 10. We now resume our examination of the gelatinous tissue, which we have as yet only examined as subintegumentary basement membrane {Carinina and Hoplonemertea), or also as intermuscular substance {Eupolia and Schizonemertea) in its further partici- pation in the muscular investment. In the circular and inner longitudinal layers gela- tinous intermuscular tissue is unmistakably present, and its presence is revealed both by the nuclei and by its peculiar homogeneous appearance, but at the same time, owing to the far greater compactness of these last named muscular layers, when compared with the outer longitudinal one of Eupolia, the position of the connective jelly is much more sub- ordinate, and its presence less easily demonstrable. Still it may be observed in the larger REPORT ON THE NEMERTEA. 67 species that, besides the distinctly granular cells included, in it, striation in nearly every direction forms a prominent feature of this intermuscular substance. Inside the innermost muscular coat the gelatinous tissue appears in the third modifica- tion which we have established for it. This modification is in many respects the most im- portant, and at the same time most liable to a considerable amount of variation according to the species we happen to examine. It has been already noticed that of all Nemertea Pelago- nemertes shows it in its fullest development, whereas the eminently muscular Carinina is only very sparsely provided with it. It will thus be safest first of all to examine it in Pelagonemertes, and to indicate by what features the other genera differ and gradually lead to grades of development as far down as Carinina. The most striking feature in examining any section of Pelagonemertes, either with low or high powers, with a special view to the gelatinous tissue inside the muscular body- wall, is the homogeneity of this tissue, the comparative scarceness of nuclei, and the uniform imbibition of the staining reagent, by which it has obtained a delicate rose colour. A comparison with the medusoid jelly, or, more distantly, with the intercellular substance of cartilage, is here inevitable. This general feature being established, the secondary characteristics are — (1) the imbedded nuclei of this jelly; (2) the fibres forming part of its substance, other fibres tra- versing it in apparently tubiform channels; (3) differences in the coloration of the jelly in different regions, and lastly, a fibrillation of the utmost tenuity, only visible by the aid of very high powers, which appears in different regions of the apparently homogeneous jelly, other and large portions, however, retaining the homogeneous aspect even with these high powers, though then not appearing wholly limpid but cloudy, perhaps an indi- cation of a yet finer fibrillation escaping the analysing power of our objectives when studied, not in the fresh state, but in Canada balsam. The different coloration of the jelly in different regions is partly arbitrary, i.e., darker- coloured patches are irregularly scattered throughout the general lighter hue. At certain places the darker staining is, however, constant, viz., contiguous to those regions where the jelly is interrupted. Thus the channels above alluded to, in which nerve-fibres and others take their course through the jelly, are marked by a double boundary line of darker colour (PI. VIII. fig. 6, n"), corresponding to the tract along which the continuity of the jelly is interrupted for the passage of these fibres. These nervous tracts being exceedingly numerous, the transversely or obliquely cut ends of similar distinctly red tubes are discovered in every section (PL VIII. fig. 3). Moreover, all round the two principal nerve-stems (figs. 6, 8), and bounding the cavity of the proboscidian sheath (PI. VIII. fig. 12, B), the blood spaces (fig. 8, bl), and the cavities in which the generative products are lodged (fig. 8), the same continuous dark red tint, which that portion of the jelly has acquired by the picrocarmine, is observed. A peculiarity which I have further to notice in the sections, is the different hue that 68 THE VOYAGE OF H.M.S. CHALLENGER. has been acquired by two superposed layers of the basement membrane of Pelagonemertes (PL VIII. fig. 13, B, B') ; the outer one being darker than the inner. I have no explanation to offer of this phenomenon, nor of the fact shown by the same figure (PL VIII. fig. 13) that the same phenomenon is repeated just below the basement membrane, the jelly between the bundles of longitudinal muscles being much darker than that which is found inside. The latter, into which the former gradually merges, is not specially indicated in fig. 13. Mention ought here to be made of what is marked dr in that figure, and what appears to be in several cases (PL VIII. fig. 5) a central cavity enclosed by peculiar cells, of which the nuclei are specially distinct. I must leave it undecided whether these structures, radially directed towards the surface of the body, are the same as the masses dr (PL VIII. fig. 4), which I see in a glycerine preparation of the integument made by Professor Moseley on board the Challenger from the fresh animals, and also whether these structures might be looked upon as glandular, and comparable to similar gland-masses in the jelly of Amphvporus moseleyi (PL XV. figs. 11, 12). That the course of the fibrds traversing the jelly is in no way strictly limited to certain directions, but that we find them now parallel to the proboscidian sheath, now to the intestinal wall, now convergingly directed against the generative ducts, is sufficiently demonstrated in PL VIII. figs. 3, 8. A very large number are, however, radially directed towards the surface, and though it was not observed in one section, it might perhaps be possible to find dorso- ventral fibres uniting both surfaces. That the nerve-stems, spring- ing from the lateral cords N (PL VIII. fig. 3), are indeed encased in a tubular space bounded by darker stained gelatinous substance, is best seen in fig. 6 of the same plate, as well as the fact that in the immediate neighbourhood cells and fibres (/) form part of that substance. In the other Hoplonemertea this continuous jelly, though much less prominent, has very much the same character as in Carinella, with the exception that no special tubi- form tracts for the passage of nerves, &c, are as distinct as they are in Pelagonemertes. The passage of radial fibrous tracts through the gelatinous tissue is, however, everywhere demonstrable (PL X. figs. 1, 2), as is also the origin of these fibres out of cells (PL X. fig. 2), and the continuity of the gelatinous stroma with that contained between the muscular bundles of the body-wall. The Challenger Nemertea not offering anything very special in this respect, I will postpone a more circumstantial discussion of this tissue for the monograph which I am preparing for the Naples series. Similarly I may rapidly call to mind that, in the Schizonemertea, where this gelatinous substance is best observed in the posterior region of the body, and better in large speci- mens than in small ones (PL XV. figs. 7, 10), it offers the same characters. Anteriorly, where the circumcesophageal blood-space is present, it does not play any conspicuous part ; pos- teriorly, however, it carries not only the tubular continuations of this space (the three longi- tudinal blood-vessels) but also the other internal organs, to all of which it is closely applied. REPORT ON THE NEMERTEA. 69 Here, again, its generally homogeneous character, with the addition of the same cellular and fibrous inclusions (intermediate stages between the two being similarly represented) as were noticed above, gives it an aspect similar to the corresponding tissue in the other groups of Nemertea. It cannot be denied, and has been already noticed above, that from a morphological point of view there are certain strong points of resemblance between the gelatinous tissue of the Nemertea and the jelly of the Medusas. I was very desirous to know whether this would also apply to the chemical constitution, and owing to Professor Krukenberg's kind aid I am now able to make definite statements on this head. Having sent him a small quantity of the Nemertean jelly, he kindly examined it and writes as follows : — " In accordance with your desire I have attempted, as far as it is possible, chemi- cally to solve the question, whether the jelly of the Nemertea is more closely allied to that of the Medusae or to the collagenous substances, such as I have with certainty demonstrated in Sipunculus nudus ( Vergl. physiol. Studien., i. 1882). As collagenous tissue is digested by trypsine, only in case of its having before been treated with acids or been boiled with water (Ewald and Kiihne), I tested the reaction of the Nemertean jelly towards this ferment. It proved to be very easily digested by trypsine, and would thus have to be regarded as a true proteid substance. Wholly in accordance with this result is the intense red colour which the tissue acquires when boiled with Millon's reagent ; whereas, on the contrary, tissues that contain more collagen are only faintly stained, and pure collagen is not stained at all by Millon's reagent. Moreover, the Nemertean jelly is not acted upon as are the collagens (' leimgebendes Gewebe') by boding water, nor does it furnish, when treated with diluted acids and soda, such reducing solutions as are the so-called mucins (Hyalogene, mihi). The Medusa-jelly consists of true proteid bodies.1 A similar substance is the vitreous body of the eye of vertebrates, and I am thus fully prepared, after the experimental results obtained, to agree with you in regarding the Nemertean jelly as an analogous product to the jelly of the Medusas, and similarly of an albuminoid nature." We now pass to a discussion of the muscular layers. We wdl first discuss the general distribution of muscular tissue, and reserve histological remarks for the conclusion. In discussing the muscular layers of the body-wall of the Nemertea, considerable confusion still obtains in the writings of most of the older authors, and I must confess that it took me a long time to see my way to a safe basis of comparison of the layers, that may sometimes be only one, sometimes no less than five in number, three or two being also very common occurrences. An outline of the homologies of the muscular layers was given by myself in the article Nemertea in the recent edition of the Encyclopaedia Britannica, and in the views there propounded I have no changes to make, only certain further additions have been rendered possible by the aid of the Challenger material. 1 Krukenberg, Vergl. physioL Studien., ii., 1 Abtli., pp. 23-34. 70 THE VOYAGE OF H.M.S. CHALLENGER. In all Nemertea, to whatever order or genus they may belong, there is one muscular layer that is always present. This layer may, indeed, be looked upon as identical through- out the whole series, and consists of longitudinal muscle-fibres. It is the longitudinal layer of Cephalothrix, in which genus definite or conspicuous circular layers sometimes appear to be absent (PL XL ) ; it is the only longitudinal layer of the Carinellidse — the longitudinal layer of muscle-fibres in the integument of Carinoma not being here taken into account, as indeed belonging to another system — and it is the so-called inner longi- tudinal layer of Polia, Valencinia, and of all Schizonemertea. In PL XI. this layer has uniformly been lettered a. I hold it to be the most primitive of all the Nemertean muscular layers, both on account of its constant presence and on account of the fact that in the posterior region of the body, where growth in length of the animal takes place, it often appears before the other layers that are present in addition to it in the anterior region of the body. The layer second in importance to it (morphologically speaking) is a layer of circular fibres marked ft, of very varying thickness, and which in the Carinellidse and the Hoplonemertea is immediately subjacent to the basement membrane, and external to the layer a. The very outermost fibres of this circular layer often take a differ- ent course, making an angle of about 45° with the longitudinal body axis, instead of being perpendicular to it. This, then, being the case in two directions, a decussation of this exterior portion of the layer, especially in Hoplonemertea and Carinellidse, is often noticed. Outside of the circular layer ft there is in the Schizonemertea and in Polia and Valencinia the outer longitudinal layer y, sometimes exceeding in thickness the two layers just noticed, and offering very varying conditions as to the compactness of its bundles. In most cases it remains entirely distinct from the two thin epiblastic muscular layers (see pp. 57 and 60) that make their first appearance in Carinoma, and are very generally present in Polia, Valencinia, and the Schizonemertea (PL VII. figs. 5, 9, ef; PL XIII. fig. 6, Ilcm). In some of the latter, however, a fusion occurs between the outer bundles of the longitudinal muscular layer y, and those that are decidedly of integumentary origin and significance, as was already noticed in discussing the integu- ment. It needs no explanation that these latter species offer more difficulties in rightly interpreting the relations between muscular system and integument than many others (PL X. fig. 7 ; PL XII. fig. 10). The difference in compactness just alluded to is often dependent upon the degree of development of the deep glandular layers of the integument. Sometimes these glands penetrate the whole depth of the muscular layer, reaching as far down as the nervous stratum (PL XII. figs. 2, 10); sometimes the muscles are kept further apart by the gelatinous ground substance, as was more fully discussed in a preceding paragraph. Eupolia may on the whole be cited as an example in which the degree of compactness of REPORT ON THE NEMERTEA. 71 this muscular layer remains at the lowest level (PI. VII. figs. 2, 5), in Cerebratulus corrugatus I found it very compact, and composed of very delicate fibres (PL XIII. fig. 6). Another additional muscular layer, which is not found in all but only in certain Nemertea, is marked S in the figures of PI. XI. It is an inner circular layer, and in the more primitive types [Carinina, Carinella, and Carinoma) it may even become exceedingly massive. It is directly applied against the longitudinal muscular layer a ; it forms at the same time the dorsal wall of the proboscidian sheath, the exceedingly thin ventral wall of which is in these species formed by fibres of the same inner circular layer that branch off, and are directed inwards between the space for the proboscis and the oesophagus or intes- tine, thus creating a floor to that proboscidian space. The inner circular layer is continued ventrally, and embraces the other internal organs as well. In Carinoma, where the layer has such a considerable thickness in the proboscidian and oesophageal region, and where it has disappeared in the posterior region of the body, leaving only the longi- tudinal and outer circular layers, the conclusion is of course tempting that the special development of this layer is in a certain functional connection both with proboscis and oesophagus. And if we then find that in the Schizonemertea this layer is absent, but that, on the other hand, there is a circular muscular coat to the proboscidian sheath and that this sheath has been raised to greater independence, and remains dorsally connected with the rest of the muscular body-wall in exactly the way it would be if it were the modified remnant of a restricted portion of the inner circular layer, we are led to the further hypothesis that these two may indeed be homologous. Thus all the transverse sections of the dorsal body- wall of Schizonemertea on PL XL, were they to be completed by adding the circular muscular layer of the proboscis-sheath immediately applied against them in the median line, would very strongly resemble the figures given of Carinoma and Carinina. I will not at present venture to decide whether any of the muscular layers of the oesophagus, noticed both in Eupolia and Cerebratulus (PL VI. fig. 9, oe.m ; PL XIII. fig 6, mto), may also be looked upon as derivatives of this inner circular layer, but will only add that in Hoplonemertea such a musculature is hardly developed ; whereas, on the contrary, the circular muscles of the proboscidian sheath have attained a very high importance, and are even more independent of the dorsal muscular body-wall than they are in Schizonemertea (PL IX. figs. 1-9 ; PL X. fig. 1). Here, too, I would be tempted to hazard a comparison between the absent inner circular layer and the musculature of the proboscidian sheath. The detailed histology of the Nemertean muscular system is hardly in its place here, and may perhaps be more fitly reserved for the monograph that will shortly appear in the Naples series. One point must, however, be mentioned, as its definite establishment seemed im- 72 THE VOYAGE OF H.M.S. CHALLENGER. portant to the Hertwigs,1 viz., the question whether the delicate longitudinal fibres composing the longitudinal muscular layers, and which in transverse sections are often arranged in rings (PI. III. fig. 6 ; PL XIII. fig. 6), have their matrix cells between them or not. Having very often succeeded in demonstrating an evident nucleus in the midst of this ring of cut fibres, and having constantly observed a difference in the effect of the staining reagent upon this central space and upon the connective tissue surrounding the muscle bundles, I must answer this question in the affirmative. Finally, two points deserve a short notice in this place. First, that the layer a of the longitudinal Nemertean muscles is very often separated into two, a right and a left half, by a dorso-median, sometimes also by an additional ventro-median solution of con- tinuity. It is very marked in the primitive Palseonemertea, especially in those cases where this layer is the preponderating component part of the body musculature (cf. PL XL) ; it is very rarely wholly absent in Ev/polia and the Schizonemertea ; it is less marked or even absent in the Hoplonemertea. In how far this separation may have general morphological significance, will be discussed in the chapter devoted to general considerations. The second point on which I shall offer speculative remarks in that chapter (cf. p. 127) has again reference to the same muscular layer. Sometimes it was observed that in this layer darker patches of contracted fibres alternate with lighter ones in which these contractions were absent (PL XV. figs. 9, 10). As this phenomenon of contraction was not wholly local but stretched all round the body in rings, the question must be considered whether we have here successive waves of contraction preserved at the moment of death, or whether the phenomenon has a deeper significance, is more permanent and indeed allows of direct comparison with myotomes. The comparative rarity of the phenomenon for the present prevents us from very emphatically advocating the latter view. How the circular layer of the Schizonemertea only stretches to the posterior brain-lobes, how the longitudinal muscles decussate in all directions in the head, how the develop- ment of the cephalic musculature is ontogenetically separated from that of the body musculature, are points already known to former investigators of the anatomy and embryology of the group. The fact of their having found ample discussion and mention in other monographs, and the Challenger material not having furnished new points of interest, will explain my silence in this Report on these and other points (such as the muscular dissepiments, the musculature of the cephalic slits, &c.) connected with the muscular system. 1 Die Coelomtheorie, p. 37. REPORT ON THE NEMERTEA. 73 NERVOUS SYSTEM. With respect to the nervous system, I cam indebted to the Challenger collection for very valuable additional data. In former publications ( IX, X ) I have dwelt at length on the peculiar arrangement of nervous tissue in the Nemertea as I had found it to exist in specimens that were obtained at Naples ; I am now enabled to give a more exhaustive description of this important system, and will commence by a short account of certain points in the latest investigations into the nervous apparatus of the lower forms of animal life, in order the better to explain the bearing upon questions of general morphology which the arrangements as we find them in the Nemertea may happen to have. The general and important conclusions arrived at by Kleinenberg in his classical Memoir on Hydra, conclusions which have since found their way into handbooks and textbooks as Kleinenberg's Theory of the neuro-muscular cells, have of late years been emendated by 0. and E. Hertwig. These investigators have propounded a general hypo- thesis on the phylogenetic development of the nervous system, which in their treatise Das Nervensystem und die Sinnesorgane der Medusen (Leipzig, 1878), is formulated (p. 170) as follows : — " We assume that in all Metazoa the ectoderm from which the (animal) nervous system, with its motor and sensory terminal apparatus, has originated, was primitively constituted of a simple layer of homogeneous cells in the same way as may be noticed everywhere in the earliest ontogenetic stages. We further assume that these cells, or at least part of them, have at an early period entered into mutual connection by protoplasmic processes, and have thus formed a more closely connected cell-stratum. According to our hypothesis, and on the principle of division of labour between the cells thus connected, there has been gradually developed a primitive nerve system out of this connected stratum. Whilst certain of these cells secreted contractile substance, others were provided on their surface with tactile hairs, and a third set acquired very numerous connections, the simple epithelium cells of the one-layered ectoderm thus becoming gradually and more or less simultaneously differentiated into epithelial muscle- cells, sense-cells, and ganglion-cells. Their protoplasmic connections, modified into specific nerve substance, have pari passu become converted into a plexus of nerve fibrils. When, later on, the ectoderm became constituted of more than one layer, the ganglion- cells were the first (of all the three elements just mentioned) to separate from the surface epithelium and to acquire a deeper situation." Balfour, in his Comparative Embryology (vol. ii. p. 333), accepts the leading features of this important hypothesis, partly substituting it for the earlier suggestion of Kleinenberg. The latter, in his latest publication,1 revindicates his original theory against the 1 Zeitschr. f. wiss. Z00L, Bd. xliv. p. 204. (ZOOL. CHALL. EXP. — PART LIV. — 1887.) IHlll 10 74 THE VOYAGE OF H.M.S. CHALLENGER. Hertwigs' objections, and maintains that these naturalists have furnished arguments in favour of his hypothesis rather than of their own interpretation (loc. cit., p. 205). Kleinen- berg holds that the naked nerve-cells of Hydra, that are in mutual and direct communica- tion, may transmit a stimulus by contact without the intervention of a delicate network of inter-cellular protoplasmic threads forming a network. He, moreover, holds that the epithelial cells had all of them the double significance of nerve-cells and muscle-cells, i.e., were true neuro-muscular cells before further division of labour set in, whereas the Hertwigs maintain that this division of labour took place between epithelial cells that were not yet physiologically so far differentiated. A nerve plexus, which covers a very large surface, was actually demonstrated by the brothers Hertwig not only in Actinia and other Ccelenterata but also in the Chsetognatha. Of the latter 0. Hertwig says :T — " By the fact of the nerve-fibres crossing and decussating in the most complex and diverse ways, there is formed a nerve plexus which spreads over the whole surface of the body, and in which the above described nerve-stems represent the single collecting tracts." A more or less similar plexiform arrangement of nerve-tissue has since been demon- strated in nearly all the lower groups of invertebrates, Annelids 2 and Arthropods excepted. Thus in the works of Loven, Greeff, Teuscher, Ludwig, and Carpenter the nervous system of the Echinoderms is described as offering many analogies with the type propounded as the most primitive by the Hertwigs. Nemertea, Turbellaria, Trematodes, and Cestodes can now be very fully compared, as far as their nervous system is concerned, with Hertwig's starting point, when we consider the results obtained by myself (IX, X) — which were afterwards confirmed (II) by Dewoletzky — for Nemertea; by Lang, Graff, and Pintner for Turbellaria, Trematodes, and Cestodes. Among aberrant forms one of the most striking examples of a thick epiblastic nerve- plexus with longitudinal collecting tracts is offered by Balanoglossus, as described by Spengel and more lately by Bateson. We shall have occasion again to refer to this interesting nervous system further on. For Mollusca, remnants of a more or less plexiform arrangement were found to exist in the Amphineura by myself3 (Proneomenia) and by Haller4 (Chiton), and also in other groups of Mollusca by Semper,5 Simroth,6 and others. 1 Die Chaetognathen, p. 34. 2 Lately Fraipont (Archives de Biologie, 1884, p. 274) has demonstrated the presence of an intermuscular nervous plexus in Polygordius, Protodrilus, and Saccocirrus, and thus opened the possibility of also bringing the Annelids within the region of comparison so far as this point of their organisation goes. Bergh describes a nerve-plexus in the larval Aulostoma (Arbeit. Zool. Zoot. Inst. Wurzburg, Bd. vii. p. 238). As to Arthropods there are facts which also point in the same direction, e.g., that Hoek mentions " a continuous network of ganglia and nerves " on the inner surface of the in- tegument in Pycnogonida (Zool. Chall. Exp., pt. x. p. 11G). 3 Niedcrland. Archivf. Zool., Suppl. Band, 1881. 4 Zool. Anzeiger, No. 76. 6 Archivf. Mikr. Anat, Bd. xix., p. 124, 1877; Arbeit. Zool. Zoot. Inst. Wurzburg, Bd. iii., 1877. 6 Zeitschr. f. iviss. Zool., Bd. xxxii. p. 304. REPORT ON THE NEMERTEA. 75 For Vertebrates a plexiform arrangement is known to exist in the embryonic stages of Amphibia, since the researches of Eemak and Strieker, and has lately been fully com- mented upon by Gcette and Baldwin Spencer. The last writes •} — " There may be said to exist in the Amphibian embryo a complete superficially-placed nervous sheath, out of which not only the central nervous system but all the sense organs of both head and trunk are formed, and which gradually disappears as these reach their full development." And further on : — " Aloug certain lines the cells of the nervous layer proliferate, and it is by this proliferation that the rudiments of the cranial nerves are laid down" (cf. p. 133). The significance of this plexiform arrangement of the embryonic Vertebrate nervous system will be discussed in the chapter devoted to General Considerations, and also the peri- pheral plexus of the adult Amphioxus, which lately has been more fully described by Eohon.2 Hence, since my former publications above cited, the necessity has grown more and more obvious of not looking upon the brain-lobes and the lateral nerve-stems of the Nemertea as the nervous system, but, though recognising their significance as more highly developed centres, to admit the presence of a most complicated and intricate network of nerve-tissue, originally — and in the more primitive species still — belonging to the integument. This network is most fully developed in the Schizonemertea. In Carinina its situation in the integument makes it more difficult to observe ; still I succeeded in demonstrating it both here (cf. p. 54) and in the other Carinellidae. In the Hoplonemertea the plexus has been replaced for the greater part by distinct nerves, of which the majority show a metameric arrangement. We will now pass to a more detailed description of this network, thereby purposely inverting the natural order by reserving the centres for the last. This apparent dis- crepancy disappears, however, when we look upon the network as the most ancient nervous arrangement, in which the centres have only gradually come forward. When once the eye has been trained by repeated observation to notice this particular nervous tissue of the Nemertea, it is comparatively easy to distinguish it from the surrounding tissue.- The peculiar punctate striation, the yellowish tint of the fibrous elements, the very pale carmine hue of the nuclear ones, immediately reveal the presence of nerve-tissue in sections, longitudinal or transverse, that have been made through specimens stained with picrocarmine. And when we take for our starting point, and as a basis for further description, one of the Schizonemertea of the Challenger, e.g., Cerehratulus corrugatus (PL XIV. figs. 3, 4 ; PI. XIII. fig. 6, pi), we observe in all transverse sections that the two lateral nerve-stems are in continuous connection with each other by nervous tissue that spreads out all round the circular muscular layer /3, both dorsally and ventrally. Immediately outside of 1 Some Notes on the Early Development of Eana temporaria, Quart. Jour Micr. Sci, Suppl., 1885. 2 Denkschr. d. k. Akad. d. Wiss. Wien (math.-nat. CI.), vol. xlv. 76 THE VOYAGE OF H.M.S. CHALLENGER. this nervous layer the longitudinal muscular coat y is situated. There is, moreover, present a third longitudinal nerve-stem, also situated, as are the two lateral ones, in this plexus, but medio-dorsally in the vertical plane that passes through the animal. It is this nerve which I have in a former publication (IX) proposed to call the proboscidian- sheath nerve, but of which I will, in the chapter devoted to General Considerations, offer a modified interpretation (p. 131) and which I will henceforth call the medullary nerve. The nerve-plexus uniting the three longitudinal nerve-stems, as a cylindrical coat of tissue between the longitudinal and the circular layers of muscles, cannot be separated or spread out flat, nor can we succeed in getting horizontal sections of it, just because of this cylindrical curvature. A portion of it may, however, be contained in the few consecutive sections passing in a horizontal plane through the medio-dorsal nerve or elsewhere, tangential to the cylindrical surface of the nerve-plexus. From such horizontal sections figs. 2, 3, and 4 of PI. XIII. and fig. 1 of PL XIV. have been taken, and where the plexus (n. pi) is touched right and left of the medio-dorsal nerve m, it has wholly the appearance of a dense network, the meshes of which are more especially due to the fact that radial bundles of contractde tissue — by which the muscular layers and the integument are held together, and which may even pass from the dorsal to the ventral body-wall of the animal— pierce the nerve-plexus. The longitudinal dorsal nerve stands out very boldly in the midst of the plexus. It is extremely important, and may be verified in any other surface section of the nerve- plexus, that from this dorsal nerve spring, both right and left, at more or less regular distances, thicker tracts of nerve-tissue (tr. n), also forming part of the plexus, but being straio-ht instead of tortuous, and having altogether the character of metamerically arranged nerve-stems that are not yet recognisable as independent structures, but that are fairly on the way to special differentiation as so many chief conducting tracts of nervous energy in the midst of the plexiform nervous tissue which binds them together. The presence of these transverse stems may also be noticed in transverse and longitu- dinal sections as a local thickening of the plexus, but as the whole stem is rarely attained in one transverse section, this thickening may be followed in consecutive sections, and is found stretching from the medio-clorsal down to the lateral nerve-stems.1 How far these transverse stems may be said to be metamerically arranged, everyone may judge for himself by consulting fig. 1 of PL XIV. The chief tracts are certainly symmetrical, i.e., spring from the longitudinal dorsal medullary nerve at opposite points, and about the same distance may also be seen to separate each successive pair from the foregoing. Other transverse bundles, some thinner, some thicker, some more obliquely placed, &c, but all similarly forming part and portion of the plexus, are, however, visible between the 1 Von Kennel, who Las so considerably advanced our knowledge of the Nemertea, appears to have observed, as early as 1879 (Die in Deutschland gefundenen Landplanarien, p. 39), the presence of certain of these transverse dorsal nerve-stems (commissures, v. Kenn.). He did not, however, notice or describe the nerve-plexus, nor the fact of the existence of a ventral connection, both by means of the plexus and of ventral metameric stems. REPORT ON THE NEMERTEA. 77 principal ones, and it is this fact that more or less obscures the metamery here alluded to (PI. XIII. figs. 2, 4). This metamery in the nervous plexus is of the same character as the metamery that is noticed in the intestinal arrangement, in the nephridia, in the generative organs, and in the blood vascular system of the Nemertea ; it may indeed be called incipient. To its significance, for the important question of the origin of segmen- tation, we shall have occasion to return by and by. Amongst the forms in which I found the metamery to be very distinct, Cerebratulus angusticeps stands foremost (PL XIV. fig. 1). And I must here call attention to the fact that the transverse stems here described are not only dorsally, but also ventrally, most regular and conspicuous, uniting the longitudinal nerve-stems below the intestine by a regular series of transverse commissures in the plexus, which is the primary connecting medium. It is important to note that there is no ventro-median longitudinal stem in Nemertea opposite the dorso-median one ; and not less important, that the same favourable species just named enables me to establish with certainty that the ventral transverse stems reach much further forwards than might originally be expected. The mouth alone inter- feres with their course; they are, however, found immediately before as well as behind it, and whilst in front of the mouth the lateral stems very soon merge into the lower brain- lobes, it is clearly seen that the transverse commissures are still recognisable, i.e., that the lower brain-lobes are united by thin ventral commissures, separated by a very short distance, till close up to the massive ventral commissure that has been hitherto regarded as the only ventral connection between the brain-lobes. The thin commissures just described are, however, not directly connected with the fibrous core of the brain-lobes, which is, on the contrary, directly continued into the massive inferior commissure, but they seem to derive their fibres from the outer cellular coating of these lobes. They pass underneath the two vagus stems, where these spring from the lower brain-lobes, and where these are in their turn, in front of the mouth, united by transverse commissures, as was noticed above (p. 38, 45 ; cf. PI. XIV. fig. 5). The histological description of the plexus may be very short, and has already been touched upon in the beginning of this section. Fibrous and cellular nerve-tissue are very regularly intermixed, the direction of the fibres follows that of the tracts in which they are found, and the fibres are, on the whole, closer together than they are often found in other Platyelminthes, where the designation of the nerve-stems — before they were recognised as such — as " spongiose strands" (spongiose Balkenstriinge) was current, and not inadequate. The nerve-fibres, however, are not so closely bound together, that the bundles are not very frequently found to be pierced by radial contractile fibres, as was noticed above, and is rendered evident by comparison of PI. XIII. figs. 3, 4, rf. That this intermixture is indeed a primitive character may safely be concluded, if we observe that Lang in his monograph on the Polyclada (XVIII) specially mentions similar features in the nervous arrangement of that group of Turbellaria, and also if we remem- 78 THE VOYAGE OF H.M.S. CHALLENGER. ber that in Carinella the passage of strong contractile fibres, even through the substance of the brain, was already known (IX). Of the cellular elements enclosed in the plexus the nuclei alone are conspicuous, and it is rare to find, either in the plexus or in the medullary nerve, distinct cell outlines (multipolar or other) round these nuclei, such as they are very often found in the brain. The nuclei characteristic of medullary nerve and plexus have exactly the same dimen- sions and shape as those that constitute by far the greater portion of the cellular coating, both of the brain-lobes (PI. XIII. fig. 1; PL XII. figs. 1-4) and of the lateral nerve-stems. The direct continuity between the nerve-fibres of the plexus, and those forming the axis of the lateral nerve-stems, can be demonstrated in all well-preserved sections, at any rate in those species where the plexus is well developed (PI. XII. fig. 2). Nor is the continuity with the fibres of the medullary nerve subject to any doubt (PI. XII. figs. 3, 4). This medullary nerve, a dorso-median thickening in the plexus, may be traced back- wards down to the hindmost extremity of the body, forwards up to the brain-lobes, and even in front of these. A section of that foremost extremity of the medullary nerve can hardly be distinguished from that of an ordinary cephalic nerve, but for its median situation, and greater size and distinctness. It is here independent, i.e., not enclosed in the plexus, which does not stretch further forwards than the brain-lobes, or than the layer of circular muscles.- The latter is known to cease in the region of the brain. The connection of the brain-lobes with the plexus, and with the medullary nerve, is much more intimate than I was hitherto inclined to believe. Certain specimens of Cerebratulus col- lected by the Challenger (PL XII. figs. 7, 8 ; PL XIII. fig. 1) permit me to form a definite judgment on this question. We there see that the anterior prolongation of the medullary nerve bends downwards in the region of the dorsal commissure of the brain-lobes, and enters into connection with a nervous stratum which may, in this region, be either con- sidered as a median portion of the brain, or as an anterior thickening of the plexus. Large ganglion-cells can be detected in it, also fibrous nerve-matter, both of them in the most intimate connection with the nerve-cells and nerve-fibres of the brain-lobes (PL XII. figs. 7, 8). From this anterior thickened region of the plexus, in which a transverse core of fibres — the dorsal commissure of the brain-lobes — takes its course (PL XIII. fig. 1), other fibres are seen to start in the direction of the body-axis and to arrange themselves into a longitu- dinal tract, which is also provided with nerve-cells, and which becomes the medullary nerve (PL XIII. fig. 2). My former statement (ix), that the medullary or proboscidian sheath nerve emerges from the dorsal commissure (loc. cit., pi. i. fig. 1), although exact, must thus be amplified in the way just described. I may add that a direct passage of fibres of the medullary nerve into those of the commissure, though sometimes noticed, is not always a constant phenomenon. Fig. 1 of PL XIII shows a state of things in which the fibres belonging to the plexus and medullary nerve appear to be more or less REPORT ON THE NEMERTEA. 79 independent from those of the brain commissure that is seen to pass under it, and to have a different texture and arrangement. Our observations on the nerve-plexus would not be complete if we did not allude to the very elaborate branches that pass out from it into the superposed muscular layers which they innervate. Some of them can even be traced as thick radial nerves piercing these muscles, and spreading out into the integument (PL XIII. fig. 6, n). Similarly the underlying muscular layers receive fine nerve-twigs out of the plexus, which are thus directed inwards as well, and first penetrate into the circular layer /3. For this reason they are best seen in longitudinal sections. The peripheral nerve system of the Schizo- nemertea has thus — as was already fully indicated in a former publication (x) — a totally different character from that of the Hoplonemertea. The profusion of radial nerve-stems springing from the plexus, every transverse section showing a great number of them, may convince us of the high degree of elaboration to which the nerve system of this group attains, and of which the great sensitiveness and quickly reacting movements of the worms themselves are the outwardly visible tokens. Nor may we omit to record the important fact, which was first observed in a Challenger specimen of Cerebratulus corrugatus, that in the region of the long slit-like mouth and oesophagus (behind the region where the very strong nerve, to which the name of vagus-nerve has been given (V, IX), leaves the inferior brain-lobes on its way to inner- vate the oesophagus) we can observe that from the plexus distinct nerves become detached, pierce the circular and inner longitudinal muscle layers (/3 and a), cross the circum- cesophageal blood-space and enter the tissue of the wall of the blood lacunae and of the oesophagus to assist in innervating these important organs. The morphological significance of this fact will be further insisted upon later on (cf. pp. 134, 142). The phenomenon is figured on PL XIV. figs. 3, 4. We have now traced the facts concerning the plexus and the medullary nerve. In a general way these descriptions may be said to be applicable to the plexus of Carinina, which, however, as was already noticed, is a less favourable object for study. It would seem as if in this species the nervous tissue, passing inwards amidst the muscles, again spreads out into a second plexiform arrangement between the muscular layers a and 8. This phenomenon, however, requires confirmation in more specimens than the two that have been available for the present investigation. One point alone requires a few words of further elucidation before we can pass from the nerve plexus to another paragraph, viz., the question as to whether the name of proboscidian sheath-nerve, formerly given by me to what I now propose to call the medullary nerve or the Nemertean medulla, must for the future be dropped altogether. It certainly must, if we wish to retain it for the longitudinal nerve originally so called ; but, curiously enough, I have now been able to make out the presence of another longi- tudinal nerve to which the name may very properly apply. 80 THE VOYAGE OF H.M.S. CHALLENGER. This nerve is furnished with fibres directly passing downwards out of the medullary nerve (PI. XII. fig. 9 ; PI. XV. fig. l); it is situated below this, and is entirely parallel to it. In one case of a very large specimen of Cerebratulus it appeared in its turn to be splitting up into two parallel nerve-stems. This proboscidian sheath-nerve more especially deserves its name because of its situation immediately above the muscular wall of the proboscidian sheath, into which it may be seen to give off fibres. It is not noticed in Carinina, Carinella, or the Hoplonemertea, but it is in Carinoma, Eupolia, and all the Schizonemertea. Its absence in the two first-named genera would appear unaccount- able if we did not remember that in both of them the proboscidian sheath is of hardly any importance, being extremely thin- walled (cf. PL II. figs. 4-7). And in this case it is all the more natural that in the oesophageal region of Carinoma it has become specially developed, being here even thicker than the medullary nerve, and about as thick as the lateral nerve-trunks of this species (PI. XL fig. 6). This is another example of sudden increase of a portion of the nervous system, and at the same time of the existence of a very marked degree of supremacy to which certain apparently subordinate parts of the organism may all at once attain. This unexpected change of size of the proboscidian sheath-nerve in one species is certainly a valuable fact for a hypothesis that will in a further chapter be enunciated (p. 133), according to which the possibility of a decrease in size of the lateral nerve-trunks is supposed to have been accompanied by an increase in significance of the medullary nerve. The fact that in this region of Carinoma the proboscidian sheath-nerve comes into the foreground so strongly that it might easily be mistaken for the medulla, may probably be ascribed to the massive development of the inner circular muscular layer S, which in Carinina, Carinella and Carinoma acts at the same time as part of the wall of the proboscidian sheath. The fact was already noticed as a peculiar feature of the species by M'Intosh (XXIV), when he first described Carinoma (under the name of Valencinia armandi). That a proboscidian sheath-nerve is wholly absent in the Hoplonemertea is still more easily accounted for. From the moment the brain and longitudinal trunks of the ances- tral Hoplonemertea were no longer lodged in the midst of the muscular tissue of the body-wall, but have come to be situated within the gelatinous tissue that fills up the space inside this muscular body-wall, not only has the plexus disappeared and been replaced by the remaining metameric nerves described above, but at the same time the innervation of the proboscidian sheath has altered. This innervation is now brought about by the peripheral and metameric nerves, which, in favourable cases (Pelagonemertes, &c), may be seen to send fine twigs into the muscular tissue of that wall. With this freer develop- ment of the peripheral nerve-system, the special arrangement by which the innervation of the proboscidian sheath is brought about, as long as the nerve-sheath is the source from which all peripheral nerve-fibres take their origin, has at the same time disappeared. REPOET ON THE NEMEETEA. 81 It is certainly all the more remarkable that in the Hoplonemertea we nevertheless find such very distinct traces of the medio-dorsal medullary nerve, notwithstanding the disappearance of the plexus. And more remarkable still that this remnant — not dis- tinctly traceable in only one specimen of Ampliiporus, whilst other specimens of the same species still have it, and whilst it is even very conspicuous in Drepanophorus and others — should occupy the same position as it does in the most primitive Palseoneniertea, i.e., in or even outside the basement membrane of the integument. This is another argument for directly deriving the Hoplonemertea from the Palseonemertea. Cephalothrix may be said to fill up part of the distance which separates Carinina and Carinella from the Hoplonemertea as far as the situation of the nerve-system is concerned (cf. PL XL fig. 5), whereas Eupolia may be said to do the same with respect to the ciliated grooves on the head, and partly also to the posterior brain-lobe, its glandular investment, and the long duct leading from it to the exterior. We must now pass on to the description of the brain-lobes and the lateral nerve-stems. It is known that these offer the lowest degree of specialisation in the more primitive genera of Palaeonernertea, e.g., Carinella. For this genus the brain- and nerve-stems have been sufficiently described before (IX), and, in comparing this with what we find in A.T. PL Fig. 5.— Side view of the brain of Carinina in outline, reconstructed from the sections. The fibrous core is indicated by a dotted line. A. I., anterior lobe ; P. I., posterior lobe ; Co., ventral commissure ; Ln, Lateral nerve-stem. Carinina, the latter genus must be recognised as representing in this respect a somewhat higher scale in the developmental series. This higher development finds its expression in the presence of a posterior brain-lobe, comparable to the same lobe of the Schizonemertea which was often designated as the side organ, although it is formed of nerve-substance directly merging into that of the brain. Carinella inexpectata has been formerly shown (VIII ) to possess a ciliated passage leading into the brain-substance, without any special differentiation of that portion of the brain into which this ciliated channel pene- trates. In Carinina such a differentiation has set in, and the brain-substance, into which a ciliated canal leads, has become a separate lobe.1 In consequence of this we are, moreover, enabled to draw a general — though by no means a sharply defined — distinction between the portion of the brain-mass with which this accessory lobe is in contact, and 1 Chapuis has lately noticed posterior brain-lobes in a Ceplmlothrix (Arch. d. Zool. Exp., vol. iv. p. xxi., 1886). His description is, however, very incomplete. (ZOOL. CHALL. EXP. — PAET LIV. — 1887.) Hhll 11 82 THE VOYAGE OF H.M.S. CHALLENGER. that which is continued into the lateral nerve-stems, i.e., an incipient distinction between a pair of upper and a pair of lower lobes, respectively limited by the dorsal and ventral brain commissure ; the whole forming a ring round the proboscis and its sheath. A side view of the brain of Carinina — reconstructed from a series of sections — is given in the accompanying woodcut, in which the very thin dorsal commissure is not indicated. A comparison with figs. 4 and 7 of PL V. will at once show the relation of this stage of differentiation to that to which Eupolia has attained. In the Schizonemertea the separation between upper and posterior lobes is more marked still than in Eupolia (PI. XIV. fig. 6) ; in the Hoplonemertea they are definitely separated, and only con- nected by one or more nerve-strands (PI. IX. fig. 10), their situation being then some- times behind, and even sometimes before the rest of the brain. The ciliated canal penetrating into the posterior lobe of Carinina is simple (not divided in two as in certain Hoplonemertea), and provided with a high, ciliated epithe- lium of its own. It is figured in figs. 1 to 3 of PL VI. Certain glandular cells, gl. br, are seen in this same figure to have become specially developed in connection with this posterior lobe and its ciliated canal. Similar glandular cells also form a characteristic feature of the posterior brain-lobes of Schizonemertea and Hoplonemertea. In Carinina it is evident that these glandular cells are derivatives, or at least morphological equiva- lents, of the deeper glands Gi, of the integument. The nerve- cells themselves, out of which the brain is built up, still undoubtedly belong to the integument, and it is exceedingly difficult, if not impossible, to draw a sharp distinction between the outermost brain-cells and the surrounding integument-cells. In the figures just cited this difference has been artificially very much accentuated in order to bring out more distinctly the outline of the brain ; for the same reason, the integu- ment in this figure was on purpose not fully worked up. The inner core of the brain is fibrous, so is the core of the longitudinal stems, where, however, the attempt at a distinction between nerve-cells and cells of the surrounding integument is equally hazardous (PL III. fig. 8). This fibrous core is in direct continuity with the nerve-plexus, that spreads out in the deeper integumentary layers. A vagus nerve passing from the lower portion of the brain on both sides towards the oesophagus is also distinct in Carinina (PL VI. fig. 1, Nv). The passage of contractile fibres through the brain-substance is unmistakable, though less evident than in Carinella. Passing on to the description of the nerve-centres of Eupolia, we immediately recognise the difference resulting from the fact that here the brain is imbedded inside the muscular layers, as is also the case in all Schizonemertea. Our description of the brain of Eupolia maybe based upon the figures of PL V., which were obtained not cle visu, but by reconstruction from a series of sections.1 Figs. 1-4 1 For the making of these and many other series of sections, and for assistance in the reconstruction above alluded to, I am indebted to the kindness of Dr. Oudemans, my former assistant, now director of the Zoological Garden at the Hague. REPORT ON THE NEMERTEA. 83 and 8 represent the whole of the brain tissue, figs. 5-7 and 9 the fibrous core, as it is enclosed by the nerve-cells, the limit of this cellular investment being given in outline in the latter figures. It must, however, from the first be remarked, that this outline should be completed by the plexus, and by the median medullary nerve. They are not indicated in these figures, although both of them are found along the whole length of the lateral nerve-stems, and reach forward as far as the region of the dorsal brain commissure. It is seen at a glance that the fibrous core repeats the external folds and prominences of the brain-masses, that the lateral nerve-stem is continued into the lower lobe, and that the upper lobe is distinguished by a prominent fold of its surface, a gyrus (fig. 1, SL), into which a separate knob of the fibrous core is seen to pass, and by two other fibrous projections — the one stretching towards the blunt end of the posterior lobe, the other running forwards and accompanying the ciliated canal, which is also marked in outline in fig. 6, and (in red) in fig. 5. The canal cc, in figs. 2, 3, 4, 8, is the exterior portion of this duct. The different thickness of dorsal and ventral brain commissure may be gathered from figs. 1, 3, 8, 9 ; from the latter two, the fact that the nerve-fibres are very strongly preponderant in these commissures over the cells. Close behind the ventral commissure the nerve for the oesophageal wall, vg, the so-called vagus uerve, is seen to leave the common fibrous core of the brain, whereas the nerves for the proboscis (pn) spring from the inner surface of the ring, where the fibrous core turns up from the ventral to the dorsal commissure (figs. 5, 9). The vagus nerve is soon after its origin connected by transverse fibres with its opposite neighbour ; this vagus commissure is sometimes repeated ; it will be again referred to in the general considerations on the nervous system. The cephalic nerves that leave the brain and innervate the head are only very imperfectly rendered in these figures ; their number is far greater than might be concluded from figs. 5, 6, an. The aspect of several portions of the brain of Ewpolia, in transverse section, is repre- sented in PI. VI. It will there be noticed that fig. 4 represents an anterior section through the inferior brain commissure and the point of innervation of the proboscis, fig. 5 one just behind this, cutting the dorsal commissure and the vagus root at the same time. The exact situation of these sections will be best understood by comparing them with PI. V. fig. 9, where the respective positions of the commissural ring, the proboscidian nerve, and the vagus are clearly indicated. Fig. 7 is a transverse section lying further backwards, almost in the level where the dotted line, SL, in PL V. fig. 4 terminates, whereas the section fig. 8 lies again somewhat behind this, at a point where the "gyrus" of the superior brain-lobe actually divides the central fibrous nerve-substance into an upper and a lower portion. These sections, at the same time, show the difference in size between the brain-cells and the glandular elements partaking in the constitution of the brain, along the superficial part of what I have called the posterior brain-lobe (side-organ, auct.). 84 THE VOYAGE OF H.M.S. CHALLENGER. The brain of the Schizonemertea was fully discussed in a former publication (ix). The Challenger Schizonemertea all conform to this type, with the additional facts alluded to above in connection with the medullary nerve. The difference in the size of the eanerlion cells in different regions of the brain, as it appears in PI. XII. figs. 7, 8, and PI. XIII. fig. 1, is much more marked in certain species of Cerebratulus than in others. The larger sized nerve-cells appear to be princij>ally peripherally and anteriorly situated ; that they are absent, or less numerous, in certain other species, may be seen by comparing PL XII. fig. 1, with the above mentioned figures. The relative distribution of fibrous and cellular nerve-matter in the brain need not be any further described in detail after our foregoing description and figures of Eupolia. The size and shape of the posterior lobe is, however, somewhat different in the Schizonemertea. This will be obvious by com- paring pi. i. fig. 1 of the treatise referred to (ix) with our present figures of Eupolia. Not having been able to study any of the Challenger species alive, we should have to be content with reconstruction from section series, if I were to enter more fully into the discussion of the respective differences, and for that reason I wish to restrict myself to these general remarks. One other point connected with the posterior lobe and its ciliated duct deserves special mention, viz., the observation I was able to make that the duct which leads from the bottom of the cephalic sbts into the nerve-tissue of the posterior brain-lobe (inside the brain-lobe it very generally has an S-shaped, and, at the same time, a spiral twist, thus being very often as in PI. XIV. fig. 6, cut in three places, all in one section), and which is clothed in the neighbourhood of its external opening with an epithelium directly con- tinous with, and similar to, that of the outer surface, not only shows certain differences in its epithelium, as we pass further inwards (PL XIV. fig. 11), but also offers certain complications, which we have now to consider. These complications very distinctly con- cern the participation of deeper cellular layers of the integument. As indicated by gl in PL XIV. fig. 11, these deeper layers segregate and form a ring-shaped or cushion-shaped addendum to the simple epithelial tube. It must be doubted whether they communi- cate with the exterior, as do the deeper glands of the integument, although this deserves special attention, because of the glandular significance which must be attached, according to Dewoltezky (il), to the strongly refractive cells present on the posterior surface of the hinder brain-lobe (cf. p. 94). The epithelium has undergone still more con- siderable alteration when it passes inside the posterior brain-lobe. Its nuclei are distinct (PL XIV. figs. 6, 7, 8), but instead of direct cell partitions we may observe a fine striation vertical to the axis of the ciliated canal (fig. 8). This feature, known to former observers (IX, figs. 35, 36), may here be more especially alluded to, because in Hoplonemertea (PL XIV. fig. 10) we find that the discharge of glandular products from the deeper gland- cells takes place between the interstices of this striated region. This discharge into the lumen of the canal is a point that is put beyond doubt by numerous Challenger sections. REPORT ON THE NEMERTEA. 85 The fact that the canal is single in the Schizonemertea, whereas it is double in the Hoplonemertea, was known before (ix). It was also found to be confirmed in all the Challenger species ; the bifurcation of the canal taking place in such a manner, that the one branch passes through the distinct nerve-cells, forming the greater mass of the lobe, whereas the other one immediately penetrates — more peripherally — amongst the much larger glandular cells overcapping the foregoing. Carinina corresponds with the Schizo- nemertea in having the canal single. Having considered the central fibrous substance of the brain in the Palseonemertea and Schizonemertea, we have only to add that the Challenger Hoplonemertea have also con- firmed the fact that here this fibrous core is less complicated, the brain-lobes being at the same time more compact, the cephalic nerves very numerous. In Cerebratulus angus- ticeps (PI. XIV. fig. 6) the fibrous core is very massive and conspicuous also. As to the innervation of the numerous eyes, I have no new observations to record {cf. V and IX), nor as to that of the proboscis, with the exception of the fact that in Drepanophorus and Amphiporus I could distinguish numerous nerves springing from the brain-ring and corresponding to the numerous longitudinal trunks in that organ. This point, which was left in doubt by v. Kennel (XVI), is thus definitely settled. The phenomenon was parti- cularly distinct in one specimen of Amphiporus moseleyi that had retained its proboscis. It has only been partly figured in PI. IX. fig. 10, where only two are indicated, so as not to obscure the diagrams. As to the innervation of the oesophagus, little need be said as far as the Schizo- nemertea are concerned, the well-known strong and double vagus nerve being constantly met with. Distinct nerve-branches are seen to take their course in the walls of the oesojDhagus (PI. XIV. figs. 3, 4) ; it was already noticed above (p. 79) that these may be partly traced to separate branches springing independently from the nerve-plexus, whereas for the other part they are ramifications of the so-called vagus. Nerves to the intestinal canal, very easily detected in the oesophageal region, could not be traced with the same accuracy and distinctness in the post-cesophageal region of the intestine, most probably owing to the extreme tenuity which these fine and delicate nerve- twigs may here have obtained. It cannot be determined at present whether this portion of the intestine also receives branches from the oesophageal vagus system or only directly from the plexus, now that the existence of such a double method of innervation {Cerebra- tulus corrugatus) has been actually demonstrated for the anterior regions of the intestine. On a priori grounds, I look upon the latter arrangement as by far the most probable.1 The course of the vagus is somewhat modified in Drepanophorus, and perhaps in Amphiporus. I find the strongest nerve-stem, connecting the brain with the oesophagus, in Drepanophorus, running forwards instead of backwards (PI. IX. fig. 10). Other smaller 1 It should here be noticed that Klemenberg (loc. cit., p. 114) has also failed to detect visceral nerve-branches to the endodermal intestinal epithelium of the Annelid, Lopadorhynchus. 86 THE VOYAGE OF H.M.S. CHALLENGER. stems leave the brain in corresponding regions of the lower brain-lobe, i.e., along the surface turned towards the proboscidian sheath, and run in the direction of the oesophageal epithelium. This secondary innervation, though different in morphological aspects, is more or less homologous with the facts above disclosed in the case of Cerebratulus corru- gatus. That the vagus proper — the massive and thick stem — is here turned forwards may be a consequence of the change in the situation of the mouth, which in the Hojdo- nemertea is no longer behind the brain, but in front of it. This gradual change of position may very possibly have drawn the vagus-stem with it. In concluding our remarks upon the brain I have only to add that the well-known difference between Schizonemertea and Hoplonemertea with respect to the connection between anterior and posterior brain-lobes (side organs) also obtains in the Challenger specimens. The latter are connected in Drepanophorus and Amphiporus with the brain by one or more fibrous commissures. Another difference several times observed between the fibrous brain-tissue of these two Hoplonemertean genera on the one hand, and Cerebratulus, Eupolia, &c, on the other, is a marked increase in compactness of the fibres, so much so that the fibrous character of the central portions of the brain has often more the aspect of Ley dig's " Punktsubstanz," and even shows a still more delicate and more compact texture by the appearance in this " Punktsubstanz " of regular patches with very faint outlines, which apparently are still more compact regions of this tissue. . The longitudinal nerve-stems, which are the posterior continuations of the lower brain-lobes, hardly need any special mention. It must only be insisted upon that in them, as well as in the brain-lobes, there is no absolute distinction between the cellular envelope and the fibrous core, but that inside this core nuclei are invariably scattered, which bear testimony to the absence of any such definite boundary. Still, there is generally a homogeneous and very thin layer between the cellular coating and the fibrillar core, a kind of membranous neuroglia, through which the fine processes of nerve-cells may be seen to take their course in groups, which then become lost amongst the fibres of the core. Then, again, certain favourable sections (PI. XII. fig. 2) very distinctly show the course of nerve-fibres inside this fibrous core that are not longitudinal, and thus puncti- form in transverse sections, but that are interwoven at right angles with the latter and continue their way into the nerve-plexus. The transparent sheath of the fibrous core of the nerve-stem is more distinct in Cerebratulus than in either Pakeonernertea or Hoplonemertea. It is rarely encountered in the brain, where fibrous and nervous elements are more intimately interwoven (cf. PI. VI. figs. 4-8; PI. XII. figs. 7, 8; PI. XIV. figs. 7, 8). Outside of the stems there is another accumulation of homogeneous connective tissue arranged as a protecting envelope round the nerve-stems. This is much more conspicuous in the brain-lobes, and more so in the Hoplonemertea (PI. XII. fig. 5) than in the Schizonemertea (PI. XII. fig. 2 ; PI. XIV. fig. 2). Still in the latter it is far more REPORT ON THE NEMERTEA. 87 conspicuous than in Carinina, where we have already pointed out the apparent absence of any sharp or distinct boundary line between the cellular brain-tissue and the sur- rounding cellular tissue of the integument. We cannot pass on to the description of the peripheral nerves without first referring to the terminal portion of the lateral nerve-stems, known to terminate at the posterior end of the body, right and left of the anus in the Schizonemertea. but also known to meet in a connecting commissure above the anus in several Hoplonemertea (IX). This commissure was found by me in several Challenger species, but at the same time I was able to verify the unexpected fact that in Eupolia the fibrous cores of the longitudinal nerve-stems are also posteriorly united by a commissure. What most especially deserves attention in this posterior commissure of Eupolia is, that it is found beloiv the anus, the longitudinal stems and the commissure, together with the brain, thus forming an immensely elongated ring round the intestine, whereas in most of the Hoplonemertea alluded to, all the portions of the nerve-system may be said to remain above the intes- tine. This is, indeed, very emphatically the case in Amphiporus moseleyi, where we find (PI. IX. fig. 4) not only the brain and the anal commissure above the intestine, but also the longitudinal stems, that take their course above the intestinal caeca. Nevertheless, in Drepanophorus the anal commissure is above the intestine, although here the longitudinal stems are diametrically opposite in position, i.e., below the intestinal cseca. They were for this reason considered (IX) to furnish a transition stage to the ventral cord and circumcesophageal ring of Annelids and Arthropods, a consideration which derives very strong support from the existence of transverse commissures that will hereafter be described. At all events, these very curious differences — the anal commissure of Eupolia is figured on PL VII. fig. 8 — furnish another proof of the extraordinary plasticity which we meet with in the group of the Nemertea, with respect to the morphology of the most important components of the system ; a plasticity and diversity which are at the same time indicative of the primitive and low scale on which the Nemertea may be said to find themselves. Coming now to the peripheral ' nerve-system, I may note that I have already, some years ago (X), stated that it is difficult to apply this name in its generally accepted significance to the arrangement which we find in Schizonemertea and in Palseonemertea. It is, however, applicable to that of the Hoplonemertea. Here only we find distinct metamerical peripheral nerves leaving the longitudinal nerve-stems at regular intervals, and innervating the body musculature, the integument, the internal organs, &c. In the Schizonemertea and Palseoneniertea the cephalic nerves, starting from the brain, are directly comparable with those of the Hoplonemertea, but the rest of the peripheral system is here represented by the plexus and its innumerable branches and twigs, which are directed upwards and downwards, serve for the same purpose, and render the peripheral arrangement in this group so primitive and so important. Still, in very large specimens 88 THE VOYAGE OF H.M.S. CHALLENGER. of Schizonemertea (PL XIV. fig. 2) I could observe that from the thickened part of the plexus, which forms the longitudinal stem, fine nerve-branches also take their origin, and pass directly to the periphery. The essential difference between Schizonemertea and Hoplonemertea in this respect nevertheless remains the same as above formulated. We may now turn to the Hoplonemertea. The different species contained in the Challenger collection confirm the well-known facts about the metamerically placed pairs of peripheral nerves of the Hoplonemertea, some of which are turned dorsally, others ventrally, and which, dividing dichotomously, finally spread out in very numerous bundles of nerve-fibrils, serving for the innervation of the environing regions. In addi- tion they also furnished me with certain important new points. To begin with the latter, I will first draw attention to the two longitudinal nerve-stems of Drepanophorus lankesteri, which are situated, as is characteristic of this genus, below the intestinal caeca (PI. IX. figs. 1, 5, 6), about midway between the lateral margin of the body and the median ventral line. These stems in transverse section very much resemble those of other Hoplonemertea. One of them is figured on PI. XII. fig. 5 ; from this it may be seen that the nerve-cellular coating is generally not distributed as a sheath all round the fibrous core, but as a double band applied upon this core at two diametrically opposite points. The participation of this cellular coating in providing the outgoing peripheral nerves with delicate nerve-fibres is distinctly seen in this section, as is also the direct continuity of other portions of the peripherally directed nerve-fibres with those of the core. When in Drepanophorus lankesteri I followed some of these peripheral nerves in their further course, by examining the consecutive sections in which they are continued, I was struck by the very remarkable fact, never noticed before, that some of them did not dichotomise — or at least very rarely — and did not taper towards the periphery, but passed directly under the intestine from the one longitudinal nerve-stem into the other, a distance in this specimen of 1^ mm. This was an unmistakable com- missure, which could in no way be compared to the well-known commissure above alluded to, which connects the two longitudinal stems above the anus. And not only was one such ventral commissure present, but on closer inspection I found a great number of them, and by registering the respective distances at which they were present, the one behind the other (about ^ mm.), I was forced to the conclusion that we here have before us a system of very regular metamerically placed commis- sures between the longitudinal stems, and forming a nerve-ladder (PL IX. fig. 10), which is very directly comparable to that of Sabella and other species among Annelids, and to that of Proneomenia and Chiton anions; Molluscs. In a few of these commissures I detected dichotomy and fusion of one of the branches thus formed with the fore- going or with the following commissure, a peculiarity also known to exist in Chiton, but evidently of rarer occurrence in Drepanophorus lankesteri. Moreover, I may also men- REPORT ON THE NEMERTEA. 89 tion that in certain commissures it was clear that fine nerve-twigs spring from them and serve to innervate the surrounding tissues, their significance thus not being solely com- missural. The different facts just recorded are represented semi-diagrammatically on PI. IX. fig. 10, which was reconstructed from the very large number of sections which I have of this species. It is, moreover, seen in this reconstruction how other peripheral nerves spring from the longitudinal stems as well, some being directed upwards or downwards, some towards the side or inwards. These peripheral stems are metamerical, as are the commissures, a metamery which, though not absolute, and sometimes broken by certain irregularities, is still more advanced towards perfect regularity than is the incipient metamery which we observe in the nerve-tracts that are noticed inside the plexus of the Schizonemertea, and that were more fully described above. The transverse commissures between the lateral stems may be noticed to go up quite close to the brain- lobes, as indicated in the diagram on PI. IX. fig. 10. Both in this respect and in the fact of their existence, they call to mind the ventral commissural tracts in the plexus of Schizonemertea. I have no doubt that the two systems are homologous, the commissures having subsisted in Drepanophorus lankesteri although the plexus has disappeared. Finally, it must be mentioned that as yet I have looked for them in vain in other species of Drepanophorus, or in other Hoplonemertea. My other specimens of Drepanophorus are, however, less well preserved than is the one specimen of Drepanophorus lankesteri. Another peculiar feature of the peripheral nerve-system of the Challenger Hoplo- nemertea, which has also remained hitherto unnoticed, is most favourably observed in Amp>hiporus marioni, although I afterwards noticed it in other Hoplonemertea. It is figured on PI. X. fig. 1, ne, and consists in the fact of a peripheral stem, which has taken its outward course away from the longitudinal nerve-trunk, and which has penetrated amongst the pennate fasciculi of longitudinal muscle-fibres of the layer a, spreading out in a plane parallel to that of the body-surface, and thus forming a kind of local plexus between the muscular layers a and /3. It must for the present remain an open question whether this arrangement, which can be noticed in different regions of the same section, and which in no section was absent, must be regarded as a primitive feature connected with the plexiform arrangement which must have obtained in the ancestral forms of the Hoplonemertea, or whether it is merely a special adaptation, having arisen in certain Hoplonemertea, and being in some way subservient to the innervation of the muscular investment or the integument. At all events it is a peculiar arrangement, and, as such, deserves special mention. How intricately and yet how regularly the peripheral nerve-system of the Hoplo- nemertea may be said to be distributed can also be gathered from Moseley's figure of Pelagonemertes, which we have copied on PL I. fig. 23, where the peripheral nerves are seen to spring, two at a time, from the lateral trunks, which here, too, are united posteriorly by a commissure above the anus. (zool. CHALL. EXP. — part liv. — 1887.) Hhh 12 90 THE VOYAGE OF H.M.S. CHALLENGER. One peculiarity, finally, deserving special mention is the presence in the majority of well-preserved specimens of Hoplonemertea of a medio-dorsal longitudinal nerve homo- logous to the medullary nerve described above (cf. p. 81, and PL XL fig. 8). Its connection with the rest of the nervous system could not be satisfactorily made out, although traces of a connection with the dorsal brain commissure were not wanting in many specimens. Its presence is, however, significant, and its retention in the Hoplonemertea, where the arrangement of the nerve-system has so considerably deviated from the primitive Palseo- nemertea and Schizonemertea, must prevent us from underrating its morphological significance. This will be more fully entered into in the chapter of General Considera- tions at the end of this Report. SENSE-ORGANS, ACCESSORY GLANDULAR STRUCTURES, AND ORGANS OF UNKNOWN SIGNIFICANCE. The most conspicuous sense-organs of the Nemertea are without doubt the eyes. Although eyes are absent in very many genera and species, and although in some species pigment spots at the tip of the snout are regarded as such, other genera have very well- marked and numerous eyes, with a hyaline hemispherical refractive body, a layer of visual rods, and an optic nerve connecting the eye with the brain-lobes. These more highly- developed eyes were previously known to occur in the Hoplonemertea, and the Challenger material has confirmed these general conclusions. The most primitive of the Palaeo- nemertea, Carinina, is not provided with eyes. Nor do I find traces of eyes in those species of Eupolia which were contained in the collection, and of which the head was studied in sections. Eupolia giardii is among these. However, it is known from other researches (VTl) that different species of Eupolia have often very numerous eyes, increasing in number with the growth and the age of the animal, and, moreover, that these eyes resemble those of the Hoplonemertea in many respects. Nor were the Schizo- nemertea of the Challenger provided with eyes that revealed their presence in the microscopic sections, although I would not venture to affirm the total absence of eye- like structures or pigment spots. In this respect the fresh animal often shows at a glance what is very difficult to demonstrate in the series of sections, e.g., the number and disposition of the eyes or pigment spots. As, moreover, these data can hardly be of any taxonomic value for the determination of the Schizonemertea, I think I may pass on to the discussion of the eyes of the Challenger Hoplonemertea. They have the characteristic histological arrangement already described and figured by myself in a former publication on the subject (IX, fig. 42). A posterior layer of rod- like elements, which is in direct connection with the optic nerve, is enclosed by pigment, REPORT ON THE NEMERTEA. 91 and the whole completed into a more or less oval cup-form of which the anterior part shows very distinct cellular elements with nuclei. Distinct lenticular structures, which were formerly noticed in Mediterranean Nemertea, could not be certainly made out in the Challenger specimens. What varies most is the pigment coating the posterior hemispherical surface. Sometimes this pigment is intensely black, and so extremely fine that it looks almost homogeneous (Amphiporus moseleyi), whereas in other cases [Amphi- porus marioni) the pigment granules are uncommonly coarse and large sized, the colour being in this case a brownish-green rather than black. In discussing the further sense-organs of the Nemertea, a great significance must certainly be ascribed to the sensory elements distributed in the skin, and primarily serving for tactile functions. With respect to this organ of sense, the spirit specimens at my disposal have, however, revealed no new facts of importance. I have only convinced myself — as has been already noticed both when describing the cellular integument and its innervation — of the presence of distinct sense-cells in all parts of the integument. They have the well-known form of the sense-cells described by the Hertwigs, by Lang, and by others in the Ccelenterata, in the Platyelminthes, &c, and it is to them that the extreme delicacy of the tactile sense, which is revealed in living specimens of Nemertea, must be ascribed. Bateson's fig. 77 1 of Balanoglossus comes very close to what was observed in the Nemertea in connection with these sense-cells. Their direct innervation by nerve-fibres, starting radially from the plexus in the two more primitive groups, has been noticed above (cfl PL XIII. fig. 6; PL XIV. fig. 2). Sense-cells with distinctly longer and stiffer hairs, such as I have been able to observe in living specimens from the Bay of Naples, have not come under my notice in the Challenger sections. The similarity with Balanoglossus, just alluded to, is increased if we consider Bateson's figs. 70, 75, and 79, and compare them with the integument of Carinina. The similarity is significant. In sections of Balanoglossus made by myself I was very much struck by this resemblance, reaching from the cilia down to the nerve-plexus and subjacent muscles. Another question I wish to allude to here, and which has been pressed upon me by certain of the Challenger specimens, is whether the terminal transverse furrow which is encountered at the tip of the head in Carinina (PL I. fig. 1-3 ; PL II. fig. 1), which is also distinctly seen in certain Amphipori (PL IX. fig. 9), has not also primarily a tactile significance. And, in addition to that, I wish to ask whether we might not look upon this terminal groove, which lies more or less in a horizontal plane passing through the animal, as having preceded the paired longitudinal cephalic furrows which form the distinctive feature of all the Schizonemertea. When considering the probability of this suggestion, the following points should not be lost sight of — (l) that in certain Cerebratuli these cephalic furrows do meet at the tip of the head (PL I. figs. 13, 14, 18, 19) ; (2) that the furrows in this case are comparatively short (figs. 13, 14) ; 1 Quart. Journ. Micr. Sci., June 1886. 92 THE VOYAGE OF H.M.S. CHALLENGER. (3) that in Carinina the opening through which the proboscis is everted lies ventrally to this furrow ; (4) that in most Cerebratuli the proboscidian opening would similarly lie ventrally to the point of meeting of the lateral cephalic furrows if we supposed these to meet at the tip of the snout ; (5) that in a few cases the opening appears to have shifted into the central part of the fissure ; (6) that in other Cerebratuli the fissures, though continued on the anterior and apparently truncated portion of the snout, do not wholly fuse (PI. I. fig. 12), but that just in the interval the proboscidian opening is situated; (7) that again in other Cerebratuli this anterior truncated portion is wholly devoid of any continuation of the fissures, which in their turn may be exceedingly deep and long (PI. I. fig. 16); (8) that in the Amphipori alluded to (PI. IX. fig. 9) the common opening for proboscis and digestive cavity is also situated ventrally to the terminal fold ; (9) that in Ewpolia transverse and very shallow cephalic fissures are found, which very strongly resemble those of the Hoplonernertea (even in the presence of short and numerous secondary grooves perpendicular to the principal groove), and which similarly contain the opening that leads into the posterior brain-lobe, as is the case in Carinina and the Hoplonernertea, and that even yet in certain Eupolise a trace of a terminal horizontal furrow has been retained (PL I. fig. 7). Tabulating these different facts, the case would appear to stand thus : — Carinella annulata, one terminal shallow horizontal groove, two transverse lateral ones, no ciliated canal. ,, inexpectata, (VIII) terminal groove uncertain, transverse lateral grooves, ciliated canals into the brain-substance opening out into these grooves. Carinina, . . . terminal groove present, lateral grooves, with ojjenings of ciliated canal leading into a separate posterior brain-lobe. Amtphiporus, . . terminal groove present, lateral grooves, with openings of ciliated canal leading into a separate posterior brain-lobe. Ettpolia, . . . hardly a trace of terminal horizontal groove, lateral grooves as in the Carinellidas and Hoplonernertea. Valencinia, . . no grooves at all, simple round opening for ciliated canal. Schizonemertea, . two longitudinal (never transverse!) cephalic grooves which in some cases are wholly separate, in other cases meet at the tip of the snout, and might then in their entirety be compared to a terminal horizontal groove such as that of Carinina. If the latter conjecture be true, i.e., if we may suppose the lateral furrows of the Schizonemertea to be derived from an ancestral phase, in which a terminal groove like that of Carinina was separated into two halves which deepened and widened on both sides of the head, reaching down as far as the opening of the ciliated canal into the REPORT ON THE NEMERTEA. 93 posterior brain-lobe, then, indeed, Carinina may be said to represent — also with respect to the fissures and grooves on the head— a stage of development in which both the characteristic features of the Hoplonernertea and the Schizonemertea are still present in combination. In deviating from the original arrangement, the Schizonemertea would have gone a longer way than would the Hoplonernertea, some of which still answer to the original type of structure found in Carinina. I now wish to consider more closely certain details of these different cephalic fissures, which cannot be discussed in a more appropriate place than in the section treating of the sense-organs, although their exact significance must for the present remain unsolved. In spirit preparations the ciliated coating of the body is generally never better preserved than in these furrows ; and the cells carrying the cilia, as well as their nuclei, are in most cases exceedingly distinct. However, in a few cases it is only the superficial layer of the integument that is thus continued as a clothing of the inner surface of these furrows. In most cases I could observe the deeper layer of integumentary glands (Schizonemertea) to be continued, although less compact, along the whole inner surface of the cephalic slits. Generally these deeper glandular layers appear to have undergone some special modifica- tion in connection with the canal that opens out at the bottom of the furrow, and leads into the brain-substance, a modification which may already be noticed in so primitive a genus as Carinina, which, however, with respect to this apparatus, may be said to be more differentiated than the allied genus Carinella. We find in Carinina, (PI. VI. figs. 1-3) that all round the bottom of this groove (Cg) there is a marked increase of the number of nuclei in the integumentary tissue ; and although these nuclei can scarcely be said to belong to the layer of the deeper glands (cf. PL IV. fig. 1 E, and the paragraph on the integument), but rather to the one exterior to this, the fact of their accumulation in this marked way, just along the inner surface of the cephalic groove, is a most rehable indication that the integument is in some way modified in adaptation to the significance of these grooves. In PL VI. fig. 1 a distinctly pointed shape is assumed by this wedge-like or horse-shoe-shaped accumulation of nuclei, and a fibrous band connecting them with the intermuscular tissue is even visible, more or less clearly, separating — at least in this section — a posterior brain-lobe Br', into which the canal passes, from the anterior brain-mass Br (cf. woodcut, p. 81). The glandular layer is, however, not indifferent or neutral during these changes in the exterior nuclear one ; and although the two specimens at my disposal do not permit me to unravel the whole of the modifications it undergoes, I may still be permitted to observe that in figs. 2 and 3 of PL VI. its participation (Gi and gl.br) cannot be denied, whereas a comparison of all the three figures here given makes it appear very probable that these glandular elements (gl.br), derived from the deeper layer of the integument (Gi), play a part with respect to the posterior brain-lobe of Carinina, which may best be compared to the glandular investment of the posterior brain-lobe, as it is encountered in all the other Nemertea 94 THE VOYAGE OF H.M.S. CHALLENGER. (cf. PL VI. figs. 7, 8 ; PL XIV. figs. 6-8, 11). What the physiological meaning of this glandular investment may be must remain unsolved for the j)resent, although we will return to this question further on. Carinina demonstrates that, as the brain-lobes are direct derivatives of the integument, so is the glandular investment of the posterior one. I may here note that, in studying the development of Lineus obscurus, I have (XIV) been able to determine the fact that the glandular investment and the nerve-cells of the posterior brain-lobes also arise in that species out of the same mass of embryonic cells. Moreover, I must add that the glandular significance of this investment was for the first time more emphatically brought forward by Dewoletzky in a short notice on the Nemertea (II). Although I cannot accept all the conclusions to which this naturalist arrives with respect to the apparatus of which we are here treating, and must demur when he rejects the specially adapted respiratory significance which the brain canal must necessarily have in very many species, still it is only fair to call attention to his inquiry into the nature of this cellulo-glandular investment. The only point which has still to be noticed, and which is partly a repetition of what has been already described in the paragraph on the nervous system, is the fact of the actual observation by myself in DrepanopJiorus lanhesteri, from the Challenger collection, of the passage of the contents of part of these glandular investing cells into the lumen of the canal (PL XIV. fig. 10). Moreover, one point should not be lost sight of, viz., that between the glandular cells that form an actual investment of the posterior lobe (and which in Carinina could be identified with the deeper glandular structure of the integu- ment) and the actual integumentary glands, there exists in Schizonemertea a constant and considerable difference, even with respect to the affinities for staining reagents, and still more in the general aspect. The case of Carinina is on this account all the more interesting. We must only be careful not to look upon the exceptional case which I was able to observe and to figure (PL XIV fig. 11, gl), in which an additional glandular (1) ring surrounds the ciliated canal after it has passed out of the brain-lobe on its way to the exterior, as one of transition. I hardly think that this special adaptation, which has been already noticed above (p. 60), pertains to the layer of the deeper skin-glands ; and though I am not prepared to offer a definite opinion, I am much more inclined to compare this curious accumulation of distinct and nucleated cells with a similar accumulation which we have also noticed in Carinina, and have there encountered more peripherally but still surrounding the ciliated canal (PL VI. figs. 1-3). In either case the physiological significance of the arrangement cannot at present be decided. And to a certain extent this may also be said to be the case with respect to the whole of the posterior brain-lobe. As long as it went by the name of side-organ — which, how- ever, did no justice to its intimate connection with the brain — it was generally regarded as a specific sense-organ of unknown function (Quatrefages, MTntosh, &c). Later on I published a paper (IX) in which the attempt was made to show that, in a very large number REPORT ON THE NEMERTEA. 95 of cases, the ciliated canal is adapted to give the oxygenated sea-water access to the hsemoglobiniferous nerve-tissue. At the same time (loc. cit., p. 35) I did not deny all sensory significance to the organ, but repeated that we had not found any specialised sensory epithelium in it, and could not judge of what kind the sensory impressions might be that were carried by the apparatus to the animal's sensorium. Since then I have been able to fix the exact mode of origin of the apparatus in at least one species of Nemertea (XIV, XV), and may here recall to mind that the central lining of the canal most decidedly takes its origin as an invagination of the epiblast. This invagination second- arily coalesces with the brain. In these two ontogenetic data we have only very vague indications. They allow of a comparison both with olfactory and with auditory pits. Strange as it may seem, I do not see that the first comparison has many more a priori arguments for it than the second. Otoliths, it is true, have not been found in this lobe, but who can tell what purpose the minute concretions, formed by the ensheathing gland-cells, and sometimes accumulated inside the lumen of the canal, may serve ? We have, however, to suspend our judgment. Graeffe's, Keferstein's, and Claparede's observations on the existence of special otolith capsules in the Nemertea require further confirmation. They may, perhaps, have mistaken the highly refractive globules in the gland-cells of the posterior brain-lobe for otoliths. One more point may be mentioned, viz., that the comparison of the cavity of these lobes with a branchial slit of Balanoglossus, &c. (cf. Bateson, loc. cit.), which I tenta- tively attempted in a former paper (IX, p. 33) has to be definitely abandoned, now that the epiblastic origin of the cavity has been indubitably shown by myself and afterwards by Salensky (XIV, XXX), and since on this point the statements of earlier authors as to the hypoblastic origin of this cavity (Barrois, &c), which led to my former suggestion, have to be definitely abandoned. If the comparison with a gill-slit is no longer tenable on morphological grounds, this in no way changes my views as to the physiological importance that must be attached to the direct respiratory function of the nerve-tissue, which can nowhere be so perfectly accomplished as in the posterior canalised lobe. I have no doubt, however, that in some species — -more especially of Hoplonemertea — its significance as an organ of sense may supersede its importance as a respiratory chamber, the haemoglobin, though present in these species, being there much more diluted, at any rate colouring the brain less intensely red, and the connection between posterior and anterior lobes being at the same time less intimate. Having now discussed all those parts of the organism which we have any — though sometimes even questionable — right to consider as sense-organs, I must pass on to those which are of a still more dubitable nature, and which fall under the head for which the further part of the superscription of this section was intended. Since in some cases I find them in the head and directly innervated by the brain, since 96 THE VOYAGE OF H.M.S. CHALLENGER. in another case they are at the sides of the head stretching backwards for a not incon- siderable distance, but always exactly at the lateral margin, and because a glandular epithelium plays an important part in their constitution, there is indeed some, though of course very distant, analogy with the glandular parts of the posterior brain-lobes which we have considered before. I do not wish to attach any importance or significance to this analogy for the present, for we have no sufficient data ; but I mention it by way of explanation as to how I come to intercalate the description of these parts of the organism in this place. The first structure which I have to mention occurs in the head of Drepanophorus lanlcesteri, and something analogous to it was noticed by former observers (XVI) in Geonemertes palaeensis. In the Challenger specimen the horizontal sections through the precerebral region demonstrate, when viewed with low powers, the presence of a lobulated mass which is imbedded in the gelatinous ground-substance and partly traversed by contractile fibres that radiate through the head in so many directions. When higher powers are applied, this mass is dissolved into groups of cells enclosing more or less circular free spaces, which, being present in consecutive sections, represent a system of canals coated by the cells just mentioned, the whole forming a kind of spongy tissue. A branch by which this canalicular system communicates with the exterior could not be made out in my specimen, although von Kennel has found such an opening in his Geonemertes, and thus I do not wish to lay too much stress on the fact of my being unable to rediscover it in the only Challenger specimen in which I found this structure. The cells are much more granular and at the same time larger than the surrounding cells of the intermuscular gelatinous mass ; the nuclei are large and distinct (PL XV. fig. 13). It must be noted that the character of the cells and the aspect of the organ differ very essentially from von Kennel's description. It is the situation that is correspondent. A similar precerebral glandular lobulated organ was found by me in Drepanophorus rubrostriatus from the Mediterranean, though not in the Challenger specimens of this species, in which, as was remarked above, the head was deficient. A special innervation by nerve-fibres belonging to the cerebral nerves was in both cases made out. The second structure to which I alluded as occurring in Challenger Hoplonemertea was found by me in the different specimens of Amphiporus moseleyi, both in transverse and horizontal sections. It may shortly be characterised as being an accumulation of short, saccular tubes, blind posteriorly and opening to the exterior by a distinct neck, which pierces muscular layers as well as basement membrane, its internal epithelium then fusing with the integument. These short flask-like sacs sometimes internally coalesce with each other, the same interior cavity then communicating outwards by more than one duct. This, however, appears to be more or less exceptional. They are very numerous, though short, at the tip of the head (PL X. fig. 3, gl.s). They become larger when we follow them further backwards, where we find them situated laterally in that REPORT ON THE NEMERTEA. 97 region of the body where the dorsal musculature passes into the ventral, and where the intervening layer of muscular tissue is either very thin or even (sometimes locally) absent, at least as a special layer. In every transverse section some three or four of these glandular flasks are simultaneously cut (PI. XV. fig. 11). They show no further variety of structure, but are not found along the whole length of the body. In the hinder half at least I have not detected them. The histological constitution of these flask-shaped glands is difficult to make out in the spirit specimens. The epithelium clothing them had a very much changed and deteriorated aspect, and only in certain favourable regions could I definitely establish the fact that it was built up of largish cells with large nuclei (PI. XV. fig. 12), the contents of the cells being granular like those of the cells of the median precerebral organ before described. Innervation was not so clearly traced here, though in the head these flask-shaped organs certainly receive fibres from the very numerous cerebral branches there present. Having once determined the presence of the organs just described and their situation, it was found that their presence is even externally traceable in the specimens of Amphi- porus moseleyi (see PI. IX. fig. 8) by a white line running backwards from the tip of the head along the comparatively sharply edged margin of the body (see pp. 20, 21). When a transverse section is made in this region with a razor, the naked eye can trace this white line extending inwards for a short distance as if the pigment occasioning it were very thickly applied. Viewed with the microscope, it is then easily seen that this white spot breaks up into the accumulated flasks as above described, which are surrounded and supported by the gelatinous tissue. In such a section the latter tissue is much more transparent and homogeneous, thus bringing out the glands as white lateral spots in this transparent imbedding mass, in which also the rest of the internal organs may be seen to be suspended, and which is dorsally and ventrally limited by the body musculature and the integument. It may finally be noticed that I have never succeeded in finding the necks of these flasks that lead to the exterior wholly free and open, as I have the canal of the posterior brain-lobe or the excretory duct of the nephridia. These very numerous necks of the flask-shaped organs are always filled with a mass that has a streaked and fibrous appear- ance. I mention this because it partly contributes to establish my conviction that the physiological significance may indeed be glandular, and that the secreted, viscid mass, passing out perhaps more copiously when the animals are immersed in spirit, remains fixed in this passage to the exterior. If it be not premature — as I consider it to be — to establish any comparison between these organs and parts of the organism of Vertebrates, one would certainly be reminded of those canals in the head which are known as the " Sehleimcanale," and which are continued along the sides of the body as the lateral line. The significance which of late has been more and more definitely assigned to these organs in the Vertebrates, as (ZOOL. CHALL. EXP. — PART LIV. 1887.) Hbh 13 98 THE VOYAGE OF H.M.S. CHALLENGER. accessory to the sense-organs forming the system of the lateral line, a sensory epithelium being protected by and combined with them, has no direct parallel in Arwphvporus moseleyi, although it cannot be denied that the flask-shaped glands are in the immediate vicinity of and on a level with the lateral nerve-trunks. The absence of these glandular structures in the Palseonemertea and Schizonemertea hitherto observed, renders the suggestion of any close homology very hazardous. Still, I would not wholly refrain from pointing out the distant kind of parallelism which may be noticed, and which has certainly contributed to induce me to consider these organs in the paragraph devoted to the sense-organs, a proceeding which future inves- tigations may perhaps show to have been wholly unfounded. The significance of this parallelism will once more be discussed, when, in the chapter of General Considerations, there is further scope for speculation. PROBOSCIS AND PROBOSCIDIAN SHEATH. Concerning this important organ, so very fully described by MTntosh in his Monograph on the group (XIX), the Challenger material has not revealed any startling peculiarities. Nevertheless, it deserves some closer consideration, because certain points, e.g., the exact mode of the anterior attachment of the proboscis in these worms, could be studied more favourably by me in certain of the Challenger sections than ever before. Moreover, the Russian naturalist Salensky has lately1 propounded certain views concerning the proboscis and its sheath which deserve consideration and refutation. I will first describe the facts with respect to the proboscis which we notice in the Challenger Nemertea. Carinina has a proboscis which, in transverse section, reveals the remarkable peculiarity that the primitive order of succession, according to which in the body-wall we meet with (1) integument, (2) longitudinal nerves, (3) musculature, also obtains in the proboscis, the innervation of which takes place through the intervention of two longitudinal nerves, which are so situated as to be enclosed by the internal cellular epithelium (PI. II. figs. 11, 12), just as is the body nerve-stem in a section of the trunk. This fact, though it cannot be looked upon as a direct confirmation of the hypothesis advocated by me after I had become acquainted with Graffs Monograph on the Rhabdoccela, viz., that in the Nemertea also the proboscis should be looked upon as a gradual derivative of an original continuation of the body-wall, which has become introvertible like the snout of the Rhabdoccela proboscidea, still throws a very favourable light on these views. And this is further the case when we notice that in many Schizo- nemertea there is also an order of succession of the layers in the proboscis-wall which is 1 Zeitschr. f. wiss. ZooL, Bd. xliii. p. 509. REPORT ON THE NEMERTEA. 99 similarly more or less a repetition of the same arrangement in the body-wall (PI. XV. fig. 2, a', £', npl, y'). In addition to the peculiarity just described, there is another morphological consider- ation which tends to show that this interpretation of the significance of the proboscis is indeed the right one. When we consider a horizontal section through the region of insertion of the proboscis in the head (PI. III. fig. 5), we see that in Carinina the mode of fixation of the proboscis is exceedingly simple, its longitudinal muscular coat being in direct continuity with the longitudinal muscle-layer of the body-wall. Somewhat in front of the transverse cephalic grooves, about on a level with the anterior brain-lobe, we see certain of the fibres of this longitudinal coat, instead of pursuing their course onwards towards the tip of the head, bending inwards, traversing the space which I have termed ( XIII, XV) the archicoele, and then running backwards as the longitudinal fibres of the proboscis. Other fibres, parallel to those just referred to, do not contribute towards the formation of the proboscis, but continuing in their original direction, take part in the formation of the muscular wall of the head (PI. III. fig. 5). It certainly deserves remark that the same comparatively simple arrangement is met with in the much more highly differentiated Hoplonemertea, as a glance at fig. 3, PI. X. will show. There, too, the longitudinal musculature (a) of the body-wall is partly continued towards the tip of the head, where it partly bends round and largely contributes to the formation of the muscular layers of the proboscis. I suppose this way of stating the facts is more in accordance with their actual relations, than to say that the longitudinal musculature of the proboscis is internally inserted upon that of the body. Here also the direct continuity of body -wall and proboscidian-wall, the latter appearing merely as an inverted portion of the former, is forced upon our attention, as is in the same way the direct continuity of the exterior integument J, through that of the rhynchodseum Rh to that which clothes the proboscis itself, and which on the eversion of that organ forms the exterior surface. We must now consider these different parts more in detail. Commencing with the rhynchodseum {of. p. 8), we find in the Palseonemertea and Schizonemertea its walls bathed by the blood-spaces in the head, as may be gathered from a comparison of the figures in Oudemans' paper (XXVIl) on the blood-vascular system. This is no longer the case in the Hoplonemertea, where these blood-spaces are replaced by the distinct vascular loops. The proboscidian walls, fusing anteriorly with the musculature and the external epithe- lium of the head, are different in the different subdivisions. Contractile fibres and cellular elements, the materials of which the rhynchodaeum is built up, are present in Carinina in quite a different relation from that in which they occur in Cerebratulus or Amphiporns. In Carinina, as a glance at PL III. fig. 5 will show, it is the cellular elements (APe) that are extremely preponderant. These cells are vacuolated, more than one layer thick, different in aspect from the true proboscidian epithelium, and held together by a fibrous 100 THE VOYAGE OF H.M.S. CHALLENGER. coating, which is attached by radial fibrous bands to the cephalic musculature. The rhynchodaeum is thus suspended in the cephalic blood-space, as was formerly (IX) described by me in Carinella. The histological difference between the vacuolated cells of the rhynchodaeum of Carimna and the cells of the outer integument, is less than that between the former and the epithelium of the proboscis proper. A comparison between figs. 3, 4 of PI. III., fig. 1 of PL IV., and figs. 1-3 of PI. VI. will sufficiently demonstrate this, vacuolated cells playing a very prominent part in the outer strata of the integument in Carinina. Still the three epithelia (external, rhynchodseal, and proboscidian) are immediately contiguous, the passage from the one to the other being gradual and only in the latter case relatively abrupt. In the rhynchodasum of the Schizonemertea and Hoplonemertea the cellular and ciliated layer of the rhynchodasum of the earbest Palaeonemertea has been relegated to the back- ground, and the whole has become more a muscular sheath, in which the muscles have, however, a different arrangement in the first and in the second group. The increase in muscularity is in the Schizonemertea more a regular thickening of the fibrous investment, whereas in the Hoplonemertea it is much more massive in one region than in the other. In this way an annular and massive muscular sphincter (as it may be adequately termed) arises in the posterior part of the rhynchodaeuin (PI. X. fig. 3, Sp. Pr.). In this muscular sphincter longitudinal and circular fibres are very intimately inter- woven, more or less in basket fashion, as indicated in the figure. Moreover, the connection with the general musculature of the head is again brought about by radial bundles, also visible in the figure. Of the Schizonemertean rhynchodaeum no special figure is given ; it answers to the short description which was given above, it can be well observed in several figures in MTntosh's monograph, and, like the rhynchodasum of the other groups, it reaches back- wards just as far as the implantation of the proboscis in the musculature of the head. The epithelium of the rhynchodasum is in most cases distinctly ciliated. The proboscis itself has been the subject of so detailed study and so elaborate description by M'Intosh and other investigators, that I must necessarily restrict myself to those few points on which the Challenger material furnishes certain deviations or additions. The inner epithelium of the proboscis of Carinina shows considerable differences according to the region under observation. In front there are papillae of a more or less arborescent shape, on which a coating of fairly large cells, with distinct nuclei, and partly vacuolated, is present. Posteriorly the cells are lower and more closely set (PI. III. figs. 1, 2, Pe). This may be partially the result of a different state of contraction, by which the anterior portion is thrown more into folds. The difference is, on the other hand, further accentuated by the presence of a con- REPORT ON THE NEMERTEA. 101 striction in the proboscis, which I observed in the longitudinal sections of the only specimen of Carinina that had retained its proboscis. I cannot affirm that this constriction was natural, i.e., that it would also be found in fresh specimens. Since, however, I have formerly described (VII) similar constric- tions in the proboscis of Carinella and Valencinia, there are many a priori grounds for also accepting its normal occurrence in Carinina. At the same time it more or less coincides with the change in the character of the epithelium just noticed. The epithelium of the proboscis of certain Schizonemertea was known to be charac- terised by the presence of nematocysts. In a former publication (VII) I showed that the observation of Max Muller, who first noticed urticating elements in Cerebratulus urticans, might be extended to nearly all Schizonemertea, although the size of these elements is generally considerably below that of the type species just mentioned. Muller has given good figures of the shape of the elements in his species ; of the others no figures have hitherto been given, and fig. 2 of PI. XV. is intended to show the situa- tion of packets of urticating elements, batteries, as they might be called, in a transverse section of the anterior part of the proboscis, rather than to furnish particulars concerning the histology of these nematocysts. They are seen to be situated close to the free surface of the cells, and to be of different sizes on the dorsal and on the ventral surface of the proboscis. Three batteries are figured lying free in the lumen of the proboscis ; when seen from the side they have the aspect of a brush with close hairs seen in the same way, when seen from above they appear to be more or less circular, and each of the elements composing the battery is then found to be represented by a fine dot instead of by a straight line, as was the case in the side view. In the spirit specimens of Cerebratulus more than these general facts could not be ascer- tained. I may add that fresh specimens from the Mediterranean showed that each of the elements out of which such a battery is composed has a spindle-shaped form, being more or less pointed at both ends and somewhat bulging in the middle, and that from one of the pointed ends — which in its natural position is directed away from the proboscidian epithelium — the fine urticating thread may be observed to issue. This thread is, in most cases, comparatively short. Whde in Cerebratulus urticans there is hardly any doubt that each urticating element may act independently of the others, it is not improbable that in some species, as the one here described, they remain connected in batteries by whose joint action, when the proboscis is projected, delicate animals may be wounded or para- lysed upon extrusion of the proboscis, and may thus fall an easy prey to the proprietor of this formidable weapon. As to the more detailed histology of this epithelium, I wish to withhold further remarks till I am enabled to publish the observations on the fresh specimens examined at Naples. Similarly the curious and very adhesive epithelium of the foremost portion of the Hoplonemertean proboscis which I have formerly described (IV, V), which was well 102 THE VOYAGE OF H.M.S. CHALLENGER. known to M'Intosh, Graff, and other observers, and which I have again met with in the Challenger Hoplonomertea, cannot be more circumstantially described. Moreover, the stylets in the different Hoplonemertea did not offer any remarkable deviation from the well-known type, and though the transverse sections gave very clear detads regarding the arrangement of the muscle-fibres in the muscular bulb, about the epithelium of the glandular duct that conveys the probably venomous secretion of the posterior cavity to the base of the stylet, &c, these are only confirmations of facts already known and need not be recapitulated here. The shape of the stylets was mentioned when the species were described; those of Drepanophorus, though not obtained from an actual Challenger specimen, are represented in the woodcut on p. 16. The muscular walls of the proboscis differ in the various genera, and these differences speak for themselves when we compare figs. 11 and 12 of PI. II., and figs. 1, 2, and 5 of PI. III. (Carinina), fig. 11 of PI. VI. (Eupolia), fig. 7 of PI. VIII. (Pelagonemertes), fig. 6 of PI. XII. (Amphiporus) , and figs. 2 and 3 of PI. XV. (Cerebratulus). These latter show the muscular layers of the proboscis of the Schizonemertea to be a repetition of the muscular layers of their body-wall : a circular layer between two longitudinal ones, the circular layer giving off fibres at diametrically opposite poles to the external membranous sheath (b), and moreover, a nervous plexus (n.pl), which is also situated between the outer longitudinal muscular coat y' (that just below the epithelium), and the circular one ft'. This nerve-plexus does not go all round, at least it cannot be distinctly made out except throughout one-half of the circumference. It is also traversed by radial fibres, and is again replaced by definite longitudinal stems when we examine a transverse section of the proboscis further back (PI. XV. fig. 3). These longitudinal stems are characteristic of certain species of Cerebratulus, and a plexus, even a far more complete and cylindrical one than the one figured (PI. XV. fig. 2) for Cerebratulus macroren, is characteristic of others. The nerve-stems enter the proboscis at its point of insertion, and spring from the right and left extremity of the ventral brain commissure. In the posterior regions of the proboscis of Eupolia, of Pelagonemertes, and of nearly all the other species, the musculature appears to be reduced to a simple longitudinal layer, carrying the epithelium on one side, and being held together by an ensheathing membrane on the other (PL VI. fig. 11, and PL VIII. fig. 7). In Carinina there is an additional circular layer, and the remarkable fact, which has been already noticed above, of the situation of the nerves still enclosed in the epithelium. In Amphiporus, Drepanophorus, and Pelagonemertes (anterior portion) the proboscis-wall exhibits the notable complications corresponding with the curious dis- position of the nerves in the proboscis, and which was described in sufficient detail by M'Intosh (beaded layer) (XIX, XX), myself (IX), von Kennel (XVI), and Graff (HI). To von Kennel the merit is due of having definitely established the nervous significance of the parts in question. The innervation may here be REPORT ON THE NEMERTEA. 103 said to represent a plexus with numerous longitudinal thickened portions or stems. The passage of the nerves from the brain into the proboscis can very rarely be well observed, because the proboscis is nearly always extruded and torn off when the animal is killed. I may, however, repeat what was noticed above (p. 85), viz., that in Amphiporus moseleyi, more particularly, this doubt has now been dispelled. I can observe in my sections that, instead of two strong nerves innervating the proboscis, as in the Palaeonemertea and Schizonemertea, a much larger number of branches leave the brain-ring and enter the proboscis in the region of its attachment. That these may dichotomise and give rise to a larger number of longitudinal stems has been already stated by von Kennel (XVI). This nervous plexus and the longitudinal stems subdivide the longitudinal muscle- layer into an outer and an inner portion, the latter (when the proboscis is everted) being again subdivided into as many longitudinal columns as there are nerve-stems in the proboscis (PI. XII. fig. 6). Outside and inside of this longitudinal layer there is a circular layer of fibres, outside of the exterior one of these the epithelium. As to the nerve-stems and the plexus, one specimen of Amphiporus marioni showed very distinct cellular accumulations just between each nerve-stem, as if a longitudinal tract of nerve-cells alternated with one of nerve-fibres in the plexus. For the study of this phenomenon fresh specimens will be absolutely necessary. The phenomenon itself has been already noticed, but has not yet been wholly understood, either by von Kennel (XVI) or by Graff (ill). Just as it has been necessary to curtail our observations on the proboscis because of the detailed information already available concerning this important organ, the probos- cidian sheath need not be treated at any length in view of the data that are already furnished by others. It is known to be a closed space surrounding the proboscis, having in the majority of cases its own muscular wall, by the contractions of which the fluid contained in the space is driven against the anterior proboscidian attachment. The muscular sheath thus serves to protrude the proboscis as far as the length of the posterior portion — acting as a retractor-muscle — will allow it. There can hardly be any doubt, when we take into consideration all the morpholo- gical data at our disposal, that the muscles composing the proboscidian sheath gradually took their origin by the increase and modification of pre-existing muscular elements, which belonged to the body-wall and to the body-parenchyma before the proboscis, modified from a tactile organ, as it appears to have primitively been, had yet become evolved, through the growth inwards of the anterior tip of the body, into an aggressive weapon, with stylet or nematocysts, &c. We find the shorter proboscides, and the less significant proboscidian sheaths among the more primitive genera of Nemertea. Carinella has a short proboscis ; the dorsal wall of its sheath is still a component part of the musculature of the body-wall ; the ventral wall is thin, and only composed of a 104 THE VOYAGE OF H.M.S. CHALLENGER. few fibres. So it is in Carinina (PI. II. figs. 3, 6, 7, 9, 10 ; PI. IV. fig. 6, Ps). In Eupolia the proboscis is longer, but the sheath is still most insignificant, as may be gathered from the figures (PI. VI. figs. 9, 10 ; PI. VII. fig. 10). It is a space having internally a cellular coating very similar to that of the blood-spaces, the cells of this internal epithelial covering often more or less projecting into the lumen of the sheath. Outside of these cells a few circular fibres are seen to have developed ; outside of these there is again the body-paren- chyma, with the enclosed blood-lacunas. There is no doubt that from sections of Eupolia alone nobody would be inclined to look upon the cavity of the proboscidian sheath as a very independent cavity, nor is it possible to affirm that the mode of the protrusion of the proboscis, as it was sketched above, is indeed fully developed in Eupolia and Carinella. There is no doubt that of all Nemertea observed alive, these two were never seen to protrude their proboscis spontaneously, and very often even preserved them in death, when the Hoplonemertea always forcibly expel and even spontaneously detach their proboscis. There is, on the other. hand, no evidence at all which would justify us in regardiug the arrangement of these Palseonemertea as secondary or degenerated from a higher differentiated stage. The participation of the body musculature in bringing about the movements of the proboscis in these lower forms renders this more intelligible. Only in the more highly differentiated Schizonemertea, and especially in the Hoplo- nemertea, the muscular walls of the proboscidian sheath undergo a very rapid increase in bulk, and at the same time become more and more, and in the last-named group even wholly independent of the body musculature. This increase of an organ so eminently mesoblastic as the proboscidian sheath, by gradual addition of new fibres that are even arranged in multiple layers, can thus be traced in all its various stages in the different genera of Nemertea. Salensky would probably not have made his startling hypothesis above alluded to,1 based on ontogenetical observations of a scission in the proboscidian wall, by which (l) a muscular proboscidian sheath surrounding the proboscis becomes separated from, and independent of, the musculature of the proboscis itself, and (2) an isolated ccelome — the proboscidian cavity — is originated, if he had been as well acquainted with the comparative anatomy of the animals about which he writes as he is with certain details of their ontogeny. Granting even that the development may, in the species observed by him, follow the paths he has sketched (my own observations on the ontogeny of Lineus obscurus (XIV) have led to wholly different results on this head), it is not yet per- missible to base upon those two ontogenetic observations phylogenetic speculations wholly at variance with all the facts that are furnished by a comparison of the different living genera. The woodcuts given by Salensky, in which a Rhabdoccele proboscis and a Nemertean one are put side by side, look very tempting, but cannot be accepted by me. 1 Archives de Biologie, vol. v. p. 561 ; Zeitschr.f. wiss. Zool, Bd. xliii. p. 508. REPORT ON THE NEMERTEA. 105 While fully recognising the importance of Graff's observations for our own interpretation of the Nemertean proboscis, and the genetic relation of this organ to that of the Rhab- docceles (not direct but collateral), I must as emphatically reject the proposed derivation of the proboscidian sheath advocated by Salensky. We must, indeed, represent to ourselves the gradual evolution of the proboscis as that of an epiblastic organ reaching further and further inwards in successive genera- tions, and strengthened and completed by a mesoblastic musculature ; and outside of this the free and independent development out of other mesoblastic elements (primarily belonging to the body-wall) of the sheath. It has been abeady noticed elsewhere (XIV, XV), that if these mesoblastic structures could be traced down to amoeboid mesoblast cells derived in loco out of the subjacent hypoblast, an ontogenetic homology between the tissues constituting the proboscidian sheath and those forming the notochord of Vertebrates would be established. Returning to the proboscidian sheath of the Schizonemertea, we find it to consist of an outer layer of circular fibres and an inner one of longitudinal (PI. X. fig. 8, mPrs; PL XV. fig. 1). The former sometimes, when the sheath is thick and contracted, shows a wavy line. Radial fibres, piercing the two fibrous layers, insert themselves against the inner epithelium, which covers the whole inner surface, looking towards the cavity of the proboscidian sheath. Between this epithelium and the muscular layers there is a broad band of transparent basement tissue (PL XV. fig. 1, b) following the numerous longitu- dinal folds of the epithelium just mentioned. These folds disappear when the proboscidian sheath is in distension (PL X. fig. 8), a phase that may repeatedly be noticed, even without any extrusion of the proboscis, e.g., as a consequence of complicated coi lings of the proboscis inside its sheath. It is easily understood that during such distension the thickness of the subepithelial homogeneous basement layer and of the muscular layers is considerably reduced. A maximum degree of distension is figured on PL X. fig. 9, where the epithelium was no longer separately visible, and even the oesophageal epithelium has been flattened out, together with the proboscidian-sheath wall. As will be seen from PL XV. fig. 1, we find outside of the outer circular layer of the sheath the gelatinous body-parenchyma, a thin layer of this even separating the probos- cidian sheath from the longitudinal muscular layer a, in the midst of which we notice the true proboscidian-sheath-nerve (pr.sn). In addition, I think it is not unimportant to remark, that just below this layer of longitudinal fibres, there are strands of circular fibres which do not apparently belong to the proboscidian sheath, and which, after having been closely applied against the dorsal musculature in the middle line of the back, radiate amongst the parenchyma and the intestinal caeca. It is these fibres (and perhaps in addition to them the circular layer of the sheath itself) which may possibly be looked upon as representative of the layer S in the Carinellidge (cf. PL XL), and which there takes such a conspicuous part in the dorsal delimitation of the proboscidian sheath. (ZOOL. CHALL. EXP. PART LIV. — 1887.) Hllh 14 106 THE VOYAGE OF H.M.S. CHALLENGER. In many Schizonemertea the proboscidian sheath is thus constituted ; in others I find that a phenomenon, which receives its more final expression in certain Hoplonemertea, is not wholly absent — I mean the presence of varicosities, in which the inner space of the sheath is bulged out, without the musculature following. In this way more or less irregular appendages are brought about, generally along the side or the lower corners, having the aspect of accessory reservoirs. In Cerebratulus sp. inc. (PL XV. fig. fi), from Japan, it seems as if these appendages in the posterior region of the body even surpass in size the sheath itself, which is not a very significant organ in that region of the body, and, moreover, as if these two caeca are filled with a sub- stance of the nature of which the available spirit specimens do not enable me to judge. I do not wish to discuss here the significance of these facts, not having for the present sufficient material to study them more fully ; it is only my purpose to call the attention of future investigators to the phenomenon, which may be so significant for a correct inter- pretation of the posterior, often semi-rudimentary portion of the proboscidian sheath. It has been observed that in certain Hoplonemertea the phenomenon just noticed finds a more definite and more regular expression. The first observation of this is due to M'Intosh, who detected in the proboscidian sheath of Drepanophorus regular metameri- cally placed openings, by which the space inside the muscular sheath communicated with other cavities outside of it, that had no muscular walls (XX). While M'Intosh supposed these accessory cavities to establish a communication between the cavity of the proboscidian sheath and the blood-vascular system, I have since demonstrated (VII) that no such communication exists, but that Dre-pano-pliorus possesses closed membranous sacs communicating with the proboscidian sheath, and probably serving as reserve spaces for the fluid contents of the proboscidian sheath during the very powerful contractions and distentions which the organ may undergo. In the Challenger specimens the same phenomenon was observed, and I have even ventured to assign all those specimens in which these regular paired appendages of the proboscidian sheath were found to the genus Drepanophorus, even when I have not succeeded in deter- mining the armature of the proboscis so characteristic of the genus. Two sections through the proboscidian sheath of Challenger Drcpanopliori are figured on PI. X. figs. 4, 5. The curious arrangement of circular and longitudinal fibres, having the appearance of basket-work in the transverse section, may be understood from these figures, even without any further description. Between the musculature and the inner epithelial layer there is again a homogeneous membrane, with longitudinal folds indicative of the contracted state in which the sheath here figured was at the moment of its preservation. In neither of the two was the section quite vertical, thus only one of the lateral diverticula is cut, instead of the pair that are REPORT ON THE NEMERTEA. 107 * opposite each other. The epithelium which coats the internal cavity of the sheath is seen to be continued uninterruptedly in these lateral spaces, whereas the musculature is deficient, and these caeca may thus with propriety be called membranous. There is a decidedly thicker epithelial coating in the lateral sacs of fig. 4 {Drepanophorus lankesteri) than in those of fig. 5 [Drepanophorus serraticollis). On the other hand, the musculature of the proboscidian sheath of the latter is much more massively developed than that of the former. One remarkable detail concerning the lateral appendages in Drepanophorus lan- kesteri, is the fact tbat I found a few of the anterior ones connected by a short longitu- dinal communicating tube at their distal extremity, this connection being thus parallel to the proboscidian sheath itself. Simdar connections were not noted further backwards, nor in any other species of Drepanophorus. Whde the proboscidian sheath of Amphiporus marioni (PI. X. fig. 1) is built on the same plan as that of Drepanophorus, that of Pelagonemertes is seen to be much simpler (PI. VIII. fig. 7). Both are quite freely suspended in the gelatinous tissue, and only connected with the body musculature in the head (PI. X. fig. 3). DIGESTIVE APPARATUS. The digestive canal of the Nemertea cannot be said, from a morphological point of view, to be very complicated. Communicating with the exterior by a ventral mouth close behind the tip of the head and by a terminal anus, it stretches along the whole length of the body, and only two rather sharply defined regions may be distinguished in it : the oesophagus and the hind gut or intestine proper. Still, even the mouth is not always an independent structure, as it is known to become confluent with the opening through which the proboscis pro- trudes, i.e., the terminal opening of the rhynchodaeum,1 in at least two genera [Amphiporus and 3IaIacobdella2). In that case this common opening is either terminal or nearly so (PI. IX. fig. 9), and generally larger than the separate openings in other Hoplonemertea. This feature is clearly not primitive but derived from that condition in which the mouth lies behind the brain-lobes on the ventral surface, as it does in the most primitive 1 Rhynchodaeum (see p. 8) is the name that may conveniently be given to the passage stretching from the point of insertion of the proboscis in the head to the level of the exterior opening on the surface of the body through which the proboscis is seen to be thrust forth. Its walls are marked APe in fig. 5 of PI. III.; Rh. and Sp. Pr. in fig. 3, PI. X. * Salensky has lately (Biologisches Centralblatt, 1883, p. 740, and Archives de Biologic, vol. v.), in publishing embryo- logical researches on a certain species of Nemertea, imagined that the feature here alluded to was then and there discovered by him for the first time, and necessitated the creation of a new genus {Monopora). Although his attention was drawn to the superfluity of this proceeding (XTV. p. 41), he still retains the name in a later publication (Zeitsclir. f. vriss. Zool, Bd. xliii. p. 481). Still, I am afraid this will not extend its longevity, as all the other anatomical characters most decidedly conform to Amphiporus. 108 THE VOYAGE OF H.M.S. CHALLENGER. Carinellidse and in the Schizonemertea, in fact in all Neniertea, with the exception of the Hoplonemertea. In the latter the mouth is always in front of the brain ; it has thus shifted forwards, the extreme range of this shifting process being reached when the mouth becomes confluent with the opening of the rhynchodseum just recorded. Another difference between the Hoplonemertea and the two other orders of the class, with respect to the digestive system, is found in the relative position of the oesophagus and hind-gut. While in the two last-named groups these two subdivisions of the intestine pass into each other along a straight line and do not overlap, we see that such an overlap- ping does occur to a more or less considerable extent in the Hoplonemertea. In a number of transverse sections the hind-gut is cut when the oesophagus is also still present in the section, showing that the latter overlaps the former. Still, I should be inclined to adopt the view that the gradual process by which this came about was not so much a further extension backwards of the oesophagus, as a tendency of the hind-gut to spread out and to reach forwards below the oesophagus. This would seem to be indicated by the fact that in these Hoplonemertea the intestinal caeca, that properly belong to the hind- gut, but that have come to be situated below the oesophagus (PI. XV. fig. 20), may even reach so far forwards as to become situated close to the brain-lobes, a phenomenon which is never observed in the lower groups, where the whole length of the oesophagus separates the brain-lobes from the hind-gut. Possibly the shifting forwards of the hind-gut and its diverticula may be a phenomenon that runs parallel with (if not due to the same cause as) the disappearance of the lacunar blood-spaces round the oesophagus, and the substitu- tion for them of cylindrical blood-vessels communicating by transverse ducts with the medio-dorsal vessel. The latter arrangement is also typical of the region of the hind- gut in the Schizonemertea, where, however, the circumcesophageal portion of the blood- system is eminently lacunar. These speculations need not, however, be further insisted upon, and we may now pass to a description of the oesophagus in the Challenger Palaeonemertea. Here, again, Carinina offers features of interest. In the first place, the exceedingly close application of the oesophageal epithelium against the muscular body-wall below and the thin muscular layer of the proboscidian sheath above is peculiar (PI. IV. fig. 3). There is no gelatinous connective tissue between the cells and the bundles of circular muscles, not even a base- ment membrane, and strong powers are wanted to demonstrate any intervening tissue between the bundles themselves, so strongly are they interwoven, and so dense are the muscular layers in this region of the body-wall. Anteriorly there is a sharp bend down- wards where the mouth is situated, and in front of this a short bulging out forms a prostomial extension to the oesophagus, which is seen to be cut through in PI. II. fig. 3. The cells of the oesophagus, as seen from the section figured, are finely granular, and below those which actually clothe the lumen there are sometimes seen others also with large nuclei (PI. IV. fig. 7) but less granular and with less distinct boundaries. This REPORT ON THE NEMERTEA. 109 figure at the same time reveals the presence of a conspicuous cuticula covering the free surface of the oesophageal cells. It is not streaked, as the figure erroneously indicates, but homogeneous. On it the cilia are implanted. Concerning the behaviour of the intestine in the posterior body-regions of Carinina, nothing can be said, as only anterior fragments were preserved. In Eupolia the oesophagus has become more independent of the body-musculature than in Carinina, and in addition to this a separate oesophageal musculature — at least in Eupolia giardii — is present (PI. VI. fig. 9, oe.m). In this oesophageal muscular investment an inner longitudinal and an outer circular layer may be distinguished ; between the latter and the body -wall there is the gelatinous tissue, only locally inter- rupted by the lacunar spaces of the vascular system (in which the nephridia [wp] are suspended), by the dorsal blood-vessel, and by the proboscidian sheath. I mention this, because there is no evidence that this splanchnic musculature has directly evolved out of hypoblastic elements, whereas the evidence that it is enclosed in one continuous stroma with the " somatic" musculature — which in Carinina was the only musculature noticed — and not separated from this by a body-cavity, is very complete. It should, moreover, be remarked that where . this more prominent intestinal musculature makes its appear- ance (certain Schizonemertea, Ewpolia giardii, &c), the internal circular muscular layer of the body- wall of the Carinellidae (8, PI. XL) is no longer present in that -situation. How far these two may be considered as homologous, must be left undecided as long as we do not possess more complete ontogenetical data. This oesophageal musculature was not noticed in all species of Eupolia, and it is certainly curious that it should be present in Eupolia giardii, where the body musculature is so exceptionally thick, and might be expected to serve the purpose of compressing and dilating the oesophageal wall quite as efficiently as we must suppose this body musculature to do in the Carinellidse. M'Intosh, who detected it in Cerebratulus corrugatus, suggests (XXII) that the oesophageal musculature might assist in a partial protrusion of the oesophagus. In Eupolia nipponensis and Eupolia australis this special musculature is absent, and we find, on the contrary, a much more considerable development of the deeper cellular layers of the oesophageal wall (PL VII. fig. 12), and a comparatively sharp demarcation between the internal ciliated epithelium (Je) and this thick cellular coating (Jm), this demarcation even sometimes rising to the importance of an apparent basement membrane (B). Between these cells radial fibres, starting from the body musculature, penetrate, and solitary tangential fibres may be observed, but a separate muscular investment of the oesophagus can never be shown to exist. I am not inclined to believe that this difference may be caused by the age or the size of the individual, one of the specimens of Eupolia nipponensis being indeed of considerable size and nearly as thick as Eupolia giardii. The same deep cell-layer is met with in the oesophagus of the Schizonemertea (PL XIII. fig. 6), and is no doubt of glandular significance. Generally many of the 110 THE VOYAGE OF H.M.S. CHALLENGER. component cells are flask-shaped, the thinner extremity shoving in between the pallisade- shaped inner ciliated epithelium. Moreover, among the Schizonemertea there appear to be differences in the development of muscular tissue in the oesophageal wall, sometimes the circumcesophageal blood-lacuna directly bathing this cellular coating, sometimes (e.g., Cerebratulus corrugatus) a special muscular investment of conspicuous development (PI. XIII. fig. 6, into) being again present together with very strong nerves (nv). The passage from the oesophagus to the sacculated intestine is more or less gradual, in the absence of any forward extension of the latter below the former, as was noticed for the Hoplonemertea. Macroscopic dissection enables us, nevertheless, to make a clear distinction between these two portions of the gut, although microscopic investigation of transverse sections shows that, histologically speaking, the passage is tolerably gradual. The cell layers of the posterior portion of the intestine have been more than once sufficiently described (see von Kennel (XVI), pi. xviii. fig. 11), and it is not always easy to show them to be provided with a nucleus or with cilia. Still I do not hesitate to declare that the whole of the intestine is ciliated, both the central passage and the lateral, generally sym- metrical caeca. But this ciliation is often rendered inconspicuous by the fact that the very elongated cells, composing the wall of this portion of the gut, are so overfilled with small spherical globules as not only to render the ciliation invisible, but even to efface the traces of the boundaries between the cells, so that in certain cases — both amongst Schizo- nemertea and Hoplonemertea — it would seem as if the intestinal wall were replaced by a compact mass of those globules enclosing the intestinal lumen between them. Similar phenomena were observed by Lang (XVIII) in the Polyclada, and have been described by other naturalists for other groups of Invertebrates. I will not here enter upon the question of the relation of this phenomenon to the process of intracellular digestion, which on a priori grounds may also be presumed to exist in the Nemertea, but will only add that the nuclei of these high and elongated cells may in favourable specimens be discerned, and are deeply situated, far away from the surface. Whilst strong vertical muscle-fibres pass from the dorsal to the ventral body- wall in a lamellar arrangement, thus constituting what I have termed in a former publication (V) the muscular dissepiments, placed alternately between the intestinal cseca, these cseca themselves are destitute of any special musculature. The muscular lamellae just men- tioned, together with the general body musculature, appear to be sufficient to bring about all the contractions in the intestinal wall needed for the progress along this channel of the food swallowed. The intestinal epithelium itself is thus directly implanted upon the gelatinous tissue, and this phenomenon is no less clear in the Schizonemertea thanm the Hoplonemertea (PL VIII. fig. 3; PI. IX. figs. 1-6; PI. XV. figs. 7, 10). Among the latter Pelagonemertcs is the most striking example of this, because of the preponderance of the gelatinous tissue. It has been already noticed in a preceding REPORT ON THE NEMERTEA. Ill paragraph tbat this jelly appears to be somewhat denser, and that at any rate it more strongly imbibes staining reagents, all round the circumference of the intestinal epithelium where this is implanted upon it. As to the oesophagus of the Hoplonemertea, I wish to observe that it is less thick and massive than that of Eupolia and the Schizonemertea, and more resembles the simple arrangement of Carinina. Its wall is generally only one cell-layer thick, and a distinct cuticula, as was noticed in Carinina, may also be often observed here. Fig. 1, Oe, of PI. X., representing a section of the oesophagus of Amphiporus marioni, gives a very fair representation of it. Outside the oesophageal epithelium there are indicated in this figure a layer of flattened cells which I at first expected to form an outer tunic to the oesophagus. Closer investigation revealed the presence of these cells in all the tissues — they may also be seen in the basement layer B of the same figure — and at the same time convinced me that these unicellular bodies are parasitic organisms. They infest all the tissues of their host, and are more abundantly heaped together just outside the oesophageal epithelium, where nutritive substances may be expected to be more plentiful. Curiously enough, they were also noticed in the smaller specimens of Amphiporus marioni. Similar cases of specimens of Nemertea infected with unicellular parasites were noticed by me on other occasions. There is another case amongst the Challenger material (see p. 48), but there the parasite is much larger than in Amphiporus marioni, and altogether differently constituted. It agrees with the former only in the fact of its presence in all parts of the tissues. These parasites differ from others which are found in the lumen of the intestine, and which have already been noticed by former observers. Another view of the Hoplonemertean oesophagus is given (in longitudinal section) in PI. XV. fig. 20. Here, too, the comparative thinness of the walls is conspicuous, and the connection with the posterior gut portion which stretches forwards under it, is clearly indicated. In Pelagonemertes the last mentioned phenomenon could not be observed. The sec- tions through the mouth and oesophagus were, however, not intact, because of the macro- scopic dissection to which the specimen had previously been subjected by M'Intosh. As to the caeca of the posterior body region little remains to be noticed, but that they are more regularly distributed as we approach the tail, i.e., the region where new caeca are being continually formed. Their mctameric and paired arrangement is here more evident than further forwards, where the degree in which they are filled with food par- ticles may be more or less different, and may thereby become the cause of a partial, but not very common, asymmetry. The innervation of the intestine was for the, greater part described in the paragraph treating of the nervous system. It may be remarked here, in addition, that in transverse sections of the foremost portions of the oesophagus it is very easy to detect the consider- 112 THE VOYAGE OF H.M.S. CHALLENGER. able nerve-stems, both of the vagus nerve and of the visceral nerves, by which this is brought about, and that in longitudinal sections also (PL XIV. fig. 5) the vagus nerve may often be followed uninterruptedly for a very considerable distance backwards, being applied upon the outer surface of the oesophagus, and only gradually dichotomising and sending delicate nerve-fibres amongst the oesophageal epithelium (cf. PL XVI. fig. 1). Of the two sources of innervation of the intestine, the one by the nerve-stem directly issuing from the brain-lobes (the so-called vagus nerve) is the most conspicuous, and can be demonstrated in all species from Carinina to the more specialised Hoplonemertea without exception. In Carinina it is represented in PL VI. fig. 1, Nv; in this species the visceral nerves (PL XVI. fig. 1, vi.sy ; PL XIV. figs. 3, 4, vi.n) have not been definitely demonstrated as yet in either of the two available fragments. These latter nerves are more easily detected in larger Schizonemertea, where the thick nerve-plexus is itself so much more conspicuous. Cerebratulus corrugatus especially answers this jiurpose. In the Hoplonemertea too the vagus is very evident, and already represented on Quatrefages' figures (XXVIII) ; in Drepan(ypliorus lankesteri I saw its principal stem running forwards towards the anterior oesophageal portion that passes under the brain. In this species we find numerous thin nerves, both from the lower brain-lobes and the lateral stems, further participating in the innervation of the oesophagus — a state of things which may be directly compared to the mixed innervation described above for Cere- bratulus corrugatus. I cannot as yet supply any definite statement regarding the innervation of the posterior region of the intestine. NEPHRIDIAL APPARATUS AND BLOOD-VASCULAR SYSTEM. Our knowledge of the nephriclia of the Nernertea is only of a comparatively recent date. Though discovered by Max Schultze as early as 1851 in a Tetrastemma (XXXIl), the observations of this naturalist concerning the Nernertean nephridia were for a long time wholly unjustifiably disregarded, and this general scepticism made me very careful in formulating any definite opinion, when I also discovered in Schizonemertea separate lumina (IV), which I could hardly account for in any other way than by regarding them as parts of a nephridial system. This was afterwards more emphatically done by von Kennel (XVI), to whom is due all the credit of having rediscovered the nephridia, and of having described their histo- logical appearance in several different genera. His results were later on confirmed by other authors (II, III, XI). I have afterwards observed and described (xil) a special modifica- tion of the nephridial system in Nernertea, in which an indubitable internal opening is REPORT ON THE NEMERTEA. 113 present. This was found by me in the nephridial system of Carinoma, and these nephridia thus establish a communication between the internal blood-spaces (archiccelome) and the exterior. In other genera, Carinella excepted, similar internal communications were, however, sought for in vain. A very notable and exhaustive contribution to our knowledge of the Nemertean nephridia was then furnished by Oudemans (XXVII), and the Challenger material, which was partly made subservient to that publication, only furnishes a few additional data here worth recording, most of the peculiarities having been already mentioned by Oudemans. I hold it to be one of the principal results arrived at by this author, that he definitely demonstrated the presence in numerous species of Nemertea of a very large number of exterior openings, connected by short transverse branches with the principal and longitudinal canals of the nephridial system, and that, at the same time, he noticed that these transverse canals were paired and opposite, and showed an arrangement which might most assuredly be compared to an incipient stage of metamery in the nephridial system. This fact is most distinctly borne out by the Schizonemertea. The Pakeonemertea (at any rate the CarinellidaB) and Hoplonemertea show an arrangement which presents different features, although, again, certain Hoplonemertea (Arnphiporus lactifloreus) answer very well to the Schizonemertean type. A very remarkable form of nephridia is found in Carinina grata. The two fragments of this species that form part of the Challenger collection both contain this important structure in toto, so that I am able to figure both transverse and longitudinal sections. These figures are brought together on PI. IV., and will first have to be discussed. Two portions may be distinguished in the nephridia of Carinina : in the first place, a glan- dular canalicular portion in which numerous delicate tubes appear to be closely applied together into a larger lobulate mass, which is situated right and left in the blood-space, not being freely suspended in it, but applied on one side against the muscular layers (figs. 5, 6), and on the other side, offering a free surface towards the cavity of this blood space. PI. IV. fig. 4 gives a more enlarged view (Nsp) of a longitudinal section through this portion of the nephridial apparatus, and, at the same time, enables one to judge of its extension, the whole of this glandular spongy portion having been reached in this section. It is, of course, also present in other sections, but does not stretch either further forwards or backwards. The furthest blind ends of the anterior caeca of the digestive canal are found to pene- trate between the lobes of this glandular mass (Jc, fig. 4). These lobes are, moreover, subdivided by bundles of muscle-fibres detached from the inner circular layer (fig. 5), against which the whole apparatus is so closely applied (fig. 4, Cm). The cells composing the glandular portion are filled with a granular protoplasm, and have very distinct large nuclei. However much I have looked out in my different sections for a definite opening by which the canalicular system here described might enter into communication with the (ZOOL. CHALL. EXP. PART LIV. 1887.) Hhll 15 114 THE VOYAGE OF H.M.S. CHALLENGER. blood-space (archicoelome) surrounding it, I did not succeed in demonstrating oue. Con- sidering that I had found such openings in Carinoma, and that Oudemans had afterwards demonstrated them in Carinella, I expected they would also be present in Carinina. For the present, however, the result of a very attentive search is that they are absent ; at least no opening is visible which can be said to prove a similar communication beyond all doubt and by which it might be demonstrated once for all. Questionable points of communication, which might eventually be interpreted as such, I have not allowed to influence my testimony, so that, for the present, I must answer the question in the negative. The case stands in a similar light with respect to the Schizonemertea and Hoplonemertea, as will be seen hereafter. If this communication with the blood-spaces is thus not demonstrated, that with the cavity of the second part of the nephridial system is subject to a much less degree of doubt ; and though I did not actually see the lumen at the point of communication, I did see the communication itself as represented in fig. 4. It is then seen that this second portion is distinguished from the one just mentioned by the presence of a spacious cavity. This cavity, which may be called the nephridial canal, is first found ventrally to the glandular spongy portion (figs. 4-6), but then gradually bends upwards as it passes further backwards along the animal, until it becomes a narrow channel with a very distinct and ciliated epithelium (figs. 1, 2), which passes at a very strongly inclined angle (fig. 1) through the successive muscular layers, then makes a very sharp bend towards the exterior surface, and traversing also the basement membrane and the integument, opens on the exterior. This exterior opening has not been figured, but is found in the sections following upon that which is represented in fig. 1. The two exterior openings of the nephridial system lie on the dorsal surface of the animal, and at the same time mark the point where the nephridial system reaches furthest backwards, the glandular portion of it stretching forwards towards the head. That this nephridial canal may, at all events in its proximal part, be more or less folded, is seen both in figs. 4 and 5, Nc, in the latter figure the lumen having an appearance as if it were doubled. The nephridial system of our second Challenger genus of Pakeonemertea, Eupolia, is, as was already known from Oudemans' researches, more comparable to that of the Schizo- nemertea than to that which has just been described in Carinina. It offers certain peculiarities which deserve special mention. Here, too, we may distinguish, as we may throughout the whole class of the Nemertea, longitudinal and principal nephridial ducts situated in the blood-spaces or enclosed by the gelatinous tissue (Hoplonemertea), and transverse or deferent ducts placed perpendicular to the foregoing, varying in number and somewhat in size, and brineing about a communication between the ducts before mentioned and the exterior. Of the aspect and situation of these two portions in a transverse section, a comparison of figs. 9 on PI. VI. and 3 on PL VII. may convince us. Each of them represents REPORT ON THE NEMERTEA. 115 about one-half of the thickness of the body-wall of the animal, and though somewhat differently magnified, the position of the nerve-stem and of the layer cm may easily guide us how to combine them. It is then seen that in PL VI. fig. 9 the longitudinal trunk, situated in the circum- oesophageal blood-space, is not only cut through, but that even more than one nephridial lumen (nep) appears in this section, showing that there is a doubling or at least a branching of the principal nephridial duct in this region. Internal openings of this system, funnels or anything comparable to them, were not detected by me in Eupolia giardii or any other Eupolia ; there were, however, very definite bends at right angles, piercing first the inner longitudinal and then the circular muscular layer, next the nervous plexus, and then arriving in the outer layer of longitudinal muscles. In this position the deferent branches of the nephridial system are seen in fig. 5, PI. VII., which, moreover, reveals the important fact that sometimes more than one of them is found at exactly the same level. I could not make out whether this duplication is in any way related to that of the longitudinal tube ; I can hardly conceive it to be so, the increase of the number of deferent vessels being also noticed in other cases, though hardly as a regular phenomenon. In Eupolia giardii, too, I find it to be exceptional in this sense, that most of the deferent ducts are single ; one section of another species ( Cere- bratidus truncatus) contains the double duct on both sides, making the exceptional phenomenon at the same time symmetrical. The number of deferent ducts observed in the Challenger specimen of Eupolia giardii is seven on the right and five on the left side, the latter being opposite to and symmetrically placed with five out of the seven on the right. These numbers, however, only apply to that portion of the trunk behind the head which was transversely cut, and belongs to the same series. I have not followed up the nephridial apparatus to its posterior portion, but we may feel assured, on the authority of Oudemans' researches, that it will on the whole answer to the diagram given on pi. i. fig. 11 of his treatise. One point deserving mention is that the first trace of the nephridial apparatus of this Eupolia is visible in about the ninetieth section behind the tip of the snout. When I add that the upper brain-lobes occupy sections 25 to 45, the forward extension of the nephridia can more easily be imagined. Nothing special can be said of the deferent or of the principal ducts but that their epithelium is distinctly one cell thick, nucleated, and unmistakably ciliated. Nor have I any special discoveries to record with respect to the nephridial system of any of the Challenger Schizonernertea, and may refer the reader for certain specific peculiarities, number and disposition of deferent ducts, &c, to the description of the species where I have embodied these details when they did not appear to have any general significance. I should like, however, to refer a little more fully to the conspicuous development at which the main, longitudinal canal of the nephridia has arrived in a certain species of Cerebratulus (Cerebratidus macroren), where its walls are unusually massive (PI. 116 THE VOYAGE OF H.M.S. CHALLENGER. XIII. figs. 7-9), although, when these walls are distended by the contents, their thickness correspondingly diminishes (PI. XIII. fig. 8). In this species, moreover, it could be observed that towards the anterior end of this massive nephridial canal it subdivides into smaller branches (PI. XIII. fig. 9), applied against the wall of the blood lacuna, and nearly escaping observation amongst the epithelial cells of these lacunaa. Here, again, no internal opening could be demonstrated ; whereas, at the posterior end, the nephridial canal, making a sharp curve, passes outwards above the nerve-stems. We thus observe differences in the nephridial system of the Schizonemertea, which may be classified under the following heads : — (a) The nephridial canal may be massive and single throughout the greater part of its course, only ramifying anteriorly (Cerebratulus macroren, &c). (b) It may be very copiously subdivided, every transverse section showing a consider- able number of lumina, the whole thus forming a kind of network, with certain principal longitudinal ducts, and being suspended against the wall of the circumcesophageal blood lacuna (Eupolia). (c) The meshes of this network may be situated in the region between the probo- scidian sheath and the longitudinal nerve-stems (most Cerebratuli), or they may stretch ventrally to these nerve-stems {Ev/polia). (<£) The ducts leading to the exterior may be one on each side, and these generally terminal (posteriorly). (e) There may be two on each side, and they may then communicate with the chief longitudinal canal about its middle. (/) They may be more numerous, often in various phases of distension, and arranged more or less metamerically. The histological characters are fairly uniform, the lumen is never excavated in a row of cells as is the case in so many Platyelminthes, in the Discophora, &c, but is always bounded by a certain number of cells in every section. Each of these cells has a large and distinct nucleus ; its protoplasm is granular, and the free surface, which is turned towards the lumen, is ciliated. We have now to examine the nephridia of the Challenger Hoplonemertea. In Oudemans' paper several points concerning them have been already noticed ; other details referring to the situation and the number of the deferent ducts were mentioned above when the species were described. Hence it may here be sufficient to call attention to the fact, made specially palpable by certain of the Challenger preparations, that whereas the nephridia of the Hoplone- mertea are no longer suspended in blood lacunae, but wholly surrounded by the gelatinous tissue, they are at the same time much more intricately coiled and ramified, as can be very easily seen from a comparison of fig. 1 (Nep), PI. X., with figs. 7, 9, PI. XIII. This involves, however, no important change in the histology; what we have noticed under this head in the Schizonemertea holds good for the Hoplonemertea as well. REPORT ON THE NEMERTEA. 117 Even less remains to be said when we consider the blood-spaces of the Challenger Nemertea, be they the closed longitudinal and the metameric transverse vessels of the Hoplonemertea and of the posterior body-region of the Schizonemertea, or the circum- cesophageal and circumrhynchodaeal lacunae of the latter, or the two spacious lateral longitudinal cavities of Carinina. All these points have been dwelt upon at length by Oudemans (XX VII), and the Challenger material furnishes a general confirmation of his results. Doubt cannot be any longer entertained that these spaces are all clothed by an epithelium, — at any rate by a special and continuous cellular coating, either applied against the muscular tissue when this surrounds (PI. IV. figs. 5, 6 ; PL XIII. fig. 6 ; PI. XIV. fig. 4), or traverses (PI. VII. fig. 10) the cavities or against the gelatinous tissue, when it is in this that the vessel takes its course (PI. X. figs. 1, 8). In the latter case the structure of the vessel is still more complicated, so that, as described by Oudemans, there is constantly a tubular, denser layer of the homogeneous tissue just outside the inner epithelial coating ; and outside this tubular layer, which might be termed the basement layer of the vessel, we find a second layer of generally more spherical cells, amongst which a layer of fibres, specially belonging to the blood-vessel, may be seen to make its appearance (PI. XV. fig. 1, dv). This same description holds good for the longitudinal vessel, as far as it takes its course inside the proboscidian sheath (PL X. fig. 8). The gradual narrowing of the circumcesophageal lacunar space into the two ventral vessels shows the passage of the epithelium of the one into that of the other very clearly. The difference in the distance along which the medio-dorsal vessel is enclosed in the proboscidian sheath was mentioned and figured by Oudemans (xxvil) for different species ; my own observations on the Challenger specimens fit into the same general outlines ; a few additional data concerning these points are contained in the systematic description of the species. GENERATIVE ORGANS. Certain not unimportant additions to our knowledge of the generative organs of the Nemertea are due to the Challenger specimens. Among these facts I wish successively to record : (a) The irregular distribution in certain species of very numerous generative sacs enclosed in the gelatinous tissue, and having each its separate external opening, which are consequently neither paired nor metameric. (b) The comparatively late period at which the definite external opening is formed, although long before that time the sac is characterised by a pointed projection reaching between the muscular tissue and foreshadowing the definite openings, dehiscence of the body- wall being certainly not the normal way of exit of the generative products. 118 THE VOYAGE OF H.M.S. CHALLENGER. As to the first point, it has been hitherto the current idea that in the Nemertea the generative sacs, alternating with the intestinal caeca, are paired and more or less meta- meric. This is no doubt the case in the very large majority, and relieves us from the duty of further describing the position of the generative sacs in JEupolia, the Schizonemertea and most Hoplonemertea. In other Hoplonemertea, however, we find a multiplicity of generative sacs in one transverse section (PI. IX. fig. 4) which cannot possibly be made to answer to the type just alluded to, and this irregularity reaches its extreme expression in Amphiporus moseleyi. The specimens of this species are literally full of sacs, which I was able to notice in their first stages, as well as in their later development and ripest stages. In each transverse section (PL IX. fig. 4) a great number of them may be seen; in horizontal sections (PI. IX. fig. 7) the same irregular multiplicity is met with. Externally I did not notice the openings, but it must be remarked that only in very ripe specimens are these distinctly present. Now this aberrant distribution is not wholly limited to Amphiporus moseleyi. There are even reasons for considering it as an arrangement which has been retained in this species, but which was common in the more primitive ancestral species of both the Hoplonemertea and Schizonemertea. At least I find it in a similar condition in Carinella annulata, with this difference, (1) that it is here only in the dorsal half, above the intestine, that the gene- rative products are found ; and (2) that the external openings are generally very distinctly visible as numerous irregular whitish dots on the dorsal surface of the animal, in the dark coloured space between the dorsal median white line and two successive transverse ones. Whether in Carinina similar conditions exist could not be verified, because the preserved fragments contain no generative products. At any rate the fact is fully established for one of the most primitive genera of Palaeonemertea, and this may justify our insisting upon its having an archaic significance. With the exception of Amphiporus moseleyi, Drepanophorus lankesteri is perhaps the only Hoplonemertean in which still another trace of it is preserved, at least if we may consider the fact that in this species (PL IX. fig. l) two ventral generative sacs are present on each side, as a reduced phase of the phenomenon, which we have found represented in Amphiporus moseleyi, instead of looking upon it as a secondary dupli- cation of primarily simple generative caeca. For myself, I should feel inclined to take the first view as rather the more probable. The other Hoplonemertea have not this peculiarity. The generative sacs are paired and metameric, and if in ripe specimens two of them are cut in one section (PI. IX. fig. 3) it is generally an indication that the section was not perpendicular. There is then also found an adequate distance between the external pores. This same figure clearly illustrates the fact that the generative products (in this in- stance ova) may attain a considerable development and closely approach the ripe condition before the generative sac itself communicates by a pore with the exterior, as indicated REPORT ON THE NEMERTEA. 119 in the second of the two propositions above enumerated. At the same time we find the spot where the pore will appear indicated by a pointed projection of the sac between the muscles, and it is a fact very worthy of notice that the more we approach this preformed outward duct, the less ripe are the ova (PI. XV. fig. 14). The same fact is noticed in many other Nemertea ; Pelagonemertes even shows traces of it (PI. VIII. fig, 8, ov.) In Carinella I found it persisting even after the definite pore is established, and the most plausible explanation appears to me to be this, that the deeper inwards the developing generative products are situated, the more they are surrounded by the gelatinous tissue, and the better their conditions of nutrition must be ; whereas those which are observed close to the duct, piercing the muscles, wdl only gradually increase in size, when in their turn similar favourable conditions are offered to them. The fact of the occurrence of definite preformed ducts without pores was observed by me in Amphiporus moseleyi, Amphipoms marioni, Drepanophorus lankesteri, and Drepanophorus serraticollis, and of several of these I also had ripe specimens, in which the fact could be determined that it is, indeed, in the places where these pointed projections are found that the pores afterwards appear (PL IX. figs. 5, 6). In this place I may mention that according to embryological observations (XIV) the generative sacs are primitively in connection, at least in the species investigated, with the epiblast by means of a strand of tissue which is not indicative of the ultimate duct. These strands, however, are situated on the other side of the nerve-stem, and thus are in no way identical with the projections which were here described and discussed as preced- ing the definite generative pores. The duct which pierces the muscles to afford a passage to the generative products is generally shortened as maturity advances, through the distension of the body by the ripening of ova or spermatozoa, and the consequent decrease in thickness of the mus- cular body-wall. Still, in Cerebratulus macroren this duct presented an uncommon feature, which must here be mentioned, and which is figured on PI. XV. fig. 19. After having pierced the circular muscular layer /3, it distends in the layer 7 to a second sac-like expansion, which in its turn communicates with the exterior by a small opening. This was not a local disturbance, but was met with in very numerous generative ducts, both on the right and left sides of the animal. In Eupolia and Schizonemertea the ripening sacs, developing in pairs between each successive pair of intestinal caeca, are often wedge-shaped, with the sharp edge turned outwards, and the broad end between the two intestinal caeca, where these emerge from the principal longitudinal cavity of the intestine (see woodcut fig. 6, p. 120). And, in addition to this, another fact is very marked in Eupolia giardii, viz., that dorsally and ventrally the generative casca become lobulated or arborescent, sending out short lobes of indented and irregular shape. It would, however, lead us too far, and hardly offer any additional interest, to discuss 120 THE VOYAGE OF H.M.S. CHALLENGER. the changes in shape which the developing generative sac may undergo ; accordingly we will now consider a few diverging peculiarities of the generative products themselves, which I was able to observe in the Challenger Nemertea. That they are developed from the epithelium of the generative sacs has long been known ; this is figured on PL XV. fig. 14, in a Hoplonemertean. It was particularly evident and very distinctly visible in Cerebratulus longifissus, where it was at the same time also demonstrable that there is a very sudden decrease in ripeness of the generative products close to the tail-end, where growth in length of the animal is going on, and where new generative sacs are being formed between the intestinal cseca. In this region all the most different stages of ripeness of the ova may be studied side by side in the same longitudinal section. The ova of Amphiporus marioni FlsGecUo7ofra' "cereLaud^ff (^ -^- **§• 15) are characterised by the presence, in addition to the laterat'^ffica t ^f^the'genitoi nucleus> °f a round or reniform body, which is stained dark red by ^rM^klgtmLXand'Tus^ picrocarmine, and but for this offers a certain analogy to the od-drop in fish eggs, being also more refractive than the nucleus, though not quite so highly as the latter. This paranucleus was already observed in the youngest eggs (PI. XV. fig. 15, a) ; at that time its relative size, when compared to that of the whole egg, was much more considerable. A second smaller specimen of Amphiporus marioni was distinguished by the same peculiarity, which may thus in certain cases help to identify the species, as I did not find the same feature in any other species of Hoplonemertea. Of the eggs of Pelagonemertes it has already been recorded that they are distinguished by an investment of follicle cells (PL VIII. fig. 11) ; the development of this could also be traced downwards to early stages, which were present in the same specimen (fig. 10) side by side with the riper stages. In the Schizonemertea two facts deserve mention, although their significance cannot well be discussed as long as fresh specimens are not available. The one is the pre- sence round the ripe eggs in Cerebratulus, sp. inc. (PL XV. fig. 18), and Drepanophorus serraticollis (PL XV. fig. 17) of a hyaline, apparently mucous layer, which surrounds each egg separately, and which is pressed into a polygonal shape when many ripe eggs are enclosed together in the same sac. The layer is comparatively thick. The second fact was observed in Cerebratulus parkeri, where the peripheral proto- plasm was much more darkly stained and more coarsely granular, all the eggs having thus the aspect as figured on PL XV. fig. 16. As to the spermatogenesis I have no new observations to record, spirit specimens alone rarely presenting favourable material for such researches. These phenomena have, moreover, been recently fully studied by Sabatier (XXVIIIa). Hermaphrodite specimens were not encountered by me in the Challenger collection. GENERAL CONSIDERATIONS. In venturing at the close of this Report on the Nemertea, collected by H.M.S. Challenger, to leave the region of demonstrated facts and actual observations, and to enter upon that of speculation and suggestion, I gladly avail myself of the permission for so doing granted to me by the editor, Mr. John Murray. I thought it necessary to ask for that permission, because general speculations on the ancestry of the Chordata hardly appeared to me to fit into the framework of these Eeports. My desire in this case to deviate from a rule which I held to be salutary, was due to the fact that of late these speculations have been conducted along very varying channels, an entirely new one having only very lately been opened by Bateson's important series of papers on Bal- anoglossus. An attempt to give more depth to one of these channels, and thus to lead into it the attention of a greater number of my fellow-workers, especially commended itself to me, since it was my conviction that the lines laid down by myself in former publications derived considerable support from the Challenger material, and were thus entitled to a renewed and full consideration. I would formulate the proposition, to the further development of which this chapter is to be devoted, as follows : — More than any other class of invertebrate animals, the Nemertea have preserved in their organisation traces of such features as must have been characteristic of those animal forms, by which a transition has been gradually brought about from the archi- ccelous Diploblastic {Coelenterate) type to those enteroccelous Triploblastica, that have afterwards developed into the Chordata {Urochorda, Hemichorda, Cephalochorda, and Vertebrata). It will be seen that this statement excludes the idea of any direct ancestral relations between Nemertea and Chordata. If any such relation were proposed, it might with good reason be asked — considering the very extensive variation which is met with amongst Nemertea — which species or which genus was more particularly pointed to. The question in itself condemns the proposition which leads to it. It will, moreover, be seen that this statement accepts the outcome of Bateson's researches and speculations, in so far as the points of agreement between Balanoglossus and Amphioxus are fully recognised. A provisional link between these two, and an (ZOOL. CHALL. EXP. PABT LIV. — 1887.) HLh 16 122 THE VOYAGE OF H.M.S. CHALLENGER. arrangement of Balanoglossus amongst the Chordata, appears to be quite as justifiable as the elevation of the Urochorda to their new dignity in zoological classification. There is, however, a great difference between looking at Balanoglossus as a low type amongst the Chordata (in which I fully agree with Bateson) and rejecting the signifi- cance of the Nemertean type as one of transition in the way above indicated. There is no doubt that the Nemertea represent a more primitive phase than the Enteropneusta (Hemichorda). They have no enteroccele, and they have no gill-slits ; but their nervous system shows certain unexpected analogies with that of the higher Chordata of more intrinsic value than those that obtain between Balanoglossus and the Chordata in general. Also for the important question, which is so vital in any considera- tion of the ancestry of the Vertebrates, viz., the origin of metameric segmentation, it appears to me that the Nemertea offer points very worthy of consideration. The question of the proboscis and its sheath, as comparable to hypophysis and notochord, was fully treated by me in another paper, and will here only be very briefly touched upon. In my opinion, this comparison is all the more forced upon us, now that in other respects (nervous system, &c.) new evidence of genetic relationship is here brought forward. The first point I wish to consider is that of metameric segmentation. It has been specially treated of late years by various authors of renown, with whom I do not wish to enter at this moment into any lengthy controversy, but will briefly state what may be gathered for the theory in general, from a careful consideration of the incipient metamery of the Nemertea. If we start from a more or less radiate ancestor of the earliest diploblastic type, in which neither a radial nor a serial repetition of organs or organ systems has yet come about, and which may indifferently be considered to resemble either a more flattened Trichoplax or a more spherical gastrula, we may assume that in the course of the development of other internal organs (towards the formation of which the secondary accumulation of cells between the two primary layers often so largely contributes) the radial symmetry may either be further accentuated or may be replaced by a tendency towards bilateral symmetry. In the latter case we are inclined to ascribe the first impulse towards this bilateral symmetry to a preference, which slowly establishes itself in the animal mechanism, for moving in one direction rather than in any other, i.e., for generally stretching forward, when moving about, one particular portion of the body. One impulse of this sort will suffice to lead us to understand, or rather to deduce, a very considerable number of consequences, which cannot fail to make their appearance under the influence of natural selection acting upon the organisms that have inherited this tendency in different degrees. Thus we may understand the narrowing and lengthening of an animal that moves in one direction in preference to any other ; and similarly the development in the nervous system of a centralisation not far away from the anterior extremity. REPORT ON THE NEMERTEA. 123 All this has already been stated by Balfour in clearer terms in bis Comparative Embryology (vol. ii. pp. 308, 311), where he describes the gradual steps by which a radiate medusa-like animal may have passed into a bilateral worm-like form, with two longitudinal nerve-stems, which are regarded by Balfour as the stretched nerve-ring of the Medusa. I fully endorse these views ; only, with respect to the nervous system, I hold it to be safer to leave out of comparison the already specialised nerve-ring of the Medusa, and rather to go back to a Ccelenterate nervous system as primitive as that of the Actinia, where the plexus, both of the epiblast and the hypoblast, with an increase in density in the region of the mouth and the tentacles, may be said to be the fair representative of one of the lowest starting points. In this the plexiform arrangement predominates. Now we find in all the lower invertebrates various though distinct nerve tracts that are being speciabsed in this plexiform nerve-tissue according to the modes of motion of the animal, and according to the general shape of the body. Thus in the Medusaa, which move about in the water by annular contractions of the lower portion of the bell-shaped body, one of the nerve-rings already alluded to was demonstrated by the Hertwigs to innervate the musculature by which this is brought about. In the Ctenophora the nerve-system is less satisfactorily known, but still Lang does not hesitate to bring them into genetic relationship with the Polyclada (XVIIl). Among the latter, Gunda, with its two longitudinal lateral stems, may be looked upon as an extreme term in this series. Another series may indeed be supposed to have derived longitudinal stems from a ring which became extended to form lateral cords, as the animal passed from the radial to the bilateral symmetry, in the way suggested by Balfour. Still, even in this case, a nerve-plexus may be expected to be co-existent with or to have preceded the nerve-ring. The longitudinal stems originating from the anterior thickenings of the plexus that innervate the sense organs and the tip of the head (specially sensitive in connection with the forwardly-directed movements of the body), would all the more probably be preserved and increase in development, as during this forward movement they form a right and a left centre for the reception of outward stimub. In the same way those of the radially-arranged stems of the Polyclada that are parallel to the longitudinal body-axis, and mark out right and left, are more strongly developed than the others, presumably on account of their importance in connection with the well-directed movements of the body in response to external agents. In the ancestral Mollusca, I think we may assume with great probabdity the presence of four longitudinal stems — two latero-dorsal, and two latero-ventral ones ; in the ances- tral forms of Annelids and Arthropods two, which have gradually coalesced ventrally, as was first suggested by Gegenbaur. Again, in Nematodes differently situated longitudinal 124 THE VOYAGE OF H.M.S. CHALLENGER. stems in what was originally a uniform plexus are preserved ; whereas, in ancestral Nemertea, two lateral longitudinal trunks in the plexus were undoubtedly characteristic features. That one medio-dorsal stem in this plexus, in which all the impressions made by out- ward agencies on both halves of the body might be concentrated, and from whence the corresponding movements might be regulated, will more fully answer the purpose than two lateral stems, however they may be united by an intervening plexus, is a priori probable, and would explain the first impulse towards the formation of such a longitu- dinal concentration in the uniform plexus. And when once such a dorso-median stem is present, in addition to two lateral ones, a struggle for supremacy, presided over by natural selection, may lead to a diminution of the lateral stems, and to an increase of the dorso-median one. This, in my opinion, as will be more fully developed below, was the case in the ances- tors of the Chordata, traces of this struggle and of the competing structural elements being duly preserved. If we suppose the bilateral symmetry to be estabhshed in one of the lower re- presentatives of the Metazoa, and the type to go on increasing in length in the course of generations ; then this increase, indeed, exposes it to very different, and perhaps more numerous dangers and enemies than would threaten it were the same bulk concentrated in a spherical or radial circumference. And if, even in the latter case, injuries to the specimen might prove fatal were it not provided with strong powers of regeneration (cf Star-fishes, Ophiurids, Crinoids, &c), stdl it needs no comment that, when bilateral symmetry and increase in length so considerably enlarges the surface which is open to attacks, and so enormously facilitates the rupture of the specimen, or the severing of parts by rapacious enemies preying upon it, siindar regenerative powers are none the less required to insure the persistence of the type. These dangers, continually threatening the existence of the specimens, and thus in- jurious to the species, counteracted as they are by regenerative processes (power of repro- duction of lost parts), I hold to be at the base ofcdl those cases of metamery in the animal kingdom which do not fall under the head of strobilation, the latter being comparatively rare with respect to the former. Incipient metamery, once established by this cause, may further differentiate in the most diverse directions (heteronomous segmentation, &c), even after the absolute cessation of the causes that in the first instance have provoked it. The explanation has, moreover, the advantage of being applicable to radial as well as to serial metamery. These propositions must now be more fully developed. The power of reproduction of lost parts comes, without doubt, under the general laws of formation and growth. We find it even in the lowest Protozoa. If the material which heredity has accumulated, either in such a unicellular being or in the egg of a Metazoon, and out of which the elements of REPORT ON" THE NEMERTEA. 125 the different organ systems will gradually develop, is hereditarily so disposed that a compensation for the loss of important parts is facilitated, this wdl, of course, constitute an advantage. Such a compensation may, e.g., be obtained where the generative pro- ducts are developed in very many separate centra, and not in one closed sac. Injury to the latter will, ceteris pa/ribua, be more fatal than an equivalent injury destroying one or more of the former. The same holds good for diffused instead of concentrated nervous centra, for the case of liver saccules to the intestine, instead of one compact liver, for numerous apertures and deferent ducts to the nephridial system instead of one, &c. And all this is still more evident when we have before us a long, bilaterally symmetri- cal animal, which is easily snapped in two. In this" case it must be of pre-eminent importance, that the remaining halves, which may in their turn be severed by the same cause into smaller parts, possess sufficient power of reproduction to repair the damage. Now, it cannot be doubted that an equal distribution of the important components of the organism (nervous centra, generative organs, nephridia, intestinal appendages, &c.) throughout the whole length of the animal meets this requirement. Any severed portion will then be provided with these more important parts, and will be more or less adapted for a separate and individual existence. The formation of a new mouth and of new brain-lobes in a fragment of this descrip- tion remains, of course, quite as wonderful and inexplicable as before, but still we cannot fail to see that such an arrangement as here indicated must somehow be beneficial to the species, and that we need not stop short with Bateson,1 when he says that " the repetition of various structures is one of the chief factors in the composition of animal forms The reason for their appearance is as yet unknown, and the laws that control and modify them are utterly obscured." Obscurity is not exchanged for broad daylight, but some- thing is gained when we can see that a growth of the principal organ-systems in separate and more or less independent batches, which in an elongated and bilaterally symmetrical animal insensibly passes into the phenomenon of incipient metamery, may be of the highest value for the persistence of the species. Now this is actually the way in which we find the important organ-systems distributed in the lower Nemertea. And out of this more irregular distribution a gradual metamery, in some incipient, in others more complete, is seen to evolve within the boundaries of the class. Even the nephridial system, in the primitive forms provided with only one opening to the exterior, participates in this tendency towards metamery, and acquires a greater number of apertures, serially arranged in pairs, thereby also tending towards a diminution of damage when artificial division into two takes place in the nephridial region. The meta- mery, the regular and serial repetition of parts, is thus seen to be of great importance in aiding towards repair after damage to a lengthened bilateral form, in the same way as the radial repetition of parts facilitates repair in the Echinodermata. In both cases the 1 Bateson, The Ancestry of the Chordata, Quart. Joum. Micr. Sci., vol. xxvi. pp. 545, 546, 1886. 126 THE VOYAGE OF H.M.S. CHALLENGED. destruction is only partial, the other homonodynamic portions temporarily ministering, thanks to their more independent relation to the injured region. When the faculty of repair of damage, occasioned by the severing of the animal into two or more portions, has in the course of generations become more and more complete, it can be readily understood that it becomes at the same time a defensive instead of being only a curative process. An animal that at the approach of danger can separate in two or more parts, each of them capable of reproducing an entire new animal, evades this danger very effectively by doing so ; whereas another that is attacked in the same way and does not possess this faculty, is laid hold of, shaken about, and wholly or partly swallowed. So in the Nemertea there is indeed a very strong faculty of spontaneous division combined with the faculty of repair1; and anybody who has observed a fresh and living Cerebratulus, with its extremely delicate sense of touch, commence to rupture into two, in preference at the spot where it was grasped with the forceps, cannot fail to see in this a defensive action. This mode of self-defence may in quite another respect be useful to the species, because at the same time it serves for propagation. Thus we see that the passage of this defensive process to that of reproduction by fission is so gradual, that it would be impossible to decide in every case what name should properly be applied to it. It cannot well be denied that in all probability ours is only a special case, in which the power of reproducing the species by a process of fission, reaching down as far as the unicellular ancestors, has come to be regulated by other motor forces than growth and — if it may not be called voluntary fission — still may be regarded as sudden and spontaneous fission, brought about by external influences, of a threatening nature to the further existence of the specimen. This regulation is no doubt a consequence of selection. Schizogony having once been established, it must have been further beneficial to the species, on the grounds that were developed above, that the internal organs should be present in multiple num- bers, and this having once come about it is easy to understand, that a regular, rigorously metamerous arrangement of this multiple material, still more fully answers the same purpose, and is gradually evolved under the influence of selection. Thus we may be said to be able to follow the appearance of metamery in a bilateral animal, along all the gradual steps by which it is evolved, and many of these steps have remained fixed and permanent in different Nemertean genera. The last system that will participate in this metamery is the muscular system, and a rash conclusion — such as is not rare in these days of ontogenetic fetichism — might lead to the rejection of the views concerning metamery here developed, on the consideration that it is exactly the metamery of the muscular system which appears first of all in the 1 Both M'Intosh and Barrois have observed and described very peculiar cases of repair in Nemertea, where the head, brain, side-organs, &c, were reproduced on a headless trunk-piece. These experiments are well worthy of careful repetition. It might be that only those fragments in which a portion of the oesophagus was retained were capable of repair of the head. REPORT ON THE NEMERTEA. 127 ontogenetic development of Vertebrates. I will not circumstantially refute this argu- ment, but will only remark that in Polygordius and other Chaetopods, which are representatives of a group of animals in which segmentation reaches such a very hio-h degree of perfection, the longitudinal muscular layer of the body-wall is as yet continuous in the adult, and not divided into metameric sections, as it is in certain Arthropods and in Vertebrates. Now let us consider contractions of the inner muscular layer a of the Nemertea, the only layer that is common to all of them, from Carinella to Cerebratulus and from Cephalotkrix to Pelagonemertes. This layer also corresponds with the longi- tudinal muscular layer just alluded to of other lower worms, such as Polygordius, and, as was noticed in our paragraph on the muscular system {cf. p. 72), its contraction is sometimes very distinct in favourable sections. We then see the contraction marked out as so many successive blocks of contracted, thickened fibres, separated by intervening parts of non-contracted fibrous tissue (PL XV. figs. 9, 10). The sections demonstrate that the phenomenon persists throughout the whole breadth of the animal, i.e., that successive rings of contractile tissue alternate with intervening rings in which no contraction is observed. This phenomenon is thus in a certain degree comparable to an arrangement in distinct myomeres. It is not unimportant that it was especially noticed in the fundamental muscular layer, and it may at the same time be remarked that it appears, from what I have as yet been able to observe myself, that the number of these rings in a given length of the animal, is the same, or a multiple of the number of intestinal caeca and transverse nerve-tracts in the plexus ; in other words, that the incipient metamery of the internal organs is in a definite relation to these phenomena — which might also deserve the name of incipient metamery — in the muscular layers. For the present the fact is, however, not yet definitely demonstrated that these successive blocks are indeed present as such in the living animal. The possibility is still open that they may be waves of contraction which have been fixed at the moment of the immersion of the animal in the preserving fluid. For this reason I will not lay any undue weight on this observation. The ideas concerning the origin of metamery here expressed, and advocated for several years in my university lectures, differ from those of Lang (XVIII) and Sedgwick,1 in so far as they do not recognise the primary importance of the so-called ccelomic sacs — the paired archenteric diverticula of Amphioxus — for the solution of this question. The question of the Vertebrate ccelome, so full of obscurities and difficulties, is purposely left out of consideration here, where the relation to archiccelous ancestral forms is discussed, and where an attempt is made to show that it is indeed probable that the impulse towards the establishment of metamery is due to forces for which the 1 A. Sedgwick, On the Origin of Metameric Segmentation, Quart. Joum. Micr. Hci., vol. xxiv. p. 43, 1884. 128 THE VOYAGE OF H.M.S. CHALLENGER. archenteron was not the only, nor perhaps the most important part of the organism to act upon. Still more different are they from those advocated by Perrier 1 and Cattaneo,2 who have adhered to and extended the idea already held by others, but by them most actively defended, " that the metamery of Arthropods, Vertebrates, and a great many Vermes, has originated out of the multiplication by transverse fission of very simple primitive worms which were not metamerous. The products of this transverse fission remaining connected together have then formed a chain of individuals, or a linear colony; later on the unity of the chain has become more definitely established, the single individuals at the same time becoming different both in form and in function, and the foremost individual thus becoming the head of the series. Each segment (metamere) thus repre- sents a reduced individual ; a metameric (segmented) animal is the result of the more or less complete fusion of single individuals into an individual of higher order." Emery, from whose paper 3 1 have translated the foregoing sentence, has very success- fully combated these propositions. This author, however, adheres to Lang's views in ascribing to the archenteric pouches, the " gemmation " as Emery calls it (loc. cit., p. 18) of the intestine, the most important and initial significance for the first origin of metamery, " the sexual glands and excretory canals being in relation to the intestinal diverticula," and following the lead. I have above explained why I cannot adhere to this argumen- tation, which brings the ccelome and the sacculated intestine too strongly into the fore- ground, and why I rather suppose incipient metamery to have been antecedent to either of these {e.g., Carinella). On the other hand, many views contained in Emery's im- portant paper coincide with my own. Thus he writes (loc. cit., p. 11), speaking of that interesting marine Triclade, Gunda segmentata : — " The metamery of Gunda is thus manifestly the consequence not of the ' symbiotic ' fusion of a colony of equivalent 'parts' (meridi), but of the 'autobiotic' differentiation and perfectioning of one ' part ' (meride);" and further on (p. 15) : — " When I consider the facility with which certain worms break into one or more pieces even spontaneously, it appears to me that this capacity for rupture may well have been the origin of the reproductive purpose of transverse scission in similar elongated organisms. The rupture, in the first instance accidental, could have contributed to the more rapid multiplication of the organism, being followed by the regeneration of the parts that were deficient in the separate fragments. This process of rupture might further have been so perfected that the spot best adapted for rupture, with a view to the best condition of the fragments, was prepared in advance. In the more perfect evolutional phases of the process, which are at the same time those that have till now been more carefully investigated, the new head is formed anteriorly to the rupture, or at least its essential parts are pre-established." 1 E. Perrier, Les colonies animales, Paris, 1881. 2 G. Cattaneo, Le colonie lineari e la morfologia dei Molluschi, Milano, 1883. 3 C. Emery, Colonie lineare e metameria, Kapoli, 1883. REPORT ON THE NEMERTEA. 129 My own views emphasise the presence of a peculiar process of development of the internal organs, running parallel to this predisposition for rupture in a particular spot — the spot which will correspond to the outwardly visible demarcation between the future segments. They thus go one step further — and, in my opinion, a very essential step — in the attempt to explain the origin of metamery in the lower Platyelminthcs, these bilateral descendants of radiate Coelenterata, and at the same time predecessors of both Chordata and Appendiculata.1 This view of the origin of metamery also affords an explanation for the very different degrees in which we find metamery or segmentation expressed in the different divisions of the animal kingdom. The incipient metamery which we have traced (and which we have pictured to ourselves as arising through natural selection amongst those forms, which, while developing in length, find metamery to be a protective peculiarity) imme- diately creates, by the fact of its existence, new and variable material for selection, again to be acted upon. And whilst metamery develops in one direction in one line of descendants, the other line brings to the foreground a different set of advantageous combinations, each of them again the stock of new and varied forms. In other words, metamery once established in its most primitive form, and intimately connected with spontaneous fission under the influence of external agents, has been of very great moment in the brineine about of new and endless variations of animal life. And it is irrational, when we have before us, say one of the lowest Vertebrata, in which nobody will deny the presence of distinct metameric segmentation, to conclude that this metamery must necessarily be in many respects reduced, and that in the ancestral forms it must have been far more complete, must have stretched forwards along the whole of the head, must have been more forcibly expressed than it is now — in all the cephalic nerves, in the nephridia, the gill-slits, &c. ; — all this on the presumption of the existence of an ancestor so completely and exemplarily segmental as to throw no light on the origin of segmentation and metamery, unless by the aid of Perrier's and Cattaneo's exaggerations. Such conclusions must, however, necessarily be made by those who follow Dohrn's and Semper's lead concerning the phylogeny of the Chordata. Bateson, in taking Balanoglossus as his starting-point, finds the acknowledged points of resemblance in the metamerous gill-slits, &c, and adds to them important data con- cerning the metamerous ccelomic diverticula. Still, for a general view on the origin of metamery, Balanoglossus offers no points that we do not find more strongly represented and more forcibly expressed in the Nemertea. It certainly deserves mention that long 1 Gegenbaur, in his Grunrlriss der Vergleichenden Anatomie(1878), hints at similar explanations to those advocated by Emery and myself, when he says (p. 64) : — " Die Metamerie .... lasst Zustande des Beginnes und der nicht aus- gefuhrten Beendigung mannichfach erkennen .... In dem Maasse als ein Metamer die Abhangigkeit vom Gesammt- organismus durch die Ausbildung seiner eigenen Organe aufgiebt emancipirt er sich vom Ganzen und gewinnt die Befahigung freier Existenz." Further on he speaks of incipient metamery as " eine stellenweise, fur den Organismus praktisch werdende Ausbildung" of the different organ systems. (ZOOL. CHALL. EXP. PART LIV. — 1887.) Hllh 17 130 THE VOYAGE OF H.M.S. CHALLENGE!!. before Bateson drew renewed attention to the numerous points of agreement between Balanoglossus and the Chordata, M'Intosh had done the same for Balanoglossus and the Nemertea, a separate paragraph of his monograph (XIX) being devoted to the discussion of these homologies. Sedgwick (loc. cit.) holds the unsegmented worms to be wholly " negligeable quantities," at any rate superfluous links in the chain that connects the Chordata with the antecedent Diploblastic stages. In my idea both these authors, valuable as certain of their suggestions are, have not been thoroughly aware of the necessity that, in all discussions on the origin of metameric segmentation, we must attempt to grasp at data that give a clue to the possible action of natural selection in the gradual evolution of metamery. This clue appears to me to be far more distinctly contained in the views here advocated than in the other hypotheses. It may further be remarked, now that we have once more alluded to Bateson's phylogeny of the Chordata, that even this naturalist does not feel justified in wholly rejecting the Nemertea from the Vertebrate pedigree. Whilst in the text of his article (loc. cit., p. 566) he does seem to prefer this negative alternative; still, in the subjoined diagram of the general relationships of Urochorda, Hemichorda, Cephalochorda, and Vertebrata, the Nemertea are introduced — with a point of interrogation, however- — as a side branch lower down on the common parent stock. Now, this being concordant with my owTn views of the Chordate phylogeny, — the point of interrogation excepted, — it is necessary to inquire why there is this discrepancy between Bateson's speculations in the body of his treatise and the hypothetical pedigree at the end of it. It appears to me that this is due to his hesitation (loc. cit., p. 555) in accepting the views hitherto entertained and advocated by myself as to the phylogenetic connection between the Nemertean and the Vertebrate nervous system. For this hesitation Bateson has good reasons, and while I appreciate the soundness of them, I hope in the remainder of this chapter to remove the reluctance of him and others to accept the phylogenetic significance of the Nemertea, thanks to new light that may be thrown on the evolution of the central nervous system of the Chordata by the observations above recorded on the nervous system of the Challenger Nemertea. It is to these speculations on the nervous system that we now have to turn our attention. As will be seen from the terminology introduced in the paragraph on the nervous system (p. 76), and as it is now time more fully to develop, I am inclined to attach considerable morphological importance to the arrangement of the different constituent parts of the nervous system in the Nemertea. In former publications (X, XI«) I have repeatedly insisted on the significance of certain points in the anatomy of the Nemertea, when considering the general question of the relationship of the Chordata to their REPORT ON THE NEMERTEA. 131 unknown invertebrate ancestors, and I have insisted not only on the possibility of the homology between the Nemertean proboscis and the hypophysis cerebri of the Verte- brates, but I have, even earlier still, attempted to show that the nerve-system of these two groups might be considered in a common light, as was first indicated by Harting in his Leerboek van de Grondbeginselen der Dierkunde of the year 1874. Further reference to the hypothesis here alluded to is found in Balfour's Monograph on the Elasmobraneh Fishes (pp. 170-172), in my own publications (IX, X), and in Balfour's Comparative Embryology1 (vol. ii. p. 258). I will not here enter upon this hypothesis more fully, but will briefly state that it attempted to consider the central nervous system of the Vertebrates as a possible median coalescence of two nerve-trunks, that were lateral in the primitive ancestors of the Vertebrates, in the same way as the coalesced ventral nerve-cord (Bauchmark) of Annelids and Arthropods may be considered with Gegenbaur as having arisen out of a double lateral trunk, which in certain, still more highly differen- tiated forms have fused ventro-medially. A strong argument against the first-mentioned hypothesis is the fact that the spinal cord ontogenetically always makes its appearance as a median unpaired plate or thicken- ing, a very faint trace of a possible double origin of this plate being hitherto only observable in one species of Amphibia, Triton tseniatus ; whereas in all other vertebrates, Amphioxus and the Cyclostomata not excepted, the unpaired origin is most evident. The bilateral symmetry of the full-grown brain and spinal cord is a much later feature, and can hardly be regarded as the expression of a primary coalescence of two separate halves to form a median whole. I am the more inclined to abandon this hypothesis, because I will attempt to show that we can establish phylogenetic comparisons between the Chordate and the Nemertean nervous system on a much more simple basis ; comparisons which at the same time cover a far more extensive ground than did those of Harting, Balfour, and myself, which I have just alluded to. Since in the nervous plexus of all the Nemertea a median longitudinal tract, some- times of comparatively large size, has now been detected, since even in the Hoplo- nemertea, where the plexus has disappeared, the same medio-dorsal nerve-tract has in most cases been preserved, and, finally, since from this dorso-median stem metameric and paired nerve-tracks may be seen to emerge in Palseonemertea and Schizonemertea, we must inquire, in how far the direct comparison of this medio-dorsal nerve-stem with a primitive spinal cord may be said to hold good. In order to do this we must first consider the relation of this stem; to which we have given the name of medullary nerve or medulla, to the rest of the nervous system, more 1 It may here be remarked that Balfour has omitted to mention that Harting was the first to bring forward this hypothesis: it is well to be reminded of this when Beard, Bateson, and others similarly ignore this claim to priority of my venerated predecessor. 132 THE VOYAGE OF H.M.S. CHALLENGER. especially the brain-lobes.1 In a former publication (IX), where the medullary nerve was for the first time noticed and described as the proboscidian-sheath-nerve, I traced its origin to the dorsal commissure between the two lateral halves of the brain (he. cit., pi. i. fig. 1). Thanks to certain very favourable specimens in the Challenger collection, I have now been able to add new data to this statement. Sections through the brain of Cerebratulus macroren, Cerebratulus corrugatus and Gerebratulus angusticeps (PL XII. figs. 1, 7, 8; PL XIII. fig. 1) show that the condition of things is indeed less simple than this original statement would imply, — that the medullary nerve is not an eminently fibrous cord springing at right angles from the eminently fibrous upper brain-commissure, but that the nerve-tissue constituting the foremost and uppermost portions of the upper brain-lobes spreads out over a far more considerable surface than the fibrous tract which is known as the dorsal commissure. This expansion of nerve-tissue, in which the cellular elements are no less conspicuous than the fibrous, is posteriorly directly continuous with the plexus above described, laterally with the brain-lobes, anteriorly with the cephalic nerves springing from these lobes. It attains its fullest development just before and behind the region where a transverse bundle of fibres uniting the fibrous core of the lateral brain-lobes forms the well-known dorsal brain-commissure. This commis- sure is a transverse fibrous tract forming part of a more extensive nerve-plate. To this expansion of nerve-tissue the presence of nerve-cells gives a more primitive, at any rate a less specialised, character. These nerve-cells and nerve-fibres are directly continuous with those of the medullary nerve and (backwards) with those of the nerve-plexus, of which this nerve is only the median longitudinal thickening. There is even more reason to look upon the fibres of this medullary nerve as a tract of the general fibrous stroma not necessarily connected with the fibres of the brain-commissure. In other cases a more direct continuity between the commissural and the medullary nerve-fibres was however observed. In order clearly to understand the relative importance of the different parts of the nervous system here noticed, the primitive Palseonemertea offer the best starting-point. Thus in Carinella we find the brain-lobes not yet separated into distinct upper and lower lobes, nor do we find a posterior lobe (side-organ). The brain is a double lateral and anterior thickening in the nerve-plexus, situated like it and like the lateral nerve- stems outside the muscular body-wall in the deeper strata of the integument. The only difference between the medio-dorsal medullary nerve in this species and the lateral nerves with their anterior enlargements (the brain-lobes) is its position and its greater tenuity (PL XVI. fig. 1), which, however, does not prevent its being very clearly observable in every transverse section (PL XL figs. 3, 4). Its connection with the brain-commissure was already described (IX, p. 25), and figured by me (he. cit., pi. iii. fig. 31). It must, 1 In the course of these considerations a certain amount of repetition of facts already noticed in the paragraph on the nervous system cannot well be avoided. REPORT ON THE NEMERTEA. 133 however, be remarked that in these most primitive Pak-eonemertea, the anterior dorsal brain-commissure is less significant than in the Schizonemertea, and hardly anything else than the foremost of those numerous transverse metameric tracts in the plexus (dvr, PI. XVI fig. 1) which connect the lateral stems with the medullary nerve (dorsally) and with each other (ventrally). These important metameric nerve-pairs are most distinctly observed in Carinelku Here, as in the Schizonemertea, the medullary nerve is also continued forwards in front of the brain thickenings. This continuation sometimes shows a short bend just on the level of the commissure, so that both the medullary nerve and its anterior continuation may be seen in one section. This explains at the same time the arrange- ment traced on PL XII. fig. 8. Posteriorly the medullary nerve can be followed down to the hindmost extremity of the body. In Eupolia and the Schizonemertea the arrange- ment remains the same, the metamery of the transverse stems is perhaps more clearly expressed, the whole plexus and the longitudinal stems are no longer in the integument, but between the muscular layers. Still the whole of the nervous system also answers to the general type as represented in the diagrammatic fig. 1 on PI. XVI. We have now seen enough of it to understand that a comparison with the central apparatus of the Vertebrate nervous system cannot indeed be called a strained comparison. On the contrary, the comparison is much less artificial than was the one which Balfour was inclined to adopt, and which, as noted above, rendered necessary the acceptance of the phylogenetic development of the Vertebrate medulla out of a double cord. And so I do not hesitate to proclaim the medullary nerve of the Nemertea to be a very important link in the phylogenetic chain, of which the Vertebrate spinal cord is the outcome. Like the Nemertean medulla, the Vertebrate spinal cord is median, unpaired, and composed of nerve-cells and nerve-fibres ; like the Nemertean medulla, it is a thicken- ing in a nervous plexus, originally wholly epiblastic, of which, among Vertebrates, the Amphibian embryos offer such a striking example. This instructive and suggestive case was known to Remak and Strieker (as the " Nervenschicht" of the frog embryo), it was more carefully studied and elaborately described by Goette (his " Grundschicht" of the epiblast, in his Entwickelungsgeschichte der Unke), and it has been again recently brought into the foreground by Baldwin Spencer, in his latest paper on the subject.1 The latter author compares the Amphibian plexus with that of Pakeonemertea and Schizo- nemertea (loc. tit., p. 134), as had already been done before him by my friend Professor Ray Lankester, with whose suggestion I at that time (1880) did not yet venture fully to associate myself. The numerous data that have since been accumulated for a direct comparison of Nemertea with lower Vertebrates appear, however, now to fully justify that comparison 1 Baldwin Spencer, Some notes on the early Development of Rana temporaria, Quart. Journ. Micr. Sci., vol. xxv. Suppl., p. 123, 1885. 134 THE VOYAGE OF H.M.S. CHALLENGER. which was first expressed in a footnote to a former paper of mine (X, p. 438). There can hardly be any doubt as to the existence, consequent upon natural selection, of a constant tendency in the different component parts of living organisms towards simplifi- cation and increased efficiency (Roux's Kampf der Theile im Organismus). This fact enables us to understand the gradual supremacy of the median cord in the Nemertean plexus over the two lateral ones. It seems as if it were mathematically demonstrable that for the delicate adjustment of the impressions from the exterior to the co-ordinated movements thereby occasioned, one longitudinal central stem in bilateral, lengthened animals, would be more efficacious than two lateral ones. And if we ask if, at the final stage of this struggle for supremacy between three longitudinal stems, any remnants of the lateral cords are yet detectable in the Vertebrate embryos, perhaps even in the adults, I am inclined to answer in the affirmative. Here I must be allowed to insert a reference to the three figures on PL XVI., which will facilitate the exposition of the further conse- quences of the hypothesis I am here developing. Fig. 1 represents the chief points in the nervous system of the Nemertea. The brain-lobes are simple lateral swellings of the longitudinal stems, as in Carinella ; plexus, medulla and transverse stems, together with brain-lobes and lateral stems, may be considered as forming part of the integument (cf. Carinina). A double innervation of the respiratory portion of the intestine is indicated ; one due to visceral branches (vi.sy) springing from the plexus (or from its transverse tracts), the other to the more specialised nerve (v), which has above been indicated as the Nemertean vagus nerve. The plexus and its innumerable radial fibres, both sensory and motor, are not indicated in this figure, nor are the nerve-stems which, when present (PI. XIV. fig. 2), pass from the lateral stems directly to the integument, This figure must now be compared with the two others. Of these, PI. XVI. fig. 2, diagrammatically represents the chief points that may be considered as characteristic of the nervous system of a lower Vertebrate, in which the dorsal and ventral roots of the spinal nerves (dr and vr) are as yet separate nerve-tracts, in which the sympathetic nerve system is as yet only represented by visceral branches given off by these dorsal roots {vi.sy), and in which the polymerous character of a primitive vagus ( Vag) is established. PL XVI. fig. 3, stands for Amphioxus, as far as we know its nervous system, more particularly through the researches of Rohon and others. It differs from the foregoing by the absence of a distinct brain swelling, and of a polymerous vagus. A number of spinal nerves are considered as homologues to the vagus of Vertebrates by Rohon. The commissural connections between the successive spinal nerves form a plexus, which is peripherally even much more elaborate, according to Rohon's figures. This plexus does not reveal the presence of any distinct lateral longitudinal nerve, nor any ganglia of spinal or cephalic nerves. The latter (en) may be said to be three in number. Visceral branches (vi.sy) are given off by the dorsal nerves (dr). The ventral ones, springing from the lower edge of the medulla, are here represented as shorter stems (i<)-). REPORT ON THE NEMERTEA. 135 The opposite half of the system, seen in transparent perspective, as given in the two other figures, is purposely omitted here, because of the asymmetry of Amphioxus in this respect. Now a glance at these figures will convince us that the situation of the Nemertean medullary nerve in its plexus, and with its set of transverse nerves, is directly compar- able to the Vertebrate medulla and spinal nerves. The nerve-plexus filling up the intervening spaces in Nemertea is present as a transitory structure in Amphibian embryos. The ulterior appearance of an anterior enlargement forming the Vertebrate brain ; the higher complication attained by the brain and spinal cord when its mass increases, but not its dorsal expansion, by the appearance of medullary ridges ; and the formation of a neural canal by infolding of the neural plate, all these are important developmental facts which do not in any way weaken the grounds for comparison of the two structures. They may be looked upon as adaptations to the much more considerable efficiency and concentration that is gradually attained by the central nervous system as we ascend higher in the scale of the animal kingdom. The fact that the neural ridge in so many Vertebrata precedes the appearance of the spinal nerves, and is inserted along the top of the folds that bend together to form the neural tube, may be thus interpreted, that during the phylogenetic process of infolding, the transverse nerve-tracts (dorsal spinal roots) remain attached in the same way to the medio-dorsal collecting trunk as they did in the ancestral forms, and are dragged upwards by the infolding process. The ventral roots must be phylo- genetically linked to the plexus as well ; inasmuch as the musculature originally lies inwards of the nervous plexus, their deeper situation is not surprising. In the points hitherto enumerated there is entire coincidence between Amphioxus and the other Vertebrata, as far as their comparability with the Nemertean diagram goes. Another point of coincidence is the way in which the foremost portion of the intestinal canal and adjacent blood-vessels are innervated by visceral nerve-stems, indicated in all the three diagrams by vi.sy. The claims to validity of the comparison here made between the spinal nerves of the Chordata and the transverse stems of the Nemertea, should be again insisted on, now that the researches of Rohon,1 Freud,2 Schneider, Eansom, and d'Arcy Thompson 3 1 V. Rohon, Untersuchungen iiber Amphioxus lanceolatns, Dcnkschr. d. k. Akad.d. Wiss. Wien, Bd. xlv. 2 S. Freud, Ueber Spinalganglien und Riiekenmark des Petromyzon (Sitzungsb. math.-nat. cl. k. Akad. Wiss. Wien, Bd. lxxviii.,' Abth. 3, 1878). This author says (p.154) : " Ich kann wenigstensvon den letzten Wurzeln des Oaudalmarks sagen dass ihre Selbstiindigkeit so gross ist, dass man von vorderen und hinteren Spinalnerven, anstatt von vorderen und hinteren Wurzeln reden konnte"; and Wiedersheim in the 2d edition of his Lehrbuch der Vergleichenden Anatomie (p. 321): "Vieles spricht dai'iir dass die Vorfahren der heutigen Wirbelthiere getrmnfe dorsale und ventrale Nervenwurzeln besessen haben miissen." 3 W. R. Ransom and d'Arcy W. Thompson, On the spinal and visceral nerves of Cyclostomata, Zool. Anzciger, No. 227, July 1886. 13G VOYAGE OF H.M.S. CHALLENGER. have established for the lower Chordata (Cephalochorda and Cyclostomata) that the typical chordate spinal nerve is not originally provided with a double root, but that this double root appears to have arisen by the coalescence of what were primitively in the groups just mentioned, separate and alternating dorsal and ventral nerve-tracts. With these so much simpler spinal nerves the transverse nerve-stems of the Nemertea undoubtedly offer points of comparison. These Nemertean nerves specially differ from the Vertebrate spinal nerves in two respects : (1) they give off nerve-fibres in different directions, which are probably motor as well as sensory and visceral, according to the different organ systems they terminate in ; and (2) they go round ventrally, each of them forming a loop all round the body. As to the first point of difference just alluded to, it is the expression of a low and primitive degree of differentiation, and when a step forwards is made, differentiation of labour will tend to develop certain tracts more particularly con- taining sensory and visceral nerve-fibres, which are more especially directed towards the epithelia (the primitive dorsal or posterior roots), and others more particularly containing motor nerve-fibres, and more especially directed inwards towards the muscles (the primitive ventral or anterior roots), because the musculature, as was already mentioned, is originally situated internally to the nervous system. For the present we can only hold it to be established that the fibres of these three categories are blended in the Nemertean plexus, without being able to determine in how far the specialisation therein observed, of the appearance of transverse metameric nerves, may at the same time be accompanied by a commencement of differentiation, such as has just been alluded to. We may, in other words, not yet assume that among these metameric stems there is already a tendency to an alternation between such as have sensory and visceral, and such as have motor predispositions. Only in a few cases may we be justified in saying that certain nerve-tracts belonging to the Nemertean peripheral s)Tstem are more especially sensory (PI. XIV. fig. 2) or visceral (PL XIV. fig. 4),- and these no doubt offer important analogies in their situation and connections to similar nerve-tracts of the Vertebrata. The second point of difference, viz., the continuity in the ventral median line of the transverse tracts of the Nemertea, is no doubt a consequence (a) of their origin in a perfectly continuous plexus, (b) of the cylindrical arrangement of the muscular layers, which in most cases are uninterrupted both in the dorsal and in the ventral median line. It is all the more important to notice, that more especially in the primitive Carinellidse, the tendency is very marked towards a scission of this muscular body-wall into a right and a left half (n, Innervation for the proboscis. an, Cephalic nerves. Figs. 1-9. Eupolia giardii, n. sp. Reconstructions of the brain-lobes from a series of sections. Figs 1, 2, 3, i, 8 represent the outward aspect, viewed from different sides. Figs. 5, 6, 7, 9 represent the fibrous core ; the extension of the ganglion cells enclosing this core being indicated by faint outlines. Figs. 1, 5. Seen from above. Figs. 2, 6. Seen from below. Figs. 4, 7. Seen laterally and outwardly. Fig. 3. Seen from behind. Figs. 8, 9. Seen laterally and inwardly (after section of the commissure and removal of the left half). In figs. 5 and 6 the course of the ciliated canal is specially indicated ; in the former figure by a red outline. The Voyage of H.M.S Ctiallenyei Nemerle;i I'l V dcrn vcm. jgy sl \ LN . M**i . dcm '^•B**.^ L.V " ' M-t,-*1 ** cc SL f SlBfl^ff; - dcm vein cc tttm£t''^ f' ■ ■ ,% ■■ - E U P OLI A. LithPW.vrl WeijeHJLr PLATE VI. PLATE VI. Fig. 1. Carinina grata, n. gen. et sp. The brain as situated in the deeper strata of the integument, seen in horizontal section (cf. woodcut fig. 5, p. 81). Br, Br, the anterior and the posterior brain-lobes. The inner fibrous core of these lobes white ; in the anterior lobe traversed by radial fibres, in the posterior one containing the ciliated canal ec, that opens out in fig. 2 into the cephalic groove Cg ; Ngc, the nerve-cells of the brain-lobes (stretching outwards as far as the red tint is applied in the figure); . VL" ■ "•: ■* « .■'-",■' J^.7. brv >.' ijlljjM % Fivt 3. ■ Fiq. ¥. u Fiy.5. Q-lrn »oc •'.V H,^ ^ o Im \? m fee ■•■> ■.,., ■■■■■ * : WMi Bet fPtlis B 1 ^.. tf ^ 7 & ^-£S£^' . ;MM ft , . ; 0 }■■'. ■■■ ■'■ ■■■■ es^-h : * . J %-'^W Fia. W Fur. 9. — o.ln U Bet 14< - .... rfs . ■ . . . Flq.il. vim. Jm, "CI ■.-■ ■ ■ ' ■;•" ' • • ■ • ■ ■ ■ ■ - :■, - ■ . '•.••■ ■ ■ ,.. ' . ■ Hulrecht et de G-root ''del FM,;,iih'r« E U P 0 L 1 A PLATE VIII. PLATE VIII. Figs. 1, 2. Pterosoma plana, Less. After Lesson, supposed by Moseley to represent a species of Pelago- nemertes. Fit;, i. Seen from above. Fig. 2. Seen from below. Fig. 3. Pelagonemertes rollestoni, II. N. M. A transverse section, supposed to be not quite vertical to the longitudinal body axis, so that to the right one of the digestive, to the left one of the generative cseca was touched in the section. Ep, external epithelial layer; B, basement membrane; LM, layer of longitudinal muscles; J, intestine and its branching diverticula; N, longitudinal nerve-stem; bl, two lateral blood-vessels; Gc, external opening of the generative caecum; P.S.C, the cavity of the proboscidian sheath; Ps, the wall of that cavity. All the internal organs are surrounded by and imbedded in a wholly continuous gelatinous ground substance, in which a few cells and numerous fibres can be detected, and which has more strongly imbibed the staining reagent in the immediate vicinity of the different organs and tissues that traverse it. Fig. 4. Pelagonemertes rollestoni, H. N. M. A horizontal aspect of the muscular layers of the body- wall, from a preparation made of the fresh animal by Professor Moseley. Im, the longitudinal muscles ; em, the sparse circular muscular fibres, external to the foregoing ; dr, granular patches, eventually glandular structures; ct, the gelatinous connective tissue visible between the muscle fibres. Fig. 5. Pelagonemertes rollestoni, H. N. M. A transverse section of what most probably corresponds to one of the granular patches of fig. 4, furnishing arguments for looking upon the latter as glandular. " A central lumen (or fibre1?), with nuclei surrounding it can be detected. Fig. C. Pelagonemertes rollestoni, H. N. M. The longitudinal nerve-stem iv", in transverse section. A branch n gives off smaller nerve-twigs n' and n". Nuclei are imbedded in the fibrous nerve substance. The gelatinous connective tissue is more deeply stained all round, and at ff has a distinctly fibrous appearance. Fig. 7. Pelagonemertes rollestoni, H. N. M. The posterior region of the proboscis, in transverse section. II, the longitudinal muscular fibres, externally invested by a basement mem- brane ; E, the epithelium, of which no details could be made out. Fig. 8. Pelagonemertes rollestoni, H. N. M. An empty genital caecum, in transverse section. In the lower narrower part the epithelium is high and very distinct and a couple of epi- thelium-cells are becoming converted into ova, ov. N, the nerve-stem; hi, the blood- vessel ; Gc, the empty cavity ; etc, connective tissue cells in the gelatinous ground- sabstance ; /, fibrous tracks in the same. Fig. 9. Pelagont mertes rollestoni, H. N. M. Isolated transversely striated cells from the wall of the cavity, Gc, in fig. 8. Fig. 10. Pelagonemertes rollestoni, H. N. M. A young ovum. Fig. 11. Pelagonemertes rollestoni, H. N. M. A larger ovum, surrounded by its follicle cells (cf. fig. 3). No distinct nucleolus, but numerous chromatic granules inside the nucleus. Fig. 12. Pelagonemertes rollestoni, H. N. M. Portion of the proboscidian sheath, in transverse section, under higher power. B, the inner homogeneous limiting membrane; Im, longitudinal; cm, circular muscular layer; ct, the outer sheet of the gelatinous tissue immediately applied against the muscles and again more deeply stained. Fig. 13. Pelagom mertes rollestoni, H. N. M. Portion of the outer layers of the body-wall, under still higher power. The external cellular epithelium has not been represented. B, the thick basement membrane, below this epithelium, traversed radially by apparent glandular ducts («•/'. figs. 4, 5); B', a deeper portion of the same, less affected by the staining solution, and carrying connective tissue cells; cm, isolated circular, LM, thicker bundles of longitu- dinal muscular fibres ; dr, probable glandular ducts in the gelatinous ground-substance penetrating through the muscles to the exterior. The Voyage of HM. S."Challenger" Fig. 1. Nemertea Pl.VIII Fiq. Z. Fiq f- HutrecM et de Groot del FHuft, lift' E Jin' PELAGONEMERTES PLATE IX. (ZOOL. CHALL. EXP. PART LIV. 1887.) Hhll. PLATE IX. J+B. Integument and basement membrane. M. Body musculature. Ot. Gelatinous tissue. Prs. Proboscidian sheath. LN. Lateral nerve-stem. Int. Intestine. Gs. Genital sacs, in many cases not yet in open com- munication with the exterior. Prs.div. Lateral diverticula of proboscidian sheath. br. Blood-vessel. Figs. 1-6. Diagrams of different Hoplonemertea to elucidate (1) the relative extent of integument, muscular body-wall and internal gelatinous tissue; (2) the situation of the genital glands and their respective openings to the exterior. Fig. 1. Drepannphorus lankesteri, n. sp. Middle of the body, diverticula of proboscidian sheath included in the section. Fig. 2. Drepanophorus lankesteri, n. sp. Towards the extremity of the tail, between two pairs of diverticula of the proboscidian sheath. Fig. 3. Amphiporus marioni, n. sp. Fig. 4. Amphiporus moseleyi, n. sp. Numerous genital sacs, both dorsal and ventral, contained in one transverse section. Fig. 5. Drepanophorus serraticollis, Hubr. Fig. 6. Drepannphorus serraticollis, Hubr. In a further advanced stage of ripeness of the genital products (with distinct genital openings to the exterior); the diver- ticula of the proboscidian sheath not touched in this section. In all these sections the longitudinal blood-vessels are indicated, the median one below the proboscidian sheath, the lateral ones close to the lateral nerve-stems. In Amphiporus moseleyi the lateral nerve-stems are seen to lie above, in Drepano- phorus, below the intestinal caeca. Fig. 7. Amphiporus moseleyi, n. sp. Diagram of a horizontal section through the body. The intestine and its caeca are dark grey, the generative casca light grey. The latter are seen to be very numerous and in no way regularly or metamerically arranged. Fig. 8. Amphiporus moseleyi, n. sp. A specimen with flattened ventral surface. Natural size. A whitish line from the tip of the snout backwards along the lateral margin marks the exterior openings of the lateral glands (cf. PI. XV. figs. 11, 12). Fig. 9. Amphiporus moseleyi, n. sp. Head, seen from below. Longitudinal slit both for the intestine and the proboscis ; terminal transverse sensory groove and lateral bent grooves into which the cavity of the posterior brain lobe opens. Fig. 10. Drepanophorus lankesteri, n. sp. Diagram of the principal features of the nervous system. B, brain-lobes; p. Br, posterior brain-lobes (side organs) with their cavity opening to the exterior at e.o; CepJi.ne, numerous cephalic nerves to the tip of the head, the eyes, &c., only a few of them are here indicated in outline ; Pm, nerves for innervation of the pro- boscis (they are more numerous than is here indicated); Va, outline of vagus nerve springing from the lower brain-lobes and running forwards towards the oesophagus. The latter passes beneath the brain-lobes and their double commissure, but above the ladder commissures (Comm), which metamerically unite the longitudinal nerve-stems (LN) below the intestine ; pe.ne, peripheral nerves springing from these nerve-stems. Pig 11. Amphiporus moseleyi, n. sp. Stylet and accessory darts. A and h,V, the central stylet and its two accessory sacs, in position ; b", bases of two accessory darts, viewed laterally (figure to the left), and perspectively (figure to the right). The Voyage of HM.S.'Xhalleiiger" Fy. J. Nemertea. PI IX- COTT1771., FRulKbtVEdm1 ' AMPHIPORUS. D REPA NOPHO RUS PLATE X. PLATE X. Fig. 1. AmpJiiporus marioni, n. sp. Part of a transverse section through the oesophageal region, indi- cating the relative importance of the musculature with respect to the integument and the gelatinous tissue. Prs, proboscidian sheath in contracted state, the outline only partly worked up to show the interlacement of muscular fibres, the basement layer and the internal epithelium ; dv, dorsal blood-vessel surrounded by gelatinous tissue, as is the proboscidian sheath, and Oe, the resophagus with its cellular epithelium. Outside of this there is a layer of darker fusiform bodies, which are most probably unicellular parasitic organisms. Gt, the gelatinous tissue ; a, the longitudinal, B, the circular muscular coat ; at ne, a bundle of nerve- fibres spreads between a and B after having traversed the musculature between two of the larger bundles of a ; B, thick basement' membrane with only a few nuclei ; J, integument ; LX, longitudinal nerve ; Nep, nephridial tubules ; Nep.d, part of the communicating duct of the nephridia with the exterior ; inc, peculiar crystalloid inclosures of a greenish colour, irregularly distributed in the gelatinous tissue. Fig. 2. Drepanophorus lankesteri, n. sp. Part of a section through the tail end. J, the integument, with an outer layer of sense-cells and supporting cells ; granular glands leading to the exterior, a layer of nuclei and one of deep lying cells with fainter nuclei ; B, basement membrane with imbedded nuclei ; a, the longitudinal, B, the circular muscular layer ; Gt, the gelatinous tissue with nuclei and cells inclosed, certain of these being on their way of transformation into fibres ; LN, longitudinal nerve-stem. Fig. 3. AmpJiiporus moseleyi, n. sp. Horizontal section through the tip of the snout. Pr, proboscis and its musculature passing into and being the direct continuation of a, the longitudinal muscular layer ; Prs, the proboscidian sheath; Sp.Pr, muscular arrangement in the wall of the rhynchodseum constituting a sphincter; Rh, external opening of the rhynchodseum, which is internally clothed b}' a layer of cells very gradually passing into the proboscidian epithelium, and externally into ./, the integument ; B, basement membrane ; B, circular muscular layer, obliquely cut ; Gt, gelatinous tissue ; Br, Br', left and right brain-lobes ; E, eyes ; gls, lateral glands, continued along both sides of the animal (c/ PI. XV. figs. 11, 12). Fig. 4. Drepanophorus lankesteri, n. sp. Proboscidian sheath with diverticula (div.Prs). Prs.ep, epi- thelium of the sheath, separated by folded basement tissue from the muscular wall. Fig. 5. Drepanophorus serraticollis, Hubr. Proboscidian sheath with thicker muscular walls and thinner walled diverticula (div. Pr*.). Fig. 6. Eupolia giardii, n. sp. The boundary line between integument and body musculature. Jdvl, deeper layer of vacuolated cells of the integument ; B, reduced and folded primary basement membrane ; y.vl, outer longitudinal muscular layer with large vacuolated cells and rare muscle fibres. The vacuolated cells have larger nuclei than those of the integument (cf. PL VII. fig. 5). Fig. 7. Cerebratulies sp. iuc. (medullatus ?). Transverse section of medio-dorsal region. J, integument; b, secondary basement membrane ; y, outer, a, inner longitudinal, B, circular muscular layer ; Prs.ep, epithelium of the proboscidian sheath ; PrsN, longitudinal nerve of proboscidian sheath ; ne, nervous layer with median medullary thickening. Figs. 8, 9. Cerebratulus macroren, n. sp. The proboscidian sheath wall and intestinal epithelium wholly (fig. 9) and half (fig. 8) distended by the proboscis, dv, dorsal blood-vessel (in fig. 8 still within the proboscidian sheath cavity); Prs.ep, the epithelium of the proboscidian sheath, supported by a homogeneous membrane, b, and surrounded by muscular layers m.Prs. Oe.ep, epithelium of the oesophagus. The Yoyage of H.M. S." Challenger: fur 1 i «y-*- cb.v.Prs. ■■■ ■■ ' Wn ■ frs.y>. '■-•■ .• -?1 fn ! ^1- r M - . . tLLv.Prs. Fig 6 Gt. ■ ■■'■■ . t f/ M •%J ^PS -::3S&5 ':: .'--#■■: " ... /K* ^ ■ ■ " .£. ■ l I . v,^'"- Hubrech iV5 ' • § C AR IN E L L A. C ARIN 0 M A. C EPHALOT HR IX CE REBRAT U L U S fy *7- *Bf-&~- 3 F AMPHIPORUS Hntrechl ! F Hutli.Lilh'EJin1 PLATE XII. PLATE XII. Fig. 1. Cerebratulus macroren, n. sp. Transverse section through a part of the brain-lobes of the New Zealand specimen. The medullary nerve (m) is still in connection with the brain-lobes (hi); prs, the proboscis, passing between the nervous ring formed by the brain-lobes and their commissures. Superiorly two bundles of longitudinal muscle-fibres are enclosed by the brain-tissue. Fig. 2. Cerebratulus macroren, n. sp. Transverse section of the body-wall in the region of the lateral nerve-stem (Japanese specimen). a./?.y, the three muscular layers, the latter with the deeper integumentary glands embedded between the muscle-bundles (r/. PL XL figs. 10, 11); pi, the nerve plexus ; b, the secondary basement membrane with the thin layers of integumentary muscles just below it ; J, the cellular integument ; N, the lateral nerve-stem, into the fibrous core of which part of the fibres of the plexus may be seen to be interwoven. Fig. 3. Cerebratulus corrugatus, M'Int. Transverse section of the medullary nerve (m) at a point where a pair of transverse stems (of. PI. XIV. fig. 1) merge into it. n, these nerve-stems (being thickened tracts in the plexus, ^>Z); pf, nerve fibres radially emerging from these tracts and having the significance of sensory or motor peripheral twigs ; ./?, radial fibres (not nervous) piercing the plexus (cf. PI. XIII. figs. 3, 4, rf). Fig. 4. Cerebratulus corrugatus, MTnt. Transverse section of the medullary nerve of another specimen. Lettering as in the foregoing figure. Fig. 5. Drepanqphorus lankesteri, n. sp. Transverse section of the lateral nerve stem (iV), with two peripheral nerves, pf, springing from it. The fibres of the latter partly emerge from the ganglion cells, partly from the fibrous core of the lateral stem. Fig. 6. Drepanopliorus serraticollis, Hubr. A portion of a transverse section through the proboscis. /./', the longitudinal muscle-fibres, in two strata, between which lies the nerve-plexus np, of which In is one of the longitudinal thickenings (nerve-stems) ; ox, the outer ; i.c, the inner layer of circular fibres ; b, transparent basement tissue. Fig. 7. Cerebratulus macroren, n. sp. Transverse section of the superior nervous connection between the two brain-lobes (Japanese specimen). Nerve cellular elements predominate in this region whence the medullary nerve is continued both backwards and forwards, nl, the homo- geneous layer forming an investment to the nerve-tissue ; pre, the proboscidian sheath in outline. Fig. 8. Cerebratulus macroren, n. sp. A few sections further backwards. The medullary nerve m is here more distinct, its anterior continuation m' being on the point of coalescing with it. Lettering as in fig. 7. pr, outline of the proboscis ; Br, fibrous core of the brain-lobe. Fig. 9. Cerebratulus medullatus, n. sp. Transverse section of the medullary nerve (m). pi, the plexus ; )3 and a, the circular and longitudinal muscular layers ; (inner circular muscular fibres are seen to form the outer layer of the proboscidian sheath) ; prn, the proboscidian sheath nerve, receiving delicate fibres from the medullary nerve and situated just above the pro- boscidian sheath musculature. Fig. 10. Cerebratulus medullatus, n. sp. Integument and muscular body wall. Lettering as in fig. 2. dgl, the deeper glands of the integument enclosed in the musculature. The Voyage of H.M.S "Challenger." Nemertea.Pl.XII. HuWcto et 6e Groot del C E R E B R AT U L U 5 . DREPANOPHORU S T Huth, Lift* Elm1 PI ATE XITT. (ZOOL. CHALL. EXP. PAET LIV. — 1887.) — Hllh. PLATE XIII. Fig. 1. Cerebratulus corrugatus, M'lnt. Transverse section through part of the superior brain-lobes and dorsal brain commissure. Part of the fibrous core (h) is represented; its continuation into the fibres of the commissure (f.cn) being distinctly noticed on the left side. On the right this connection is only visible a few sections further. The bulk of the brain-lobe is seen to con- sist of nerve-cells of different sizes (n.cl); from these n cephalic nerve (era) is seen to emerge on the left, whereas this cellular coating is also continued over the commissure, and there forms the starting point (?») for the longitudinal dorso-median medullary nerve. Fig. 2. Cerebratulus corrugatus, M'lnt. Horizontal section of the same specimen through the medullary nerve m. In this section only a portion of the cylindrical nerve-plexus (ftp/), of which the medullary nerve is only the median thickening, is visible. The transverse nerve-tracts tr.n, paired and metamerical, being thicker than the plexus, are cut along a more considerable surface and thus stand out as separate nerves. The openings in the plexus give passage to radial, transverse, contractile fibres, as is specially indicated in fig. 4, rf. Fig. 3. Cerebratulus corrugatus, M'lnt. Portion of a horizontal section through the ventral extension of the nerve-plexus, n, the nerve-tissue of the plexus with nuclei of nerve-cells, and bundles of radial fibres rf. piercing it. Fig. 4. Cerebratulus corrugatus, MTnt. Portion of the same section as fig. 2, more highly magnified. /;, the tissue of the nerve-plexus with delicate fibres and distinct nuclei of nerve-cells, also visible in the medullary nerve m ; //, the bundles of radial, contractile fibres. Fig. 5. Cerebratulus eomigatus, M'lnt. Ventral view of the head and long mouth with rugose lips of large specimen. Natural size. Fig. 6. Cerebratulus corrugatus, M'lnt. Padial strip out of a transverse section in the oesophageal region. Jsg, the outer glandular layer of the integument (with " Schleimstabchenzellen "); b, the secondary basement membrane below this ; Jlem, the longitudinal and circular muscle-fibres peculiar to the integument; Jdg, the deep glandular layer just below these; B, the primary basement membrane of the integument radially traversed by contractile fibres (rf) and nerve-bundles (ra) going towards the integument ; y, the outer longitudinal muscles ; pi, the nerve-plexus with fibres and cells, pierced by the radial contractile bundles ;/; /?, the circular muscular layer; a, the inner longitudinal muscular layer; cos, the circumcesophageal intercommunicating blood-lacuna?, clothed by a cellular endothelium ; into, the oesophageal musculature ; wo, nerve-tissue in the oesophageal wall ; pap, an (Esophageal papilla with deeper glandular and superficial ciliated cell-layers. Figs. 7, 8. Cerebratulus macroren, n. sp. Two sections through the principal nephridial duct of a specimen from Japan. Contracted (fig. 7) and distended (fig. 8). Fig. 9. Cerebratulus macroren, n. sp. Anterior tubuliferous proliferation of the nephridia. a, /J, mus- cular layers as in fig. 6 ; Nep, nephridian tubules applied against the wall of the circum- cesophageal blood-lacuna ; Oe, outline of oesophagus wall. Tin- Voyage ,,l 11 M S Challenger tfe 'tea i I SHI . , A A n.'cl . m : - :. J!rn, ir u trii ■ ^ o* : e? ~ r; -ft si 'Si V -V. >„ f® 1 C II ?P fe-iW' -'•'.'•: ';^X O ;';:',•' ■; £& 1 & -<^>- :-J fie ff mto II ulir ('l J e fa. del. CEREBRATULUS. PLATE XIY. PLATE XIV. Fig 1. Cerebratulus angusticeps, n. sp. The medullary nerve (m) and the transverse paired metamerical nerve-stems (trn). Drawn with the camera. 1 mm. on the same scale is indicated on the left of the figure. Fig 2 Cerebratulus corrugatus, M'Int. Transverse section of the lateral nerve-steni (IN) and body-wall a 8, y, the muscular layers ; B, the primary hasement membrane, and J, the integument. in outline (cf. PL XIII. fig. 6); rf, radial fibres piercing the muscular layers ; n and pn nerve-tracts; of which one (pn) is seen to emerge directly from the lateral nerve-stem and to innervate the sensory layers of the integument ; npl, nerve-plexus. Fig. 3. Cerebratulus corrugatus, M'Int. Diagrammatic figure of a transverse section in the mouth-region to show the innervation of the oesophagus and blood-lacunas. OeE , the oesophageal epithelium; Prs, the proboscidian sheath; co.l, the circumcesophageal blood-lacuna. In the nerve-plexus, which is indicated by a black line, m marks the medullary nerve, and nl, the lateral nerve-stems; vi.n, the visceral branches springing from the plexus; n.ra, the vagus ramifications transversely cut and intermixed with the branches vi.n. Fig 4. Cerebratulus corrugatus, M'Int. Part of the same section as fig. 3, more considerably enlarged. Lettering as in fig. 3; npl, nerve-plexus ; rf, radial, not nervous fibres ; a./3.y, the muscular layers; co.l, the blood-lacunas with cellular coating. The oesophageal epithelium (oe.E) is not represented in its whole thickness. Fig 5. Cerebratulus parkeri, n. sp. Diagram, drawn with camera, of the inferior brain-lobes, fir and the vagus nerve (va) springing from them, the two stems being united by a series of trans- verse commissures co close to their origin. The brain-lobes merge into the lateral nerve- stem (N) ■ the ventral metameric connections of the latter are indicated by dotted lines, v.tr.n, those of the brain-lobes being lettered c.tr.n; Oe, outline of the oesophageal epithelium. Fig 6 Cerebratulus angusticeps, n. sp. Horizontal section of the left upper and posterior brain-lobes (Br and PBr) • col, cl, the free blood-lacuna between the bram, the proboscidian sheath, and the oesophagus, the boundaries of these' two latter being only indicated by dark shading ; a, 8, muscular layers; In, anterior portion of lateral nerve-stem, the ganglion cells of the cellular investment being here cut; cc, the outwardly directed portion of the ciliated canal inside the posterior brain-lobes; cc', cc", the deeper portion of the same being cut in two places, because of the S-shaped curve which this canal makes inside the brain-lobe. The relatively large size of the posterior brain-lobe, and the preponderance of fibrous nerve-tissue in the upper lobe, are indicated in this figure. Figs 7, 8. Cerebratulus mmroren, n. sp. Two transverse sections (a few sections apart) through the posterior brain-lobe audits investment of large granular glandular cells (g.cl). m, the outer membranous investment of the lobe; n.cl, the ordinary ganglion cells with fibrous core in the centre of the lobe ; cc, the ciliated canal with an epithelium of its own, more distinct in fig. 8 than in fig. 7. Fig. 9. DrepanopTiorus lankesteri, n. sp. Part of a horizontal section through the upper bram-lohe, with interior fibrous cote (cf.) and outer layer of ganglion cells, n.cl ; a few of which aie much larger (nCl). Fi. 10 Drepanophorus lankesteri, n. sp. Part of a horizontal section through the posterior brain-lobe^ °' Bm the outer investment of the lobe ; n.cl, the nerve-ceUs ; cc, the ciliated canal with its Epithelium ; gel, the granular glandular cells ; the anterior ones pouring their contents into the lumen of the ciliated canal. Fig 11. Cerebratulus macroren, n. sp. Part of a transverse section through the ^showing the .ciliated canal to the posterior brain-lobe in its course from that lobe (PBr) to the exterior gel, the granular glandular ceUs ; cc, the lumen of the canal coated by an epithelium of varied histological character as we pass outwards towards E, the external layer of the integument. At gl. a sort of ring-shaped cushion of peculiar cells may be said to embrace the canal (cf., PL VI. figs. 1-3). Just behind this the epithelial cells are very closely set and provided with elongated nuclei, further backwards they are seen to pass without any sudden transition into those lining the canal, inside the brain-lobe. TlueYoyage of II M S ChaHeneeT Nemertea PJ XTV p'ii:---'-' :■■■: S3*.-'-" ■ ■ s -. ■ iJiJ>; e i Ut et le troot del. ErtlX / CEREBRATULUS DREPA N OPHORUS PLATE XV. (ZOOL. CHAI.I.. EXP. PART nv. — 1887.) — Hhli. Fig. i Fig. 5 Fig. 6 Fig. 7 PLATE XV. Fig. 1. Gerebratulus longifissus, n. sp. Transverse section through the proboscidian sheath, a, the inner longitudinal muscular layer of the body wall; pr.sn, the proboscidian sheath nerve; Pr.S, the lumen of the proboscidian sheath; e, its inner epithelial lining; b, the homo- geneous basement membrane of this epithelium, traversed by radial fibres and separating the aforesaid epithelium from the muscular layers (an inner longitudinal, and an outer circular) of the proboscidian sheath ; Gt, the gelatinous tissue between the proboscidian sheath, the intestine and the body wall ; dv, the dorsal blood-vessel. Fig. 2. Gerebratulus inwrnr'-ii, n. sp. Transverse section of the proboscis, b, the ensheathing mem- brane ; a and ■/, the two layers of longitudinal, y3', the layer of circular muscular fibres, the latter connected crossways at two diametrically opposite points with the membrane b ; n.pl, the nerve-plexus between /3' and y; E, the very high inner epithelium with smaller (c) and larger (C) batteries of nematocysts. Fig. 3. ( 'erebratulus macroren, n. sp. Transverse section of the proboscis much further backwards ; the proboscis is thinner, the muscular fibres (a) only longitudinal, the epithelium (E) flattened, the nervous plexus replaced by two longitudinal nerve-stems N. Gerebratulus angusticeps, n. sp. Head, a, from below ; b, side view. Gerebratulus parkeri, n. sp. Head, a, side view ; b, from below. x l\. Gerebratulus sp. inc. Two fragments from Japan. x 2. a, dorsal view ; b, ventral view. Gerebratulus sp. inc. Transverse section of one of the fragments of fig. 6. /, the integu- ment ; a, /3, and y, the much reduced muscular layers ; N, the lateral nerve-stems ; Int., the intestinal canal and caeca, suspended in the gelatinous ground tissue Gt ; Gt', the same tissue expanded between the very sparse external fibres of the muscular layer y ; PrS, the probos- cidian sheath; bv, the blood-vessels. Fig. 8. Gerebratulus sp. inc. Lateral view of the head of another specimen from Japan (Kobe). Fig. 9. Gerebratulus longifissus, n. sp. Longitudinal section of the tail end. /, the integument ; Lit, the intestinal epithelium ; a, /?, y, the muscular layers ; npl, the nerve plexus. This figure is meant to show the general aspect of the muscular layer a, more highly magnified in fig. 10. Fig. 10. Gerebratulus longifissus, n. sp. The muscle-fibres of the layer a are seen to present a more con- siderable diameter in successive vertical planes, forming rings round the intestine and alternating with others in which this increase in diameter is not visible. Nuclei are also more numerous in the former portions than in the latter ; Ep, the intestinal epithelium ; Gt, the gelatinous tissue between this and the body musculature, with bundles of radial fibres. Fig. 11. Amphiporus moseleyi, n. sp. Transverse section of the lateral region with glandular cavities ('//), arranged along the lateral line between the dorsal and ventral musculature a, piercing the circular muscular layer /? at e, and suspended by the gelatinous tissue gt. I, integu- ment ; B, basement membrane. Fig, 12. Amphiporus moseleyi, n. sp. One of the glandular caeca, more considerably magnified, with granular epithelium and flattened nuclei exteriorly, the latter belonging to the surrounding gelatinous tissue. Drepanqphorus lanlcesteri, n. sp. A section of similar glandular caeca in the dorso-median region of the head. Amphiporus marioni, n. sp. A very unripe generative cascum, not yet opening to the exterior. /3, Circular, a, longitudinal muscular layer. The caecum reaches down into the gelatinous tissue, gt. Amphiporus marioni, n. sp. Ova in different stages of development. Gerebratulus parkeri, n. sp. An ovum. Drepanophorus snraticollis, Hubr. An ovum. Cerebratulus sp. inc. A batch of ova from a ripe generative csecurn of a specimen from Station 321. The ova are surrounded by a hyaline membrane (which has erroneously been dotted by the lithographer) and pressed into polygonal shapes. Fig. 19. Cerebratulus macroren, n. sp. Part of a transverse section, in outline. Prs, The cavity of the proboscidian sheath ; 7, integument ; y, /3, the muscular layers ; N, the longitudinal nerve-stem ; ge.c, the generative caecum ; o, its exterior opening. The duct to this opening, where it pierces the layer y, is considerably extended in bulk. Fig. 20. Amphiporus moseleyi, n. sp. A longitudinal section through the greater part of the compressed oesophagus Oe, and subjacent intestinal casca (eo). Fig. 13. Fig. 14. Fig. 15. Fig. 16. Fig. 17. Fig. 18. 'I'll.- Yovade of II M S Clialleu£< NemeTtea IM .XV tig I 1 ■ ■ Fig \ Fi'j 5, Fid -2 ': Fid (1 Fid 3 K ',TPi '' :- fi ibi 1 1 It i '■ Gtool lei CERE8IUTUUJS AMPHIPORUS Lith.D.W.G PLATE XYI. PLATE XVI. Three diagrammatic figures for the comparison of the nerve system of the Nemertea, of the Vertebrata, and of the Cephalochorda. Fig. 1. The chief points in the nervous system of the Nemertea. In, hi, the lateral nerve-stems with their anterior swellings, the brain-lobes, Lg. The latter are connected in front by a strong ring-shaped commissure, the former by much thinner but unmistakable commissures, dvr, metamerically placed both dorsally and ventrally in a continuous plexus of nerve-tissue, that ensheathes the body. This plexus is no further indicated in the figure. It moreover carries a longitudinal medio-dorsal nerve-tract m, which is also continued forwards, in front of the brain commissure, with which it is connected. Into this medullary nerve m the transverse tracts dvr converge. From the latter (sensu strictiori from the plexus) radial nerve twigs of sensory and motor significance can be traced both in centripetal and in centrifugal direction, but these are not indicated in the figure. Other similar nerve twigs, indicated by vi.sy, innervate in a similar way the wall of the oesophagus and of the blood-lacuna surrounding it (c/. PI. XIV. figs. 3, 4). The innervation of the oesophageal wall is, moreover, brought about by a paired nerve, v, springing from the brain-lobes, and not strictly separate in its peripheral distribution from that of the nerves vi.sy. en, nerves to the tip of the head ; M, mouth. Fig. 2. The chief points in the nervous system of the Vertebrata. Instead of the medullary nerve we find here in the corresponding situation the medulla spinalis m and its anterior enlargement, the brain. The spinal nerves are represented in their primitive condition, i.e., with the dorsal (sensory) roots dr, and the ventral (motor) roots vr, not yet connected into a spinal nerve of higher order. Anteriorly is represented the union of more than one metarneric transverse root into a complex polymerous vagus nerve, Vag, motor and sensory branches of whicli are represented as innervating the branchial pouches of the fore gut (the latter not specially indicated in the figure). From the vagus ganglion also springs hi, the ramus lateralis nervi vagi or nervus lateralis ; hi', the left nervus lateralis. This nerve is continued f'jrwards into other ganglionic swellings Lg, representing the series of ganglia of the cephalic nerves. The visceral branches springing from the dorsal roots and inner- vating vegetative organs are indicated by vi.sy : they are supposed not yet to have united into a longitudinal "sympathetic nerve." In front the region innervated by them overlaps and fuses with that of the intestinal branch v. of the vagus, and of other cephalic nerves. M, mouth. Fig. 3. The chief points in the nervous system of Amphioxus. The medulla spinalis m without any considerable anterior enlargement ; en, pairs of cephalic nerves ; dr, dorsal roots ; vr, ventral roots ; the former springing from the medulla at a higher level than the latter. The dorsal roots are connected together in a plexiform arrangement; they give off visceral branches (vi.sy). M, mouth. The Voyage of II M S Challenger Neiuertea 1JI XVI 1} VT ill tii.2 Ar vi- lli' VI *y Fii ,lr VI" VI IV HuWei THE VOYAGE OF H.M.S. CHALLENGES. ZOOLOGY. REPORT on the Cumacea collected by H.M.S. Challenger during the years 1873-76. By Professor G. 0. Sars, of the University of Christiania. INTRODUCTION. The Cumacea obtained during the voyage of H.M.S. Challenger form a comparatively small part of the rich collection of Crustacea procured during the Expedition, the number of species amounting to only fifteen. They belong, however, to eight different genera, and thus most of the leading types of this interesting group are in reality found to be represented in the collection. The specimens were, as a rule, preserved in the usual manner in strong spirits ; but a few forms had been immediately mounted in Canada balsam on glass slides, and the examination of these could not, of course, be so satisfactorily made as in the case of the others. Four of the species were detected by myself in examining some small samples of dredged material kindly sent to me by Mr. John Murray. The specimens were collected partly in the Atlantic, partly in the Pacific, and partly in the Southern Ocean, especially at Kerguelen Island, where no less than five of the species were procured. The greatest depth from which Cumacea were procured was at Station 246, in the North Pacific (2050 fathoms), and the least depth at Flinders Passage (7 fathoms). Two of the species were, moreover, taken at the very surface of the sea, off Samboangan, Philippine Islands. In describing the species I have made use of the terminology adopted in my Report on the Schizopoda of the Challenger collection,1 and have endeavoured to give a more detailed account of such forms as represent separate groups or families of the order. 1 Zool. Chall. Exp., vol. xiii. pt. xxxvii. p. 3. (ZOOL. CHALL. EXP. — PART LV. — 1886.) Iii 1 n 2 THE VOYAGE OF H.M.S. CHALLENGER. The eight genera represented in the collection may be referred to seven different families, viz., Cumidse, Vaunthonipsoniidae, Larnpropidae, Leuconidae, Diastylidee, Cumel- lidas and Campylaspidae, all of which are also represented in the Northern Ocean. Of the fifteen species contained in the collection, one only, Diastylis stygia, has been previously described ; all the others I regard as new, although some of them seem to be very nearly related to earlier known species. Principal Works on Cumacea. Kroyer, H., Fire nye Arter af Slsegten Cuma. Nat. Hist. Tidssb:, Bd. iii, 1841. Om Cuniaemes Faniilie. Nat. Hist. Tidsskr., Ny Easkke, Ed. ii., 1846. Goodsie, H., Description of the Genus Curna and of two New Genera nearly allied to it. Edin. New Phil. Journ., 1843. Spence Bate, On the British Diastylidas. Ann. and Mag. Nat. Hid., vol. xvii., 1856. Beneden, P. J. v., Eecherches sur la faune littorale du Belgique — Crustaees. Mem. Acad. Sci. Bruxelles, t. xxxiii., 1861. Sars, G. O., Om den aberrante Krebsdyrgruppe Cumacea og dens nordiskc Arter. Forhandl. Vidensk. Selsk. Christiana, 1864. ■ Beskrivelse af de paa Fregatten Josephine's Expedition fundne Cumaceer. K. Svensk. Vetensk. Abo/. Handl., Bd. ix., 1871. Beskrivelse af syv nye Cumaceer fra Vestindien og det Sydatlantiske Hav. K. Svensk. Vetensk. Akad. Handl, Bd. xi., 1873. — Middelhavets Cumaceer. Arehiv f. Mathem. og Naturvid., vols, iii., iv., 1878-79. Dohrn, A., Ueber den Bau und die'Entwickelung der Cumaceen. Zeitschr. d. gesammt. Naturwiss., Bd. v., 1870. Norman, A.M., Crustacea Cumacea of the "Lightning," " Porcupine," and " Valorous " Expeditions. Ann, and Mag. Nat. Hid., ser. 5, vol. iii., 1879. Boas, J. E. V., Studien uber die Verwandtschaftsbeziehungen der Malacostraken. Morphol. Jahrh., Bd. viii., 1883. MORPHOLOGY OF THE CUMACEA. The morphological relationship of the Cumacea to the other groups of Crustacea is rather obscure, and hence their systematic position has been much disputed. At first they were even considered by some of the greatest authorities, as Mdne-Edwards, Dana, and L. Agassiz, as merely larval forms of higher Crustacea. Both Kroyer and Goodsir, however, clearly showed them to be adult animals, and thus it was necessary to range them within the carcinological system. These two authors regarded them as lower forms of Macrura, whereas Dohrn and others seem to be more inclined to associate them with the Edriophthalinia (Isopoda). I think their affinity to either of these groups is so very slight as to justify the establishment of a distinct order for their reception. More recently, Dr. Boas has discussed the affinity of the different groups of the Malacostraca from a phylogenetic point of view, and has been led to the conclusion that the Cumacea are very nearly related to the Mysidae, and may have descended directly from this group of Schizopoda. With this view I cannot, however, fully agree. It is undoubtedly a matter of great difficulty, if not quite impossible, to arrange the recent groups of Crustacea in a genealogical manner, as most of them in all probability represent very diverging branches, the origin of which from one or a few ancestral forms may go back to a very remote period of geological time. The Cumacea would seem to represent such an isolated branch, and cannot, of course, in my opinion, strictly be derived from any of the recent groups. As to the external appearance of these peculiar Crustacea, one would perhaps be most inclined to associate them with the Podophthalmia, but on closer examination we find them to differ materially in many points, and even in that important character from which the name of the above-mentioned group has been derived. The anatomical detads present, on the whole, a peculiar mixture of the podophthalmous and edriophthalmous type, and the development is rather unlike that of the Podophthalmia, and evidently much more similar to that of the Isopoda. The oral parts, which are generally regarded as affording highly important characters, I find — in contradiction to the opinion of Dr. Boas — very different from those in the Mysidae, and on the whole constructed on the same general type as in the Isopoda. On the other hand, the presence of well-developed 4 THE VOYAGE OF H.M.S. CHALLENGES. natatory exopodites on a certain number of the limbs would seem to remove these Crustacea far from the Isopoda. and to bring them in closer approximation to the Schizopoda, but this character is, perhaps, the only one pointing to a genetic connec- tion between these two groups. In all other respects I find the difference so very- great as cpiite to forbid the adoption of the view of a direct descent of the one from the other. In some characters the Cumacea would seem to occupy a still more primitive position than even the Mysidae. Thus, in the development of the higher Crustacea we find the Cumacean type, as it were, imitated by the early stage preceding the Mysis-stage, and to which the name of Zoea has been applied, the tail being in this stage, as in the Cumacea, very mobile and slender, and at first without any trace of ventral limbs, and the natatory exopodites confined to the anterior limbs only. Moreover, the general, form of the body in the Cumacea, and especially that of the tail, strongly remind us of that ancient group to which Mr. Packard has given the name of Phyllocarida, and of which the recent genus Nebalia is regarded as a direct descendant. True, the limbs in the Cumacea are very different from those in Nebalia, but it is by no means proved that the limbs in all of the ancient Phyllocarida were of exactly the same structure as in the recent genus Nebalia. It has been generally admitted that the phyllopodous form of the limbs is the most primitive one in Crustacea, and that all other forms might be derived from this type. But the legs in the oldest of all known Crustacea, the Trilobita, have been stated by Mr. Walcott to exhibit a totally different form, and this fact does not seem to corroborate the general validity of the above supposition. The structure of the branchial apparatus in the Cumacea is very remarkable and quite unlike all that is observed in other Crustacea. It is true that the part to which the gills are affixed represents the epipodite of the maxillipeds, and that this part is also found in the Mysidaa, but here it always constitutes a simple membranous plate without any trace of gills, and may be assumed only indirectly to subserve a respiratory function. As is well known, we also find the same part peculiarly developed for respiratory purposes in the cheliferous Isopoda, but even here without any trace of gills. The antennae in the Cumacea are totally different in structure from those in the Mysidae or any other form of the Podophthalmia, whereas they exhibit, especially as regards their peculiar modification in the males, a certain similarity to those in Nebalia, as also to those in the Amphipoda. In conclusion, I am inclined to regard the Cumacea as representing the descendants of a very ancient form, long ago extinct, which may have combined some characters of both the Phyllocarida and Trilobita. Perhaps even some of the palaeozoic forms placed among the Phyllocarida may have formed a direct transition to the Cumacean type. REPORT ON THE CUMACEA. DEFINITION OF THE FAMILIES. All the known families, except one, the Pseudocmnidae, being represented in the Challenger collection, I have found it advisable to give here short diagnoses of the same, and to enumerate the several genera contained in each of them. It may be remarked that these families do not seem to have strictly the same systematic value as those of the Schizopoda characterised in my former Report. The latter should perhaps more properly be regarded as groups of a higher rank, or as sections. 1. Cumidje. Integuments generally strong, calcareous. Tail not sharply defined from the trunk ; in female very slender, cybndrical ; in male much stronger, and with distinct epimeral plates. Eye generally present. Antennulse very small, with one of the flagella rudimentary. Antennae in male with posterior part of peduncle distinctly biarticulate, flagellum filiform and composed of numerous short articulations. Mandibles well developed, with the anterior branch produced and armed with a dense lateral series of spines. Epipodite of maxfllipeds very large, navicular ; gill-lobules numerous, leafdike, arranged in a straight series ; terminal plate of exopodite scaledike and strongly indurated. Second pair of gnathopoda rather large, with some of the joints expanded and laminar. The four posterior pairs of legs in both sexes simple, without any trace of exopodites. Five pairs of well-developed pleopoda present in male. Uropoda with both branches biarticulate, or the inner uniarticulate. Telson cpiite wanting. Genera. 1. Cuma, Milne-Edwards. 2. Cydaspis, G. 0. Sars. 3. Stephanomma, G. 0. Sars. 4. Iphinoe, Sp. Bate. 5. Gumopsis, G. 0. Sars. 2. Vaunthompsoniid^e. Integuments thin, scpiamous. General form of body, antennulse and mandibles much as in the Cumidse. Eye present or wanting. Antennae in male with flagellum composed of very elongate and slender articulations. Epipodite of maxillipeds (in female) very narrow, with only few and digitiform gill-lobules arranged in a semicircle ; terminal plate of exopodite membranous. Second pair of gnathopoda with the joints scarcely expanded. The three anterior pairs of legs in female, and all but the last in male THE VOYAGE OF H.M.S. CHALLENGER. provided with well-developed natatory exopodites. Five pairs of pleopoda present in male, as in the Cumidse, Uropoda with both branches biarticulate. Telson wanting. 1 Vaunthompsonia, Sp. Bate. Genera. 2. Leptocuma, G. O. Sars. 3. Lampropid^e. Integuments thin, squamous. Form of body generally very slender; tail in both sexes cylindrical, without any epimeral plates. Eye present or wanting. Antennuhe with both flagella well developed and nearly equal. Antennae in male with flagellum composed of numerous short articulations. Mandibles well developed. Epipodite of maxillipeds comparatively short, with but few, digitiform gill-lobules ; terminal plate of exopodite only slightly indurated. Second pair of gnathopoda rather slender, pedi- form. The two anterior pairs or only the first pair of legs in female, and all but the last in male, with well-developed natatory exopodites. Only three pairs of pleopoda present in male, or all wanting as in female. Uropoda with the branches very slender, inner tri- articulate, outer biarticulate. Telson distinctly defined, more or less laminar, with three or more terminal spines. Genera. 1. Lamprops, G. 0. Sars. 2. Hemttamprops, G. 0. Sars. 3. Paralamprops, n. gen. 4. Platyaspis, G. 0. Sars. 5. Chalarostylis, Norman. 4. Leuconidje. Integuments very thin and pellucid, smooth. Form of body generally slender, with tail very elongate and mobile, not sharply defined from trunk. Eye always quite wanting. Antennulse with one of the flagella small, uniarticulate. Antennas in male with flagellum composed of numerous short articulations. Mandibles rather small, with the body exserted into two angular processes ; anterior branch short, wanting the usual series of lateral spines, molar tubercle very massive. Palp of first pair of maxilla) with a single terminal bristle. Epipodite of maxillipeds comparatively small, with but few digitiform gill-lobules ; terminal plate of exopodite membranous, exserted into a narrow flap. Second pair of gnathopoda rather short, with some of the bristles issuing from the outer edge very strongly developed. The three anterior pairs of legs in female, and all but the last in male, with well-developed natatory exopodites. Second pair of legs only REPORT ON THE CUMACEA. sex-articulate. Only two pairs of pleopoda present in male. Uropoda with both branches Particulate and somewhat complanate, inner spinous, outer setose. Telson wanting. 1. Leucon, Krdyer. 2. Eudorella, Norman. Genera. 3. Eudorellopsis, G. 0. Sars. 5. DlASTYLID^E. Integuments generally strongly indurated and often armed with spines. Anterior division of body more or less dilated and sharply defined from tail, which is very slender and mobile. Carapace large and vaulted, with a more or less conspicuous rostriform projection in front. Eye present or wanting. Antennulae more or less dissimilar in the two sexes, rather slender, with the flagella unequal, the smaller one (in female) tri- articulate. Antennae in female forming a simple cylindrical stem divided into distinct articulations, in male more or less elongated. Mandibles well developed, anterior branch with a dense series of lateral spines. Epipodite of maxillipeds rather large, gill-lobules arranged in a more or less distinctly spiral manner ; terminal plate of exopodite membranous, triangular. Second pair of gnathopoda with basal joint very large and curved. The two anterior pairs of legs in female, and all but the last in male, with well- developed natatory exopodites. Only two pairs of pleopoda present in male, the posterior pair sometimes imperfectly developed. Uropoda very slender and elongate, outer branch biarticulate, inner bi- or triarticulate and spinous at inner edge. Telson distinctly defined, with the outer part generally narrowed and spinous at the edges, apex with two spines only. Genera 1. Diastylis, Say. | 2. Leptostylis, G. O. Sars. 6. PSEUDOCUMID^;. Integuments generally thin and squamous. Tail more or less sharply defined from the trunk. Eye present or wanting. Antennulae small, with one of the flagella rudi- mentary. Antennae in male with outer part of peduncle indistinctly biarticulate, flagellum composed of rather elongate and slender articulations. Mandibles not very strong, anterior branch with but few lateral spines. Epipodite of maxillipeds with only slight traces of gilhlobules, terminal plate of exopodite scarcely indurated. The two 8 THE VOYAGE OF H.M.S. CHALLNGER. anterior pairs of legs in female and all but the last in male with well-developed natatory exopodites. Only a single pair of pleopoda present in male. Uropoda, com- paratively short, with the branches lanceolate, outer Inarticulate, inner uniarticulate. Telson distinctly defined, but extremely small, rounded at the tip and unarmed. Genera. 1. Pseudocuma, G. 0. Sars. 2. Petalomera, G. 0. Sars. 7. CtTMELLIDJE. Integuments more or less indurated. Tail very slender and rather sharply defined from the trunk. Eye single or double, in male much larger than in female. Antennula? quite similar in both sexes, fiagella very unequal, the smaller one uni- or Inarticulate. Antennae in male with outer part of peduncle distinctly Inarticulate, flagellum composed of rather elongate and slender articulations. Mandibles with the anterior branch slender and slightly dilated at the extremity, lateral spines few in number. Epipodite of maxillipeds with the gill-lobules in female very slightly developed, terminal plate of cxopodite exserted into a narrow tubular flap. Second pair of gnathopoda with the terminal part irregularly flexuous. The two anterior pairs of legs in female, and all but the last in male, with well-develojaed natatory exopodites. The three posterior pairs of legs extremely slender and terminating in a distinct claw. Pleopoda quite wanting in both sexes. Uropoda with outer branch Particulate, inner uniarticulate. Telson wanting. Genera. 1. Cumella, G.'O. Sars. 2. Nannastacus, Sp. Bate. 3. Spencebatea, Norman. 8. Camp yl a spice. Integuments strongly indurated and calcareous. Carapace enormously developed and highly vaulted in its hinder part. Tail comparatively short, with the segments somewhat depressed. Eye generally present. Antennulse similar in both sexes, with one of the fiagella obsolete. Antenna? in male with the flagellum filiform and composed of elongate and slender articulations. Mandibles slender, with apex obliquely truncate and dentate, molar process very narrow, styliform. First pair of maxilla? unusually large, with the masticatory lobes strongly spinous ; second pair poorly developed, forming simple, small, membranous plates. Maxillipeds imperfectly developed, terminal part REPORT ON THE CUMACEA. 9 obsolete ; epipodite narrow, with digitiform gill-lobules arranged in a semicircle ; terminal plate of exopodite membranous and exserted into a narrow flap. First pair of gnathopoda rather strong, with penultimate joint greatly swollen. Second pair pediform, only slightly differing from first pair of legs. The two anterior pairs of legs in female, and all but the last in male, with well-developed natatory exopodites. Pleopoda quite wanting in both sexes. Uropoda with the scape elongate, outer branch biarticulate, inner uniarticulate. Telson wanting. Genus. Campylaspis, G. O. Sars. LIST OF THE SPECIES. The following is a list of the species of Cumacea procured during the Challenger Expedition : — CUMID.E. 1. Cyclaspis auslralis, n. sp. 2. „ pusilla, n. sp. 3. „ exscidpta, n. sp. Vaunthompsoniid.e. 4. Vaunthompsonia, meridional 'is, n. sp. Lampropid.e. 5. Paralamprops serrato-costata, n. gen. et sp. Leuconid^e. 6. Leucon assimilis, n. sp. 7. „ tenuirostris, n. sp. 8. Eiulorella abyssi, n. sp. DlASTYLIDjE. 9. Diastylis stygia, G. 0. Sars. 10. ,, horrida, n. sp. 11. ,, erinaceus, n. sp. 12. „ mystacina, n. sp. Cumellid^e. 1 3. Nannastacus suhmii, n. sp. Campylaspim:. 14. Campylaspjis pacifica, n. sp. 15. ,, nodulosa, n. sp. (ZOOL. CHALL. EXP. — PART LV. — 1886.) Iii 2 DESCRIPTION OF GENERA AND SPECIES. Order CUMACEA. Family I. Cumid^e. The most characteristic feature of this family is the total want, in both sexes, of exopodites on the four posterior pairs of legs, a character not found in any other known group of Cumacea. These legs also exhibit a very simple structure, and possess only a scanty supply of bristles. On the other hand, the males are distinguished by the presence of five pairs of well-developed pleopoda, and there is only one other family, Vaunthompsoniidse, which agrees with the Cumidse in this respect, whereas the males of all other known Cumacea either have a less number of these limbs, or want pleopoda altogether, as in the females. As a character peculiar to this family may also be mentioned the highly indurated and strong integuments, the full development of the branchial apparatus, and the total want of a telson. Of the five genera comprised within this family, only the genus Cyclaspis is repre- sented in the Challenger collection. Cyclaspis, G. 0. Sars, 1864. Cyclaspis, G. O. Sars, Om den aberrante Krebsdyrgruppe Cumacea og dens nordiske Arter, p. 81. Generic Characters. — Carapace large and swollen, almost globular, with the dorsal line strongly curved, and the pseudorostral projection very slight ; a distinct notch on each side anteriorly beneath the latter projection. Ocular lobe more or less produced, and usually having at its end a distinct eye. Only four segments of the trunk exposed behind the carapace, the anterior rather large, the three others suddenly much narrower, and scarcely broader than the caudal segments. Tail very slender and elongate, last segment obtusely produced at the tip. Antennulse with one of the flagella very small, knob-shaped, the other biarticulate, and bearing two long sensory appendages at the extremity. Posterior Up with the lateral lobes armed at the tip with strong teeth. 12 THE VOYAGE OF H.M.S. CHALLENGER. Second pair of gnathopoda strongly developed, the basal and meral joints expanded exteriorly at the end into lamellar processes. First pair of legs with the basal joint highly indurated and expanded in the middle, terminal part very slender. Second pair unusually short, with the coxal joint broad and lamellar, terminal joint sjrinulose at the edges. Uropoda comparatively short, with both branches lanceolate, outer Inarticulate, inner uniarticulate. Remarks. — The present genus is nearly allied to Cuma, Milne Edwards, from which genus it is chiefly distinguished by the great development and peculiar form of the carapace, as also by a rather different structure of the uropoda. The genus was founded upon a peculiar deep-sea form, Cyclaspis longicaudata, obtained by the author off the Norwegian coast, and since also met with in the great depths of the Atlantic. In this form the eye is totally wanting, but in all other respects it closely agrees with the species here described. Another form, Cyclaspis cornigera, described by the author from the Mediterranean, would seem to differ much more from the typical sjiecies, and may perhaps be regarded as the type of a separate genus. No other species of the genus has hitherto been known. In the Challenger collection no less than three different species are represented, and will be described below. They may be easily distinguished by the sculpture of the carapace, as follows : — somewhat gibbous in the hinder part and provided on each side with a slight transverse keel, and another obliquely longitudinal, joining each other in about the middle part of the carapace, . Cyclaspis australis, n. sp. Carapace quite glabrous and with the dorsal line evenly arcuate, very gibbous and highly sculptured by strong keels, both trans- verse and longitudinal, limiting several deeply excavated areas on each side, ....... Cyclaspis pusilla, n. sp. Cyclaspis exsculpta, n. sp. 1. Cyclaspis australis, n. sp. (PI. I. figs. 1-20). Specific Characters. — Carapace slightly compressed, keeled dorsally and somewhat gibbous in the posterior part, with a slight curved prominence passing down the sides and joining an oblicpiely longitudinal ridge above, which limits the gastric region. Ocular lobe narrow and produced to the tip of the pseudorostral projection, its outer part swollen and provided with well-marked lenses and pigment. Tail distinctly keeled along the middle line, the three anterior segments, as also the two posterior segments of the trunk with a subdorsal keel on each side of the middle one. First pair of legs with terminal part shorter than basal, the three outer joints nearly equal in length; second pair shorter than third, last joint armed with six spines. Inner branch of uropoda about as long as scape, finely serrate along the inner edge, outer branch a little longer. Length, 10 mm. REPORT ON THE CUM ACE A. 13 Remarks. — This species may easily be distinguished from the typical form, Cyclaspis longicaudata, by the less tumid carapace and the sculpture of the same as well as that of the tail. The ocular lobe has, moreover, at the end distinctly developed visual elements, whereas in the typical species the eye is quite wanting, and the ocular lobe very short and rounded. The species is also very different from the Mediterranean form, Cyclaspis cornigera. Description. — I have been able to examine several specimens of this form, both males and females, but none of them seem to be quite full-grown. The largest specimen, represented on PI. I. figs. 1, 2, is a young male. The length of this specimen is 10 mm., whereas that of the largest female specimen (fig. 3) is only 8 mm. The form of the body in both sexes (see PL I. figs. 1-3) is very slender, with the anterior part, as in the typical form, suddenly and markedly dilated, almost club-shaped. The carapace is large and deep, though somewhat more compressed than in the typical species, its thickness being rather less than its height (compare figs. 1, 2). It is somewhat gibbous in its posterior part, the strongest dorsal curvature lying rather behind the middle, from which point the dorsal line declines almost in a straight line towards the pseudorostral projection. The latter, to which generally the very improper term of " rostrum " has been applied, is quite short and obtusely truncate at the tip, and, as usual, is chiefly conqnosed of the antero-lateral lobes of the carapace projecting in front of the antennal segment. Immediately beneath this projection there occurs on each side a distinct notch, limited by a short projecting angle, from which a short elevated ridge runs backwards over the carapace. The inferior edges of the carapace are evenly curved and join the posterior margin at an obtuse angle. On each side of the carapace, somewhat posterior to the middle, a slight crescent-shaped prominence is seen running transversely down the sides to the inferior edge. Above, this prominence joins an obliquely longitudinal ridge, that begins at the most elevated part of the carapace, limiting posteriorly the gastric region, and disappears at the sides of that area ; between this ridge and the above-mentioned short lateral keel in one of the specimens (fig. 1), a somewhat irregular forked prominence occurs, joining both in the anterior part of the carapace. The gastric area (see fig. 2) exhibits, as usual, a somewhat bell-shaped form, and is limited to each side by a distinctly marked fissure. From the extremity of this area the ocular lobe is seen projecting. This lobe exhibits a rather peculiar appearance, being continued along the whole upper side of the pseudorostral projection to its very tip ; it is (fig. 4) almost cylindrical and slightly swollen at the extremity, where an assemblage of distinctly marked and regularly arranged lenses or cornea-facets occurs, as also interiorly a dark pigment. Only four segments of the trunk are found to be exposed behind the carapace. Of these, the first is firmly connected with the carapace and almost of the same height as 14 THE VOYAGE OF H.M.S. CHALLENGER. that part, whereas the three others become suddenly much lower and scarcely at all broader than the caudal segments. These hinder segments of the trunk are moreover very movably connected with each other, flexible membranous spaces being interposed both above and below. Besides a median dorsal keel, also continued on the carapace, these segments exhibit on each side of that keel another running along the dorsal surface (see fig. 2). The epimera of the first segment are somewhat expanded and truncated at the tip ; those of the three posterior segments are narrowly rounded. The tail, especially in female specimens (fig. 3), is extremely slender, even somewhat longer than the anterior division of the body, and of a narrow cylindrical form. It has a distinct median dorsal keel running along all the segments, and, besides, the two anterior segments exhibit on each side of the median another subdorsal keel, as on the posterior segments of the trunk. Laterally each of the five anterior caudal segments is provided at the anterior margin with a small knob-like process that fits into a corre- sponding notch in the posterior margin of the preceding segment when the tail is fully extended. As to the relative size of the segments, they slightly increase in length posteriorly to the penultimate, which is much longer than any of the others. The last segment is somewhat shorter, and produced at the end as an obtusely rounded protuberance (see fig. 20). The integuments are highly indurated and calcareous, showing under the microscope partly a granular, partly a squamous or densely reticulate structure. The colour of the preserved specimens is light brown with darker shadings, especially on the anterior part of the carapace. The antennulse (fig. 5, a1) are attached close together beneath the pseudorostral pro- jection, being partly concealed by that prominence, their outer part, however, projecting freely from the anterior notches of the carapace. They consist each of a triarticulate peduncle and two very short flagella. The first joint of the peduncle is rather large and highly indurated ; it is turned obliquely outwards and has the outer edge finely ciliated, the inner, close to the end, is provided with a few short bristles. The two succeeding joints of the peduncle are much smaller, and taken together scarcely as long as the basal. The second joint has at the end on the inner side a bunch of three auditory setae, and exteriorly a small simple bristle ; the last joint is provided with a single auditory seta (see fig. 6). Of the flagella, the one is quite rudimentary, only consisting of a small knob- like joint bearing a very minute auditory seta and two simple bristles. The other flagellum is about as long as the last joint of the peduncle and Particulate, the terminal joint being the smaller, and bearing at the tip two very long and regularly segmented sensory appendages, besides two simple bristles, one of which is rather long. The latter flagellum must undoubtedly be regarded as homologous with the outer flagellum in other Crustacea, since it bears the sensory appendages, but the antennula is generally twisted in such a manner that it in reality lies inside the other (see fig. 5). REPORT ON THE CUMACEA. 15 The antennae (fig. 5, a2), originating immediately behind the antennulse, are, as usual, quite rudimentary in the female, only consisting of a broad lamellar basal part and a very small conical terminal segment, the latter representing the rudiment of the flagellum. The basal part is highly indurated, and bears on the inner (or upper) edge two very strong and densely ciliated setae ; the terminal segment, which is connected to the basal at an obtuse angle, is tipped by a bunch of very minute auditory seta?. In the young male the antennae show an appearance quite simdar to that described further on in the male of Paralamprops serrato-costata, and lie folded beneath the lateral parts of the carapace, so as not to be visible in the lateral aspect of the animal. The anterior lip (fig. 5, l) forms a comparatively small rounded lobe, projecting immediately behind the antennal segment. It is slightly emarginate at the free posterior edge, and finely cdiated within the emargination. The posterior lip (fig. 7), on the other hand, is rather large and quite membranous in structure. It consists of two somewhat subtriangular lobes, connected at the base, and terminating in a slightly inflexed angle armed with three strong compressed teeth. The lobes are, moreover, densely cdiated along their inner edge, as also at the outer part of the exterior margin. The mandibles (fig. 8), as usual, do not exhibit the slightest trace of a palp, and are wholly indurated and rather brittle, so that they easily break by slight pressure. Their body is navicular in form and shows an elegantly reticulate structure ; its posterior pointed extremity is articulated to the dorsal part of the carapace at the point where, on each side, the fissure separating the gastric area from the lateral parts begins. From the anterior part of the body two unequal branches originate, the one projecting forward nearly in the axis of the body, the other inward at a right angle to the same. The anterior branch is slightly compressed and has the tip incurved, with the cutting edge somewhat different on the two mandibles. On the right mandible it is simple and only divided into two small teeth, whereas on the left mandible it is double, forming two acute projections, each indistinctly bidentate, and the posterior one movably con- nected with the anterior. Behind the cutting edge there is a dense series of curved spines, arranged in a comb-like manner and occupying almost the whole length of the inner edge of this branch. The posterior branch, representing the so-called molar tubercle, is shorter and more cylindrical in form, and terminates in a finely fluted surface, more oblique on the right than on the left mandible. The first pair of maxillae (fig. 9) are composed of a thick muscular basal part, from the end of which originate two slightly incurving masticatory lobes, and of a peculiar recurved palp. The outer masticatory lobe is of rather firm consistency and a little narrowed towards the tip, which is truncated and armed with strong spines forming a double row, besides a few very small bristles. The inner lobe is much shorter and of a more membranous nature ; its tip is narrowly rounded, and bears four curved spines, the 16 THE VOYAGE OF H.M.S. CHALLENGER. posterior of which is the longest and finely ciliated. The palp, which is affixed to the outer side of the basal part close to the base of the exterior masticatory lobe, is very elongate and narrow, turned straight backward, and projecting within the branchial cavity. It is uniarticulate and linear in form, with a distinct muscular band running throughout its axis ; from its tip two peculiar compressed setse originate, the outer being somewhat longer than the inner, and both having one of its edges sharpened and armed with fine recurved hooks. The second pair of maxillse (fig. 10) are more membranous in structure and of a rather different appearance. The basal part is composed of two segments, the first of which is very short and filled by strong muscles. The second segment is almost triangular, and exhibits on the inner side two very slight masticatory expansions, the anterior one somewhat obliquely truncate and provided with short cdiated bristles, the posterior having a regular series of slender setae arranged in a combdike manner. From the lower side of the basal part, close to its apex, a very short biarticulate palp originates, directed anteriorly and movably articulated. The last joint of this palp is obliquely truncate at the tip and armed with a dense row of slender curved spines, and the first joint is on the inner side expanded to a lobe of quite a similar appearance. Moreover, the outer edge of the basal part forms a very thin and pellucid expansion, which may be regarded as homologous with an exognath; this expansion has the edges quite smooth, without any trace of bristles. The maxillipeds (fig. 12), which are closely applied against the other masticatory parts, and themselves partly covered by the two parrs of gnathopoda, form each a rather stout and somewhat compressed stem, contiguous with the one of the other side, and having the outer part slightly incurved. The basal part is rather large, occupying half the length of the stem, and consists of two segments, the first of which, however, is very short and rather difficult to see. At the end this part projects inwards as a rounded triangular masticatory lobe, armed at the tip with a few short and thick spines densely cdiated in their outer part. Along the inner edge there also occurs a series of five rather strong plumose setse, and another strongly cdiated seta is found at the base of the masticatory lobe. The terminal part of the stem, or the palp, is composed of but four distinctly defined joints, of which the two first are very broad and lamellar, and connected by a very oblique articulation ; the outer of these latter joints is armed along the inner edge with a series of small, curved, bidentate spines (fig. 12, «), besides numerous fine bristles, and has at the end externally a strong anteriorly curving seta. The two last joints become suddenly much smaller and more or less incurved; the penultimate is provided at the end with three ciliated setae (two on the inner and one on the outer side) ; the terminal joint is very small and tipped with three or four short bristles. At the base of these limbs the remarkable branchial apparatus is affixed (see fig. 1 ] ). REPORT ON THE CUMACEA. 17 The greater part of this apparatus constitutes a very large membranous plate of a somewhat navicular form, and curved so as to exhibit an outer convex and an inner concave surface. It terminates in a recurved acuminate lajipet, and is furnished along the inner edge with a regular series of lamellar gill-lobules, increasing in size anteriorly, the foremost of which is turned backwai'd. The whole plate is moved in the living animal by an assemblage of strong muscles partly passing from the base of the corre- sponding maxilliped. Anteriorly another smaller plate is seen extending forwards and forming an imperfect funnel ; it tapers to a narrow band-like ligament, and terminates in an indurated lamella armed with six curved setae, and encircled by a very thin and pellucid border (fig. 13). This terminal lamella, which is distinctly marked off from the plate by a transverse suture, is found projecting in front of the pseudorostral prominence in close juxtaposition to the corresponding one of the other side, and in living specimens both lamellse are seen performing a peculiar snapping movement, whereby the water is expelled at intervals from the branchial cavities, during the rhythmical strokes of the true branchial plate. Although this anteriorly directed plate would seem to be partly connected at its base with the principal plate, I am still disposed to entertain the assumption I have set forth in another work,1 that the former is the highly modified exopodite of the maxillipeds, whereas the latter represents the epipodite. Of the limbs of the trunk the two anterior pairs are closely applied against the oral parts, and do not seem to have any locomotory function, and they may, therefore, properly be termed gnathopoda. These two pairs differ materially from each other both as regards size and structure. The first pair of gnathopoda (fig. 14) form simple stems, composed of the same number of joints as the maxillipeds, but are rather more slender. The basal part is longer than the terminal and somewhat laminar, with a slight longitudinal keel running along the upper surface and terminating posteriorly in a dentiform projection of the outer edge. This part, as in the maxillipeds, chiefly consists of a single joint (the basal), the coxal joint being very small and imperfectly defined. It is furnished at the end with a rather strong ciliated seta, and has, moreover, a few bristles at the outer part of both edges, those of the exterior edge being larger and about eight in number. The first joint of the terminal part is quite short and somewhat swollen, almost globular, with a single bristle at the end on the outer side; the succeeding joint is a little longer, con- stricted at the base, and furnished along the inner edge with several ciliated bristles, and a single one at the end exteriorly. The two last joints are rather small, and beset with bristles at the end. At the base of each of these gnathopoda there is affixed, in the adult female of all Cumacea, a lamella edged with very long setaj projecting within the marsupial pouch. This lamella was, however, only slightly developed in the specimens of the present species, owing to their being not full-grown. 1 Beskrivelse af de paa Fregatten Josephine's Expedition fundne Cumaceer. (ZOOL. CHALL. EXP. PART LV. 1886.) Ill 3 18 THE VOYAGE OF H.M.S. CHALLENGER. The second pair of gnathopoda (fig. 15) are much larger, and also of coarser structure. They originate widely apart, at the outer posterior angles of the carapace, but are suddenly bent inwards so as to meet along the greater part of their length, being more or less closely applied against the ventral side of the body. The basal joint is very large, more than twice as long as the remaining part, and strongly indurated, exhibiting, especi- ally on the lower surface, a distinct nodulose sculpture ; it gradually expands towards the distal extremity, and juts out exteriorly as a large laminar projection of a somewhat trian- gular or broadly lanceolate form, fringed along the inner edge with a row of short ciliated bristles. At the base this joint bears a well-developed natatory exopodite of about half its length, and composed of a simple cylindrical stem and a five-jointed terminal part, furnished with long natatory setae. The ischial joint, not defined in the former pair, is distinctly developed, though rather short and simply cylindrical. The meral joint, on the other hand, exhibits a rather peculiar form, being expanded exteriorly to a large securiform projection. The carpal joint is subtriangular in shape, or produced interiorly to an almost right angle bearing a few simple bristles. The propodal joint is quite narrow and articulated to the outer corner of the preceding joint. The terminal joint, finally, is still more narrow, linear, and together with the propodal, admits of being impinged against the interior projection of the carpal joint. The first pair of legs (fig. 16) are constructed upon the very same type as the last pair of gnathopoda, and in function would seem to be intermediate between gnathopoda and true legs, their basal part being generally closely applied against the ventral side of the body, whereas the terminal part is freely projecting and very mobile, so as partly at least to act as true locomotory organs. They are much larger than any of the other limbs, equalling in length, when fully extended, the carapace and the two first segments of the trunk taken together. The coxal joint, which on the gnathopoda appears but very slightly indicated, is on these limbs distinctly defined (see figs. 1, 3), but so firmly con- nected with the surrounding parts as generally not to be obtained in connection with the leg by dissection. The basal joint is very large and strongly indurated, being more- over greatly expanded in the proximal part, and forming an obtuse angle at the middle of the inner edge; whereas the distal part is narrowed, without forming any expansion at the apex. The edges of this joint are quite devoid of bristles except at the ajDex, where a strong cUiated seta occurs exteriorly, and a much smaller one interiorly. The exopodite affixed to the base of this joint is somewhat larger than that of the last pair of gnathopoda, and its basal part is a little dilated at the middle, and is also more strongly indurated. The remainder or terminal part of the leg is very slender, somewhat shorter than the basal, and generally exhibits a geniculate bend at the junction between the carpal and propodal joints. The lengths of these two joints are nearly equal, whereas the two preceding are considerably shorter and thicker. The terminal joint is a little shorter and also narrower than the preceding joint, and provided at the tip with REPORT ON THE CUMACEA. 19 several slender bristles ; similar bristles are also found along the inner edge of the pro- podal joint. The four remaining pairs of legs are all very small, and, in both sexes, quite simple, without any trace of exopodites. They represent the true ambulatory legs, and for this purpose are also turned more exteriorly. The anterior of these pairs (fig. 17) are somewhat shorter than the succeeding pair, but have the coxal joint very large (see figs. 1, 2), forming an almost semicircular plate firmly connected with the corresponding segment. Of the other joints the basal is as usual the largest. The carpal joint is obliquely truncate at the apex, or produced on the outer side to a blunt process bearing two stiff bristles. The propodal joint is extremely small, whereas the terminal is rather large, compressed, and armed at the tip and aloug the outer edge with six short spines. The three posterior pairs of legs (figs. 18, 19) are quite similar in structure, de- creasing, however, somewhat in length, owing to the basal joint being shortened succes- sively. The coxal joint in all of them is very small and scale-like; the four outer joints successively decrease in size, and the terminal joint has the form of a simple biarticulate spine. A few strong bristles originate from the tip of each joint. The ventral limbs of the tail are, as usual, completely wanting in the female (fig. 3). In the young male (fig. 1) there are five pairs of imperfectly developed pleopoda, con- sisting of a stout basal segment and two short branches without distinctly developed setae. The uropoda (see fig. 20) are about twice as long as the last caudal segment, with the basal part or scape almost cylindrical or slightly dilated towards the apex. The terminal branches are both lanceolate in form and at the apex drawn out to a sharp point. The inner one is the shorter, about as long as the scape, and uniarticulate, with the inner edge finely serrulate ; the outer branch is distinctly biarticulate, with the first joint very short, and the inner edge of the second joint setiferous. Habitat. — Several specimens of this form were collected on April 1, 1874, off the entrance to Port Philip, Australia, Station 161, from a depth of 38 fathoms; sandy bottom. 2. Cyclaspis pusilla, n. sp. (PL I. figs. 21-23). Specific Characters. — %. Carapace almost globular, quite smooth at the sides, dorsal line evenly curved ; pseudorostral projection very short and obtuse. Ocular lobe only slightly produced, its outer part swollen and provided with well-developed visual elements. Tail slighty keeled dorsally, but elsewhere quite smooth. Limbs much as in Cyclaspis australis. Colour whitish, anterior part of body mottled with brownish-pink. Length of adult female only 3^ mm. Remarks. — This beautiful little form is readily distinguished from the preceding 20 THE VOYAGE OF H.M.S. CHALLENGER. species, not to speak of its much smaller size, by the smooth and almost globular carapace, the shorter pseudorostral projection, and the simple structure of the caudal segments. Description. — I have only had a single specimen of this species for examination, a full-grown female with greatly developed marsupial pouch. Its length does not exceed 3tj mm., and it thus belongs to the smaller forms of this order. The form of the body (see figs. 21, 22) is that characteristic of the genus, the anterior part being greatly dilated and almost globular, whereas the posterior part is extremely slender. The carapace is very large and cpiite smooth at the sides, about as thick as high, and with the dorsal line evenly arched. The pseudorostral projection is short and obtuse, and limited below at each side by a well-marked notch. The ocular lobe (fig. 23) is rather shorter than in the preceding species, and is provided at the somewhat dilated extremity with distinctly developed visual elements. The four exposed segments of the trunk decrease rapidly in size, and the two posterior do not exceed the caudal segments in height or thickness. The tail is very slender, cylindrical, and somewhat longer than the anterior division of the body. It has a slight keel running along the middle of the dorsal surface, and this keel is also continued along the segments of the trunk and the carapace ; but for the rest the segments are quite smooth and exhibit the usual relation in length. The colour of the animal is whitish, with small patches of brownish-pink, especially along the sides of the anterior part of the body. As to the limbs, they on the whole seem to agree fairly well with those of the preceding species. The three anterior pairs of legs were, however, partly broken in the specimen examined, and it may be that the terminal part of the first and second pairs might have exhibited some minor differences, as to the relative length of the several joints. The marsupial pouch is greatly developed and projecting below, giving the anterior part of the body an almost completely globular form. Habitat. — I found the above-described specimen among some dredged sand kindly sent to me by Mr. John Murray. The sand was procured with the dredge on September 8, 1874, at Flinders Passage (between Australia and New Guinea), from a depth of 7 fathoms. 3. Cyclaspis exsculpta, n. sp. (PL I. figs. 24-26). Specific Characters. — ?. Carapace very gibbous and strongly sculptured by elevated crests limiting irregular depressed areas ; two of the crests exceedingly strong, crossing the dorsal surface of the carapace transversely, and uniting on each side near the infero- posterior angle, being moreover connected above on each side by a thin longitudinal ridge, whereby a large depressed area of irregularly oval shape is formed on each side of the • REPORT ON THE CUM ACE A. 21 carapace. Pseudorostral projection somewhat produced, obtuse at the tip, and limited below on each side by a distinct angular notch. Ocular lobe extending along the dorsal side of the pseudorostral projection to its tip, only slightly dilated at the apex and pro- vided with but few corneal facets. Segments of trunk angular, with three longitudinal dorsal crests. Integuments highly indurated. Colour uniformly whitish. Length about 5 mm. Remarks. — The present species is at once distinguished by the remarkable sculpture of the carapace and the very strongly indurated integuments. Although the sole specimen is incomplete, it may without any doubt be referred to the genus Cyclasjiis. Description. — The specimen examined wants the whole posterior part of the body, as also the terminal part of the first pair of legs. The preserved anterior division of the body has a length of 2 mm., and the whole body may thus have been about 4^ or 5 mm. long. The carapace (see figs. 24, 25) is highly remarkable for its strongly marked sculpture. It is rather gibbous in its hinder part and exhibits two very strong transverse crests running across the dorsal surface and continued down the sides towards the infero- posterior corner, where they unite almost at a right angle (see fig. 24). Of these crests the anterior is placed immediately behind the frontal lobe and is uninterrupted in the middle, whereas the posterior is divided in the middle line by a distinct notch at about the posterior third of the length of the carapace (see fig. 25). Both crests are connected by a median longitudinal keel and two lateral ridges, whereby the middle part of the carapace is divided into four depressed areas, two dorsal and two lateral, the latter rather large and of an irregularly oval or rounded triangular form. From the anterior crest, moreover, two slight ridges run forward on each side, the one towards the base of the pseudorostral projection, the other disappearing at some distance from the antero-lateral angle of the carapace. Behind the posterior crest, finally, there occur on the dorsal surface three longi- tudinal keels, each terminating in a somewhat rounded laminar projection at the posterior edge. The pseudorostral projection is rather produced, obtusely triangular in form, and limited below on each side by a distinct angular notch. The ocular lobe (fig. 26) is very narrow and continued along the dorsal side of the pseudorostral projection to its very apex ; it is but slightly dilated at the tip and exhibits a few rather small corneal facets ; of pigment, on the other hand, no trace could be detected in the specimen examined. The four exposed segments of the trunk are at a much lower level than the carapace and exhibit a somewhat angular form, the dorsal surface being scalloped out on each side of the middle line, or provided with three longitudinal crests. As usual, the three posterior segments are rather smaller than the anterior and connected by very movable articulations. All the integuments are very strong and highly indurated, exhibiting under the microscope a closely granulated structure. 22 ' THE VOYAGE OF H.M.S. CHALLENGER. The colour of the animal is uniformly whitish, without any distinct pigmentary deposits. The limbs do not seem to differ materially in their structure from those of the two preceding species. The terminal part of the first pair of legs was, however, wanting in the specimen examined. Habitat. — The above-described specimen I found among the same sample of dredged material from which the preceding species was derived. The locality is stated as follows: — September 8, 1874, at Flinders Passage, 7 fathoms. Family II. Vaunthompsoniid^e. This family is rather closely related to the Cumidse, from which, however, it differs materially by the three anterior pairs of legs in the female, and all but the last pair in the male, being provided with well-developed natatory exopodites. From the Lampro- pidae, with which family it agrees somewhat in the general appearance and the thin integuments, it is distinguished by the want of a distinctly defined telson, and by the male having five pairs of strongly developed pleopoda as in the Cumidae. The family comprises at present two distinct genera, viz., Vaunthompsonia, Spence Bate, and Lepto- cuma, G 0. Sars. Besides, the form described by the author at an earlier date under the name of Leucon anomalus, from the West Indies, would seem more properly to belong to this family than to the Leuconidse, representing in all probability another distinct genus. Of these genera only the first is represented in the Challenger collection. Genus Vaunthompsonia, Spence Bate, 1858. Vaunthompsonia, Spence Bate, Nat. Hist. Review, vol. v. p. 203. Generic Characters. — Form of body not very slender, scarcely compressed; anterior and posterior divisions not sharply defined from each other. Integuments thin, squamous. Carapace only slightly arcuate above; pseudorostral projection very short; antero-lateral angles more or less produced. Eye distinctly developed. Five segments of the trunk exposed behind the carapace. Last caudal segment produced at the end, but without any distinct telson. Antennutae small, similar in both sexes, one of the flagella very minute, uniarticulate. Antennae in female biarticulate, with two strong ciliated setae, in male shorter than the body. First pair of legs rather elongate and slender ; second pair spinous, with the ischial joint not defined ; penultimate pair in female without any rudiment of exopodite. Pleopoda of male very large, with inner plate uniarticulate, outer biarticulate. Uropoda strongly spinous, scape rather short, both branches biarticulate. REPORT ON THE CUMACEA. 23 Remarks. — This genus is chiefly distinguished from Leptocuma hy the body being less elongate and not compressed, by the presence of a well-developed eye, by the second pair of legs having the ischial joint obsolete, and finally by the penultimate pair of legs in the female wanting the rudimentary exopodite present in the former genus. Of the genus only a single species, Vaunthompsonia cristata, Spence Bate, was hitherto known. The following new species from the Challenger Expedition is now added. 4. Vaunthompsonia meridionalis, n. sp. (PL II. figs. 1-5). Spjecific Characters. — ? . Carapace shorter than the free segments of the trunk taken together, with two short serrate crests on the posterior part of the dorsal surface, anterior part unarmed. Pseudorostral projection broadly truncate at the tip, anterior edges finely serrate and having three stronger serratures at the lower corner ; subrostral notches distinct. Antero-lateral angles of carapace sharply produced, inferior edges strongly serrate in the anterior part. Ocular lobe rather large, almost reaching to the tip of the pseudorostral projection, with two small denticles in its anterior part. First pair of legs shorter than in the typical sjsecies, with the three outer joints subequal. Posterior pro- jection of last caudal segment triangular, with five slender spines on each side. Uropoda. shorter than the two last caudal segments taken together, scape comparatively stout, inner branch not longer than outer, with last joint small, scarcely half as long as first, both strongly spinous on inner edge. Length, 10 mm. Remarks. — This new species may easily be distinguished from the typical form, Vaunthompsonia cristata, Spence Bate, by the carapace having only two short diverging crests on its posterior part, whereas the anterior part is quite smooth ; by the broadly truncate pseudorostral projection and the distinctly marked notch beneath the same ; also by the much shorter legs of the first pair, and finally by a rather different form of the uropoda. Description. — The sole specimen before us is a young female, with the marsupial pouch not yet developed. It measures in length about 10 mm. The general form of the body (see PI. II. figs. 1, 2) is nearly as in the typical species, though somewhat less compressed. The anterior division tapers gradually backwards, without being sharply defined from the posterior. The carapace is comparatively rather small, not attaining the length of the free segments of the trunk taken together. The dorsal surface is but very little vaulted, and exhibits posteriorly two short curved serrate crests diverging anteriorly, but for the rest is quite smooth, without any trace of the strongly serrate longitudinal crest observed in the typical species. The pseudorostral projection (see fig. 4) is very short, and broadly truncate at the apex, with the anterior edges finely serrate, and having moreover three much stronger denticles inferiorly. The notch limiting this projection inferiorly on each 24 THE VOYAGE OF H.M.S. CHALLENGER. side is very distinct and rather deep, and the antero-lateral corners of the carapace are produced to a sharp point. The inferior edges of the carapace are evenly curved, and in the greater part of their length strongly serrate. The ocular lobe (see figs. 3, 4) is rather large and oval in form, almost reaching to the tip of the pseudorostral projection. It is armed on the anterior part with two small anteriorly directed denticles, and seems to exhibit pretty well developed visual elements. The exposed part of the trunk is composed of five well-defined segments, the anterior of which is the shortest, the second the largest. On all the segments the lateral parts form broad laminar epimera, slightly projecting at the sides (see fig. 2). The tail is only very slightly longer than the anterior division of the body, and of the usual slender cylindrical form. Of its segments the penultimate is by far the longest. The last segment (see fig. 5) is rather short, and has the posterior edge produced in the middle to a short triangular plate arching over the anal orifice, but not defined from the segment by any suture. On each side of this plate is found a row of five slender spines increasing in length posteriorly. All the integuments are rather thin, only very slightly indurated, and exhibiting under the microscope a squamous structure. The colour of the specimen preserved in spirit is whitish, with a few stellate pigmentary patches, especially distinct on the carapace. The antennulae (fig. 4, a1) are of moderate size, and have one of the flagella very short and uniarticulate, the other triarticulate, and bearing at the tip two ribbon-like sensory appendages. The antennsa (fig. 4, a2) are very small and biarticulate, their outer part projecting from the bottom of the anterior notch of the carapace. The terminal plate of the exopodite of the maxillipeds projecting in front of the carapace (see figs. 3, 4, x) is very thin and cprite membranous, being drawn out to a narrow, somewhat tubular lappet. The first pair of legs (see fig. 1) are rather slender, though not nearly so elongate as in the typical species ; they only slightly exceed in length the carapace, and have the three last joints nearly uniform in size. The remaining pairs of legs do not seem to differ materially in structure from those of Vaunthom/psonia cristata, and as in that form only the two last pairs are simple, whereas the two preceding pairs are provided with well-developed natatory exopodites of similar structure to those on the first pair. The second pair are only slightly shorter than the first, and composed of but six joints, the ischial joint not being distinctly defined. The uropoda (see fig. 5) do not attain the length of the two posterior caudal segments taken together, and exhibit a comparatively coarser structure than in the typical species. The scape is rather thick, and armed along the inner edge with about ten strong spines. REPORT ON THE CUMACEA. 25 The branches arc about equal in length and somewhat longer than the scape, whereas in the typical species the inner one is by far the longer, and the outer scarcely as long as the trunk. The inner branch is composed of two distinctly defined joints, the first of which is much the larger and somewhat expanded at the base, sublaminar, and armed along the inner edge with a dense series of spines, some of which are more elongate than the rest. The outer joint of this branch does not attain even half the length of the preceding joint, and is also much narrower. It is armed with about eight spines, three of which originate from the apex, the rest from the inner edge ; the apical spines being by far the strongest. The outer branch is also biarticulate, but the first joint is here very short, whereas the terminal joint is rather elongate, and furnished along the inner edge with a row of strong ciliated seta?; from the apex four somewhat more elongate setae originate, and to the outer edge about six slender spines are affixed, one of which occurs at some distance from the rest, close to the apex. Habitat. — I found the above described specimen in examining a sample of dredged mud preserved in strong spirit, and kindly sent to me by Mr. John Murray. The mud was procured on January 29, 1874, off Cumberland Bay, Kerguelen Islands, from a depth of 127 fathoms, Station 149h. Family III. Lampropid^e. I have seen fit to establish this family for the reception of three northern genera, Lamprops, Hemilamprops, and Platyaspis, which present several peculiar characteristics not found in other Cumacea, and now I am able to add another generic type from the Challenger Expedition, undoubtedly belonging to the same family. The genus Chalaro- stylis, established by Norman for a Cumacean from the " Porcupine " expedition, must also, I think, be referred to this family, which thus at present comprises no less than five different generic types. The most characteristic features distinguishing this family from the two preceding are : — the fuller development of the antennulee, both flagella of which are distinctly articulated and nearly equal in size ; the less rudimentary structure of the antennae in the female; the comparatively great length of the second pair of legs; the presence in the male of but three pairs of pleopoda (in one of the genera, Lamprops, the pleopoda are even quite wanting); the distinctly developed telson; and finally, the slender form of the uropoda, the inner branch of which is triarticulate. All the known forms belonging to this family, except the one described below, would seem to be restricted to the northern ocean. (ZOOL. CHALL. EXP. — PART LV. — 1887.) E* * 26 THE VOYAGE OF H.M.S. CHALLENGER. Genus Paralamprops, n. gen. Generic Characters. — Anterior division of body well defined from posterior. Cara- pace large and sculptured with denticulated crests, dorsal surface vaulted, sides somewhat expanded; pseudorostral projection very slight, subrostral notches obsolete. Ocular lobe very small, eye wanting. Exposed part of trunk composed of five well-defined segments with laterally projecting epimera. Tail very slender. Antennuhe unusually large, with both flagella elongate and multiarticulate. Antennse in female five-articulate, with two strong ciliated setae on the middle joint. First pair of maxillae without any trace of palp. Second pair of gnathopoda very slender, basal joint not expanded at the end. The two anterior pairs of legs in female with well-developed natatory exopodites, third and fourth pairs with a small biarticulate appendage at the base. Three pairs of pleopoda present in male. Telson narrow, with the edges spinulose. Uropoda slender, inner branch longer than outer. Remarks. — This new genus is chiefly distinguished by the large vaulted carapace, sculptured with serrate crests, the total want of the eye, the strong development of the antennulae, and the absence of the usual palp on the first pair of maxillae, in which latter respect it differs from all hitherto known Cumacea. In the other anatomical details, it would seem to agree most closely with the genus Hemilamprops, though in some particulars also exhibiting a certain affinity to the genus Platyaspis. Only a single species of this genus is represented in the collection. 5. Paralamprops serrato-costata, n. sp. (PL II. figs. 6-13; PL III.). Specific Characters. — Form of body very slender. Carapace much larger than exposed part of trunk, sculptured by seven longitudinal serrate keels, one median running along the anterior half, two subdorsal on each side, occupying the posterior, somewhat gibbous part of the dorsal surface, and two lateral encircling the carapace at the sides and meeting in front. Pseudorostral projection short and somewhat upturned. The three anterior exposed segments of the trunk somewhat gibbous in the dorsal part and armed with small denticles, epimera produced anteriorly as an acute angle. Tail extremely slender and much longer than anterior division of body. Antennulae with the peduncle densely setose; flagella almost as long as peduncle, inner triarticulate, outer six-articulate. Penultimate joint of antennae in female with outer edge denticu- late. First pair of legs about as long as anterior division of body, the three outer joints very slender, penultimate the longest; second pair nearly as long as first, antepenulti- mate joint strongly spinose at inner edge. Telson shorter than scape of uropoda, slightly ddated at base, outer part sublinear, with about nine pairs of lateral denticles, the last one much larger than the rest, apex with three strong spines of equal size. Uropoda REPORT ON THE CUMACEA. 2/ very elongate, scape narrow and densely spinulose at inner edge, inner branch about as long as scape, first joint almost twice the length of the two others taken together, all three joints densely spinulose at inner edge ; outer branch with first joint very small, second elongate and slender, slightly tapering, with short bristles on both edges. Length of female, 12 mm. Remarks. — The present species being the only one as yet known, it is rather difficult to point out the characters which should be regarded as strictly specific. Meanwhile, I have thought it right to enumerate some structural details, which, from analogy with related genera, may be assigned the value of specific characters. Description. — Of this form I have been able to examine numerous specimens, both males and females. None of them, however, seem to have attained to sexual maturity, as there are neither ovigerous specimens, nor males with fully-developed antennae and pleopoda. The length of the largest female specimen is about 12 mm., that of the males somewhat less. The form of the body (see PI. II. figs. 6-8) is extremely slender and elongate, with the anterior division well defined from the posterior, and of a somewhat ovoid form. The carapace is of very considerable 'size, much larger than the exposed part of the trunk. It is somewhat expanded laterally and distinctly sculptured by longitudinal serrate keels, of which the following may be distinguished : — one median, running along the anterior part of the dorsal surface and terminating at the end of the frontal lobe ; two subdorsal keels on each side of the strongly arched branchial regions, the inner one the shorter and curved, occupying the most prominent part of that region and confined to the posterior half of the carapace, the outer keel reaching anteriorly to the base of the frontal lobe ; finally, on each side a lateral keel horizontally encircling the carapace and continued anteriorly along the edges of the pseudorostral projection to its very apex. There is also a more or less distinct but very small and interrupted keel, also serrate, on the posterior part of the carapace between the lateral and outer subdorsal keels. Beneath the lateral keel, on the ventrally inflected parts of the carapace, a slight longi- tudinal keel may moreover be noticed (see PI. III. fig. 1), but this keel, unlike the others, is quite smooth. The pseudorostral projection is very short and somewhat upturned; it is not limited beneath by any distinct notches, and there are no projecting angles marking off the inferior from the anterior edges of the carapace. As seen from above (PI. II. fig. 6) the carapace appears rather broad, on account of the expanded sides, its breadth being considerably greater than its height : anteriorly it is evenly rounded, without any distinctly marked projection in the middle. The ocular lobe (see PL III. fig. 2) is very small, and of a short conical form, without any trace of pigment or visual elements ; it does not extend, as in the above described species of the genera Cyclaspis and Vaunthompsonia , along the dorsal side of the pseudorostral projection, the lateral lobes of which meet in front of it. 28 THE VOYAGE OF H.M.S. CHALLENGER. The exposed part of the trunk is composed of five well-defined segments, the three anterior of which are somewhat gibbous dorsally and provided above with a short transverse row of small denticles. The epimera of all the segments are distinctly marked and laterally extended ; those of the three anterior segments produced anteriorly as an acute angle. Of the segments, the second is the largest, the last being very small, scarcely broader than the caudal segments. The tail is extremely slender and elongate, almost twice as long as the anterior division of the body. Its segments are cylindrical, with the posterior margin slightly emarginated above and below. They taper somewhat posteriorly and increase successively in length to the penultimate, which is very elongate. The last segment is much shorter than the others, somewhat depressed and expanded at the end. The integuments are not very strong, though somewhat more indurated than in the other forms belonging to this family, and they exhibit a slight squamous structure. The colour — to judge from a fresh specimen mounted in Canada balsam — is uniformly yellowish, without any pigmentary spots or ramifications. The antennuke (PL II. fig. 9) are remarkable for their unusual size, being almost as long as the carapace, and also for the great development of the flagella. The peduncle is rather elongate, and as usual, composed of three distinctly defined joints. The basal joint is about as long as the two others taken together, slightly curved, and provided at the end with several strong, partly ciliated bristles, and a few small denticles. The two succeeding joints are scarcely narrower than the basal and densely setose, especially towards the inner edge ; some of the setae are ciliated, some simple and very slender. The second joint, like the basal, is finely denticulate at the end, and somewhat larger than the third. Of the flagella, the outer one is almost as long as the peduncle, and composed of no less than six distinctly defined joints — a quite unusual number — each provided at the end with slender simple bristles. Of the joints, the first and penultimate are the longest, whereas the last joint is very small and might be easily overlooked. Besides the bristles, this flagellum bears at the tip two ribbon-like, densely articulated, sensory appendages, one of which originates from the end of the penultimate joint, the other from the last joint. The inner flagellum, which in the Cumacea generally presents a more or less rudimentary state, is rather fully developed, only slightly shorter than the outer, and composed of three slender joints of about equal size, and beset with bristles. In the young male the antennulae (see PL III. fig. 14) do not differ materially from those in the female, except by the peduncle being slightly thickened, and by the first joint of the outer flagellum being somewhat expanded at the base; in all probability in the adult male this joint bears a bunch of sensory appendages, as in the males of several other Cumacea. The antennas in the female (PL II, fig. 10 ; PL III. fig. 1, a2) are less rudimentary than in most other Cumacea, forming a slender stem composed of five distinctly defined joints. Of these the four first represent the peduncle, and the last the flagellum. REPORT ON THE CUMACEA. 29 The peduncle is geniculate in the middle, the two outer joints almost forming a right angle with the two inner ones. The third joint is the smallest and provided with two strong hairy setae and a little dentiform projection. The fourth joint is linear and armed along the outer edge with a row of five small denticles. The nagellum forms a very slender and narrow joint, about half as long as the peduncle, and provided at the tip with four delicate auditory bristles. In the young male the antennae (see PL III. fig. 14) exhibit the usual appearance, being strongly geniculate at the base and reflexed beneath the carapace. The proximal part of the peduncle is provided with four strong ciliated setae curving anteriorly; the distal part of the peduncle is smooth and composed of two well-defined segments, the outer one the larger. The nagellum has the appearance of a densely annulated vermiform appendage, slightly flexuous, and as yet quite smooth, terminating in an obtuse point. The anterior lip (PI. II. fig. 11) forms a semi-oval fleshy prominence, slightly emarginate at the tip, with the free edge densely ciliated. The posterior lip (fig. 12) is quite membranous, its lateral lobes rounded, triangular, and terminating in a somewhat incurved, densely ciliated corner. The inner edges of the lobes are also ciliated throughout their whole length. The mandibles (fig. 13) are well developed and similar in structure to those in the genus Cyclaspis, as described above, though somewhat stouter and with the body com- paratively shorter. The first pair of maxillae (PI. III. fig. 3) are markedly distinguished by the total want of the usual reflexed palp, of which not even the slightest trace could be found by the most careful dissection of several specimens, both males and females. No other form of Cumacea hitherto known wants this part. In other respects, however, these maxillae do not exhibit any marked peculiarity in their structure. The second pair of rnaxdlse (fig. 4) present quite the usual appearance. The maxillipeds (fig. 1, mp ; figs. 5, 6) are rather short and stout. The basal joint, as usual, juts out at the end on the inner side as a linguiform masticatory lobe, having along the inner edge a row of ciliated bristles, and at the tip several short dentiform spines. The meral and carpal joints are broad and laminar, the former provided at the outer edge with a strong ciliated seta, the latter with two somewhat smaller ones ; the inner edge of this joint is also provided with a row of curved spinules, which are denticulated at one of their borders in a comb-like manner (fig. 5, a), as also with numerous simple bristles. The propodal joint becomes suddenly much narrower and strongly incurved, bearing at the end internally two ciliated setae, and externally a bunch of somewhat more slender bristles. The terminal joint is very small and linear, with three small spines at the tip. The structure of the branchial apparatus in this form (see fig. G) would seem to corroborate the supposition set forth by the author at an earlier date, but not adopted by 30 THE VOYAGE OF H.M.S. CHALLENGER. Dr. Boas, viz., that both the exopodite and endopodite take part in the formation of this peculiar apparatus. For in this species it is clearly composed of two different parts, well defined from each other, the posterior part apparently representing the epipodite, the anterior the exopodite. The epipodite forms an irregular, oblong, membranous plate, broadest anteriorly, where it exhibits a freely projecting acute corner, and gradually tapering posteriorly, its extremity forming a lanceolate flap curving anteriorly in the form of a hook. To the inner face of this plate are affixed six digitiform gill-lobules pointing forwards, five of which form a row along a narrow inflexed lamella proceeding from the inner edge, whereas the sixth occurs at about the middle line of the plate. The anterior division of the branchial apparatus, which in my opinion represents the highly modified exopodite, forms an anteriorly directed narrow plate, suddenly constricted in front of the middle to a narrow semicylindrical neck, and terminating in a triangular expanded part, which is found protruding from beneath the pseudorostral projection. The whole of this division only serves as an efferent duct for the water introduced into the branchial cavity, and in the living animal probably does not partake in the rhythmical movements of the epipodal division. The first pair of gnathopoda (fig. l,gnl; fig. 7) are rather slender and have the ischial joint distinctly defined, though very small. At the end of the large and somewhat curved basal joint there occur at both edges several strong plumose setae. The carpal joint is densely setose on the inner edge and considerably larger than the propodal joint, which is slightly dilated at the end, where it -is densely beset with bristles. The terminal joint is narrow, conical, and provided with a strong apical spine. At the base of these gnathopoda in the female the usual setose lamella was distinctly developed, though the setae were still very short and rudimentary. The second pair of gnathopoda (fig. 1, giv; fig. 7 bin) are, as usual, provided with well-developed natatory exopodites, the terminal part of which is composed of eight setiferous articulations. In other respects the endopodite or stem has much the same appearance as that of the first pair of gnathopoda, though somewhat more elongate and with the basal joint slightly tapering towards the end, and bearing along the inner edge a dense series of strong plumose setae. The first pair of legs (fig. 8) are rather slender, and, when fully extended, about as long as the whole anterior division of the body. The basal joint, occupying somewhat more than one-third of the length of the leg, is strongly curved and setose along the whole of the inner edge and the distal half of the outer. The three succeeding joints exhibit a similar relation in length to each other as in the gnathopoda, though somewhat more slender, and only sparingly beset with bristles. The two next joints, on the other hand, are much more elongate, and form with each other and with the preceding joint distinct geniculate bends. The penultimate or propodal joint is considerably longer than the terminal, which is very narrow and linear ; both joints are furnished, especially on REPORT ON THE CUMACEA. 31 the inner edge, with scattered, slender setse, three of which arise from the tip of the terminal joint. The exopodite is of the same structure as that of the second pair of gnathopoda, except that the terminal part has one joint more ; the basal part is provided with six slender, ciliated bristles along the outer edge. The second pair of legs (fig. 9) are not very much shorter than the first, and are like- wise very slender and composed of the same number of joints. The basal joint is almost straight and not so densely beset with bristles as that of the first pair. The carpal joint is rather elongate, linear, and armed along the inner edge with a dense row of short spines. The propodal joint, on the other hand, is very small, and constricted at the base, whereas the terminal joint is elongate, linear, and provided with scattered spine-like bristles ; these two joints, taken together, about equal in length the carpal joint. These legs are in both sexes provided with well-developed natatoiy exopodites of the same structure as those on the first pair. The three posterior pairs of legs (figs. 10, 11) successively decrease in length, and have the basal joint very slender, and longer than all the others combined. In the female all these legs at first sight appear simple, without any exopodites, but on closer examination a very small and narrow appendage is found affixed to the basal joint of the two anterior pairs (see fig. 10), as in the genera Lawn/props and Hemilamprops. This appendage (fig. 10, a), which undoubtedly represents a rudimentary exopodite, is composed of two distinctly defined joints of nearly equal length, the last of which is provided with four simple bristles. In the male the two anterior pairs of these legs (fig. 15) have fully- developed natatory exopodites of the same structure as those of the preceding pairs, and the basal joint exhibits a slight dilatation at the base to receive the muscles moving the exopodite. The last pair of legs (fig. 1 1 ) are in both sexes simple and much smaller than the preceding pairs, both the basal and meral joints being considerably shorter. In most of the female specimens the incubatory lamella? forming the marsupial pouch are in the course of development at the bases of the second pair of gnathopoda and the three anterior pairs of legs, though they are still rather small, so as not to meet in the middle line (see fig. 1 ). Only three pairs of pleopoda are present in the male, as in the genera Hemila/mprops and Platyaspis, and they belong to the three anterior caudal segments (see PL II. fig. 8). They were not fully developed in the specimens examined (PI. HI. fig. 16), though exhibiting all their parts distinctly defined. The basal part or scape is narrow, quadrangular, and still without any trace of bristles or spines. The terminal plates (fig. 17) are very short, and each of them is provided at the tip with four short and thick bristles. The outer plate is distinctly biarticulate, whereas the inner is uniarticulate ; the latter is somewhat broader than the outer one, and its external edge juts out in the middle as a narrow conical process bearing two auditory bristles. 32 THE VOYAGE OF H.M.S. CHALLENGER. The telson (see fig. 12), as in the other forms belonging to this family, is distinctly developed, and defined from the last caudal segment by a well-marked transverse suture. It is about twice as long as the last caudal segment and very narrow, almost linear, with the base very slightly dilated to receive the terminal part of the intestine. Its distal part is armed on each edge with a regular series of about ten denticles, the last of which (see fig. 13) is much longer and stronger than the rest, and precisely of the same appearance as the three apical spines, which are quite uniform in size. The uropoda (see fig. 12) are extremely slender and more than half as long as the tail. The scape is considerably longer than the telson, narrow and cylindrical, with the inner edge in the greater part of its length armed with a dense row of somewhat unequal- sized spines. Of the terminal branches the inner one is the longer, nearly equalling in length the scape, and composed of three distinctly defined joints, which are spinulose at their inner edges. The first joint of this branch is by far the largest, being almost twice as long as the two others taken together ; of the latter, the terminal is the smaller one. The outer branch about equals in length the two first joints of the inner taken together, and is only biarticulate, the first joint being very short, and the last slender, and gradually tapering towards the tip, with a series of very delicate bristles on both edges, those of the inner edge being more elongate. From the tip of this branch, more- over, a somewhat stronger seta arises. Habitat. — Numerous specimens of this form were collected on January 29, 1874, off Cumberland Bay, Kerguelen Islands, from a depth of 127 fathoms (Station 149h). A sample of sifted mud from the same locality, kindly sent me by Mr. John Murray, also contained, besides numerous small Amphipoda and Isopoda, several specimens of this Cumacean, which thus seems to have occurred in great profusion at this place. Family IV. Leuconid.e. This family exhibits some very marked anatomical characters, distinguishing it from all other known Cumacea. Thus the mandibles are quite peculiar in structure, and the other masticatory parts also present certain well-marked differences from the usual structure. Of the legs, the three anterior pairs in the female are provided with fully- developed natatory exopodites, as in the Vaunthompsoniidte, but only two pairs of pleopoda are present in the male. The tail in all the forms is very slender and highly movable, but completely wants the telson, whereas the uropoda are powerfully developed. None of the forms show any trace of an eye, and their habits would seem to be chiefly fossorial, the second pair of legs having, in accordance therewith, an unusually strong develop- ment. REPORT ON THE CUMACEA. 33 The family at jjresent comprises three genera, viz., Leucon, Kroyer, Eudorella, Norman, and Eudorellopsis, G. 0. Sars, the two first of which are represented in the Challeuger collection. These two genera are easily recognised by the structure of the carapace, as follows : — Carapace compressed, with a dorsal serrate crest; pseudorostral projection prominent, antero-lateral corners of carapace more or less produced, .... very short, truncated in front, dorsal surface smooth, pseudorostral projection Leucon, Kroyer. Eudorella, Norman. Genus 1. Leucon, Kroyer, 1846. Leucon, Kroyer, Nat. Hist. Tidsskr., Bd. ii. Heft. ii. p. 208. Generic Characters. — Body very slender. Carapace compressed, with a dorsal serrate crest (sometimes obsolete in male) ; pseudorostral projection prominent ; antero-lateral angles of carapace more or less produced. Antennula? of usual structure, not geniculate, inner flagellum mostly rudimentary ; antenna? in female triarticulate, in male attaining the length of the body. First pair of legs elongate, gradually tapering, clothed with ciliated bristles; terminal joint compressed and armed at the tip with curved spines; second pair rather powerful, ischial joint obsolete, terminal more or less compressed and armed with stiff radiating spines. Last caudal segment slightly produced at the end. Uropoda rather powerful, scape cylindrical, both branches biarticulate, inner spinose, outer setose. Remarks. — This genus is chiefly distinguished by the form of the carapace, which exhibits a well-marked dorsal crest, more or less strongly denticulate, and a distinctly prominent, often very elongate, pseudorostral projection. In anatomical details it exhibits great affinity to the genus Eudorella, except that the antennulas are less powerful in structure, and not, as in that genus, geniculate. The genus at present comprises, besides the two new forms described below, nine species, seven of which are from the Northern Ocean, one from the Mediterranean, and one from the Atlantic. The two species procured during the Challenger Expedition are easily distinguished by the following characters : — Pseudorostral projection triangular, comparatively short, scarcely longer than frontal lobe. Dorsal crest of carapace densely serrate. Inner branch of uropoda shorter than outer, ....... conical, very slender and elongate, slightly ascending. Dorsal crest of carapace with scattered denticles. Inner branch of uropoda longer than outer, ....... Leucon assimilis, n. sp. Leucon tenuirodris, n. sp. (ZOOL. CHALL. EXP. — PART LV. — 1887.) Iii 5 34 THE VOYAGE OF H.M.S. CHALLENGER. 6. Lexicon assimilis, n. sp. (PI. IV.). Specific Characters. — Form of body moderately slender. Carapace with the antero- lateral corners triangularly produced, dorsal crest densely serrate, three of the denticles placed near the posterior margin at some distance from the others ; pseudorostral projection triangular, scarcely occupying one-fourth of the length of the carapace. Inner flagellum of antennulae very small and knob-like. First pair of legs slender, with penultimate and antepenultimate joints subequal ; second pair with terminal joint narrow and linear. Uropoda with inner branch shorter than outer, last joint scarcely half as long as first, both joints strongly spinulose on the inner edge. Length, 9 mm. Remarks. — This form is very nearly allied to the northern species, Lexicon nasicus, Kroyer, from which it may, however, be distinguished by the somewhat different form of the pseudorostral projection, which is comparatively shorter and simply triangular, not as in that form obliquely truncate at the tip. Descrijition. — The form of the body (PI. IV. figs. 1, 2), as in most of the species, is rather slender, though somewhat less so than in Lexicon nasicus, and slightly compressed, especially in the anterior part, with no sharply-marked definition between the anterior and posterior divisions. The carapace, occupying about half the length of the anterior division, is rather deep and distinctly compressed, with a well-marked crest running along the dorsal side. This crest is somewhat arcuate in the anterior part and densely denticulate, the denticles having the form of regular serratures, from eighteen to twenty in number, the three posterior of which are removed from the rest and placed at a short distance from the posterior margin of the carapace. Between each pair of serratures a small hair is affixed. The pseudorostral projection scarcely occupies one-fourth of the length of the carapace, and is triangularly pointed (see fig. 3), with the upper edge quite smooth and horizontal, the lower edges finely denticulate in the outer part and provided with slender bristles. The fissure limiting the frontal lobe is distinctly marked, and just above it on each side there are three small denticles. The end of the frontal lobe does not exhibit any distinct ocular prominence, and the eye is wholly wanting, as in the other species of the genus. The antero-lateral corners of .the carapace are distinctly projecting and triangular in form ; between them and the pseudorostral projection the edges of the carapace are evenly incurved and finely denticulate. The inferior edges of the carapace are strongly serrate in the anterior part, and posteriorly curve slightly upwards to the rounded postero-lateral corners. Behind the carapace five distinctly defined segments belonging to the trunk are exposed, all of them having broad laminar epimera slightly projecting laterally. Of these segments the first is very narrow, the others about equal in length, but successively diminishing in height, so that the last segment is not higher than the caudal segments, REPORT ON THE CUMACEA. 35 to which it, moreover, seems to bear a closer relation in so far that it is much more movably articulated to the penultimate segment of the trunk than to the first caudal segment. The tail is rather slender and somewhat longer than the anterior division of the body. Its segments are regulai'ly cylindrical, with the posterior edge somewhat more emarginated ventrally than dorsally, and forming on each side an obtuse angle. Of the segments the penultimate as usual is the longest. The last segment is rather small and forms posteriorly a rounded prominence arching over the anal orifice. The integuments are thin and pellucid, without any distinct sculpturing. The colour — to judge from a specimen mounted in Canada balsam — is uniformly light 5rellowish, without any pigmentary deposits. The antennulae (see fig. 1) are seen projecting in front from the sinus between the pseudorostral projection and the anterodateral corners of the carapace. They are of moderate size (see fig. 4) and have the joints of the peduncle successively diminishing a little both in length and thickness. At the end of each joint there are several slender setse, which on the middle joint form a complete whorl around its end. Of the flagella (see fig. 5), the inner one is very small and knobdike, bearing, besides three simple un- equal setae, a single very delicate auditory bristle (fig. 6) with only two diverging cilia at the tip ; a similar bristle is also found affixed to the end of the peduncle internally (see fig. 5). The outer flagellum scarcely exceeds in length the last joint of the peduncle, and is, as usual, composed of three joints, the first being twice as long as the other two taken together. The last joint is very small and bears, besides a few simple bristles, two ribbondike and closely annulated sensory appendages. The antenna? of the young male dissected (see fig. 4) are still imperfectly developed, scarcely reaching the length of the anterior division of the body, but yet exhibiting a distinct definition of the peduncle from the flagellum. The peduncle is geniculate at the base, with the proximal part irregularly constricted and bearing three anteriorly curving ciliated bristles, and the distal part cylindrical and exhibiting an imperfect articulation anterior to the middle, but stdl without any trace of bristles. The terminal part of the antenna forms a densely annulated sheath, gradually tapering to the tip, and within this sheath the developing filiform flagellum may be traced. The anterior lip (fig. 7) has the form of a small rounded prominence, the posterior part of which is somewhat dilated and slightly emarginated at the free edge, which is finely ciliated. The posterior lip (fig. 9) is very delicate and difficult to separate from the adjoining parts, but seems to exhibit the usual structure. The lateral lobes are very thin and membranous, with the inner corner somewhat projecting and incurved, but without any teeth, whereas a delicate ciliation is found along the inner and part of the outer edges of the lobes. 36 THE VOYAGE OF H.M.S. CHALLENGER. The mandibles (fig. 8) are comparatively small, and exhibit a structure rather differing from that in the above described Cumacea. Thus the body of the mandibles is not, as in these forms, navicular, but almost triangular, with two projecting angles, the outer pointed, the inner obtusely rounded. By the outer angle the mandible is fixed to the inside of the carapace at the point where the fissure limiting the frontal lobe terminates ; to the inner the strong rotatory muscles are affixed, whose points of insertion are visible as a number of lucid spots in the middle part of the carapace, symmetrically arranged on each side of the middle line (see fig. 2). The anterior branch of the mandibles is comparatively short and strongly incurved, with the cutting edge divided into two comparatively simple, dentiform projections ; instead of the regular comb-like series of curved spines occurring in most other Cumacea behind the cutting edge, only two simple bristles are found on each mandible affixed close together. The posterior branch is very thick and massive, with a large fluted molar surface. The first pair of maxilla? (fig. 10) exhibit the usual structure, with this exception, however, that the posteriorly bent palp has but a single terminal seta. The second pair of maxillae (fig. 11) are comparatively small and want the dense series of delicate bristles affixed to the inner edge of the basal part in other Cumacea, instead of which there are only found three hook-like spines. The maxillipeds (fig. 1 2) are likewise rather feeble in structure, the basal part being comparatively short and forming a more or less pronounced angle with the termiual part. The masticatory lobe arising from the basal part is almost quadrangular, and bears at the base a very strong anteriorly curved seta. Of the joints belonging to the terminal part, the antepenultimate is, as usual, the largest, and exhibits along the inner edge a series of short curved spines, provided at one of their edges with a few slender denticles (fig. 14). Moreover, this joint has on the lower side numerous fine hair dike bristles, partly arranged in fascicles, and at the outer corner a very long anteriorly curving seta, densely ciliate at the outer part. Two similar, but somewhat shorter, setae are also found originating from the succeeding joint. The last joint (see fig. 13) is very small, sublinear, and provided at the tip with three bristles, one of which is quite short but furnished with a few very long cilia on both edges. The branchial apparatus (see fig. 15) affixed to the base of these limbs, is less fully developed than in most other Cumacea. The epipodite-plate is rather narrow and has only five small digitiform gill-lobules, the largest of which is affixed at some distance from the inner edge and points posteriorly. The anteriorly directed narrow division of the apparatus, representing the modified exopodite, is considerably longer than the epipodite-plate and terminates in a somewhat dilated lanceolate plate bordered by a very thin and pellucid membrane. In conjunction with the corresponding plate this part forms an elongated tube, which admits of being exserted from the tip of the pseudo- rostral projection, and serves for the expulsion of the water introduced into the branchial REPORT 0>Q THE CUMACEA. 37 cavities. During the rhythmical movements of the epipodite-plates, this tube remains immovably exserted, but is immediately withdrawn when the movements are interrupted. The first pair of gnathopoda (fig. 16) do not exhibit any peculiarity in their structure. The second pair of gnathopoda (fig. 17), on the other hand, exhibit a form rather different from that in most other Cumacea. Thus, the basal part is rather broad and only very slightly longer than the terminal part. It has several ciliated bristles along the inner edge, and from its external corner four very strong anteriorly curving seta? originate, the outer of which especially is of quite unusual size and densely ciliate. The terminal part is provided with comparatively short bristles, with the exception of its second joint, which has externally a strong anteriorly curving seta similar to those arising from the outer corner of the basal part. The exopodite is rather powerful, with the basal part compressed, oblong, and the terminal composed of five articulations. The first pair of legs (fig. 18) are about equal in length to the carapace and the two anterior exposed segments of the trunk taken together, and are rather slender, tapering gradually from the base to the tip. The basal joint is, as usual, somewhat dilated in its proximal part, where it contains the strong muscles moving the exopodite, and is fringed along almost the whole inner edge, and the distal part of the outer, with ciliated bristles ; moreover, a row of short spines occurs along the distal part at some distance from the outer edge. The terminal part of the leg is much longer than the basal, and is furnished with scattered plumose setas, especially along the outer edge. Of the joints the antepenultimate and penultimate are the longest and about equal in size. The last joint is slightly compressed but rather narrow, and armed at the tip with about six strong, claw-like spines. The exopodite has much the same appearance as that of the second pair of gnathopoda, except that its terminal part has one joint more. The second pair of legs (fig. 19) are much shorter than the first, but rather stout, with the basal and terminal parts about equal in length. The ischial joint seems to be quite wanting. The meral joint is rather thick and armed at the end on the inner side with a strong spine. The carpal joint is about twice as long, but much narrower, and has a similar, though somewhat smaller spine at the end. The propodal joint is quite short and unarmed, whereas the terminal is rather elongate, linear, and furnished with numerous stiff spine-like bristles, forming a spreading bunch at the end of the joint. The exopodite does not differ from that of the first pair of legs. The third pair of legs (fig. 20) in both sexes are furnished with well-developed natatory exopodites. The basal part of the endopodite is more than twice as long as the terminal, and, as in the two preceding pairs, rather dilated to receive the strong muscles moving the exopodite ; along both edges there are a number of ciliated seta?. Of the joints composing the terminal part, the three first slightly increase in size, and '68 THE VOYAGE OF H.M.S. CHALLENGER. from the third (the carpal) three strong setae originate, the distal part of which is densely annulated and curved at the end in a hook-like manner. The propodal joint is much smaller and has a single seta of similar structure, and from the tip of the very minute terminal joint a slender spine originates, besides two very small bristles. The two posterior pairs of legs in the female are simple, whereas in the male the penultimate pair (fig. 21) have well-developed natatory exopodites, as on the anterior pairs. The last pair of legs (fig. 22) are quite alike in both sexes and very small, with the basal joint narrow, cylindrical, and about as long as the terminal part. Of pleopoda only two pairs are present in the male, affixed to the two anterior caudal segments (see fig. 1). These limbs (fig. 23) are not fully developed in the specimens examined, though having the basal part and the two branches distinctly defined from each other. The latter (fig. 24) are very short, and each of them is provided at the tip with four still unciliated setae ; the outer branch is biarticulate, whereas the inner is uniarticulate, and, as in the male of Paralamprops, juts out in the middle to an angular projection bearing a single auditory bristle. The uropoda (see fig. 25) are rather stout and about as long as the two posterior caudal segments taken together. The scape is cylindrical and furnished with scattered spiniform bristles, especially along the inner edge. Of the branches, the inner is the shorter and about as long as the scape. It is composed of two well-defined joints, the first of which is more than twice as Ions; as the last, and armed along the inner edge with about twelve spines, three of which are much longer than the others ; the last joint is rather narrow and has five equal-sized spines along the inner edge, and two somewhat larger ones arising from the tip. The outer branch is also biarticulate, but the first joint is here very short, whereas the last is elongate and furnished along the inner edge with a row of about ten ciliated setae, besides a few scattered bristles along; the outer edge and on the upper side. Habitat. — A few specimens of this form were taken January 29, 1874, off Cumberland Bay, Kerguelen Islands, from a depth of 127 fathoms (Station 149h). 7. Lexicon tenuirostris, n. sp. (PL V. figs. 1-4). Specific Characters. — Form of body extremely slender. Carapace with the antero- lateral corners only very slightly produced, almost rectangular ; anterior edges above the corners denticulate, with a small sinus in the middle ; dorsal crest armed with only seven denticles arranged along its anterior half; pseudorostral projection very elongate and slender, almost as long as the carapace, slightly ascending and regularly tapering to the end. Inner flagellum of the antennulse about as long as the first joint of outer. The two anterior pairs of legs nearly as in the preceding species ; the three posterior REPORT ON THE CUMACEA. 39 pairs successively decreasing in length and densely beset with ciliated bristles. Uropoda with inner branch longer than outer, first joint scarcely twice as long as second, both armed with scattered denticles at inner edge. Length, 10 mm. Remarks. — The present species is nearly related to Leucon longirostris, described by the author from an imperfect specimen procured during the Swedish Expedition in the frigate " Josephine." It may, however, easdy be distinguished by the rather different form of the pseudorostral projection, the length of which in proportion to the carapace is also much greater. Description. — The sole specimen procured is a young female, in which the marsupial pouch has not yet been developed. Its length is about 10 mm. The form of the body (see PI. V. figs. 1, 2) is extremely slender, more so than in any other of the known species. The carapace (see fig. 3) is distinctly compressed and of the usual form, with the dorsal crest well marked, but denticulate only in its anterior half, the denticles being rather few in number (seven), and hence more widely separated from each other. The pseudorostral projection is quite unusually elongate and slender, almost equalling the carapace in length, and tapers gradually to the tip, which is narrowly truncate. It is somewhat ascending and has the lateral edges provided with a few scattered bristles and a single small denticle behind the middle. The anterodateral corners of the carapace are less produced than in the preceding species and almost rectangular. The anterior edges joining the pseudorostral projection are strongly denticulate and have a distinct rounded sinus below the middle, from the bottom of which the rudimentary antennae are seen to project. The inferior edges of the carapace form a slight angle at the middle and are, as usual, strongly denticulate in the anterior half. The five exposed segments of the trunk taper successively posteriorly, and the last of them does not exceed the caudal segments, either in height or in breadth. The tail is considerably longer than the anterior division of the body and very slender. Its segments increase successively in length to the penultimate, which is by far the longest. The last segment (see fig. 4), on the other hand, is quite short and almost pentagonal, with the end produced to an obtusely triangular projection arching over the anal orifice. The antennulae (see fig. 3) are comparatively rather large, though not nearly reaching to the tip of the pseudorostral projection. They are more especially distinguished by the unusual development of the inner flagellum, which in this case cannot properly be said to be rudimentary, since it fully equals in length the first joint of the outer one ; it is, however, as usual, composed of only a single joint, which bears at the tip several stiff bristles. The two anterior pairs of legs (see fig. 1) do not seem to differ materially from those of the preceding species. This is also the case with the succeeding legs, with this difference, however, that they are much more richly supplied with ciliated bristles. 40 THE VOYAGE OF H.M.S. CHALLENGER. The uropoda (see fig. 4) are comparatively a little longer than in the preceding species and have the scape cylindrical, with but four denticles along the inner edge. Of the branches, in this species the inner one is the larger and about as long as the scape. Its first joint is scarcely twice as long as the last, slightly tapering, and provided along the inner edge with only five denticles. The last joint is linear and has three denticles along the inner edge and a much stronger one at the tip, besides another very small apical denticle and a simple slender bristle. The outer branch is considerably shorter than the inner and has the last joint provided with a series of ciliated setse along the inner edge. Habitat. — The above described specimen was taken from a very considerable depth in the North Pacific, at Station 246, July 2, 1875; lat. 36° 10' N., long. 178° E.; depth, 2050 fathoms; Globigerina ooze; bottom temperature, 35°*1. Genus 2. Eudorella, Norman, 1866. Eudorella Norman, Eep. Brit. Assoc, 1866, p. 197, note. Euilora, Sp. Bate, Ann. and Mag. Nat. Hist., ser. 2, vol. xvii. p. 457. Generic Characters. — Anterior division of body more or less dilated, almost clavate. Carapace comparatively small, not crested dorsally, truncated in front, without any pseudorostral projection, the lateral lobes ascending perpendicularly and meeting immediately in front of the median lobe, leaving between their extremities an oval orifice for the expulsion of the water. No eye. Five segments of the trunk exposed behind the carapace, rapidly diminishing in height posteriorly. Tail very slender and movable. Antennulae rather large, geniculate, outer part being reflexed and doubled upon the inner by an elbow-like flexure ; inner flagellum well developed, uniarticulate, with three strong apical bristles. Antennae in female very small, uniarticulate. Oral parts and legs nearly as in Lexicon. Telson wanting. Uropoda with both branches biarticulate. Remarks. — As regards the structure of the several limbs, this genus exhibits great affinity to the genus Leucon, though it is easily distinguished by the very different form of the carapace, which gives to the species of the present genus quite a peculiar appear- ance. The strong development and peculiar curvature of the antennulas affords another good generic mark. No less than seven different species of this genus have hitherto been distinguished. Of these, three occur in the Northern Ocean, one in the Arctic Sea. two off the eastern coast of North America and two in the Mediterranean, one of which is also found in the Northern Ocean. Besides these, an eighth species, which is described below, was procured during the Challenger Expedition. REPORT ON THE CTJMACEA. 41 8. Eudorella abyssi, n. sp. (PL V. figs. 5-12). Specific Characters. — Body moderately slender and beset with short scattered hairs. Carapace almost perpendicularly truncate in front, with the upper part rounded ; anterior edges below the middle with a distinct rounded sinus limited by two obtuse projections, the upper with three, the lower with five teeth ; anterolateral corners obtusely angled. First pair of legs, when fully extended, about as long as anterior division of body, rather stout, and provided along outer edge of terminal part with elongate ciliated setae; last joint compressed and armed with numerous claw-like spines; second pair with terminal joint sublinear and provided with stiff radiating spines. Uropoda about as long as the two last caudal segments taken together, inner branch much the longer, with a regular row of short spinules along the inner edge, last joint very small, armed at the tip with a strong spine and a slender bristle. Length, 9 mm. Remarks. — In its general appearance this species is rather similar to the northern form Eudorella truncatula, but may readily be distinguished by the rounded form of the sinus on the anterior edges of the carapace, and by the different number of teeth occurring on the obtuse projections, which limit that sinus. Description. — The sole specimen procured is an adult female with greatly developed marsupial pouch. Its length is about 9 mm. The form of the body (see PL V. figs. 5, 6) is moderately slender, with the anterior division considerably dilated, and, as seen from the side, almost club-shaped. The carapace is comparatively short, scarcely occupying half the length of the anterior division, and, as seen from the side, nearly quadrangular, the front part being almost perpendicularly truncated, without any trace of a rostrum-like projection. Not- withstanding this, the lateral lobes, usually forming the pseudorostral projection, are present, but instead of projecting in front of the median lobe as usual, they ascend perpendicularly and terminate close to that lobe, leaving between their extremities an oval orifice (see fig. 6), through which the water is expelled from the branchial cavities. The dorsal line of the carapace is almost horizontal and without any trace of a keel. The inferior edges are densely serrate in their anterior half, whereas the posterior part is smooth and obliquely ascending. The anterior edges (see fig. 7) are in the upper part quite smooth, but below the middle there occur two dentate projections limiting a rather deep and rounded sinus, from the bottom of which the antennae project. Of these pro- jections the upper is armed with three, and the lower with five strong teeth. The antero-lateral corners of the carapace are obtusely angled, and armed with an anteriorly- directed tooth, forming a continuation of the series of teeth occurring along the inferior edges. Not the slightest trace of an eye can be detected, nor is any distinct ocular lobe present. (ZOOL. CHALL. EXP. — PAET LY. — 1887.) IK 6 42 THE VOYAGE OF H.M.S. CHALLENGER. The five segments of the trunk exposed behind the carapace rapidly diminish in height posteriorly, and have all rather broad laminar epimera, somewhat projecting laterally (see fig. 6). The first segment is very narrow, almost band-like, and the last does not exceed the succeeding caudal segment either in height or in breadth. The tail is very slender, considerably longer than the anterior division of the body. The form and relative length of the segments are very nearly as in Leucon. As in that genus the last segment is rather short and obtusely produced posteriorly. The integuments are rather thin and without any distinct sculpturing, but beset with scattered short hairs, especially on the tail. The antennulaa (see fig. 5) exhibit the structure and peculiar flexure characteristic of the genus. They are rather stout and affixed to a fleshy segment lying immediately in front of the upper part of the carapace, from which segment they, as it were, hang down, forming a sharp elbow-like bend at the middle. This bend occurs between the two outer joints of the peduncle, in such a manner that the last joint together with the flagella is doubled upon the proximal part of the antennula, pointing obliquely upwards. Of the flagella the inner one is fully as long as the first joint of the outer, but only composed of a single joint bearing three strong spine-like bristles at the tip. The outer flagellum is triarticulate and provided with several strong ciliated bristles, especially along the edges of the first joint ; at the tip this flagellum bears the usual two ribbon-like sensory appendages, which, however, are rather short, scarcely exceeding the simple apical bristles in length. The first pair of legs (fig. 8) are rather strong, projecting considerably in front of the carapace, and when fully extended about as long as the whole anterior division of the body. The basal joint is slightly curved and furnished on both edges of the distal part with several ciliated bristles. The terminal part of the leg is considerably longer than the basal and exhibits the usual well marked geniculate bends ; it is furnished along the outer edge with scattered and rather long setse, which are finely ciliate. Of the joints the penultimate is the longest, whereas the terminal joint is comparatively short and much compressed, with numerous claw-like spines at the obliquely truncated apex. The exopodite does not reach to the end of the basal joint and has the terminal part divided into five setiferous articulations. The second pair of legs (see fig. 5) are comparatively stronger than in Leucon, but as in that genus only sex-articulate, the ischial joint not being distinctly defined. The terminal joint is almost linear and provided with a number of diverging spines, arranged in a double row. The third pair of legs (ibid.) have the exopodite well developed and of the same structure as in the two preceding pairs, whereas the two posterior pairs (figs. 10, 11) are simple and rather small, especially the last pair. The uropoda (see fig. 12) are rather strong, fully as long as the two last caudal REPORT ON THE CTJMACEA. 43 segments taken together, with the scape cylindrical and armed with a row of about twelve small spines along the inner edge. Both branches are Particulate, as in the genus Leucon, and on the whole agree fairly well in structure with those in that genus. The inner branch is much the longer, slightly tapering towards the tip and armed with a regular row of denticles along the inner edge. Its terminal joint is very short and has at the tip a strong spine and an elongate seta, besides four small denticles along the inner edge. The outer branch is scarcely longer than the first joint of the inner, and has the basal joint rather short and obliquely truncate at the end, whereas the terminal joint is elongate and somewhat compressed, with a row of about ten strong ciliated setae along the inner edge and at the tip. Habitat. — The above described specimen was taken with the trawl from a very considerable depth in the North Atlantic, off the coast of the United States. Station 45, May 3, 1873 ; lat. 38° 34' N., long. 72° 10' W.; depth, 1240 fathoms; bottom, blue mud ; bottom temperature, 37° "2. Family V. Diastylice. The Cumacea referred to this family are easily known by the tumid and ovate form of the anterior division of the body, which is sharply marked off from the slender and very mobile tail ; also by the large, superiorly strongly arched carapace terminating in front in a more or less prominent rostriform projection, and finally by the distinctly defined, lageniform telson, and the great length of the scape of the uropoda. The family at present comprises only two genera, viz., Diastylis, Say, and Leptoshjlis, G. 0. Sars. Of these only the first is represented in the Challenger collection. Genus Diastylis, Say, 1817. Diastylis, Say, Joura. Acad. Nat. ScL Philadelphia, 1817. Condylura, Latreille, Le Regne Animal, t. iv. p. 153, 1829 (not Illiger). Alauna, Goodsir, Edin. New Phil. Journ., vol. xxxiv. p. 130, 1843. Cuma, Krdyer, Nat. Hist. Tidsskr., Bd. iii. p. 530, 1841 (not Milne-Edwards). Generic Characters. — Form and armature of body generally rather dissimilar in the two sexes, adult male always much more slender than female. Carapace large and tumid, mostly spinous, antero-lateral corners obsolete, pseudorostral projection triangu- larly pointed. Eye present or wanting. Antennula? slender, peduncle only slightly thickened in male. Antennae in female four to five-articulate, in male very fully developed, attaining the length of the body, distal part of peduncle composed of a single large segment with numerous fascicles of small sensory appendages along the outer edge, flagellum very slender, with elongate articulations. First pair of legs more 44 THE VOYAGE OF H.M.S. CHALLENGER. or less elongate, with the three outer joints very slender ; second pair considerably shorter and rapidly tapering toward the apex ; third and fourth pairs in female without basal appendages. Pleopoda of male distinctly biramose, outer branch biarticulate, inner uniarticulate. Third and fourth caudal segments in male with four strong ciliated setse in place of pleopoda. Telson rather large, with the terminal part more or less produced and in the male forming a distinct angle with the basal part. Uropoda slender, with the scape elongate, branches comparatively short. Remarks. — This genus is chiefly distinguished from the genus Leptostylis by the antennulae being far less dissimilar in the two sexes, by the much fuller development of the male antenna?, and by the third and fourth pairs of legs in the female wanting the basal appendage (rudimentary exopodite) present in the former genus ; finally, by the somewhat different form of the caudal appendages. The genus is very abundantly represented, especially in the Northern Ocean, no less than twenty-seven distinct species having been hitherto recorded. The Challenger collection contains four species, one of which has been formerly described, whereas the three others are now added, increasing the number of species to thirty-one. The four Challenger species may be easily known by the armature of the carapace and the form of the telson, as follows : — Carapace - spinous; the spines subequal, very minute. Telson with terminal part much longer than basal, and densely denticulate at the edges, rather strong. Telson with outer part much shorter than basal, and without lateral denticles, .... unequal ; some much larger than the others. Telson comparatively short, with terminal and basal part about equal in length, ...... finely scabrous, with a serrate crest on each side running obliquely backwards from tip of pseudorostral projection. Telson with terminal part shorter than basal and armed with but one pair of lateral denticles, ......... Diastylis stygia, G. 0. Sars. Diastylis erinaceus, n. sp. Diastylis horricla, n. sp. Diastylis mystacina, n. sp. 9. Diastylis stygia, G. 0. Sars (Pis. VI, VII., VIII.). Diastylis stygia, G. 0. Sars, Om Cumaceer fra de store Dybder i Nordishavet, K. Svensk. Vetensk. Akad. Handl., voL xi. p. 6, pi. ii. figs. 4^7. Specific Characters. — Carapace about twice as long as the exposed part of trunk, moderately tumid and armed (in female) with numerous small subequal spinules, more crowded in the anterior part ; pseudorostral projection conical, scarcely longer than frontal lobe ; inferior edges of carapace denticulate in the anterior part only. No eye. Anterior margin of the three first exposed segments of trunk finely denticulate. Tail in both sexes REPORT ON THE CUMACEA. 45 smooth, without denticles. First pair of legs scarcely longer than carapace, with the three outer joints subequal ; second pair with antepenultimate joint longer than the two outer taken together. Telson rather elongate-, with terminal part much longer than basal, and densely spinous at both edges ; apical spines not larger than lateral. Uropoda with scape very slender, branches subequal, about half as long as scape, inner distinctly triarticulate. Length of female, 16 mm., of adult male, 21 mm. Remarks. — The present species was briefly described by the author in the year 1873, from a single young female specimen obtained during the Swedish Spitzbergen Expedition in 1868, from the very considerable depth of 2600 fathoms. During the Challenger Expedition several specimens, both females and adult males, were procured, and I thus have had an opportunity of making a closer examination of this form, to show the very striking differences between the two sexes, as regards both the general appearance and several of the anatomical details. Description of the Female. — The length of the adult specimens is about 16 mm., and the present form thus belongs to the larger species of the order. The form of the body (see PI. VI. figs. 1, 2) is rather stout, with the anterior division considerably dilated, nearly ovoid in form, and sharply marked off from the slender tail. The carapace is very large and deep, about twice as long as the exposed part of the trunk, and nearly as broad as high. The dorsal surface is strongly vaulted in its posterior part, declining anteriorly to the base of the pseudorostral projection. The latter is conical in form and of moderate length, about as long as the frontal lobe. The inferior edges of the carapace are somewhat irregularly arched, without forming any distinct antero-lateral corners ; they are finely serrate in the anterior part, and join the posterior edge without being defined from it by any distinct angle. The surface of the carapace appears everywhere evenly arched, without any folds or keels, but is beset with numerous small equal-sized spinules, more densely crowded in the anterior part. The exposed part of the trunk consists of five well-defined segments, rapidly decreasing in height posteriorly. The first segment is very narrow, almost band-like, and the three succeeding segments only slightly increase in length, all these segments being firmly connected with each other. The last segment is very small and rather movably connected with the preceding, a space covered by thin skin being interposed dorsally between the two segments. On all the segments the lateral parts form small rounded epimeral plates, slightly projecting laterally. Those of the last segment are somewhat produced posteriorly, without, however, forming acuminate projections as in the adult male. The first three segments have the anterior margin very finely denticulate, and this is also the case with the last segment, which also bears two small dorsal denticles. The tail is scarcely longer than the anterior division of the body, but very slender and cylindrical, with the posterior margin of the segments strongly emarginate both dorsally and ventrally, and forming on each side a slight projecting angle. The three 46 THE VOYAGE OF H.M.S. CHALLENGER. anterior segments are nearly equal in size, whereas the two succeeding somewhat increase in length. The last segment is a little shorter than the preceding, slightly expanded at the end and somewhat depressed. All the segments are quite smooth, without any trace of spines or denticles. The integuments, as in the other species of the genus, are rather strong and calcareous, exhibiting under the microscope a finely reticulate structure. The colour of the body is uniformly whitish, without any pigmentary deposits. No trace of a distinctly developed eye can be detected, though a small rounded ocular lobe is present at the end of the median lobe (frontal lobe) (see fig. 2). The antennulse (fig. 5, a1) are rather slender, projecting considerably beyond the tip of the pseudorostral projection. They are composed of a triarticulate peduncle, and two well-defined flagella. The first joint of the peduncle is very large, considerably longer than the other two taken together, and also much broader. It is somewhat flattened and finely ciliate along the inner edge ; its end is obliquely truncate, and bears at the inner projecting corner a very strong and densely hairy seta pointing anteriorly, and at the outer a much smaller unciliated bristle, somewhat curved outwards. Of the two outer joints of the peduncle, the first is the shorter, and has at the end internally a ciliated seta of moderate size. The flagella are very unequal in size, the inner one being very small, scarcely half as long as the last joint of the peduncle. Nevertheless, this flagellum is found to be composed of three well-defined articulations, the middle one of wilich is the longest ; at the tip of the last joint three unequal bristles are affixed. The outer flagellum is about three times as long as the inner, and composed of four articulations, the three first of which are nearly equal in size, whereas the last is very small, and might easily be overlooked. Besides two or three simple bristles, this flagellum bears at the tip two slender ribbon-like sensory appendages, each divided into a number of short regular segments. The antennas (fig. 5, a2) have the form of simple, slightly tapering stems, about half as long as the antennular peduncle, and somewhat curved outwards and forwards. Each stem is composed of four well-defined articulations, and bears five strong, anteriorly curved, plumose setse, two of which arise from the first articulation ; of these setae the one affixed to the penultimate articulation is considerably shorter than the others. The anterior lip (fig. 5, L) constitutes a rounded lobe projecting immediately behind the antennal segment. The free edge of this lobe is slightly emarginate in the middle and finely ciliated. The posterior lip (fig. 7) is considerably larger, but quite membranous, and, as usual, consist,- of two lateral lobes united at the base. The lobes are irregularly oval and have the inner corner slightly incurved, but without any teeth, whereas a fine ciliation is found along the whole inner edge and part of the outer. The mandibles (fig. 6) are strongly developed and rather similar in structure to those REPORT ON THE CUMACEA. 47 in the genera Cyclaspis and Paralamprops, as described above. The strong adductor muscles filling the concavity of the body are connected by a broad chitinous tendon, to which two other diverging muscles are affixed posteriorly. The rotatory muscles constitute several strong bundles arising from the inside of the carapace dorsally, and joining successively a very long and slender tendon, which is affixed to the inner edge of the body immediately behind the molar process (cf PI. VIII. fig. 1 ). The first pair of maxillas (fig. 8) exhibit the usual two masticatory lobes, springing from a rather thick and muscular basal part. The refiexed palp is distinctly developed, and about as long as the outer masticatory lobe. It bears at the tip two diverging setae of somewhat unequal length. The second pair of maxillas (fig. 9) agree on the whole rather well in structure with those in the genera Cyclaspis and Paralamprops, as described above. The maxillipeds (PI. VII. fig. 2) have the basal part rather large, and provided with a ridge running along the lower surface and strongly denticulate in its anterior part. The masticatory lobe exhibits the usual structure and has a row of ciliated setae continued along the inner edge of the basal part. Of the joints composing the terminal part, the second (carpal) is, as usual, the largest and rather expanded ; it has at the outer corner a very strong and elongate plumose seta pointing anteriorly, and along the inner edge a row of short, slightly curved spines, which are densely ciliated in the middle (fig. 4), besides numerous, partly plumose bristles. The jjenultimate joint is much smaller, and provided at the end with two strong anteriorly curving setae. The terminal joint, finally (fig. 3), is conical, with two unequal spines at the tip, and one or two simple hair-like bristles. The branchial apparatus affixed to these limbs (fig. 1) is very fully developed, and the epipodite-plate, especially, is very large and of a distinctly navicular form. The gill- lobules are digitiform, and arranged in a slightly spiral series along a thin flexuous lamella issuing from the inner edge of the epipodite-plate, and partly doubled over it. They rapidly increase in length from behind forwards, the posterior lobules being very small and knob-like ; from the bottom of the concave surface of the epipodite-plate, moreover, a single somewhat larger gill-lobule, pointing anteriorly, arises. The exopodite is scarcely half as long as the epipodite, and constitutes a narrow anteriorly directed plate, broadest at the base and gradually tapering to a narrow, somewhat curved neck, which again expands into a triangular lamella bordered by a thin narrow membrane, The inner straight edge of this lamella lies in close contact with that of the lamella of the other side, both together forming an imperfect tube, which can be exserted beyond the tip of the pseudorostral projection (cf PI. VIII. fig. 1), and serves for the expulsion of the water introduced into the branchial cavities. The first pair of gnathopoda (fig. 5) are rather slender and exhibit at the base, affixed to the short coxal joint, a quadrangular lamella, to the posterior edge of which a dense 48 THE VOYAGE OF H.M.S. CHALLENGER. row of very slender unciliated setae are affixed. These setae, only found in the adult female, successively increase in length from within outwards, and, together with those on the other side, form a broad fan projecting within the marsupial pouch. The basal joint is much longer than the other joints taken together, and exhibits on the lower side a denticulate ridge similar to that on the maxillipeds. At the end this joint bears some strong ciliated setae, and along the outer edge a regular series of bristles. The terminal part apparently consists of but four joints, the ischial joint not being distinctly defined. Of these joints the antepenultimate (carpal) is by far the longest and bears at the outer corner, like the preceding joint, a ciliated seta of moderate length, and along the inner edge a dense row of smaller setae. The penultimate joint is also densely beset with bristles along the inner edge, and has also an unusually strong and elongated seta affixed to the outer edge close to its base. The terminal joint, finally, is quite small, conical in form, and provided with an apical claw-like spine besides some small bristles. The second pair of gnathopoda (fig. 6) are much larger, and possess the full number of joints. The basal joint is more than twice as long as the terminal part, and strongly curved. Along the outer half of the inner edge there is a dense series of ciliated bristles, and from the outer slightly projecting corner four very strong anteriorly directed plumose setae arise. On the lower side this joint has a few scattered denticles, and at the inner corner a somewhat stronger dentiform spine. Of the joints composing the terminal part, the penultimate (propodal) is the longest and rather slender, whereas the terminal is quite short and armed at the tip with several curved spines. The exopodite affixed to the basal joint of these limbs scarcely exceeds half the length of this joint, and has the terminal part divided into six setiferous articulations. The first pair of legs (fig. 7) are, when fully extended, about as long as the carapace. The basal joint is very large, slightly tapering towards the end, and strongly curved at the middle, with a dense series of ciliated bristles along the outer part of both the outer and inner edges. Moreover, this joint is armed with several strong denticles, somewhat irregularly arranged on the lower side, some of them, however, forming a regular series along an elevated ridge close to the outer edge. Of the joints composing the terminal part, the two first are rather short, whereas the three outer are very slender and almost equal in length, forming with each other strong geniculate bends. The exopodite is somewhat larger than that of the last pair of gnathopoda, and, moreover, differs by the basal part being denticulate and setiferous along the outer edge, and by the terminal part having two additional joints. The second pair of legs (fig. 8) do not nearly attain the length of the first pair, and taper rapidly towards the end. The basal joint is scarcely more than half as long as that of the preceding pair, and only slightly curved. It exhibits along the lower side REPORT ON THE CUMACEA. 49 an elevated ridge, which in the anterior part is armed with four strong denticles and some ciliated bristles issuing between the denticles. Along the inner edge, moreover, a series of rather strong, cdiated setse are present, increasing in length anteriorly. The ischial joint is distinctly defined though very short, and armed on the inner edge with a strong denticle. From the inner edge of the succeeding joint two similar denticles arise. The antepenultimate joint is very slender and elongate, considerably longer than the two outer joints taken together, and only provided with a few simple bristles. The terminal joint is a little longer than the penultimate and conical in form, with some hair-like bristles, the one issuing from the tip being the longest. The exopodite is exactly of the same structure as that of the first pair of legs. The three posterior pairs of legs (figs. 9—11) are all simple, without any trace of exopodites, and of essentially the same structure. Of the joints the basal is by far the largest, in the two anterior pairs even longer than all the others taken together. The ischial joint is very short, scarcely longer than broad, whereas the two succeeding joints are more elongate and about equal in size. The two last joints are suddenly very much smaller, and taken together scarcely as long as the preceding joint. These legs are rather richly supplied with ciliated bristles, esjsecially along the inner edge. But there is a number of bristles issuing from the outer part of the legs, which differ essentially from the rest in being considerably stronger, unciliated, and terminating in a finely annulated, somewhat curved point. Five such peculiarly modified bristles are present on each leg, four of which arise from the outer corner of the antepenultimate joint, and the fifth from the penultimate one. At the tip of the terminal joint is affixed a claw- Hke spine and a simple bristle. As to the relative size of these legs, the two anterior pairs are nearly of the same length, whereas the last pair (fig. 11) are considerably smaller, chiefly on account of the shortening of the basal joint. The marsupial pouch, as in other Cumacea, is composed of four pairs of large plates arising from the bases of the second pair of gnathopoda and the three anterior pairs of legs. They were fully developed in only one of the specimens, which was somewhat crushed, and therefore selected for dissection. The marsupial pouch in this specimen contained a number of embryos, of which I have figured one in four different aspects (figs. 14—17), to be described below. The telson (see fig. 12) is distinctly defined and about twice as long as the last caudal segment. It has a rather narrow, lageniform shape, the basal part, containing the terminal portion of the intestine, being somewhat dilated and cylindrical, whereas the outer part is very narrow, gradually tapering, and nearly twice as long as the basal part. At the junction between the two, on the lower side, the anal orifice occurs. The basal part is quite smooth, whereas the terminal is armed at the edges with about nine pairs of denticles, somewhat irregularly arranged, and at the tip with two denticles of a quite similar appearance. (ZOOL. CHALL. EXP, PART LV. 1887.) Ill 7 50 THE VOYAGE OF H.M.S. CHALLENGER. The uropoda (ibid.) are very slender, about equalling in length the three last caudal segments taken together, and are articulated to the last caudal segment somewhat in front of the telson. The scape is narrow, cylindrical, slightly dilated at the end, and reaches to about the tip of the telson ; it exhibits a regular row of about sixteen spinules along the inner edge. The branches are subequal in length and scarcely more than half as long as the scape. The inner branch is composed of three distinctly defined joints, the first of which is a little longer than the other two taken together. All the joints are spinous at the inner edge, the number of spines in the specimen examined amounting to fifteen (eight on the first, four on the second, and three on the last joint) ; to the tip) of the branch, moreover, a somewhat longer spine and a simple bristle are affixed. The outer branch is only biarticulate, wdth the first joint quite short, the last very narrow and gradually tapering, with a few small hair dike bristles along the outer edge, and a somewhat longer seta issuing from the tip. Description of the Adult Male. — Whilst the young or not yet sexually mature males do not differ materially in their general appearance from the females, the adult males of this and most other species appear so very dissimilar to the females, as easily to be mistaken for distinct species, a fact that has caused much trouble in the synonymy, several spurious specimens having been introduced on account of this peculiar dissimi- larity between the sexes.1 The adult male of the present species attains a much larger size than the female, measuring 21 mm. in length. The form of the body (see PL VI. figs. 3, 4) appears much more slender, the anterior division being far less vaulted and more oblong in shape, and the tail also considerably more elongate. The carapace is about twice as long as deep, whereas its height does not nearly equal its breadth; it is consequently much more depressed than in the female, and its dorsal line appears also more evenly curved and less abruptly declining to the base of the pseudorostral projection. The latter is somewhat shorter than in the female, and exhibits above, at the base, a concave space, into the bottom of which the ocular lobe is received. On each side the anterior edges of the carapace form a somewhat projecting, broadly rounded lobe, arching over the base of the antennae. These lobes are serrate at the edge, and, moreover, beset with numerous short plumose bristles (see fig. 4). The inferior edges of the carapace exhibit in the middle a strong, almost angular curvature, and join the posterior margin at an obtuse angle. When viewed from above (fig. 4) the side-edges appear strongly arched, on account of the great development of the 1 The following is a list of such spurious species : — 1. Cuma angulata, Kroyer = Adult male of Diastylis rathkii (Kroyer). 2. Cuma brevirostris, Kroyer = Adult male oi Diastylis edwardsii (Kroyer). 3. Diastylis spinosa, Norman = Adult male of Diastylis echinata, Spence Bate. 4. Diastylis stritjata, Norman = Adult male oi Diastylis rugosa, G. 0. Sars. 5. Diastylis calweri, Norman = Adult male oi Diastylis biplicata, G. 0. Sars. REPORT ON THE CUMACEA. 51 branchial regions, and the posterior part of the carapace considerably narrower than in the female. As to the armature of the carapace, it differs materially from that in the female by the almost complete want of the spinules, of which only a slight trace can be observed in the most anterior part. Moreover, on the sides of the carapace, below the middle, a distinctly marked keel runs from the posterolateral corners forwards, limiting the greatly swollen branchial regions below. This keel, found in all adult males of the present genus, is in some species strongly denticulate throughout, and thus gives the carapace a stdl more anomalous appearance. The exposed part of the trunk is considerably narrower than in the female, and less vaulted, whereas the epimeral plates are more distinctly defined and setose at the edges; those of the last segment especially exhibit a rather peculiar form, each being produced posteriorly to an acute point, and bearing at the outer edge four very strong and densely plumose setse. Of the caudal segments, the two anterior are somewhat thicker than the others, and have, moreover, a serrate crest running along the ventral surface, the crest being also continued along the last segment of the trunk. These segments are otherwise quite smooth, without any dorsal denticles. The antennulse (PI. VIII. fig. 1, a1 ; fig. 2) are comparatively stouter than in the female, with the outer part of the peduncle less slender, and provided at the base of the outer flagellum with a dense fascicle of delicate sensory filaments. The flagella are nearly equal in length, and each of them has one joint more than in the female, a short articulation being marked off at their base. The antennae (fig. 1, a2 ; fig. 3) are greatly developed, and fully as long as the whole body. The peduncle is strongly geniculate, with the proximal part consisting of four short and somewhat irregular joints, connected together by rather oblique articulations; it bears at the base three strong ciliated bristles curving anteriorly, two of which issue from the second joint, and the third from the first. The distal reflexed part of the peduncle constitutes a single, very large and slightly compressed segment, gradually tapering to the apex, and provided along the outer edge with numerous fascicles of short sensory bristles. Within this segment several strong muscles are seen converging to the tip; by the aid of these muscles the flagellum is moved. The latter is extremely slender and composed of very elongate articulations, each provided at the outer edge with numerous fascicles of small sensory bristles of the same kind as those on the distal part of the peduncle (see fig. 4). The oral parts do not seem to differ materially from those in the female. The branchial apparatus, on the other hand, is much more fully developed (see fig. 1); the great number and peculiar arrangement of the gill -lobules being especially remarkable. The gill-lobules also differ considerably in form from those in the female, being distinctly lamellar, instead of digitiform (fig. 7). They are 52 THE VOYAGE OF H.M.S. CHALLENGER. densely crowded together, and arranged most regularly side by side in a row, forming a double spiral along the membranous lamella, doubled over tbe concavity of the epipodite- plate, so as anteriorly almost completely to fill ap the concavity. As in the female, they increase in size from behind forwards, but much more gradually (see fig. 6). Of the legs, all but the last pair are provided with strongly developed natatory exopo- dites, the basal part of which (see figs. 8— 11) is remarkably broad and lamellar. In regard also to the structure of the endopodal part of the legs some differences from that in the female may be found. Thus, in the two anterior pairs (figs. 8, 9) the basal joint is compara- tively larger and more dilated in the proximal part, to receive the strong muscles moving the exopodite. In the second pair (fig. 9) this joint is twisted in a peculiar manner, the ridge running along the lower side being more elevated and strongly flexuous. More- over, the penultimate joint of this pair of legs is markedly distinguished by its inner edge being strongly denticulate or exhibiting a serrate crest, a condition quite absent in the female. In the third and fourth pairs of legs (figs. 10, 11) the basal joint exhibits a form quite dissimilar to that in the female, not being simple and cylindrical, but considerably dilated in the proximal part, and at the base of the exopodite provided with a number of very strong hairy setae. The last pair of legs (fig. 12) nearly agree in the form and relation of the joints with those in the female, but are distinguished by two unusually strong and densely hairy setae, arising from the outer edge of the basal joint. There are two pairs of well-developed pleopoda affixed to the two anterior caudal segments (see PI. VI. fig. 3). They consist (PL VIII. figs. 13, 14) each of a narrow basal part divided into two unequal segments, and of two very short branches provided with long natatory setae. Of these branches (fig. 13, a) the outer one is distinctly biarticulate, and bears at the tip four seta?, whereas the inner is uniarticulate and provided with six setae, two of which are affixed to the inner edge; moreover, this branch exhibits at the outer edge, close to the apex, a small projection bearing two minute auditory bristles. Both pairs of pleopoda are essentially of the same structure, with this difference, however, that the basal part of the first jmir (fig. 13) is more richly supplied with bristles. At each side of the pleopoda there are on the two anterior segments a number of strong ciliated setae (see figs. 13, 14), and on the ventral surface of the two succeeding segments four strong setae likewise occur (see PL VI. fig. 3), as it were, in place of pleopoda. The telson (see PL VIII. fig. 15) is considerably more elongate than in the female, and, moreover, readily distinguished by the strong geniculate flexure occurring anterior to the middle (see also PL VI. fig. 3), whereby the basal part appears sharply marked off from the terminal part. The latter is armed on each side with about fifteen ciliated denticles, whereas in the female, as above stated, the number of denticles is only nine on each side. REPORT ON THE CUMACEA. 53 The uropoda (see PI. VIII. fig. 15) are very slender and about as long as the five posterior caudal segments taken together. The scape is remarkably elongate, and pro- vided at the inner edge with a great number of ciliated denticles, somewhat unequal in size. The branches exhibit a structure similar to that in the female, but the number of denticles occurring along the inner edge of the inner branch is far greater, being no less than twenty-seven, besides the terminal spine. Nervous System. — In dissecting two specimens of this species, a male and a female, I succeeded in isolating the greater part of the ventral ganglionic cord, and have figured the most anterior part of this cord of the female specimen (PI. VII. fig. 13), and the succeeding part from the male specimen (PL VIII. fig. 5). The central part of the nervous system in all Cumacea consists, besides the supracesophageal ganglion or brain (visible in PI. VIII. fig. 1), of sixteen ganglia, ten of which belong to the anterior division, and the six others to the tad. Of the former the three anterior (PL VII. fig. 13), innervating the oral parts, are closely crowded together and almost confluent, whereas the remaining seven ganglia (see PL VIII. fig. 5), belonging to the trunk, are widely separated and connected by rather long double commissures. The three posterior of these ganglia are, however, somewhat closer together than the four pre- ceding, and the last ganglion is also a little smaller. Between each pair of the com- missures connecting the six anterior ganglia of the trunk there is also a narrow median cord, which at first sight has the appearance of another central commissure con- necting the ganglia. On closer examination, however, this part is found to consist of a continuous vessel running along the dorsal surface of that part of the nervous cord, to which it is partly connected, and apparently representing the ventral artery. The caudal ganglia (ibid.) are far less developed and only slightly ddated, whereas the commissures are very long and slender. All the ganglia consist of two well-marked symmetrical halves and give origin on each side to two or three strong nerve-trunks, which soon ramify and enter the corresponding limbs. Another rather large nerve- trunk is seen to issue from the commissures themselves, and seems to innervate the musculature of the corresponding segment. The mandibles together with their com- plex muscles are innervated by two strong nerves issuing from the oral commissures at their junction with the ventral cord (see PL VII. fig. 13), each nerve dividing succes- sively into a great number of diverging branches (partly visible in PL VIII. fig. 1). Development. — As above stated, the female specimen dissected had the marsupial pouch filled with embryos. These were all in the same rather early stage of develop- ment (PL VII. figs. 14-17), merely forming rounded saccular bodies, lined by a thin pellucid cuticle (the egg-membrane). On a closer examination several of the limbs were found in their first stage of development, as rounded buds along the strongly curved ventral surface. On the opposite side a slight sinus occurs, indi- cating the dorsal curvature characteristic of the Cumacean embryos (fig. 14). The 54 THE VOYAGE OF H.M.S. CHALLENGER. tail (T) is as yet only represented by a slight rounded prominence limiting the above mentioned sinus posteriorly, and not yet exhibiting any trace of appendages. The anterior extremity of the embryo is considerably broader and more evenly rounded than the posterior, and exhibits most anteriorly two rounded areas (o), which, from analogy with the embryos of the Podophthalmia, may be regarded as the ocular plates, as yet wholly separated, but afterwards becoming confluent and at last cpiite obliterated. Behind these plates the anterior lip (L) can be distinctly traced, and further back the bipartite posterior lip or metastoma (I). On each side of the anterior lip two transverse prominences occur, representing the antennulae (a1) and antennae (a2), the first of which are the larger, and on each side of the posterior lip the mandibles (M) may be traced as two rounded knobs. Further back two pairs of prominences, somewhat different both from the preceding and succeeding ones, are seen, the anterior (m1) being unequally bipartite, the posterior («r) slightly tripartite at the free edge. These prominences represent the two pairs of maxillae. Then follows a regular double series of posteriorly pointing prominences, all of a quite similar appearance and unequally bilobed at the end. The number of these prominences is seven pairs,1 of which the first (mp) represent the maxillipeds, the two succeeding the two pairs of gnathopoda (gn1, gn2), and the four posterior (jp1- p^) the four anterior pairs of legs, the last pair not being formed until very late, even a considerable time after the young have left the marsupial pouch. The two rounded lobes which all of these pairs of limbs exhibit, may undoubtedly be regarded as the first indication of the two principal parts, the endopodite and exopodite, and it is rather striking that even the first pair of gnatho- poda, which never exhibit any trace of exopodites in the adult animal, do not differ in this respect from the following limbs, and that, moreover, in all the embryos the two posterior pairs (third and fourth pairs of legs) have also a similar bilobed form, although these limbs in the female are cpiite simple. Immediately above the oral region on each side a curved ridge or slight fold (C) may be traced, representing the first indication of the carapace, and within the area limited below by this fold a rounded cellular body (A) occurs, which ultimately becomes the liver, or caeca of the stomach. As may be seen from the above description of the embryo, the development of the Cumacea differs materially from that of the Mysidae and other Podopthalmia in two very important points; firstly, by the embryo exhibiting, while still enclosed within the egg-membrane, a well-marked dorsal curvature, and secondly, by the last pair of legs not being formed until a considerable time after the young have escaped from the marsupial pouch. In both these respects the Cumacea agree, on the other hand, very closely with the Isopoda. Habitat. — Eight more or less perfect specimens of this species were taken by the dredge in the North Atlantic, off Nova Scotia, from a very considerable depth. 1 In fig. 14, by a mistake, eight such pairs have been indicated instead of seven. REPORT ON THE CUMACEA. 55 Station 47, May 7, 1873 ; lat. 41° 14' N. long., 65° 45' W.; depth, 1340 fathoms; bottom, blue mud. The species has also been taken during the Norwegian North Atlantic Expedition in different places in the cold area. 10. Diastylis horrida, n. sp. (PI. IX. figs. 1—8). Specific Characters. — ? . Carapace not twice as long as the exposed part of trunk, slightly depressed ; surface somewhat irregularly grooved and armed all over with numerous spines, some of which are much larger than the others ; pseudorostral projection acute, a little longer than frontal lobe; inferior edges of carapace denticulate in their whole length. No distinct eye. Free segments of trunk with transverse rows of small spines ; epimera strongly denticulate at the edges. Tail more or less densely spiuulous. Antennula? with basal joint of peduncle armed at the tip with two strong spines, the smooth seta? affixed to its outer corner remarkably long. First pair of legs exceeding in length the carapace, terminal joint very slender and much longer than the preceding ; second pair with basal joint strongly spinous, antepenultimate joint equalling in length the two last taken together ; remaining legs rather slender, with basal joint more or less densely spinous. Telson of moderate length, terminal part scarcely longer than basal and densely denticulate at the edges ; apical spines much larger than laterals. Uropoda not very slender, branches unequal, inner shorter than outer, and composed of but two joints. Length, 11 mm. Remarks. — The present species is easily known by the body being armed all over with unequal spines, more densely crowded together on the carapace ; by the great length of the terminal joint of the first pair of legs, and finally by the structure of the caudal appendages. In general appearance it would seem to be most nearly related to the northern form, Diastylis lucifiera (Kroyer). Description. — Most of the specimens of this form contained in the collection are still immature, but there are one or two female specimens which seem to be nearly full-grown, since the marsupial pouch has been formed. The larger of these specimens measures about 11 mm. in length. The form of the body (see PI. IX. figs. 1, 2) is rather stout, nearly as in the northern form, Diastylis lucifiera, to which species it would seem to bear a rather close affinity. The anterior division of the body is oblongo-ovate, and of about the same length as the posterior, with the dorsal surface rather evenly vaulted. The carapace is comparatively smaller than in the preceding species, not nearly attaining twice the length of the exposed part of the trunk. It is somewhat depressed, the height not nearly equalling the breadth, and, as seen from the side (fig. 1), 56 THE VOYAGE OF H.M.S. CHALLENGER. of irregularly triangular form, with the dorsal line gradually declining anteriorly, and the posterior edge somewhat elevated in the middle. The pseudorostral projection is rather produced and sharply pointed, a little longer than the frontal lobe, and quite horizontal. The inferior edges of the carapace are strongly arcuate in the middle, and densely denticulate throughout ; posteriorly they join the posterior edge without forming any angle, and anteriorly they ascend more perpendicularly to the base of the pseudo- rostral projection. As seen from above (fig. 2) the posterior half of the carapace is of about uniform breadth, whereas the anterior rapidly tapers to a conical point. The surface of the carapace is somewhat irregularly grooved, a slight furrow running obliquely along each side from the frontal lobe backwards. It is densely armed all over with sharp anteriorly directed spines, some of which are much larger than the others, and rather regularly arranged on each side of the middle line, especially on the dorsal surface. In young specimens the difference in size between the spines is still more pronounced than in the adult. The eye would seem to be imperfectly developed, as neither pigment nor distinct refracting bodies could be detected within the rounded ocular lobe. The five exposed segments of the trunk are rather sharply defined and armed with scattered spines, forming more or less distinct transverse rows; on each segment two of the spines arranged in pairs dorsally are somewhat larger than the rest. The epimeral plates on all the segments are strongly denticulate at the edges; those of the last segment are obtusely rounded. The tail exhibits the usual slender form and the peculiar nodose shape of the seg- ments characteristic of the present genus. Like the rest of the body, it is more or less densely spinous, especially on the dorsal surface, the number of spines varying in different specimens. The antennulse (see fig. 3) are rather slender, with the two outer joints of the peduncle unusually elongate, and, taken together, considerably longer than the basal joint. The latter is armed on the lower surface at the extremity with two strong- spines, and exhibits, moreover, as in the preceding species, two setse, the inner pointing anteriorly, and densely hairy, the outer recurved and quite smooth, and longer than usual. Of the two outer joints of the peduncle, the first is by far the longer. The flagella are comparatively small, the outer scarcely attaining the length of the last joint of the peduncle, and the inner very minute, being about one-third as long. The antennae (ibid.) are also somewhat smaller than in the preceding species, each forming a regularly tapering stem divided into four articulations, each with a single ciliated seta at the end. The first pair of legs (fig. 4) are comparatively slender, and when fully extended, considerably exceeding the carapace in length. The small coxal joint is bordered with a row of strong denticles, and the curved and elongate basal joint also exhibits a number REPORT ON THE CUMACEA. 0/ of small denticles arranged along its outer part in three rows. The whole inner edge of this joint, as also the greater part of the outer, is moreover provided with numerous slender ciliated setas. The terminal part of the leg is considerably longer than the basal, and has the two first joints comparatively short, and the three outer joints very slender. Of the latter the two first are about equal in length, whereas the last joint is considerably more elongate, and provided along the inner edge with numerous short bristles, and at the tip with a number of curved claw-like spines. The exopodite exhibits the usual structure and has the basal part armed on the outer edge with a row of about eight spines, besides a few slender setae. The second pair of legs (fig. 5) are scarcely more than half as long as the first, and have the basal joint rather large, and armed on the lower side with numerous spines, five of which form a regular row along the outer edge. The ischial and meral joints have each two strong spines at the end. The carpal joint exhibits the usual slender form and is devoid of spines, whereas a row of simple bristles occur along its inner edge. The terminal joint is conical, and considerably longer than the propodal joint; both taken together are about as long as the carpal joint. The three posterior pairs of legs (figs. 6, 7) are rather slender, and more or less spinous in their proximal part. Of the joints composing the terminal part, the meral and carpal are the longest and about equal in length. The telson (see fig. 8) is not very elongate, scarcely more than twice as long as the last caudal segment. The basal part is regularly cylindrical, and in some specimens armed along the dorsal surface with a median row of spines. The terminal part does not exceed the basal in length, and tapers gradually to the tip, which is armed with two very elongate spines ; the edges are also densely spinous, but the spines are much smaller than those springing from the tip. The uropoda (ibid.) are about half as long as the tad, with the scape rather slender, and armed along the inner edge with about ten small denticles. The branches are comparatively short, and somewhat unequal in length, the inner being considerably shorter than the outer, and consisting of but two joints of nearly equal length. Both joints are spinulose along the inner edge, the number of spines in the specimen dissected being seven, besides the strong spine issuing from the tip of the branch. The outer branch is more slender than the inner, and exhibits at the tip three equal -sized spiniform setae, besides a row of very small bristles along the outer edge. Habitat. — Of the present form several specimens were collected on January 29, 1874, off Cumberland Bay, Kerguelen Islands, from a depth of 127 fathoms, Station 129h. (ZOOL. CHALL. EXP. — PART LV. — 1887.) Iii 8 58 THE VOYAGE OF H.M.S. CHALLENGER. 11. Diastylis erinaceus, n. sp. (PL IX. figs. 9-14). Specific Characters. — % . Carapace very large and tumid, surface evenly vaulted, and armed with numerous equal-sized spines ; pseudorostral projection of moderate length; inferior edges of carapace denticulate in the anterior part only. No distinct eye. Exposed part of trunk smooth, not attaining half the length of the carapace. Tail very slender, slightly spinulose, the three anterior segments each provided dorsally with a pair of somewhat stronger spines and four ciliated setse. Antennulse with the basal joint armed with a longitudinal row of spines, outer edge minutely serrate; second joint nearly as long as basal; third much smaller. Antennae somewhat irregularly twisted, with six strong cdiated setae arising from the middle part, and two unciliated, curved setae at the extremity. First pair of legs very elongate, with basal joint strongly spinous, carpal and propodal joints exceedingly elongate and nearly subequal, terminal joint much smaller. Second pair of legs with carpal joint twice as long as the two last joints taken together. Meral joint of the three posterior pairs of legs much longer than carpal. Telson rather elongate, basal part regularly cylindrical and much longer than terminal ; the latter gradually tapering, with two small apical denticles ; lateral edge of telson unarmed. Uropoda comparatively small, with the scape shorter than telson. Length, 13 mm. Remarks. — The present species would seem to be very nearly allied to Diastylis josephinas, described by the author, from the Swedish Atlantic Expedition. It differs, however, by the carapace being relatively more tumid, and by its spines being equal-sized, whereas in Diastylis josephinas some of the spines are much longer than the rest. Moreover, the exposed part of the trunk is not spinous as is the case in that species, and the telson, though exhibiting a very similar form, wants the numerous denticles found along the edges of the basal part in Diastylis josephinw. Description. — Of this form I have only had an opportunity of examining a single female specimen, and the broken anterior part of another. The length of the perfect specimen is about 13 mm. The form of the body (see PI. IX. figs. 9, 10) is that characteristic of the genus, the anterior division being considerably dilated and sharply marked off from the slender tail. The carapace is very large and tumid, considerably broader than high, and more than twice as long as the exposed part of the trunk, with the dorsal surface considerably arched and gradually declining both anteriorly and posteriorly. The pseudorostral projection is of moderate length, horizontally projecting, with the upper surface somewhat depressed at the base, and the tip sharply pointed. The inferior edges of the carapace are strongly curved at the middle, and finely denticulate in their anterior part. As seen from above (fig. 10), the greatest breadth of the carapace occurs anterior REPORT ON THE CTJHACEA. 59 to the middle, the posterior part being slightly narrowed. The surface of the carapace is rather evenly vaulted, without any keels or folds, but armed all over with numerous spines, pointing for the most part anteriorly. These spines are all of nearly equal size and, as in the preceding species, also continued along the lateral parts of the pseudorostral projection. The eye seems to be quite wanting, and the ocular lobe is rather small and depressed. The exposed part of the trunk is considerably narrower than the carapace, and quite smooth ; the epimeral plates are evenly rounded and Ukewise unarmed. The tail is very narrow, and armed with scattered spines both laterally and dorsally. The first segment has a strongly serrate keel running along the ventral surface, and moreover exhibits dorsally, as do also the two succeeding segments, two somewhat stronger spines arranged in pairs, and four cdiated setse issuing from the posterior edge. The antennulaj (fig. 11) are rather slender, with the basal joint of moderate length, and armed with a longitudinal row of strong spines running along the lower side towards the inner corner, to which a strong ciliated seta is affixed ; the inner edge of this joint is minutely denticulate in the anterior part, but wants the usual unciliated bristle at the end. The second joint of the peduncle is very slender and almost as long as the basal, whereas the third joint is much smaller. The flagella exhibit the usual structure. The antennas (figs. 12, 13) present at first sight a structure rather different from that in other species, the joints being twisted in a peculiar manner so as only with difficulty to be counted. Moreover, the structure and arrangement of the seta? are rather anomalous. Thus, the two outer setae have assumed the character of strongly chitinous filaments curved in a peculiar manner, and quite devoid of cilia, whereas the six other setae are densely ciliated and crowded together in the middle part of the antenna, all but one apparently issuing from a single joint (the second). The first pair of legs (see fig. 9) are of very considerable length, exceeding, when fully extended, half the length of the body. The basal joint is strongly spinous and densely supplied with ciliated setaa. The carpal and propodal joints are exceedingly slender and elongate, and about equal in size, whereas the terminal is much shorter. The second pair of legs have the carpal joint very slender and nearly twice as long as the two last joints taken together. The two succeeding pairs of legs are rather powerfully developed, about equal in length, and with the meral joint unusually large, and more than twice as long as the carpal joint. The last pair of legs are much smaller and of the usual structure. The telson (fig. 14) is rather large, about as long as the two last caudal segments taken together, and exhibits in the greater part of its length a regularly cylindrical form, the posterior third part only projecting beyond the anal orifice, and constituting the terminal division. The latter is somewhat flattened and regularly tapering towards 60 THE VOYAGE OF H.M.S. CHALLENGER. the apex, which is narrowly truncate and armed with two small denticles. Excepting these apical denticles, the telson is quite unarmed, and only provided with a few very minute hairdike bristles at the edges. The uropoda were partly broken in the specimen examined. They seem to have been comparatively small, the scape not nearly attaining the length of the telson. Habitat. — The above described specimen, as also the broken example mentioned above, was taken in the North Atlantic from a very considerable depth. Station 79, July 11, 1873 ; kit. 36° 21' N, long. 23° 31' W.; depth, 2025 fathoms ; bottom, Globigerina ooze ; bottom temperature, 35°-9. 12. Diastylis mystacina, n. sp. (PI. X. figs. 1—3). Specific Characters. — ? . Anterior division of body greatly swollen. Carapace exceedingly large and deep, with the doi'sal surface boldly vaulted in the middle ; pseudorostral projection of moderate length, bordered on each side by an elevated and strongly dentate crest, continued along the anterior part of the carapace as a slightly flexuous serrate keel ; inferior edges of carapace densely serrate in the anterior half; surface finely scabrous from very small microscopic spikes. No distinct eye. Anterior edge of the two first exposed segments of ti-unk minutely denticulate ; penultimate segment with two longitudinal serrate keels dorsally. Epimeral plates of the anterior segments finely serrate at the edges. Last segment of trunk and the three anterior caudal segments each with a pair of small dorsal denticles. First pair of legs of moderate length, terminal joint nearly as long as the preceding. The three posterior pairs of legs very slender, with meral and carpal joints about equal in size. Telson with terminal part as long as basal, slightly tapering and armed with but one pair of lateral denticles. Uropoda comparatively small, inner branch shorter than outer, and composed of two articulations only. Length, 10 mm. Remarks. — This very distinct species seems to be most nearly related to Diastylis insignis, described by the author from the Swedish Atlantic Expedition, though easdy dis- tinguished by the peculiar dentate crest on each side of the anterior part of the carapace, as also by the want of the strong lateral projections occurring on the carapace of that species. Description. — The sole specimen contained in the collection is an adult female with fully developed marsupial pouch. It measures about 10 mm. in length. The general form of the body (see PI. X. figs. 1, 2) is rather similar to that of Diastylis insignis, the anterior division being greatly swollen and very sharply marked off from the slender tail. The carapace is exceedingly large and deep, about twice as long as the exposed part of the trunk, with the dorsal surface boldly arched in the middle, and declining rather abruptly posteriorly, and more gradually anteriorly. The pseudorostral projection is of REPORT ON THE CUMACEA. 61 moderate length and deeply excavated above, being bordered on each side by an elevated crest, which is continued along the sides of the anterior part of the carapace. These crests are somewhat flexuous and strongly dentate in their anterior part, some of the denticles being considerably larger than the rest, and the two anterior occupying the tip of the pseudorostral projection. The inferior edges of the carapace are strongly arcuate in the middle, and finely serrate along their anterior half. As seen from above (fig. 2), the carapace appears very tumid, the greatest breadth occurring a little posterior to the middle, and distinctly exceeding the height. The surface is evenly vaulted, and everywhere minutely scabrous from numerous very small microscopic spikes. The eye seems to be quite wanting, though the rounded ocular lobe is distinctly seen at the end of the bell-shaped frontal lobe (see fig. 2). The exposed segments of the trunk are rather sharply marked off from each other, their dorsal part being more or less elevated. The anterior edge of the two first segments is very finely denticulate. The antepenultimate segment has a slight transverse keel above, armed with a few small denticles, and on the dorsal surface of the penultimate segment two short longitudinal serrate keels occur. The last segment is very small and provided with a pair of dorsal denticles. The epimeral plates of the four anterior segments are evenly rounded and finely denticulate ; those of the last segment are obtusely produced posteriorly. The tail is very slender, and about as long as the anterior division of the body. It is quite smooth, with the exception of a pair of small dorsal denticles occurring on each of the three anterior segments. The first pair of legs (see fig. 1) are of moderate length, with the basal joint strongly spinous, the three outer joints rather slender and slightly decreasing in length. The second pair of legs have also the basal joint densely armed with spines and seem on the whole to agree in form with those of the preceding species. The three posterior pairs of legs are very slender, with the meral and carpal joints subequal. The telson (see fig. 3) about equals in length the two last caudal segments taken together. Its basal part is regularly cylindrical ; the terminal part, projecting beyond the anal orifice, is nearly of the same length, and gradually tapers to the apex, which bears the two usual spines. Of lateral spines there are in the specimen examined only a single pair at a short distance from the apex, besides a very minute denticle occurring on the left side immediately in front of the lateral spine. Theuropoda (ibid.) are not very large, scarcely exceeding the length of the three posterior caudal segments taken together. The scape does not nearly attain the length of the telson, and wants every trace of denticles on the inner edge. The branches are rather unequal in size, the inner one being considerably shorter than the outer, and about half as long as the scape. This branch is composed of but two joints of about 62 THE VOYAGE OF H.M.S. CHALLENGEK. equal length, the last conical in form and terminating in a strong spine ; of lateral denticles there are only three (one on the first and two on the last joint). The outer branch exhibits the usual form, its terminal joint being rather slender and provided with a few small hairdike bristles on both edges, and a somewhat stronger apical seta. Habitat. — The above described specimen was taken together with the preceding form at Station 79, July 11, 1873; lat. 36° 21' N., long. 23° 31' W.; depth, 2025 fathoms ; bottom, Globigerina ooze ; bottom temperature, 35°'9. Family VI. Cumellid^e. The Cumacea comprised within this family are generally very small, and at once distinguished by the very slender form of the three posterior pairs of legs, which more- over terminate in a strong curved claw. They also differ considerably from the Cumacea belonging to the preceding families, by the male wanting every trace of pleopoda. The family comprises two distinct genera, easily recognisable by the very different structure of the visual organ, viz., Cumella, G. 0. Sars, and Nannastacus, Spence Bate. Moreover, I am inclined to regard the genus Spencebatea, established by Norman, on a single male specimen procured by the " Porcupine " expedition, as belonging to this family. Of these three genera only the genus Nannastacus is represented in the Challenger collection. Genus Nannastacus, Spence Bate, 1865. Nannastacus, Spence Bate, Carcinological Gleanings, No. 1, Ann. and Mag. Nat. Hist., 1865. Diops, Paulisona, Tzsljedovanija rakoobraznjikh krasnago raorja, p. 128, 1875. Generic Characters. — Carapace large, not cristated above, with the antero-lateral corners more or less produced. Eyes distinctly separated, with well-marked visual elements. Antennulaa with one of the flagella rudimentary. Antennas in female very minute, uniarticulate, and provided with three strong hairy setae, in male shorter than body. First pair of legs with terminal part slender and strongly geniculate; second pair much shorter. The three posterior pairs of legs very slender and nearly equal in length; terminal claw very strong. Uropoda with the branches very unequal, inner much larger than outer, and composed of a single joint, strongly spinous at the inner edge. Remarks. — This genus, established by Spence Bate for the reception of a peculiar little Cumacean — described by the same author at an earlier date as Cuma unguiculata — is chiefly distinguished by the eyes being widely separated, and never, as in the other Cumacea, confluent. In other respects it comes very near to the genus Cumella. REPORT ON THE CUM ACE A. 63 Besides the typical species, Nannastacus unguiculatus, Spence Bate, another very- distinct species, Nannastacus longirostris, has been described by the author from the Mediterranean, and a third species from the Challenger Expedition is now added. 13. Nannastacus suhmii, n. sp. (PI. X. figs. 4—5). Specific Characters. — Form of body rather stout, especially in female. Carapace very large, more than twice as long as the exposed part of trunk; dorsal surface evenly vaulted, without any spines or projections; posterior part somewhat gibbous; antero- lateral corners greatly produced, but terminating in a blunt point; pseudorostral projection in female tubular and obliquely ascending, in male very short and rounded. Eyes with three well-defined corneal facets, much larger in male than in female. Legs nearly as in Nannastacus unguiculatus. Uropoda with the scape very short, inner branch elongate, outer exceedingly minute. Length scarcely exceeding 2 mm. Remarks. — This form is very nearly related to the typical species, Nannastacus unguiculatus, Spence Bate, but may readily be distinguished by the female wanting every trace of the peculiar laminar spines adorning the carapace, as well as the epimeral plates of the exposed segments of the trunk in that species. Moreover, the antero- lateral corners are not sharply pointed but blunt at the tip, and these corners in the male are comparatively more produced than in the male of Nannastacus unguiculatus. Description. — All the specimens of this form contained in the collection, having been mounted together in Canada balsam on a glass-slide, a closer examination of the anatomical details could of course not be instituted. Among tbe specimens there is but a single female, the rest being males. The length of the body does not exceed 2 mm. The form of the body (see PL X. figs. 4, 5) is rather short and stout, though a little more slender in the male (fig. 5) than in the female (fig. 4), with the anterior division somewhat longer than the posterior and rather ddated, especially in the female. The carapace is somewhat differently shaped in the two sexes. In the female (fig. 4) it is very large and tumid, with the dorsal surface boldly arched and distinctly gibbous in the posterior part. The pseudorostral projection is rather prominent, forming a narrow tubular process, which ascends obliquely in front and is tipped with small bristles. The antero -lateral corners of the carapace are greatly produced in the form of conical processes, terminating in a blunt point. Between these processes and the pseudorostral projection there is a broad shallow emargination, from the bottom of which the antennulse are seen projecting. The inferior edges of the carapace form together with the posterior an uninterrupted oblique curve, and are finely denticulate in the anterior part. In the male (fig. 5 ) the carapace is comparatively much shallower, with the dorsal line almost horizontal, and the pseudorostral projection far less 64 TFE VOYAGE OF H.M.S. CHALLENGER. prominent and obtusely rounded. The antero-lateral corners are also less prominent, though distinctly projecting, and the inferior edges are irregularly flexuous, and through- out smooth. The surface of the carapace in both sexes is quite even, without any trace of crests or spines, except a few small denticles on the posterior gibbous part, more distinct in the female. The eyes, as in the two other known species of the genus, are widely separated, and located on each side of the frontal lobe at the base of the pseudorostral projection. Each of them exhibits a distinct ocular pigment and three strongly projecting corneal facets, much larger in the male than in the female. The exposed part of the trunk is scarcely more than half as long as the carapace, and composed of five well-defined segments, the first of which, however, is very narrow. The epimeral plates are evenly rounded and quite smooth. The tad is very slender in both sexes, with the penultimate segment the longest. The last segment is produced in the middle posteriorly to an obtuse angle. In the female a few very small denticles occur on the dorsal surface of each segment, whereas in the male all the segments are quite smooth. As to the colour of the specimens freshly mounted m Canada balsam, the female specimen is rather pellucid, with only a few dark brown patches, one of which occurs on each side of the carapace close to the inferior edge, another across the penultimate caudal segment (see fig. 4). In the male the whole carapace, as also most of the body, is of a much darker brownish hue. In both sexes a dark transverse band is observed on the first pair of legs occupying the meral joint, and also on the second pair of gnathopoda a similar band occurs. The antennulse are quite similar in the two sexes, and rather slender, with one of the flagella very small and knob-shaped. The antennae of the adult male (see fig. 5) are much shorter than the body, and scarcely reach beyond the third caudal segment. The distal part of the peduncle is distinctly biarticulate and densely beset along the lower edge with rather slender sensory bristles. The flagellum is very narrow and composed of much elongated joints. The first pair of legs are about as long as the carapace, with the carpal and pro- podal joints rather slender and subequal, and the terminal considerably shorter. The second pair of legs are much shorter than the first, with the terminal joint linear and provided with a few stiff bristles on both edges and at the tip, the latter beino; the longest. The exopodites of these two pairs of legs, as also of the second pair of gnathopoda, are remarkably small, especially in the female, with the basal part not at all ddated. The three posterior parrs of legs are in the female simple, without any trace of exopodites, and very slender, considerably longer than the second pair, and about equal in length. They are almost quite naked, and have the carpal and propodal joints very REPORT ON THE CUMACEA. 65 elongate. The terminal joint, together with its apical spine, forms a strong Inarticulate curved claw. In the male the two anterior of these pairs of legs are provided with well-developed natatory exopodites, as in the preceding legs, and hence the basal joint of these legs is considerably dilated to receive the strong muscles moving the exopodite. The last pair of legs, on the other hand, exhibit a slender form quite similar to that in the female. The pleopoda are quite wanting in both sexes. The uropoda have the scape unusually short and the branches very unequal, the outer one being exceedingly small, though consisting of two well-defined joints. The inner branch, on the other hand, is rather elongate, especially in the male, uniarticulate, and provided with several small denticles along the inner edge, and a slender spine at the tip. Habitat. — No less than eleven specimens of this little Cumacean were collected by the late Dr. v. Willemoes Suhm, on October 27, 1874, off Samboangan, Philippine Islands, and preserved by him in the above mentioned manner. All the specimens were taken at the surface of the sea with the aid of the tow-net. Family VII. Campylaspid^e. This family in many essential points differs considerably from the other families, both as to the external appearance and the anatomical details. The enormous development of the carapace affords a good external character. Of the anatomical details the very peculiar structure of the oral parts may be noted, as also of the maxillipeds and the first pair of gnathopoda. As to the legs, the first pair are not, as in other Cumacea, distinguished by any considerable length, being on the whole, both in size and structure, very similar to the second pair of gnathopoda. Also in the internal organisa- tion there would seem to exist essential differences from other Cumacea. Thus, I have found the liver to consist of only a single pair of very large caeca, instead of three pairs, as in the genus Diastylis. The family comprises as yet but a single genus, viz.: — Genus Campylasjns, G. 0. Sars, 1864. Campylaspis, G. 0. Sars, Oni den aberrante Krebsdyrgruppe Cumacea og dens nordisk Arter, p. 75. Generic Characters. — Carapace strongly vaulted above, with the anterior part more or less depressed and produced ; pseudorostral projection very short, obliquely truncate ; inferior edges of carapace unarmed. Ocular lobe more or less produced along CZOOL. CHALL. EXP. — PART LV. — 1887.) I" 9 66 THE VOYAGE OF H.M.S. CHALLENGER. the upper side of the pseudorostral projection ; visual elements imperfectly developed. Exposed part of trunk very short, though composed of five well-defined segments. Segments of tail keeled laterally. Antennae in male attaining the length of the body. First pair of legs comparatively short, not projecting in front of carapace ; second pair rather stout, with terminal joint very elongate and linear ; the three posterior pairs of legs successively decreasing in length and only sparingly supplied with bristles. Inner branch of uropoda considerably larger than outer, with inner edge spinous. Remarks. — The species of the present genus are easily known by the exceedingly large and vaulted carapace, which in most of the species exhibits a very marked sculpture, partly in the form of keels and partly as variously shaped protuberances. As to the limbs, the comparatively small size of the first pair of legs also furnishes a marked and characteristic feature. No less than ten species belonging to this genus have been described, seven of which occur off the Norwegian coast, two in the Mediterranean, and one in the West Indian Archipelago. Two new species are now added from the Challenger Expedition. 14. Campylasjns pacifica, n. sp. (PI. X. fig. G). Specific Characters. — $. Carapace enormously developed, occupying more than half the length of the body, and being produced posteriorly so as to arch over the exposed part of the trunk ; dorsal surface boldly vaulted; anterior part only slightly produced ; pseudorostral projection very short ; subrostral notch rather shallow. Sur- face of caraj^ace without any keels or protuberances, but slightly uneven dorsally. Ocular lobe rather broad, containing a whitish eye-pigment. Penultimate caudal seg- ment entire, not subdivided. Uropoda very elongate, attaining the length of the four posterior caudal segments taken together, branches scarcely half as long as scape. Length, about 3 mm. Remains. — The present species seems to be most nearly related to the Mediterranean form, Campylaspis glabra, G. 0. Sars, from which it however differs by the somewhat more slender form of the body, by the carapace being slightly uneven dorsally, by the shorter pseudorostral projection, and by the presence of a distinct, though rather shallow subrostral notch. Description. — I have only had an opportunity of examining a single female specimen of this form. The specimen, which had been mounted in Canada balsam, is fairly well preserved, and measures in length about 3 mm. The form of the body (see PI. X. fig. 6) is comparatively somewhat more slender than in most of the other species, with the anterior division greatly dilated and some- what longer than the posterior, if the uropoda are excepted. The carapace is of quite a prodigious size, occupying more than half the length of REPORT ON THE CUMACEA. 67 the body, and being produced posteriorly so as to arch over nearly the whole exposed part of the trunk. The dorsal surface is boldly arched, with the strongest curvature occurring in the posterior part, whence the dorsal line declines quite gradually anteriorly, and more abruptly posteriorly. The pseudorostral projection is very short and obliquely truncated, with the upper corner slightly projecting ; beneath it, on each side, occurs a distinct though rather shallow sinus, limited below by an obtuse angle. The inferior edges of the carapace are quite unarmed and rather curved in the middle, their posterior part ascending obliquely and without any angle joining the posterior edge, which is also oblique. The surface of the carapace wants any trace of keels or distinct pro- tuberances, but appears somewhat uneven in the dorsal part ; it exhibits under the microscope a finely granular structure produced by numerous small depressed pits. The ocular lobe is rather broad and contains within it an accumulation of whitish eye- pigment, whereas distinctly developed refracting bodies seem to be quite wanting. The exposed part of the trunk is very short, and is to a great extent hidden by the posterior projecting part of the carapace. It seems to be composed, as in the other species, of five segments, but only the three posterior segments appear quite distinct in their whole circumference. The epimeral plates of all the segments are evenly rounded and slightly projecting laterally. The tail is rather slender, with the segments sharply marked off from each other and provided with a slight lateral keel. They increase successively in length to the penultimate, which is the longest. The last segment is considerably shorter, and pro- duced at the end to an obtuse angle. The colour of the freshly mounted specimen is a uniform brownish-red. As to the limbs of the anterior division, a closer examination could not of course be made in the sole specimen before us. They, however, seem to agree fairly well with those in the succeeding species, as described below. The uropoda are rather elongate, about equalling in length the four posterior caudal segments taken together. The scape is very slender, more than twice as long as the branches, and quite smooth. The inner branch is composed, as in the other species of the genus, of a simple segment, armed on the inner edge with a few denticles, and at the tip with a strong spine. The outer branch is distinctly biarticulate, but consider- ably smaller than the inner, and provided at the tip with a slender seta. Habitat. — The above described specimen was taken by the late Dr. v. Willemoes Suhm, at the surface of the sea, on October 27, 1874, off Samboangan, Philippine Islands, and was mounted together with the specimens of Nannastacus suhmii on the same glass-slide. 68 THE VOYAGE OF H.M.S. CHALLENGER. 15. Campylaspns nodulosa, n. sp. (PI. XL). Specific Characters. — $ . Carapace scarcely attaining half the length of the body ; posterior part not produced over the anterior segments of trunk ; dorsal surface evenly vaulted in the posterior part and oblkpiely declining anteriorly ; pseudorostral pro- jection rather produced ; subrostral notches distinctly defined. Surface of carapace everywhere provided with rounded nodiform protuberances. Ocular lobe oval. Pen- ultimate caudal segment subdivided by a transverse sulcus. First pair of maxdlge with the palp bisetose ; second pair with eight setae. Terminal joint of second pair of legs scarcely longer than carpal joint. Uropoda comparatively short, about half as long as the tail, scape rather strong, slightly serrate on both edges ; inner branch only slightly shorter than scape, serrate on inner edge, and armed with five spines, two lateral and three apical ; outer branch a little shorter than inner, with two unequal setae at the tip. Length, nearly 5 mm. Remarks. — Of the earlier known species the Norwegian form, Campylaspis verrucosa, G. 0. Sars, would seem to be the one most nearly related to the present species, at least as regards the sculpture of the carapace. In other respects, however, the two species are very easily distinguishable from each other. Description. — I have had for examination four specimens of this form, all of exactly the same size, and having the general appearance of females. But in all specimens the last pah- of legs are quite wanting, a fact that would seem to prove the specimens to be still immature, although their size is not inferior to that of adult specimens of other species, and considerably larger than that of the form described above, which un- doubtedly is full-grown. In dissecting one of the specimens, I have found all the remaining limbs well developed and of a structure very similar to that in adult specimens of other species. The length of the specimens is nearly 5 mm. The form of the body (see PI. XL figs. 1, 2) is that characteristic of the genus, the anterior division being greatly dilated and somewhat longer than the posterior. The integuments, as in other species, are very strongly indurated and exhibit on the several limbs a more or less pronounced squamous structure. The carapace is very large, though not nearly attaining the enormous development of the preceding species. In a lateral view (fig. 1) it exhibits a somewhat irregular oval form, with the dorsal line forming a bold curve rather abruptly declining anteriorly. The pseudorostral projection is somewhat more produced than in the preceding species, and the notch beneath it deeper and limited below by a distinctly projecting process. The inferior edges of the carapace are evenly curved in their anterior part, whereas the posterior part is slightly sinuous and obliquely ascending. The posterior edge is almost perpendicularly truncate and joins the inferior without forming any distinct angle. As seen from above (fig. 2) the carapace exhibits a rather regular oviform shape, the REPORT ON THE CUMACEA. 69 greatest breadth, which about equals the height, occurring posterior to the middle. The surface of the carapace is evenly vaulted, without any keels or excavations, but exhibits a number of well-marked, rounded, nodiform protuberances, distributed rather evenly, dorsally as well as laterally, the most anterior part alone being nearly smooth. The ocular lobe is rather large and of oval form, but seems to want distinctly developed visual elements. The exposed part of the trunk is not, as in the preceding species, arched over by the posterior part of the carapace, all the segments being distinctly visible in their whole circumference. They are rather sharply defined, with the posterior edge much elevated. The last segment does not differ either in size or in form from the anterior caudal segments, and, like these, is without any trace of limbs. The tail is somewhat shorter than the anterior division of the body, with the segments marked off from each other by deep constrictions. They are slightly keeled laterally, and the penultimate segment exhibits, moreover, posterior to the middle, a transverse sulcus, so that it is, as it were, subdivided into two segments. The last segment (see fig. 21) is, as usual, somewhat dilated posteriorly and of pentagonal form, the posterior edge being produced in the middle as an obtuse angle. The colour of the specimens, preserved in spirit, is whitish, with patches of brownish- red, but may have been in the fresh specimens more unifonnily reddish, as several northern forms acquire a quite similar spotted appearance by the action of the spirit, though originally more uniformly coloured. The antennuloa (fig. 3), which project anteriorly from the subrostral notches (see fig. 1), are of moderate length, with the first joint of the peduncle strongly curved and somewhat shorter than the two others taken together. Of the latter the first is con- siderably larger than the second. The outer flagellum (see fig. 4) is well developed and composed of three joints, the last of which, however, is exceedingly minute and might be easily overlooked. It is provided at the tip with the two usual ribbon-like sensory appendages, besides a slender seta ; moreover, a small auditory bristle is found affixed to the last joint. The inner flagellum is exceedingly minute, but on closer examination it is found to be composed of two well-defined joints, the last of which is provided at the tip with two diverging auditory bristles. The antennas (figs. 5, 6, a2) are cpiite rudimentary, constituting a simple, strongly indurated, conical prominence, divided into two segments and without any trace of bristles or srjines. The anterior lip (figs. 5, 6) is comparatively large and of a rather peculiar shape, forming a broad quadrangular plate projecting anteriorly between the antennae as a rounded nodulose protuberance, and having the posterior edge sharp and broadly incised in the middle. The posterior lip (fig. 7) is scarcely larger than the anterior, and of a very delicate 70 THE VOYAGE OF H.M.S. CHALLENGER. and fragile structure, so as only with great difficulty to be separated from the sur- rounding parts. It consists as usual of two membranous lobes connected at the base, each lobe exhibiting an irregular oval form, with the inner corner produced to a small somewhat incurved point. The inner edge of the lobes and part of the outer are finely ciliated. The mandibles (figs. 8, 9) exhibit a structure essentially differing from that of all other known genera of Cumacea. They are not very strong, with the body oblong and on the inner side forming a rounded expansion, to which the chitinous tendon of the rotatory muscles is affixed. The anterior branch (see fig. 10), forming the direct continuation of the body, is slightly incurved and somewhat compressed, terminating in a strongly dentate, oblique, cutting edge. On the right mandible this edge is simple and armed with six teeth, increasing in size internally ; on the left mandible, on the other hand, the edge appears double, a secondary bidentate plate occurring in juxtaposition to the principal plate. At some distance posterior to the cutting edge there is a densely crowded series of strong denticulate spines, six on the right and four on the left mandible. The posterior branch or the molar process is very unlike that in other Cumacea, constituting an exceedingly narrow and somewhat flexuous styliform projection, terminating in a sharp point and bearing a few small denticles along the anterior edge. The first pair of maxillae (fig. 11) are unusually strong, with the basal parts very massive and partly connected at the base. The outer masticatory lobe is rather broad and strongly incurved, with the extremity abruptly truncated and armed with numerous powerful spines, some of which are finely dentate on one of the edges. The posterior masticatory lobe is considerably smaller and somewhat tapering, with the extremity narrowly truncate and armed with five spines, the three median rather strong and densely ciliated. The palp, which as usual points straight backwards, is very slender and elongate, and bears at the tip two unequal setae. The second pair of maxillae (fig. 12), unlike those of the first pair, are poorly developed, forming simple membranous plates, without any trace of the usual bifid palp. The inner slightly produced part, representing the masticatory lobe, is provided with eight bristles, the posterior of which is somewhat removed from the rest and densely ciliated. The maxillipeds (fig. 14) are likewise imperfectly developed, the terminal part appearing at first sight to be represented by only a single oval and lamellar joint, finely ciliated along the outer edge, and provided along the inner with a few email bristles. On closer examination, however, a very small terminal joint is found affixed to the inner corner. This joint (fig. 14a) is conical in form and provided with only a single small apical bristle. The basal part, which about equals in length the terminal, exhibits at the end, on the inner side, an imperfectly developed, rounded, masticatory lobe, with two ciliated setae at the end, and is, moreover, provided on the inner edge REPORT ON THE CUMACEA. 71 with two small hooks, whereby the two maxillipeds become, as it were, clasped together. The branchial apparatus affixed to the base of these limbs (fig. 13) is, as usual, composed of two well-defined principal parts, the one pointing posteriorly, the other anteriorly, the first representing the epipodite, and the second the exopodite. The epipodite constitutes a rather narrow plate, concave on the upper surface and with the posterior part curved in a hook-like manner. The inner edge of this plate is somewhat expanded in the middle, and exhibits an almost semicircular soft rim, to which is affixed a series of eight digitiform gill-lobules, decreasing in size posteriorly, the two posterior being quite rudimentary. Moreover, a single rather large gill-lobule, pointing posteriorly, arises anterior to the others from the bottom of the plate. The exopodite is nearly as long as the epipodite, and forms a narrow band-like plate, somewhat irregularly twisted and expanding at the end into a very delicate and pellucid membranous flap, which admits of being exserted from the pseudorostral projection, forming together with the one of the other side an imperfect tube, through which the water, introduced into the branchial cavities, is expelled at regular intervals. The first pair of gnathopoda (fig. 15) are rather stout, with the basal part comparatively broad and muscular, and provided at the extremity on the inner side with a very long and densely ciliated seta pointing straight forwards, and at the outer corner with a much smaller simple bristle. The terminal part, which about equals in length the basal, is apparently composed of only four joints, the ischial joint not being distinctly defined. The first of these joints is provided at the end, on the inner side, with a rather strong ciliated seta, and the second with two simple bristles at the inner edge. The third joint is somewhat swollen and of oval form, slightly larger than the preceding, and has at the end externally a slender seta, and internally two very small bristles. The terminal joint, finally, is exceedingly minute, and armed at the tip with three unequal spines, the middle one being considerably shorter than the other two. The second pair of gnathopoda (fig. 16) are quite pediform, and nearly twice as long as the first pair. The basal joint is rather large, slightly curved and somewhat tapering, with the outer edge finely ciliated and the inner smooth. The end of this joint is not expanded, and bears exteriorly a very strong and elongate ciliated seta, and anteriorly another much smaller seta. The terminal part is about as long as the basal, and composed of five well-defined joints, the first of which (the ischial), however, is very short, and armed on the inner edge with two denticles. The second or meral joint, on the other hand, is rather large and expanded, and very obliquely truncate at the end. The outer corner of this joint is prolonged to a sharp dentiform projection, and bears, moreover, a strong ciliated seta ; the inner edge is serrate, and provided with two small bristles. The succeeding joint is much smaller, and has the outer edge serrate, being, moreover, provided at the outer corner with a cdiated seta, similar to those on the 72 THE VOYAGE OF H.M.S. CHALLENGER. preceding joints, and at the inner edge with three smaller bristles. The penultimate joint is oval in form and unarmed, with two bristles on the inner edge. The terminal joint, finally, is rather small and narrow, with four curved, claw-like spines at the tip. The exopodite is about as long as the basal joint, and has the proximal part simple and cylindrical, the distal is of about the same length, and composed of but three joints, the two last being exceedingly small ; each of the joints is provided with two natatory setae, one of which, however, arising from the first joint, is very small. The first pair of legs (fig. 17) are only very slightly larger than the second pair of gnathopoda, and of a very similar structure, though comparatively a little more slender. The basal joint is armed at the end externally with a small dentiform projection and a ciliated seta, and internally with two rather strong denticles. Two similar denticles are also found at the inner edge of the very small ischial joint. The succeeding joints, on the other hand, are quite unarmed, the two first bearing each at the outer edge a single ciliated seta, the two last only simple bristles. They successively decrease in size, the terminal joint being rather small and linear. The exopodite exhibits a structure similar to that of the second pair of gnathopoda, with this exception, however, that its terminal part has an additional joint. The second pair of legs (fig. 18) are scarcely smaller than the first and comparatively of stronger structure. The basal joint is somewhat dilated at the middle and comparatively shorter than that of the first pair. It has at the end on the inner side a small dentiform projection, and on both edges a simple bristle. The ischial joint is exceedingly small and unarmed. The rneral joint is obliquely truncate at the end, and provided on the inner side with a bidentate angular projection. The carpal joint is a little longer and armed at the end internally with two strong denticles. The propodal joint is very small and unarmed, whereas the terminal joint is rather elongate, sublinear, and provided with a number of stiff bristles, two of which issue from the outer edge, and three from the tip. This joint, together with the preceding, is generally directed outwardly, forming a geniculate bend with the carpal joint. The exopodite is exactly of the same appearance as that of the first pair of legs. The two succeeding pairs of legs (figs. 19, 20) are quite simple and rather slender, with the terminal part more or less curved exteriorly. Of the joints the basal is, as usual, the largest, though considerably shorter in the posterior pair (fig. 20) than in the anterior (fig. 19). The two succeeding joints are rather short and nearly equal in size, whereas the carpal joint is much more elongate. The two outer joints rapidly decrease in size, and the last joint has the form of a small Particulate claw. Each of the two preceding joints bears a strong curved seta at the end exteriorly. Elsewhere these legs are only very sparingly supplied with simple bristles. The last pair of legs, as has been stated above, were quite wanting in all the specimens. REPORT ON THE CUMACEA. 73 The uropoda (see fig. 21) are comparatively much shorter than in the preceding species, scarcely attaining half the length of the tail. The scape is slightly serrate on both edges, and exhibits a very marked squamous sculpture, the squamulse being rather large and triangularly pointed. The branches exhibit the structure characteristic of the genus. The inner branch is only slightly shorter than the scape, uniarticulate, with the inner edge serrate, and moreover armed with two slender spines ; from the tip of this branch three unequal spines arise, the middle one being by far the strongest. The outer branch is a little smaller than the inner, and composed of two well-defined joints, the first of which is quite short, whereas the last is rather elongate and gradually tapers to the apex, bearing a rather long terminal spine, and a much smaller one on the inner edge close to the apex. Habitat. — The specimens of this species were detected by the author in examining a parcel of dredged mud preserved in strong spirits, and kindly sent by Mr. John Murray. The mud was obtained on January 29, 1874, off Cumberland Bay, Kerguelen Island, from a depth of 127 fathoms (Station 129h). (ZOOL. CHALL. EXP. — PART LV. — 1887.) Iii 10 TABLE OF CONTENTS. Introduction, . Principal "Works on Cumacea, Morphology of the Cumacea, Definition of the Families, . List of the Species, Description of Genera and Species, CuMIDiE, . Cyclaspis, G. 0. Sars, australis, n. sp., . pasilla, n. sp., exsculpta, n. sp., Vaunthompson IID.E, Vaunthompsonia, Spence Bate, meridionalis, n. sp., LaMPROPIDjE, Paralamprops, n. gen., serrato-costata, n. sp., Leuconid.e, Leucon, Kroyer. assimilis, n. sp., tenuirostris, n. sp., Eudorella, Norman, . abysei, n. sp., Diastylhxe, Diastylis, Say, styyia, G. O. Sars, horrida, n. sp., . erinaceus, n. sp., mystacina, n. sp., Cdmellid^;, . Nannastactts, Spence Bate, suhmii, n. sp., CaMPYLASPIDjE, Campylaspis, G. 0. Sars, pacifica, n. sp., . nodidosa, n. sp., . (zool. chall. exp. — part lv. — 1887.) PAGE 1 2 3 5 9 11 11 11 12 19 20 22 22 23 25 26 2G 32 33 34 38 40 41 43 43 44 55 58 60 62 62 63 65 65 66 68 Iii 11 INDEX OF SYSTEMATIC NAMES. Synonyms and Species incidentally mentioned are in italics. PAGE Diastylis — continued. page Alauna, . . . 43 sphiosa, . . . . .50 Campylaspis, . 9, 65 strigata, . 50 glabra, 66 stygia, . . 2, 44 nodulosa, . 68 DlASTYLID-E, . . 7, 43 pacifica, . 66 DlOPS, 62 verrucosa, 68 Edriophtualmia, 3 Campylaspid^e, 8, 65 Eudora, 40 Chalarostylis, 6, 25 Eddorella, 7, 33, 40 CONDYLURA, . 43 abyssi, 41 Cum a, 5 12, 43 truncatula, 41 angulata, . 50 Eudorellopsis, . 7, 33 brevirostris, 50 Hem i la m props, 6, 25, 26, 31 unguiculata, 62 Iphinoe, 5 Cumella, 8, 62 ISOPODA, 3, 4, 54 CuMELLIDiE, 8, 62 LAMPROPIDiE, . . 6, 25 CUMIDJ3, 5, 11 Lamprops, 6, 25, 31 Cumopsis, 5 Leptocuma, . 6, 22, 23 Cyclaspis, 5, 11 , 21, 27 , 29, 47 Leptostylis, . 7, 43, 44 australis, . 12, 19 Leucon, 1 , 33, 40, 42, 43 cornlgera, 12, 13 anomalus, 22 exsculpta, 12, 20 assimilis, . . 33, 34 longicaudata, 12, 13 longirostris, 39 pusilla, 12, 19 nasicus, . 34 Diastylis, 7 , 43, 65 tenuirostris, . 33, 38 biplicata, . 50 Leuconid^;, . 6, 32 calweri, . 50 Macrura, 3 echinata, . 50 Malacostraca, 3 edwardsii, 50 Mysid^;, 3, 4, 54 erinaceus, . 44, 58 Nannastacus, . . 8,62 horrida, . 44, 55 longirostris, 63 insignis, . 60 suhniii, . 63, 67 josephinx, 58 unguiculatus, 63 hid/era, . 55 Nebaua, 4 mystacina, 44, 60 Paralamprops, 6, 26, 38, 47 rathkii, 50 serrato-eostata, . 15, 26 rugosa, 50 Petalomera, 8 78 THE VOYAGE OF H.M.S. CHALLENGER. PAGE PAGE Phyllocaeida, 4 Stephanomma, 5 Platyaspis, . 6, 25, 26, 31 Trilobita, 4 PoDOPHTHALMIA, 3, 4, 54 Vaunthompsonia, 6, 22, 27 PSEDDOCOMA, . 8 cristata, . . 23, 24 PsEUDOCUMIDiE, 7 meridkmalis, 23 ScHIZOPODA, . . 3,4 Vaunthompsoniid-e, 5, 11, 22 Spencebatea, . 8, 62 PLATE I. (ZOOL. CHALL. EXP. — PART LV. — 1887.) Hi. PLATE I. PAGE Figs. 1-20. Cyclaspis australis, n. sp., . . .12 Fig. 1. Young male, seen from the left side ; magnified about fifteen diameters. Fig. 2. Same, seen from above. Fig. 3. Young female, seen from the right side. Fig. 4. Ocular lobe, seen from above. Fig. 5. Antennal segment, viewed from below, together with antennulee (a1), antennae {a"), and anterior lip (L). Fig. 6. Outer part of an antennula; more highly magnified. Fig. 7. Posterior lip. Fig. 8. Mandibles, seen from the front. Fig. 9. Maxilla of first pair. Fig. 10. Maxilla of second pair. Fig. 11. Left maxilhvped, together with the corresponding part of the branchial apparatus, seen from above. Fig. 12. Left maxilliped, isolated, seen from below; more highly magnified. Fig. 12a. One of the marginal spines from carpal joint of same maxilliped ; highly magnified. Fig. 13. Terminal lamella of the exopodite of same. Fig. 14. Gnathopod of first pair. Fig. 15. Gnathopod of second pair. Fig. 16. Leg of first pair. Fig. 17. Leg of second pair. Fig. 18. Leg of third pair. Fig. 19. Leg of fifth pair. Fig. 20. Last caudal segment, together with the uropoda, from above. Figs. 21-23. Cyclaspis pusilla, n. sp., . . . 19 Fig. 21. Adult female, seen from the right side; magnified about twenty-six diameters. Fig. 22. Same, seen from above. Fig. 23. Ocular lobe, seen from above. Figs. 24-26. Cyclaspis exsculpta, n. sp., . . 20 Fig. 24. Anterior division of a female specimen, seen from the left side ; magnified about twenty-six diameters. Fig. 25. Same part, viewed from above. Fig. 26. Ocular lobe, viewed from above. The Voyage of H.M.S "Challenger/ 3umacea,Pl.l GO Sai Fig! 1-20, CYCLASPIS AUSTRALIS.n.sp. Fig 5 21-23, CYCLASPIS PUSILLA,n sp Fig! 24-26 CYCLASPIS EXSCULPTA, n.sp. 1,1th V? SoHachtei PLATE II. PLATE II. PAGE Figs. 1-5. Vcmnthom'psonia meridionalis, n. sp., . . 23 Fig. 1. Young female, viewed from the right side ; magnified about fourteen diameters. Fig. 2 Same, seen from above. Fig. 3. Anterior part of carapace, seen from above ; more highly magnified. Fig. 4. Same part, viewed from the right side, a1, antennula ; a2, antenna ; x, terminal membranous part of branchial apparatus. Fig. 5. Last caudal segment, together with right uropod, seen from above. Figs. 6-13. Paralamprops serrato-costata, n. gen. et sp., . 26 Fig. 6. Female, viewed from above ; magnified about twelve diameters. Fig. 7. Same, seen from the left side. Fig. 8. Young male, seen from the right side. Fig. 9. Right antennula of female, viewed from below. Fig. 10. Right antenna of female. Fig. 11. Anterior lip. Fig. 12. Posterior lip. Fig. 13. Mandibles, together with their adductor muscles, seen from the front. The Vovage of H.M.S "Challenger.' Cumacea.Pl II G.O Sars, del Fi^s1-5,VAUNTH0MPS0NIA N1ERIDIONAUS n.sp. Fig? 6-13, PARALAMPROPS SERRATO-COSTATA, n.sp. PLATE III. , epipodite. Fig. 7. Gnathopod of first pair, together with the basal lamella. Fig. 7 bis. Gnathopod of second pair. Fig. 8. Leg of first pair. Fig. 9. Leg of second pair. Fig. 10. Leg of third pair. Fig. 10a. Rudimentary exopodite of same leg ; highly magnified. Fig. 11. Leg of fifth pair. Fig. 12. Extremity of tail, together with telson and uropoda, viewed from above. Fig. 13. Terminal part of telson ; more highly magnified. Fig. 14. Antennal segment, together with left antennula and antenna of a young male, seen from below. On the right side only the bases of each are represented. Fig. 15. Leg of third pah' of same specimen. Fig. 16. Anterior pair of pleopoda of same. Fig. 17. Outer part of a pleopod ; more highly magnified. The Voyage of H M.S."Chal] CumaceaPl.ffl. G.O.Sai LithW Schlachter, Stockholm PARALAMPROPS SERRATO COSTATA, n.sp. ( Anatomy J. PLATE IV. PLATE IV. PAGE Lexicon assimilis, n. sp., . . . . 34 Fig. 1. Young male, viewed from the right side; magnified about eighteen diameters. Fier. 2. Same, seen from above. Fi Branchial leas well developed, scarcely projecting beyond the edges of the carapace, endo- podite narrow, indistinctly jointed, exopodite forming a broad rounded plate, epipodite very large, expanded at both extremities, ....... projecting far beyond the edges of the carapace, endopodite very elongate and slender, almost pediform, exopodite also rather produced and ending in a narrow point, epipodite exceedingly small, nearly obsolete, Nebalia, Leach. Paranebalia, Claus. imperfectly developed, lamelliform, endopodal and exopodal parts only slightly indicated as small triangular lobes, epipodite well defined, . Nehaliopsis, n. gen. DESCRIPTION OF GENERA AND SPECIES. Order BRANCHIOPODA. Suborder PHYLLOCAEIDA. Family Nebaliid^;. Paranebalia, Claus, 1880. Paranebalia, Claus, Grundziige der Zoologie, ed. 4, vol. i. p. 576. Generic Characters. — General appearance much as in Nebalia. Carapace smooth, oval, lateral parts or valves produced posteriorly to broadly rounded lobes separated above by an angular incision. Rostral plate well developed. Trunk scarcely longer than pleon. Tail rather sharply defined from pleon, with the first segment smaller than the succeeding. Eyes well developed, with distinct visual elements. Antennuke rather stout, with last joint of peduncle produced anteriorly to a serrate lamella; fiagellum comparatively short, greatly dilated in male. Antennae slender, with last joint of peduncle elongate and geniculate at the base, fiagellum simple, cylindrical, of similar structure in both sexes. First pair of maxillge with the outer masticatory lobe narrowly produced and abruptly truncate at the end; second pair with palp and exognath well defined but comparatively small. Branchial legs projecting far beyond the edges of the carapace; endopodite very elongate and slender, almost pediform; exopodite also very elongate and produced to a narrow point; epipodite exceedingly small, upper end produced. Pleopoda much as in Nebalia. The two succeeding pairs of rudimentary caudal limbs both inarticulate. Caudal rami much as in Nebalia. Remarks. — This genus, founded by Professor Claus upon the description given by the late Dr. v. Willemoes Suhm of his Nebalia longipes, is chiefly distinguished from Nebalia by the structure of the branchial legs, as also by the male having only the antennular flagellum peculiarly modified, whereas the antennas are cmite alike in both sexes. More- over the maxillae exhibit some structural differences apparently of generic value. In all other respects it seems to agree very closely with the typical genus. Only one species is as yet known. (ZOOL. CHA.LL. EXP. — PART LVI. — 1887.) Kkk 2 10 THE VO.YAGE OF H.M.S. CHALLENGER. 1. Paranebalia longipes, Willemoes Suhm (Pis. I., II.). Nebalia longipes, Willemoes Suhm, On some Atlantic Crustacea from the Challenger Expedition, iii; Trans. Linn. Soc. Lond., ser. 2, vol. i. p. 26, pi. vi. 1879. Paranebalia longipes, Claus, Grundziige der Zoologie, ed. 4, p. 576. Specific Characters.— Car 'apace as seen laterally, oval in form, covering besides the trunk also more or less completely the sides of the pleon ; rostral plate oblong, narrowed in front and terminating in a spiniform projection. Posterior abdominal division or tail scarcely longer than pleon and suddenly much narrower, the two middle segments denticulate at the posterior edge. Eyes slightly curved and tapering to the end, upper side strongly denticulate. Antennulse with the terminal expansion of the peduncle broadly lanceolate and armed with about twelve serrations along the outer edge ; flagellum not attaining half the length of the peduncle, and composed of but five articulations, the first by far the longest. Antennae with last joint of peduncle longer than the preceding and armed at the anterior edge with two recurved denticles, flagellum shorter than peduncle and six-articulate. Epipodite of last pair of branchial legs produced above to a slender flap. The three posterior pairs of pleopoda with the basal part strongly serrate at outer edge. The two pairs of rudimentary caudal limbs nearly equal in size. Caudal rami about as long as the three last caudal segments combined, outer edge minutely spinulose, inner setose, apex blunt, tipped with a fascicle of exceed- ingly long and slender setas. Length of adult female 6 mm. Remarks. — In the above diagnosis I have given some characteristics, to which I am inclined to believe should only be attributed the value of specific marks. As, however, the above named species is the only representative of the genus as yet known, a strictly specific diagnosis cannot of course be drawn up at present. Description. — Among the material sent me for examination no male was found, all the specimens being either immature or adult females, partly laden with eggs or embryos. But the late Dr. v. Willemoes Suhm has shortly described and figured the adult male of this form, which exhibits some very interesting features not found in the male of Nebalia. Unfortunately the male specimens examined by that naturalist have been lost, and thus my investigations have been solely restricted to the female sex. The length of the adult female does not exceed 6 mm. and hence the present form is rather inferior in size to the known species of Nebalia. All the specimens having been preserved in strong spirit, the body has become rather opaque and of a dark brownish colour, showing however in most of them, at the anterior part dorsally, a peculiar marking in the form of two symmetrical, slightly ramified patches of a whitish colour (see PI. II. fig. 1). Whether this marking is due to a merely external colouring or indicates some underlying internal tissue, I have not been able to ascertain. By carefully preparing a specimen with absolute alcohol and REPORT OX THE PHYLLOCARIDA. 11 oil of cloves, and then mounting it in Canada balsam, I have succeeded in rendering the whole body sufficiently pellucid to admit a closer examination of the several parts in their natural connection with each other (see PI. I. fig. 1). As to the external appearance (see PL I. fig. 1; PI. II. fig. 1) the present form exhibits on the whole a great resemblance to the common species of Nebalia, with this exception, however, that the terminal part of the body, comprising the four last segments, is comparatively more slender and also more distinctly marked off from the preceding- part than is the case in Nebalia. In describing the body I find it advisable to distinguish four principal divisions, viz., the cephalic part, to which the carapace is connected, the trunk, the pleon, and the urosome or tail, the two latter parts answering to the so-called abdomen in the higher Crustacea, and to the abdomen and thorax in the Copepoda. All these four divisions are pretty well defined in the present form, and are of about equal length, but differ both in form and in the structure of their respective limbs. The cephalic part is somewhat flattened, and along its whole dorsal surface firmly connected with the carapace. It bears the following appendages, enumerated from before backwards : — The eyes, the antennulae, the antennae, the anterior lip, the mandibles, the posterior lip, the first pair of maxilla;, and the second pair of raaxilke. Of these appendages only the eyes, the antennulae, and the antenna; project to a certain extent beyond the free edges of the carapace, all the other parts being wholly concealed within its valves. The trunk is almost cylindrical in form, and only slightly longer than broad. It is composed of eight short segments, well defined in their whole circumference, and nearly equal in size, each bearing a pair of branchial legs, the outer part of which projects beyond the free edges of the carapace. The pleon is at the base fully as broad as the trunk, but tapers gradually, so as scarcely to be more than half as broad at the end. It is composed of four well-defined segments, each bearing a pair of powerfully developed natatory appendages or pleopoda. The last segment is provided with distinct, rounded, epimeral plates, partly covering the base of the corresponding pair of pleopoda. The urosome or tail is rather slender, and very movably connected with the pleon, so as often to form an angular bend with it (see PL I. fig. 1). It is nearly cylindrical in form, and composed of four segments, the first of which is the smallest, and, properly speaking, corresponds to the last thoracic segment in the Copepoda. This and the suc- ceeding segment bear each a pair of rudimentary caudal limbs, whereas the third segment is without any trace of liinbs. The last segment terminates in two slender, diverging branches, evidently answering to the so-called "furca"in the Copepoda, and to the caudal lamella; in the Branchipodidae. The two middle segments of this division are denticulate at the posterior edge, and the last segment projects at the end, beneath the 12 THE VOYAGE OF H.M.S. CHALLENGEK. terminal branches, as two triangular, pointed processes, having between them the anal orifice (see PI. II. fig. 10). The carapace (see PI. I. fig. 1 ; PI. II. fig. 1) is rather large, covering the greater part of the body, including the whole cephalic part, the trunk, and part of the pleon. It is highly compressed, the free lateral parts extending in the form of two valves perpen- dicularly along the sides, so as wholly to cover the oral parts, and to include between them the basal parts of the two pairs of antennae, as also the greater part of the branchial legs. Above, the two valves pass immediately into each other by an even curve, without any intervening hinge. As seen laterally (PI. I. fig. 1), the carapace exhibits a somewhat oval form, with the dorsal line slightly arched, the anterior edges strongly curved, and joining the inferior without any intervening angle. Posteriorly, each valve forms an obtusely rounded lobe, advancing to a more or less extent over the side of the pleon, without however covering the dorsal surface of that division. These lobes are separated above by a deep emargination, at the bottom of which a small and narrow incision is se3n (see PI. II. fig. l). Owing to this emargination, the posterior edges of the carapace appear in a lateral aspect (PL I. fig. l) obliquely truncate, and joining the dorsal line at an obtuse angle. The so-called rostrum (PL I. fig. 1, R; fig. 2) forms an oblong tongue-shaj)ed plate, movably articulated to the carapace in front, above the insertion of the eyes. It is slightly arcuate, with the upper side convex, the lower concave, and terminates in a sharp spiniform projection issuing from a slight longitudinal eleva- tion running along the concave side of the rostrum. As seen from above or below (fig. 2) this plate appears broadest near the base and tapers slightly towards the end, whereas in Nebalia it is more regularly oblong — oval in form. AVhen lowered and applied against the anterior part of the carapace, it almost reaches to the inferior edges, thus, as it were, closing the anterior aperture of the carapace as an operculum. The carapace is connected with the body dorsally along a rather restricted space, extending from the base of the rostral plate to about the middle of its length. The limits of this space are faintly traced both in the lateral and dorsal aspect of the animal (see PL I. fig. 1 ; PL II. fig. 1), and according to its position, this space may properly represent the gastric region in the higher Crustacea. Throughout this limited dorsal area the carapace forms the immediate body-wall, whereas elsewhere it only loosely covers the body, which admits of being moved to a certain extent within it. On each side of the anterior part of the carapace, immediately beneath the above-named region, and at a short distance behind the man- dibles, an assemblage of lucid spots, forming together a well-defined, rounded, oval area (PL 1, fig. 1, Cm) may easily be distinguished. This area is produced by the insertion of the strong adductor muscle, by the aid of which the valves of the carapace admit of being approached to each other to a certain extent. Both in form and position this muscle entirely agrees with the strong adductor muscle of the shell met with in the bivalved Phyllopoda, thus giving the carapace an evident phyllopodous character. As REPORT ON THE PHYLLOCARIDA. 13 to structure, the carapace is rather thin and pellucid and very flexible, not at all cal- careous, and without any trace of external sculpturing. It is composed of two distinct layers, an outer chitinous coat, and a soft membrane lining the inside of the carapace. Between the two layers there is a system of hollow spaces anastomosing with each other, and forming together a complicated network of canals, in which the blood circulates. The eyes (PI. I. fig, 1, 0 ; fig. 3) are movably articulated to a short segment, lying in front of the antennal segment, and are partly covered by the rostral plate, projecting obliquely at each side (see PL II. fig. 1). They are rather large, nearly as long as the rostral plate, slightly curved, and somewhat tapering, and not as in Nebalia, expanded at the end. The upper edge is slightly convex and densely denticulate, the denticles increasing somewhat in size toward the tip of the eye ; the lower edge is almost straight and quite smooth. The eye-pigment is black, and does not completely fill up the outer part of the eye, forming a narrow, oblong central mass, from which the small, rounded crystalline cones project all around. There is no distinct facetted cornea similar to that found in higher Crustacea. The antennulse (PL I. figs. 1. a1; fig. 4) are rather powerful organs, almost pediform in character, and more or less projecting in front of the carapace, immediately below the eyes. They consist each of a strong, four-jointed peduncle, geniculate at the middle, and two terminal appendages movably articulated to the same. The first joint of the peduncle is rather firmly connected with the antennal segment and of a nearly quadrate form, without any spines or bristles. The second joint is almost twice as long, and rather movably articulated to the first, forming with it a more or less distinct elbow-shaped flexure. It is slightly dilated towards the end, which exhibits on the upper side an obtuse projection, at the base of which several slender bristles are affixed ; from the lower side, close to the end of the joint, moreover, four rather strong and recurved ciliated setae arise. The third joint is somewhat shorter than the second, constricted at the base, and likewise very movably articulated, so as generally to form with the preceding joint a strong geniculate bend. At some distance from the extremity there is a slight promin- ence, and between this prominence and the terminal edge a great number of slender, diverging bristles occur. The fourth joint is much shorter, and has only a very restricted mobility in relation to the preceding joint. It is produced along the anterior side to a compressed lanceolate expansion (fig. 6) freely projecting from its end, and strongly serrate along the outer edge. The serrations, from twelve to fourteen in number, are pretty regular, and increase in size toward the tip of the expansion, each of them being finely denticulate at the upper edge. At the base of this expansion a few slender bristles are affixed to the inner surface of the joint. Of the two terminal appendages, the outer one has the form of an oblong-oval lamella clothed along the anterior side and apex with a great number of unequal-sized slender bristles, forming together a dense brush. The inner appendage, representing the true flagellum, is more cylindrical in form, and also 14 THE VOYAGE OF H.M.S. CHALLENGER. considerably more elongate, though scarcely exceeding half the length of the peduncle. It slightly tapers towards the end, and is composed of five articulations, the first of which is nearly twice as long as the four others combined. All the articulations are densely beset with bristles, especially along the anterior edge, where the bristles are arranged in distinctly defined fascicles increasing in length toward the apex. On closer examination, some of the setae are found to be densely ciliated along one of their edges, and besides there is found in each fascicle an exceedingly delicate and pellucid ribbon- shaped appendage (fig. 5), answering to the so-called olfactory cilia in other Crustacea. According to the statements of the late Dr. v. Willemoes Suhm, the flagellum in the adult male has a very different appearance from that in the female, being greatly expanded, and without any distinct articulations. In the male of Nebalia, the flagellum is also some- what different from that in the female, but the difference here consists chiefly in this part being more elongate and having dense clusters of sensory bristles on the proximal joints. The antennae (fig. 1, a2 ; fig. 7) are somewhat more elongate and slender than the antennulae, and affixed in close approximation to the latter, so as partly to cover them at the base. They are composed of an elongate, triarticulate peduncle, geniculate at the middle, and a single multiarticulate flagellum. The first joint of the peduncle is short and thick, somewhat similar to that of the antennulae and without spines or bristles. The second joint is also quite smooth, but considerably more elongate, cylindrical in shape, and, like the antennuke, forming a more or less distinct, elbow-shaped bend with the basal joint. The last joint of the peduncle is still more elongate and close to its base abruptly curved, so as to render the peduncle distinctly geniculate at the middle. The anterior edge of this joint is somewhat irregularly flexuous, forming above, at the curvature of the joint, a rounded, knee-shaped protuberance, beset with short bristles ; two similar but less prominent setose protuberances follow further below, and, besides, two somewhat recurved dentiform projections are found at the same edge, the one nearly in the middle, the other at the end. The outer part of the anterior edge is, moreover, beset with slender ciliated setae, rapidly increasing in length towards the end, and continued transversely on the inside of the joint, so as partly to project also on the posterior side ; these setae are very elongate, almost reaching to the end of the flagellum. The posterior edge of the joint has only a single seta at the middle, but this seta is very strong and more densely ciliated than the others. The flagellum is considerably shorter than the peduncle, cylindrical in form, and composed of six distinctly defined joints, the first of which is by far the longest, and about equals the three succeeding in length. At the end of each of the joints anteriorly, a dense fascicle of bristles occurs, and in addition the first joint has three similar fascicles along its anterior edge. According to the statements of the late Dr. v. Willemoes Suhm, the antennae in the adult male do not differ materially from those in the female, whereas in the male of Nabalia the flagellum is extremely elongate, slender and filiform. REPORT ON THE PHYLLOCARIDA. 15 The anterior lip (fig. 8) forms a rounded fleshy prominence issuing from the antennal segment and projecting backwards so as to cover the masticatory parts of the mandibles. It is slightly emarginate at the free terminal edge and finely ciliated on each side of the emargination. The posterior lip I have not succeeded in isolating from the surrounding parts. In Nebalia it constitutes a rather small bifid lobe, limiting the oral orifice posteriorly. The mandibles (fig. 1, M ; fig. 9) have the body comparatively small and tapering above to a narrow point reaching rather far up the sides of the body (see fig. 1). The masticatory part is strongly incurved so as to form a right angle with the body. It has the cutting edge quite rudimentary, only forming a small, undivided dentiform projection (see fig. 10), whereas the molar tubercle is well developed, cylindrical in form, and terminating in a large fluted molar surface. No difference wdiatever can be detected in the armature of the two mandibles. The palp (fig. 11) is very fully developed, being twice as long as the mandible (see fig. 9) and generally pointing obliquely anteriorly so as to reach the inferior edge of the carapace (fig. 1, Mp). It is rather slender and composed of three well-defined joints, the first comparatively short and unarmed, the second very elongate and somewhat curved, bearing at the posterior edge four slender bristles, and at the opposite side near the end another recurved seta. The last joint is nearly as long as the preceding and strongly compressed, forming an oblong vertical lamella, slightly dilated at the middle and provided along the outer part of the posterior edge with a double series of densely ciliated bristles arranged in a comb-like manner and increasing in length towards the tip. As to structure, the mandibles in the present form agree fairly well with those in Nebalia, excepting that the cutting edge is still more rudimentary, and that the last joint of the palp is somewhat differently formed. The first pair of maxillae (fig. 1, m1; fig. 12) exhibit two well-defined and rather dissimilar masticatory lobes, and a slender recurved palp. The outer masticatory lobe is rather narrow and strongly incurved, with the apex abruptly truncate and armed with a double row of strong equal-sized spines, besides a few slender bristles, one of which, affixed at a short distance from the apex anteriorly, is somewhat stronger than the others and finely ciliated. The inner masticatory lobe is considerably shorter than the outer and has the form of a rounded lamella, finely ciliated along the free edges and, besides, provided with a row of about twelve strong ciliated setae, curved anteriorly and forming together a broad fan. The palp, which is movably articulated to the end of the basal part, at the base of the outer masticatory lobe, is very elongate and slender. It exhibits at the base two imperfectly defined segments, forming together a more or less sharp bend and followed by a very narrow setiform terminal part, projecting obliquely posteriorly along the sides of the trunk (fig. 1, m1). The latter part is provided along the lower edge with about fifteen extremely slender unciliated bristles, the posterior of which are rather regularly arranged, two issuing from the tip. In the structure of these maxillae, 16 THE VOYAGE OF H.M.S. CHALLENGER. which has been wrongly represented by the late Dr. v. Willemoes Suhm, the present form exhibits some well marked differences from Nebalia. Thus, the outer masticatory lobe in the latter genus is very dissimilar, being not nearly so much produced, and its armature is also rather different ; moreover the inner lobe is somewhat smaller, whereas the palp is comparatively more strongly developed. The second pair of maxillae (fig. 1, m2; fig. 13) are composed of a somewhat lamellar basal part, to the end of which two appendages are movably articulated, the inner one representing the palp, the outer the exognath. The basal part is divided into two segments, and juts out internally as three densely setose masticatory lobes, the posterior of which is by far the largest. In front of these lobes there is a slight expansion, bearing three slender ciliated setae, the outer of which is very elongate. The palp is shorter than the basal part and rather narrow, Particulate, with the first joint smooth, and the second tipped with three slender cibated setae. The exognath arising close outside the palp, has the form of a narrow lamella, somewhat shorter than the palp, and provided along the outer edge and apex with about nine finely ciliated setae. The maxillae above described differ from those in Nebalia, chiefly by the far inferior development of both the palp arid exognath, which in the latter genus are considerably longer than the basal part, and provided with a much greater number of bristles. The branchial legs (PI. I. fig. 1, brp; PI. II. figs. 2-4) in the present form are, as above stated, modified in a peculiar manner, so as at first sight to appear very different from those in Nebalia, this modification being apparently to make them more adapted for direct prehension of the food, whereas their original function as respiratory organs seems to be much less pronounced than in the typical genus. This is chiefly effected by the excessive prolongation of the endopodite, accompanied also by a peculiar transformation of the exopodite, and a considerable reduction of the epipodite. The branchial legs in the present form thereby acquire an appearance strongly reminding us of the true legs in some of the higher Crustacea, especially those of the Euphausiidae. As is also the case in Nebalia, these limbs are considerably more elongate in the adult females than in the young- animals and in the males, and are moreover distinguished by the great development of the bristles affixed to them. In both sexes they project considerably beyond the free edge of the carapace (see PL I. fig. 1, brp), whereas this is not the case in Nebalia. All the legs present a rather uniform appearance, forming together a densely crowded double series along the ventral side of the trunk, and being extended straight downwards, parallel to each other, and with their outer projecting parts more or less distinctly curved. Their movements in the living animal are undoubtedly performed in a simultaneous and rhythmical manner as in Nebalia. As to structure (see PI. II. figs. 2-4), the same principal parts as in Nebalia are easily found, though rather modified in form. On the main stem may be distinguished a somewhat expanded laminar basal part, and a slender terminal part or endopodite, the latter being more or less curved and divided into a REPORT ON THE PHYLLOCARIDA. 17 number of segments. These segments are, however, very imperfectly defined, and hence can only be exactly counted with great difficulty, except on the posterior pairs (fig. 4), where they appear more distinct, amounting to six in all. The inner edge of the stem does not show any trace of lateral lobes or endites, but is bordered by a regular series of rather elongate and slender ciliated setae. Besides, at a short distance from the edge along the posterior surface, another similar series of slender setae occurs, and this series on the last pair (fig. 4) successively passes over from the posterior surface to the outer edge, whereby the terminal part of the endopodite becomes densely setiferous on both edges. Finally, a third series of very minute hair-like bristles is found close inside the latter series along the posterior surface of the legs. The setae of the basal part in all the legs are considerably shorter than those affixed to the endopodite, though arranged in a similar manner. The apical setae in most of the legs (figs. 2, 3) differ somewhat from the rest, at least in the female, one of them being densely plumose, whereas the others are quite unciliated, very slender and strongly curved at the tip. On the last pair (fig. 4), however, the apical setae do not differ materially from the lateral, but on close examination a very small spine, some- what reminding one of the terminal claw in higher Crustacea, is found at the tip between the setae. The exopodite, issuing from the base of the endopodite on the outer side, exhibits an appearance very different from that in Nebalia, having the form of a slender plate, somewhat shorter than the endopodite, and produced at the end into a very narrow flap. It is fringed along its whole outer edge with a single regular series of slender ciliated bristles, one of which issues from the tip. In the middle pairs (fig. 3) the exopodite is very elongate, and beyond the middle more or less distinctly geniculate, with an approach to a division into two parts, a basal and a very narrow terminal part, thus acquiring a certain resemblance to the natatory branch in higher Crustacea. In the posterior pairs (fig. 4) the exopodite becomes considerably shortened, and of a more simple form. The epipodite, which in Nebalia is exceedingly large and expanded both superiorly and inferiorly, is in the present form reduced to a very small appendage affixed externally to the end of the basal part. It has the form of a narrow recurved plate, very delicate in structure and finely ciliated along the outer edge. In the anterior pair (fig. 2) this plate is slightly bilobed, the inferior lobe being, however, very short and rounded; in the succeeding pairs (fig. 3) it is more simple and oblong in form, and in the last pair (fig. 4) the plate is considerably more produced above than in the other pairs, forming a narrow, slightly flexuous flap, which projects upwards beyond the base of the leg (see PI. I. fig. 1). The four succeeding pairs of limbs (PI. I. fig. 1, pi; PI. II. figs. 5-7), affixed to the pleon, constitute very powerful natatory organs or pleopoda, which admit of being moved with great force from before backwards, acting in much the same manner as the swimming legs in the Copepoda. They consist each of a broad and somewhat flattened basal part, to the end of which two diverging linear branches are affixed. The basal part is strongly muscular, and composed of two unequal segments, the proximal quite short, the distal (ZOOL. CHALL. EXP. — PAET LVI. 1887.) Kkt 3 18 THE VOYAGE OF H.M.S. CHALLENGER. oblongo-quadrangular in form, and in the three posterior pairs having the outer edge regularly serrate. Of the branches the outer one is the shorter, and uniarticulate with the outer edge, and the apex strongly spinose, the inner edge being bordered by a regular row of ciliated setse, two or three of which issue from the apex inside the terminal spines. The inner branch is somewhat more slender and distinctly Particulate, the proximal joint very short, the distal elongate and slightly tapering to the apex, which juts out as a short dentiform projection and, inside the latter, bears a strong spine. Along both edges of this joint a dense series of ciliated setse increasing successively in length toward the tip is affixed, and from the inner side of the proximal joint a linguiform lateral lobe (fig. 6a) arises, armed at the tip with three strongly chitinised hooks (fig. 66). This lobe meets the one on the other side, and by the aid of the above-mentioned hooks both pleopods become thus, as it were, clasped together, so as only to admit of a quite simultaneous movement. In comparing the four pairs of pleopoda with each other, we find some slight differences in their armature and in the relative length of the branches. Thus, in the first pair (fig. 5) the outer edge of the basal part is quite smooth, without any trace of the regular. serrations occurring in the other three pairs. Moreover, the outer branch in this pair is considerably shorter as compared with the inner, and the spines of the outer edge are much more numerous and also arranged in a different manner, being only in the distal part disposed in fascicles, whereas those of the proximal part form a dense comb-like series. In all the other pairs (figs. 6, 7) the spines of the outer branch are arranged in five distinct fascicles, two in each fascicle, besides a small seta, and to the obliquely truncate apex three somewhat stronger spines increasing successively in size are affixed. The last pair of pleopoda (fig. 7) are somewhat smaller than the preceding pairs, and have the inner branch only very slightly longer than the outer, and the spines of the latter branch are also considerably smaller. The two pairs of rudimentary caudal limbs (PI. I. fig. 1, cp1, cp2; PL II. figs. 8, 9) succeeding the true pleopoda are both of a very simple structure, forming narrow uniarticulate plates, scarcely at all movable. They are setose at the inner edge and armed at the obliquely truncate apex with three spines. Those of the first pair (fig. 8) are somewhat larger and more densely beset with bristles than the posterior pair (fig. 9), with the apical spines successively increasing in length, whereas on the latter pair the middle spine is the longest. The caudal rami (PL I. fig. 1, ur ; PL II. fig. 10) have the form of two more or less diverging narrow plates, movably articulated to the end of the last segment, and about as long as the two posterior caudal segments taken together. They are minutely spinous along the outer edge, and along the inner they are provided with a row of short ciliated bristles, besides a series of minute spines. Prom the obliquely truncate apex a dense fascicle of very long and slender setse arises, which, however, were mostly broken in the specimens examined. REPORT ON THE PHYLLOCARIDA. 19 Tlie internal organs I have not succeeded in isolating by dissection, owing to the small size of this form. In the previously mentioned specimen, rendered pellucid by preparation in Canada balsam, and represented on PI. I. fig. 1 , the intestine can, however, be easily traced running through the axis of the body and terminating in a strongly muscular rectum, which traverses the last caudal segment (see also PL II. fig. 10). At the sides of the intestine the ovaries (PI. I. fig. 1, Ov) appear very distinctly, owing to their being rather more opaque than the surrounding parts. They have the form of two very elongate and narrow tubes running through the whole trunk and pleon, and, moreover, projecting anteriorly to some extent within the cephalic part and posteriorly almost reaching to the end of the second segment of the tail. They were each filled by only a single series of ovarial ova, each with a very distinct germinal vesicle in the centre. The musculature of the body may also be rather distinctly traced in the specimen. Thus, in the cephalic part several strong muscular bundles are seen passing from the dorsal side to the several appendages belonging to that division, and in the succeeding part of the body, besides the muscles moving the respective limbs, there is another group of very powerful muscles running parallel to the axis, and by the aid of which the body admits of being moved in relation to the cephalic part or to the carapace, as does also the tail upon the pleon. Of these muscles the dorsal, or extensores, are the more numerous, passing from the one segment to the other and apparently forming several layers, whereas the ventral musculature is chiefly restricted to two strong muscles running backwards beneath the intestine, and in the tail dividing into separate bundles for each segment. The heart, distinctly visible in living animals from its rapid pulsations, cmite escapes attention in dead specimens owing to its very thin and pellucid walls, and the nervous system, as also the cseca of the intestine are very difficult to see even in fresh specimens. That all these parts on the whole may agree with those in Nebalia, I cannot but believe, as these two genera are otherwise very nearly related. Development. — As above stated, some of the specimens in the collection were laden with eggs and embryos. On PL II. figs. 11, 12, I have represented one of the embryonic stages in a ventral and lateral aspect, having found it somewhat different from the corresponding stage of Nebalia, as figured by Metschnikoff. The length of this embryo is nearly 1 mm. As may be seen, it is still provided with the so-called larval cuticle, forming a pellucid homogeneous sheath investing the greater part of the body, and terminating in a slightly bilobed extremity. The anterior part of the body, constituting the cephalic division, is very considerably dilated, almost globular, and to a great extent filled up with the remainder of the yolk, whereas the succeeding part gradually tapers posteriorly. The latter does not exhibit any trace of the strong dorsal curvature found in the corresponding stage of Nebalia, according to the statement of Metschnikoff, but is cmite straight, or with the terminal part even slightly curved 20 THE VOYAGE OF H.M.S. CHALLENGER. ventrally (fig. 12). Of the several divisions of the hody which are distinguished in the adult animal, the anterior or cephalic part is, as above stated, very massive and about as long as the succeeding division or trunk, which exhibits, within the larval cuticle, all its segments well defined. The posterior part of the body, on the other hand, including the pleon and the tail, is still rather imperfectly developed and scarcely longer than the trunk. It consists of only five segments besides the caudal rami, and if the four anterior of these segments be referred to the pleon, the tail will then only be represented by a single segment. The free edges of the developing carapace (fig. 12, C) can easily be traced on each side as a curved line extending backwards from the base of the eyes and meeting above at the most anterior part of the trunk. The carapace is thus in this stage chiefly confined to the cephalic part, the greater portion of the trunk being exposed behind it. The rostral plate (R) is also readily detected as an obtuse protuberance curving interiorly in front and not yet marked off from the carapace. The several appendages belonging to the cephalic division are all visible, though still rather imperfectly developed. The eyes (0) constitute two rather large recurved prominences, which, however, as yet show no trace of either pigment or visual elements, and are also quite smooth, not as in the adult animal denticulate. The antennulse (a1) and antennas (a2) are of a very similar appearance, forming simple digitiform recurved processes projecting freely from beneath the anterior part of the head, the former slightly diverging the latter extending straight backwards along the ventral surface. Between the base of the antennas a rounded prominence (L), still invested by the larval cuticle, is seen, representing the anterior lip. Somewhat posterior to this prominence, three pairs of rather small and closely crowded processes occur, the two anterior pairs exhibiting a slender terminal appendage, distinctly marked off from the proximal part, and somewhat recurved. Of these processes the anterior pair (M) represent the mandibles, and their terminal appendage the mandibular palp; the succeeding pair (m1) are the first pair of maxillas, and their terminal appendage, which is considerably narrower than that of the preceding pair, is easily recognised as the slender recurved palp of these maxillae; the third pair (m2), finally, are as yet quite simple, conical in form and represent the second pair of maxillae. On each of the eight segments of the trunk there are a pair of bilobular appendages (brp) pointing posteriorly and each partly covering the one succeed- ing it. These appendages represent the developing branchial legs and are all exactly alike and slightly extended laterally. They are succeeded by three pairs of appendages (pi), which are also distinctly bilobular at the tip, but much smaller and quite concealed by the larval cuticle. These appendages are the three anterior pairs of pleopoda. No trace can as yet be detected of either member of the fourth pair of pleopoda, or of the two succeeding pairs of rudimentary caudal lirabs. The caudal rami (ur) form simple obtusely conical processes, lying within the symmetrically formed bilobular extremity of the larval cuticle and being well defined from the last segment, but as yet without any REPORT ON THE PHYLLOCARIDA. 21 trace of spines or bristles. The young, when fully developed and ready to escape from the incubatory cavity of the mother, have much the same appearance as the adult animal, with this exception, however, that the last pair of pleopoda, as is also the case in the young of Nebalia, are not yet developed, but only indicated as a slight projection of the last segment of the pleon. Habitat. — According to the statement of the late Dr. v. Willemoes Suhm, the present interesting form was obtained during the stay of the Expedition at the Bermudas, by Mr. John Murray, who took it in Harrington Sound, a bay which only communicates with the sea through a narrow passage. Afterwards the late Dr. v. Willemoes Suhm also collected the animal in the same bay, where it occurred not rarely under stones and Corals. But among twenty females he only succeeded in finding two male specimens. Nebaliopsis, n. gen. Generic Characters. — Carapace distinctly sculptured, very large and produced along the dorsal line so as to cover over the whole trunk and the greater part of the pleon ; valves not expanded posteriorly nor separated above by any emargination. Rostral plate very small, tongue-shaped. Trunk much larger than pleon and rather dilated. Tail very small with the segments subequal. Eyes rudimentary, without pigment or visual elements. Antennulse not very strong, last joint of peduncle produced anteriorly to a narrow projection, terminal lamella comparatively small, narrowed at the tip, with scattered spines at the edge, flagellum well developed. Antennae with last joint of peduncle divided into two segments, flagellum slender, multiarticulate. Mandibular palp very strong, subpediform. Second pair of maxillae with palp and exognath imperfectly developed and not defined from basal part. Branchial legs very small and simple in structure, forming oblong lanceolate lamellae, setose on the inner edge and slightly lobular at the outer, endopodal and exopodal parts only very faintly indicated, epipodite distinctly defined, narrow, elliptical, upper extremity more produced than lower. Pleopoda well developed, outer branch lamellar, expanded, inner lanceolate. The two succeeding pairs of rudimentary caudal limbs subequal, uniarticulate. Caudal rami lamellar expanded, outer edge spinous, inner setose. Remarks. — The present new genus is chiefly distinguished by the peculiar form and sculpture of the carapace, and by the structure of the branchial legs. In both these respects, as also, it would seem, in the structure of the oral parts, it differs very consider- ably from the two other known genera of recent Phyllocarida. The relative development of the several divisions of the body is likewise rather different, especially when compared with the typical genus Nebalia, the trunk occupying the far greater part of the body, whereas the terminal part or tail — greatly developed in Nebalia — is very much reduced in size. In the form of the carapace, the present genus shows a certain resemblance to 22 THE VOYAGE OF H.M.S. CHALLENGER. the Devonian genus Dictyocaris, Salter, in which a distinct external sculpturing has also been described ; but whether this resemblance is only accidental or points to some nearer relationship between the two genera, it is impossible at present to say with any approach to certainty. 2. Nebaliopsis typica, n. sp. (PI. III.). Specific Characters. — Carapace thin and flexible, somewhat compressed, with a slight dorsal keel running along the middle, its form, as seen laterally, oblongo-triangular, dorsal line very faintly curved, anterior extremity narrowly rounded, posterior obtusely produced above, inferior edges strongly curved in front of the middle. Eostral plate exceedingly small, oval in form, with a slight carina running along the middle and terminating in a blunt point. Surface of carapace sculptured by elevated narrow ridges anastomosing with each other and forming together a somewhat irregular open network, less distinct on the anterior part, where a somewhat flexuous transverse elevation passes down the sides for a short distance from the anterior edge. Trunk very large, about as long as pleon and tail combined. Tail scarcely longer than the three posterior segments of pleon taken together, very narrow, with the posterior edge of the segments indistinctly denti- culate. Eyes exceedingly small, cylindrical, smooth, apex rounded. Antennuke with the two first joints of peduncle rather strong and subequal in length, third much smaller, lamellar appendage about as long as first joint of flagellum. Antennas with the peduncle scarcely longer than that of the antennulae, first joint the largest. Pleopoda with basal part smooth, outer branch shorter than inner and having the outer edge strongly curved and armed with a single row of small denticles. Caudal lamellae about as long as tail, expanded in the middle, extremity tapering and slightly incurved, outer edge apparently armed with seven spines, increasing in size posteriorly, tip with two spines (or strong setas). Length reaching about 40 mm. Remarks. — The specific diagnosis given above is of course only provisional, the present form being the only one of the genus as yet known. Most of the characters mentioned above may, however, from analogy with other forms prove to be really specific in value. It is apparently this form that was mentioned by the late Dr. v. Willemoes Suhm in a letter to Professor v. Siebold,1 as a gigantic Ostracode. This strange mistake may be readily explained by the incompleteness of the first specimen obtained, of which only the carapace and a small fragment of the body was brought up in the dredge. Description. — Of this remarkable form there is in the collection an apparently young specimen with most of the limbs preserved, and also the carapace of a much larger and in all probability adult specimen, together with a fragment of the cephalic part. The former specimen has a length of 22 mm. excluding the caudal rami, which are wanting 1 Zcitschr. f. wiss. Zool, BJ. xxiv. p. xiii. REPORT ON THE PHYLLOCARIDA. 23 in the specimen. Its carapace measures 16 mm. in length. The carapace of the other specimen is fully 29 mm. long and 16 mm. high, and thus the total length of this specimen may have been about 40 mm. — a very large size, indeed, as compared with that of the other known recent Phyllocarida. The form of the body (see PI. III. figs. 1, 2) differs considerably from that in the species of Nebalia and Paranebalia, not to speak of the peculiar shape of the carapace, by the large size of the middle division, the trunk, whereas the posterior part of the bod)-, including the pleon and the tail, is comparatively far less developed and scarcely at all longer than the trunk. The carapace is exceedingly large and of a form distinctly deviating from that met with in the other two genera. It is slightly compressed, though somewhat less so than in Nebalia, and covers the greater part of the body, including the cephalic part, the trunk and the greater portion of the pleon, without, however, being connate with the body except along a very limited space in front. As seen from the side (figs. 1, 5) it exhibits an oblong triangular form, with the dorsal line only very slightly arched, the anterior extremity narrowly rounded, and the posterior extremity produced above to an obtuse angle. The inferior edges exhibit somewhat in front of the middle a strong, almost angular curvature, and behind this curvature they ascend obliquely, without forming any projecting lateral lobe, but joining immediately the posterior edges by a very slight curve. The free edges meet above at an obtuse angle, no intervening incision or emargination, similar to that found in the other known forms, being present (see fig. 6). The rostral plate (see figs. 1, 5) is exceedingly small, but distinctly defined from the carapace. As seen from above (fig. 7) it exhibits a quite regular oval form, with a slight keel running along the middle and terminating in front in a small angular projection. In both specimens it curves straight downwards between the free anterior edges of the carapace, partly closing the anterior opening of the shell as an operculum, but in the living animal it undoubtedly may admit of being raised and lowered to a certain extent, as is the case with the other known forms. As to structure, the carapace is very thin and flexible, almost membranaceous, and semitransparent, but provided with a well-marked external sculpturing, especially very distinct on the carapace of the larger specimen (figs. 5, 6). This sculpturing consists of a system of narrow elevated ridges anastomosing with each other, and forming together a somewhat irregular open network, limiting a great number of angular depressed areas of different size and form. Towards the anterior part of the carapace this sculpturing becomes gradually less distinct, and at last wholly disappears, whereas at a short distance from the anterior edge a rather coarse transverse elevation passes down the sides, making at its upper part a sharp curve backwards. In the middle of the dorsal surface (fig. 6) a straight ridge or low keel runs along the carapace from its anterior extremity to the posterior, dividing it into two symmetrical halves, and at the end of the anterior fourth part of the length of the carapace a well-defined obliquely transverse line crosses the ridge 24 THE VOYAGE OF H.M.S CHALLENGES . and marks off in front the very small space where the carapace forms the immediate body-wall, at the same time dorsally defining the anterior division of the body, — the cephalic part. Immediately below this area, on each side, the insertion of the adductor muscle of the valves is distinctly seen as an assemblage of small lucid spots. The division of the body behind the cephalic part is very large, occupying more than the third part of the length of the body. It is at once seen to answer to the compara- tively small part described above in Paranebalia as the trunk, and is, as in that genus, composed of eight well-defined segments, slightly increasing in size posteriorly, and each bearing a pair of branchial bmbs. The ventral surface of this division bulges out as it were in its posterior part, so as to project further down than the pleon, whereas the contrary is the case in the two other known genera. The pleon is considerably shorter than the trunk and rapidly tapers posteriorly. It is composed of the normal number of segments, none of which, however, exhibits any traces of distinct epimera. The tail is very small and narrow, cylindrical in form, and about as long as the three posterior segments of the pleon taken together. It is composed, as in the other genera, of four segments about ecpial in size, the three anterior being shghtly denticulate at the posterior edge. The last segment is somewhat dilated at the end, and wants the dentiform projections found in the preceding genus on each side of the anal orifice. The eyes (figs. 8, 9, 0) exhibit a quite rudimentary condition, wanting, as they do, every trace of pigment and visual elements. They are very small, so as easily to escape attention, and nearly cylindrical in form, or very slightly dilated towards the end, which is obtusely rounded and quite smooth. The antennulee (fig. 8, a1) are constructed upon the very same type as in Nebalia and Paranebalia, but appear comparatively shorter and stouter. The first joint of the peduncle is rather massive, irregularly angulated, and, as in the latter genera, very firmly connected with the head. The second joint is scarcely longer than the first, dilated in the proximal part, and rapidly tapering to the end, forming with the succeeding joint a strong geniculate bend. It bears on the inner side near the end a few recurved setse, but is otherwise, like the first, quite smooth. The third joint is scarcely more than half as large as the preceding, constricted at the base and somewhat dilated toward the end, with the terminal edge forming on the outer side an obtuse angle. The fourth joint is still smaller but rather broad, and juts out anteriorly as a slightly curved narrow process, strongly compressed and very finely denticulate at the anterior edge. As in Nebalia and Paranebalia a lamellar appendage is articulated outside at the base of this process. This appendage is, however, rather small and of a somewhat fusiform shape, with a few small spines along the anterior edge, two of which are affixed to an angular projection about the middle. Of the numerous slender bristles which clothe the corresponding lamella in the two other genera, there is no trace to be detected. The fiagellum seems REPORT OX THE PHYLLOCARIDA. 25 to have been rather elongate, in all probability much longer than the peduncle, but the outer part is wanting in both specimens. The preserved proximal part is composed of eight joints, the first of which, as usual, is by far the largest and about as long as the lamellar appendage, whereas the succeeding joints are very short. The bristles, which may have clothed these joints anteriorly, were broken off in the specimen examined. The antenna? (fig. 8, a2) are somewhat more slender than the antennula?, and comparatively rather smaller than those in Nebalia and Paranebalia. The peduncle is scarcely longer than that of the antennula? and geniculate at the middle. Its first joint is the largest, irregularly quadrangular in form, and exhibits on the outer side at the base a rounded prominence, slightly overlapping the basal joint of the antennula?. The second joint is, like the first, quite smooth, and has the distal extremity obliquely truncate, forming with the succeeding joint a sharp geniculate bend. The latter joint is considerably more slender than the two preceding, and divided into two segments by a well-marked transverse suture. It is provided along the anterior edge with a number of small slender bristles, forming on the proximal segment a double row. The ilagellum in both specimens is imperfect, the outer part being broken off, but, to judge from the form of the proximal part, it may have been rather elongate and slender. The anterior lip (figs. 8, 9, L) projecting posteriorly between the bases of the antenna?, is rather large and almost pentagonal in form, the terminal edge being not, as in the two other genera, emarginate, but on the contrary produced in the middle as an obtuse angle, and without any visible ciliation. The mandibles (fig. 8, M) are very small, more so even than in Nebalia and Para- nebalia. The structure of their masticatory part could not be more closely examined, but this part seems to be far less developed than in the genus Paranebalia. The palp (Mp), on the other hand, is very large, almost pediform, and composed of three distinctly defined joints forming angular bends with each other. The basal joint is rather massive, and forms with the succeeding joint apparently a very movable articulation. Both these joints are quite smooth and nearly uniform in size, whereas the terminal joint is considerably smaller and of a narrow elliptical form ; it is moreover strongly compressed, and provided along the posterior edge with a dense row of ciliated bristles, arranged in a combdike manner. Of the maxilla? I have only been able to examine the second pair more closely, the first pair being so much damaged in the specimen examined as not to admit of an}7 satisfactory description. The former pair (fig. 10) are very small and look rather dissimilar to those in Nebalia and Paranebalia, whereas, on the other hand, they show an unmistakable resemblance to the maxilla? in certain Copepoda. As in the latter group the inner masticatory lobe is very strongly developed and does not extend in the same plane as the others ; it is, moreover, provided at the free edge with a dense row of strong spinedike bristles, which are ciliated at one of their edges and recurved, so as to form together a dense combdike fan projecting towards the oral orifice. The three other (ZOOL. CHALL. EXP. — PART LVI. — 1887.) Kkk 4 26 THE VOYAGE OF II.M.S. CHALLENGER. masticatory lobes are much smaller, and very narrow, and arise from a common lamellar expansion lying outside the principal lobe. They were devoid of any armature, but it may be that a few bristles have been originally affixed to their obtusely rounded extremities. Neither the palp nor the exognath are distinctly defined from the basal part but appear only as simple expansions of it. The palp has the form of a rather large oval projection extended in the axis of the maxilla, and provided with numerous slender ciliated setae on both edges, those of the outer edge being the longest and arranged in several rows. The apex is evenly rounded and unarmed, whereas a row of very small spinedike bristles is found along the inner edge, inside the slender ciliated setae. The exognath is exceedingly small, constituting only a very slight lamellar expansion of the outer edge, and provided with four densely plumose and somewhat fiexuous setae increasing in size towards the base. The branchial legs (fig. 3) are at first sight very unlike those in the two other known genera, and comparatively far inferior in size, as also much more widely separate from one another (see figs. 1, 2). They are very simple in structure, forming, as they do, merely delicate membranous lamellae of oblongodanceolate form and slightly lobular at the outer edge. There is no marked limit between basal and terminal part, nor are the endopodite and exopodite distinctly defined, the epipodite being the only part distinctly marked off from the plate. The inner edge of the plate forms a very slight and even curve, and is bordered by a single row of slender setae, continued also on the narrowly rounded extremity. Beyond the middle there is outside a very slight lobiform expansion, the distal end of which is somewhat produced and separated from the terminal part of the plate by a narrow incision. This expansion, which is quite smooth, may, from its position, answer to the exopodite, and the part of the leg projecting beyond the above mentioned incision of course corresponds to the terminal part of the endopodite in the other Nebaliidae. The epipodite forms a narrow elliptical lamella affixed on the outer side nearer to the base, and separated from the exopodite by another narrow incision. It is connected to the leg by a narrow neck, and has the upper extremity considerably more produced than the lower, the former even reaching somewhat beyond the base of the leg. The substance of the branchial legs is very soft, almost parenchymatous, and between the two investing cuticles there is accumulated a granular opaque mass disposed in small patches, apparently coagulated blood. All parts of the leg, indeed, seem here to be equally well adapted for respiratory purposes. Moreover, in the proximal part several thin muscular bundles are seen, partly crossing each other and disappearing at about the middle of the length of the leg. The pleopoda (fig. 4) are rather powerful and, as in the other genera, composed of a large lamellar basal part, to the end of which two unequal branches are affixed. The basal part is oblongo-quaclrangular in form and quite smooth, though projecting at the end externally as an acute angle. Of the branches the outer one is uniarticulate and REPOET ON THE PHYLLOCAPIDA. 27 $ distinguished by its lamellar structure and peculiar expanded form, the outer edge bulging out as a strong curve, whereas the inner is quite straight. The latter bears a row of short setae, whereas the outer edge is armed with numerous small denticles, arranged pretty regularly in a single series. The inner branch is considerably longer than the outer, but much narrower and lanceolate in form. It is composed of two distinctly defined joints, as in the other genera, the first quite short and bearing inside the usual linguiform lobe, the second evenly tapering to the apex, and all around the edges fringed with ciliated setae. The first pair of pleopoda only differ from the other three by the outer branch being some- what less expanded. The last pair, as in the other genera, are somewhat shorter than the preceding pairs, but otherwise of the very same structure. The two pairs of rudimentary caudal limbs suc- ceeding the pleopoda (see fig. 1) are very small, and _# quite alike, forming simple narrow plates, setose at the edges and pointing directly backwards. The caudal rami were broken off in the specimen examined, and are only conjecturally indicated in fig. 1 by dotted lines. When, however, recently preparing the specimen for retransmission, I detected at the bottom of the tube in which it was originally pre- served, a part, which, in spite of its anomalous appearance, I cannot but believe to be one of the caudal rami, detached from the specimen. In the accompanying cut I give a figure of this part, drawn to the same scale as figs. 3 and 4 on PI. III., which represent limbs from the same specimen. In its general appearance it somewhat resembles the outer branch of the pleopoda, but is very much larger, measuring nearly 5 mm. in length and thus about equalling that of the posterior division of the body, or tail. It has the form of a rather broad lamella, somewhat expanded in the middle and drawn out to a slightly incurved extremity, which is narrowly truncate. The edge, which I regard as the outer, is sharpened and evenly curved, with seven angular notches, to which as many spines may have been affixed, one of which, the most proximal, still remains. The latter is very small, but in all probability the others successively increased in size posteriorly. The inner edge is evenly convex in its proximal part, but becomes slightly sinuous towards the end. It is considerably thickened and exhibits two distinct lips, limiting a narrow Fig.- 1. — Caudal ramus of A'cbialopsis lypita. 28 THE VOYAGE OF H.M.S. CHALLENGER. sulcus, to the bottom of which a series of slender ciliated setae seems to have been affixed, the bases of most of them being still present. From the tip of the lamella a fragment of what may have been either a strong spine or a very elongate seta projects, and external to this another similar spine or seta may have been affixed. The whole lamella exhibits a delicate parenchymatous structure similar to that of the branchial legs. As the younger specimen is rather pellucid, some of the internal parts can also be faintly made out through the integuments. Thus, in a lateral view (fig. 1), a dark string is seen running from the cephalic part through the whole trunk and part of the pleon, at some distance from the dorsal surface. On examining the animal from the dorsal side, this string is found to be composed of two symmetrical narrow tubes filled with an opaque granular mass, and having between them another tube somewhat wider and more transparent. It therefore seems evident that the string referred to must represent the intestine, together with two elongate caeca accompanying it in the greater part of its length. But, besides, the anterior part of the body contains another internal organ of far greater dimensions, constituting a large opaque mass slightly tapering posteriorly and extending through the greater part of the trunk at a short distance from the ventral surface. The significance of this body I am unable to state with certainty. It cannot represent the generative organs, since it apparently forms an unpaired mass, and, moreover, its situa- tion would seem to forbid such an assumption. I am more inclined to regard it either as a kind of liver, or perhaps more properly an accumulation of fatty deposits, answering to the adipose body which in Nebalia envelops the whole intestine together with its caeca. The ventral ganglionic cord — only with great difficulty examined in the two other genera — is here immediately visible when the animal is examined from the lower side (fig. 2), lying, as it does, immediately inside the ventral cuticle and not being concealed by the branchial legs. The ganglia of the trunk, placed in the two other genera so closely together as almost to be coalescent, are in this animal wide apart and connected by very long commissures, in close proximity to each other. The ganglia of the pleon, of which at least the anterior is very distinctly seen, are considerably larger than those of the trunk, and furnish several nerve-trunks to each side, from which numerous fine nerves arise, partly innervating the musculature of the pleon and partly entering the pleopoda. Habitat. — The first specimen obtained, which, as above stated, was only represented by the carapace and a fragment of the front part of the body, was taken with the dredge in the Southern Ocean between Prince Edward Island and the Crozets. Station 146, December 29, 1873; lat. 46° 46' S., long. 45° 31' K; depth, 1375 fathoms; bottom, Globigerina ooze ; bottom temperature, 35°'6. The other more complete specimen came up in the trawl from a very considerable depth in the South Pacific, about midway between New Zealand and Chili. Station 289, October 23, 1875; lat. 39° 41' S., long. 131° 23' W.; depth, 2550 fathoms; bottom, red clay; bottom temperature, 34° -3. HOMOLOGY OF THE RECENT PHYLLOCARIDA. In the following pages I propose to discus, more in detail the homologies of the several r L the NebaliidJ, as compared with those in other recent Crns aeea As ZZl indicated at the beginning ef this Report my view as to the reWao^np ot the .enns Nebdia to other known Crustaeea somewhat differe from that set forth by most other naturalists, and the examination ef the two related new generic types pre- Tred IV the Challenger Expedition, aad described above, has confirmed me s ill mere m theopinion at first adopted from a thorough investigation of our common northern form, " eS^hing that the great resemblanee of N**a to certain Copcpoda, Fig. %—Diosaccus tenuicornis (Clans), mare. ^ escaped the attention of most naturalists ^ZlXo^^Z examination, whereas the much more remote affinity to the n J been dwelt upon, and even Professor Clans to whom w are a* *w ed hn for an admirable work on the free living Copepoda doe not seem to ^ „f this unmistakable resemblance ; I think it may be readily toun y parison, that the afflmty of SMb to the Copepoda is "*£■£££ the sevcral Podophthalmia, both as regards the external appeurance and the struct. appendages. In some other important points it exhibi Ma * * »£ n , .agreement with the Phyllopoda, and for that reason it ought to be, y 30 THE VOYAGE OF H.M.S. CHALLENGER. placed within the same order, the Branchiopoda, though representing the type of an anomalous suborder, the Phyllocarida. In order to facilitate a closer comparison of the Nebaliidse with the Copepoda, I sub- j oin a cut of a male specimen of one of our most common marine forms of the Harpactoid group, viz., Diosaccus tenuicornis (Claus). Homology of the Body-Divisions. — In examining the body of a Nebaliid, its general resemblance to that of a Copepod, especially of the Harpactoid group, may at once be recognised. But it is at the same time readily seen that there is in the Nebaliidae a distinct division of the body which is only faintly indicated in the Copepoda, viz., the trunk, or, as it is generally termed, the thorax. What is described as thorax in the Copepoda does not at all answer to the thorax in the higher Crustacea, but undoubtedly is homologous with the anterior part of the "abdomen" in these Crustacea, or the divi- sion in the Nebaliidse described above as the pleon, whereas the so-called abdomen in the Copepoda evidently answers to only the posterior part of the abdomen in the higher Crustacea or the division in the Nebaliidte succeeding the pleon, and described above as the tail. This is especially distinctly seen in the above described form, Paranebalia longi'pes (PI. I. fig. 1 ; PI. II. fig. 1), where the latter division is very sharply marked off from the pleon, both exhibiting a form very similar to that in the Copepoda, and, more- over, quite agreeing in function, since the tail here evidently admits of being moved as a whole upon the pleon, in the very same manner as in the Copepoda. A closer com- parison between the Nebaliidas and Copepoda thus clearly shows that the terminology generally adopted in describing the higher Crustacea has been wrongly applied as regards the lower forms (Copepoda), since the divisions " thorax " and " abdomen " in the former do not answer to the similarly named divisions in the latter. This misapprehension may indeed have been the cause why the affinity of Nebalia to the Copepoda has not been re- cognised. Thus, in order to explain the supposed abnormal number of segments in the " abdomen " of Nebalia, Professor Claus has set forth an hypothesis, wdiich seems to me very unreasonable, viz., that the two last segments together with the caudal rami in Nebalia answer to the telson in the Podophthalmia, which latter part, he suggests, has been originally formed by several segments. The fact is, however, that the so-called abdomen in Nebalia does not show any similarity at all to that division in the higher Crustacea, whereas it is constructed upon the very same type as in the Copepoda, the number of segments being in full accordance with that found in a great number of these Crustacea, admitting the above given explanation of the homology of the body-divisions in both. As to the limit between the two divisions in the Nebaliidas, described above as pleon and tail, it should be remembered that the first segment of the latter division, properly speaking, answers to the segment in the Copepoda generally described as the last thoracic segment, but which in most of the forms evidently has a much closer relation to the succeeding division, the tail, or, as it is wrongly termed, the abdomen. REPORT ON THE PHYLLOCARIDA. 31 Homology of the Carapace. — As above mentioned, the carapace in Nebalia has been adduced as a character showing the affinity of this genus to the Podophthalmia, and especially the Schizopoda. On closer examination we shall, however, find that according to this character it might with quite as good reason be classed among the Phyllopoda ; for both the finer structure of the carapace and the manner in which it is connected with the body are rather more in accordance with the latter Crustacea than with the Podoph- thalmia. Moreover, the presence of a well -developed adductor muscle, never found in any Podophthalmia, gives the carapace in the Nebaliidse a very marked phyllopodous character. As to form and relation to the body, it exhibits, as it were, an intermediate condition between the carapace in Apus and the bivalved shell in Linmadia. The jointed rostral plate is a character neither found in the Podophthalmia nor in the Phyllopoda, whereas a quite similar movable rostral projection is met with in some Copepoda of the Harpactoid group, and in the latter forms, moreover, the lateral parts of the so-called cephalic segment are found to extend more or less down the sides, so as to include between them the bases of the antennae and most of the oral parts, thus assuming the character of a bivalvular carapace, though being still connate with the body along the dorsal surface. The greatly developed carapace, by which the Nebaliidse at first sight seem to be so very sharply distinguished from the Copepoda, may thus be found to have in fact its homologue also in the latter Crustacea. Homology of the Eyes. — The eyes form another character wrongly adduced to show the affinity of Nebalia to the Podophthalmia. In reality the eyes in the Nebaliidse, though properly termed stalked and mobile, differ essentially from those in the Podoph- thalmia by their much simpler structure and by the want of a distinct facetted cornea. On the other hand, they are found to agree, both as to form and structure, very closely with the eyes in a well-known family of the Phyllopoda, the Branchipodidae. Homology of the Antennulai. — These limbs certainly exhibit a structure very different from that met with in other Branchiopoda, but they' are also quite dissimilar to the corresponding limbs in the Podophthalmia, differing essentially as well by the abnormal number of joints in the peduncle, this being in all higher Crustacea invariably but three, as also by the peculiar setose lamella appended to the end of the peduncle. To compare this lamella, as proposed by some authors, to the so-called antennal scale belonging to the succeeding pairs of limbs, the antennae, in Decapods and Schizopods, is, in my opinion, quite unreasonable. Neither can it properly be regarded as homologous with the inner flagellum in these Crustacea or to the accessory flagellum in the Amphipoda, since it is affixed outside the true flagellum, which latter undoubtedly answers to the outer flagellum in other Crustacea, bearing, as it does, the characteristic sensory appendages, generally termed olfactory cilia. Thus the lamella under consideration cannot properly be compared to anything met with in the higher Crustacea, but apparently represents a characteristic feature peculiar to the Nebaliidse. I think we may better understand the 32 THE VOYAGE OF H.M.S. CHALLENGER. structure of the antennulae in the latter forms by instituting a comparison with the cor- responding limbs in the Copepoda. Thus, in the Harpactoidea we find that the anterior pair of antennas, which undoubtedly answer to the antennulse in the Nebaliidse, are composed each of two rather distinctly defined divisions, a thicker proximal part and a much narrower distal part, the former consisting pretty constantly of four articulations, the last of which forms anteriorly a lateral expansion bearing a slender, ribbon-like appendage, the latter composed of a varying number of articulations but always very sharply marked off from the proximal part. There cannot in my judgment be any doubt that the proximal part of the anterior antennas in these Copepoda answers to the anten- nular peduncle in the Nebaliidse, and the distal part to the flagellum. The lateral expansion of the last joint of the proximal part in the former is also well seen in the Nebaliidse, and, moreover, in all probability the ribbon-like appendage affixed to that expansion in the Harpactoidea is a homologue of the setose lamella in the Nebaliidse. Thus, all parts of the antennulse in the latter forms seem to be in fact represented in the Copepod antennula, and the apparently abnormal number of joints in the peduncle as compared with that in the higher Crustacea, will turn out to be quite normal in relation to the Copepoda. Homology of -the Antennte. — In comparing the antennae in the Nebaliidse with those in the higher Crustacea, we find at once their structure very different from that in any known form of the Podophthalmia, whereas they at first sight somewhat resemble the lower antennae in the Amphipoda. On closer examination, however, it is easily found that they differ materially also from those in the last named group of Crustacea by the peduncle being only triarticulate, whereas in all known Malacostraca the number of joints is far greater, generally amounting to five in all. Moreover, the peculiar geniculate bend at the middle of the peduncle is rather different from what is generally met with in the Malacostraca. I think we may also in the case of these limbs more properly derive their structure from that met with in the Copepoda, especially those of the Harpactoid group. In these Copepoda the posterior antennae are, as is well known, very small but composed of three distinctly defined segments, the last of which forms with the preceding a strong geniculate bend, and it may readily be found, by comparison, that these three segments together perfectly agree in form with the antennal peduncle in the Nebaliidse. At the tip of the last joint, moreover, in some forms a small imperfectly defined terminal joint may be distinguished, representing a rudiment of the flagellum. The accessory branch generally found affixed to the middle of the second segment is sometimes quite obsolete, whereby the accordance becomes still more pro- nounced. The peculiar modification of the flagellum in the male of Nebalia, somewhat similar to what is met with in the Amphipoda and Cumacea, does not seem to have any more general significance, since in the nearly related genus Parcmebalia there is no difference whatever to be found in the antennae of the male and female, whereas it REPORT ON THE PHYLLOCARIDA. 33 is here only the antennulse which are peculiarly modified, as is also the case in the Copepoda. Homology of the Oval Parts. — These appendages appear on the whole very dissimilar to those in other Branchiopoda, and their structure has consequently been adduced to show the closer affinity of Nebalia to the Podophthalmia. Especially is the great development of the mandibular palp certainly a very characteristic feature, since such a palp is never met with in any adult Branchiopod. In form this palp, however, exhibits, as has also been pointed out by Professor Claus, a much closer resemblance to that in the Amphipoda than to the palp in the Podophthalmia, and the structure of the mandible itself is also rather different, showing, by the reduction of the cutting edge and the great development of the molar surface, more similarity to that met with in the greater part of the Branchiopoda. In the Copepoda the mandibles are, as is well known, in most cases provided with a well-developed palp, but this palp is generally biramose or bears a so-called branchial appendage, which however is often very reduced in size and in some forms even quite obsolete, whereby the palp acquires a certain similarity to the mandibular palp in the Nebaliidse. The very slight development of the posterior lip or metastoma in the Nebaliidse is a character apparently more in accordance with the Entomostraca than with the Malacostraca, in which latter this part is always well developed and rather large. As to the structure of the maxillae, both pairs seem to me to be essentially different from those in the higher Crustacea, and especially is the structure of the first pair quite peculiar and only admits of a very general comparison with those in other Crustacea. In the structure of the second pair I find, however, a well-marked resemblance to the maxillae in a great number of Copepoda, and this resemblance is especially very striking in the case of the genus Nebaliopsis, as stated above. In the Copepoda, it is true, only a single pair of maxillae are present, but this pair I think may more properly be regarded as homologous with the second pair in the Nebaliidse, the first pair not being developed in the former Crustacea. The number of maxillae in the Nebaliidse cannot be adduced as showing their affinity to the Macacostraca, since we find at least in all Phyllopoda both pairs distinctly developed, though of a rather simple structure. Homology of the Branchial Legs. — These limbs, in my opinion, undoubtedly stamp the Nebaliidse as true Branchiopoda, agreeing, as they do, both in structure and function with the so-called branchial feet in other forms of this order. This has, however, been partly combated by some naturalists, who regard them as more closely agreeing with the thoracic legs in the Schizopoda; and, in fact, if we had confined our" examination to the structure of these limbs in the above described genus Paranebalia, such a view might perhaps to a certain extent be advocated, as in this form they certainly exhibit an appearance somewhat approaching that of the legs in the Euphausiidae. On the other hand, I think, that no one will find any trace of similarity between the simple lamelli- (ZOOL. CHALL. EXP. rART I, VI. — 1887.) Kkk 5 34 THE VOYAGE OF H.M.S. CHALLENGEK. form branchial legs in Nebaliopsis and the thoracic legs of any Schizopod or other Malacostracan. In both these genera, however, these limbs are constructed upon the same general plan, the chief agreement being found in their adaptation for solely respir- atory purposes, and in so far they very materially differ from the thoracic legs in the Malacostraca. In Nebalia the branchial legs hold an intermediate position and consequently exhibit the most typical structure, whereas in the two above named generic types they represent, as it were, the extremes of a series of modifications, apparently standing in some connection with the very different conditions of life in the two, the one being a shallow-water form, the other a very marked deep-water form. That these limbs in Nebalia are much more closely related in structure to the branchial legs in other Branchiopoda, than to the thoracic legs in the Malacostraca, cannot in my opinion be disputed. It is true that they somewhat differ from the branchial legs in the Phyllopoda by the want of the so-called endites or lateral lobes of the endopodite. But it must be remembered, that these endites are only peculiar to the Phyllopoda, whereas in other undoubted Branchiopoda, as the Cladocera, they are more or less completely obsolete. Nor can it properly be adduced in support of the assumption of the nearer relationship of Nebalia to the Schizopoda, that the " thoracic " legs in the former exhibit the same principal parts as in the latter, since these parts, viz., the endopodite, exopodite, and epipodite, may easily be distinguished in Crustacea belonging to very different groups, and in most other Branchiopoda are quite so well definable as in Nebalia. The number of these limbs in all the known recent Phyllocarida is eight pairs, and if we consider the anterior pair as homologous with the maxillipeds in the Malacostraca, the number will certainly answer to that of the thoracic legs in these Crustacea, whereas their number in other Branchiopoda is very variable and at least in the Phyllopoda is much greater. But on closer examination we shall find that the anterior pair of branchial legs in the Nebaliidse evidently belong to the trunk and not to the cephalic part, as do the maxillipeds in all the known Malacostraca, and even in the Euphausiidse, where the maxillipeds are constructed upon the very same type as the succeeding legs, the development shows them clearly to belong to the head and not to the trunk.1 It thus results that the maxillipeds are wanting in the Nebaliidse, and that the number of the limbs belonging to the trunk is in reality greater than in the Malacostraca. In the Copepoda these limbs are only represented by the so-called maxillipeds, which according to the suggestion of Professor Claus may properly be regarded as the outer and inner branches of a single pair of limbs. Homology of the Pleopoda and Caudal Limbs : — The great agreement both in form and composition between the two posterior divisions of the body in the Nebaliidse and the so-called "thorax" and " abdomen" in the Copepoda, has been stated above. This agreement becomes still more evident by a comparison of the respective appendages. 1 See my Report on the Challenger Schizopoda, Zool. Chall. Exp., pt. xxxvii. EEPORT ON THE PHYLLOCAPJDA. 35 Thus, the pleopoda in the Nebaliida? undoubtedly agree very closely with the swimming legs in the Copepoda both as to structure and number, whereas their affinity to the pleopoda in the Malacostraca is much more remote. To compare them to the caudal limbs or the so-called saltatory legs in the Amphipoda, as proposed by Professor Claus, is, I think, objectionable, since the latter limbs belong to a different division of the body and more properly answer to the rudimentary caudal limbs in Nebalia. As to the mode in wdiich the pleopoda are moved, we find also the greatest resemblance between Nebalia and the Copepoda, especially those of the Harpactoid group, the movements being not at all rhythmical nor performed in the same rapid and almost vibratile manner as in the Amphipoda and most Podophthalmia (Euphausiidae). The two rudimentary caudal limbs in the Nebaliidae likewise find their homologues in the Copepoda. The first of these pairs are thus evidently homologous with the generally very small and imperfectly developed so-called last thoracic legs in the latter Crustacea, and on closer examination we shall find that also the second pair of these limbs are represented in the Copepoda, though in a very rudimentary state. Thus, on the segment succeeding the so-called last thoracic segment in the latter Crustacea, and generally described as the first abdominal segment, there occur in most of the forms on each side, close to the genital orifice, a small tubercle provided with a few spine-like bristles. This tubercle, more distinct in the males, is undoubtedly the rudiment of a pair of limbs, evidently answering to the second pair of l'udimentary caudal limbs in the Nebaliidae. The succeeding segments in the Copepoda as also in the Nebaliidae are entirely devoid of any trace of limbs, and in most of the Harpactoidea these segments exhibit a denticulate armature on the hinder edge quite similar to that in the Nebaliidae. Finally, the caudal rami appended to the last segment in the Nebaliidge are undoubtedly homologous with the so-called " furca " in the Copepoda, as also with the caudal lamellae in the Branchipodidae, being not true limbs but more properly representing a bipartite terminal segment. These terminal appendages in the Nebaliidse are therefore not at all homologous with the so-called uropoda in the Mala- costraca, the latter being true ventral limbs. Homology of the Internal Organs. — The internal organisation of Nebalia, which I have also thoroughly studied in the northern form Nebalia bipes, may on the whole be said to be much more advanced than in any other of the known Entomostraca, but I do not find in this respect any closer resemblance to that of the Podophthalmia, whereas it exhibits some points of apparent affinity to that met with in the Amphipoda. The nervous system seems in fact to differ essentially in structure from that generally met with in the other Branchiopoda by the ganglionic cord being not double but forming a single median series of ganglia, as in most Malacostraca. It must, however, be remembered, that the peculiar double ladder-like ventral cord, though very characteristic of the Branchiopoda, does not form a universal character of these Crustacea. Thus, among the Cladocera we find the nervous system in Leptodora. wholly dissimilar from 36 THE VOYAGE OF H.M.S. CHALLENGER. that in the other forms, and also in the Branchiura the nervous cord is constructed upon a quite different type. The intestine in Nebalia is markedly distinguished by the presence of a chitinous visceral skeleton never found in any Branchiopoda ; but this skeleton does not show any resemblance to that found in the Podophthalmia, whereas its similarity to that found in the Amphipoda is unmistakable. With the latter Crustacea Nebalia also apparently agrees in the two pairs of elongate cseca arising from the gastric part and accompanying the intestine in the greater part of its length, and more especially in the presence of another pair of elongate csecal appendages originating from the terminal part of the intestine and running anteriorly along its dorsal side. On the other hand, the occurence of two short curved cseca, projecting forwards within the head, is a character pointing rather more to the Branchiopoda, in some of which, as in the genus Daphnia, we find two quite similar curved caeca arising from the anterior part of the intestine. As to the structure of the heart the genus Nebalia may equally well be compared with the Branchiopoda as with other Crustacea, since the structure of this organ in the former is very variable, being in some forms quite short and sac-like with only a single pair of lateral venous fissures, in other forms more or less elongate with a varying number of such fissures. As in the Branchiopoda the vascular system is imperfectly developed, though a pair of short arteries may be traced at each end of the heart. The generative organs in both sexes, as in most of the Branchiopoda, are rather similar and very little complicated in structure, forming merely a pair of simple cylindrical tubes extending along the sides of the intestine, and opening by a short excretory duct on the ventral surface of the last segment of the trunk. In the Podophthalmia these organs are, as is well known, much more complicated, and, moreover, always exhibit a very marked difference in their structure in the two sexes. Development. — The development of Nebalia cannot in my judgment be adduced as showing any close affinity between this form and the Decapoda, as suggested by Metschnikoff ; and this has partly also been admitted by Professor Claus, in so far, at least, that he has pointed out that the stage of the Nebalia-embrjo, termed by Metschnikoff the Zoea-stage, and on which this author chiefly bases his suggestion of the decapodous nature of Nebalia, does not in reality correspond to that stage in the Podophthalmia, but more properly to the earliest stage of the Cyclops-fovm in the Copepoda. Neither can I find that the supposed agreement between the development of Nebalia and that of Mysis points to any true consanguinity between these genera. As is well known, we find a very similar direct development also in a great number of other Crustacea not at all belonging to the Podojththahnia, as in the Cumacea and Isopoda, and even in the Phyllopoda I have recently had an opportunity of stating an instance of a quite direct development, without any metamorphosis, and agreeing, more- EEPORT ON THE PHYLLOCAEIDA. 37 over, in all essential points rather well with that found in Nebalia} The apparent agreement with the Schizopoda in the mode in which the ova and embryos are borne in Nebalia during their development, is considerably lessened by the circumstance that there is in the latter form no trace of any true incubatory pouch, the ova being simply received within the valves of the carapace and kept in place by the aid of the branchial legs. Concluding Remarks on the Phylogeny of the Nebaliidae. — Owing to the suggestion made by most naturalists, that Nebalia forms a direct transition between the Phyllopoda and Podophthalmia, it has generally been supposed that the Nebaliidae have descended from the Phyllopoda, and that, on the other hand, all the Podophthalmia should be regarded as descendants from Nebalia-\ike ancestors. In his interesting treatise on the phylogenetic relationship of the Malacostraca, Dr. Boas has sought to strengthen the latter supposition by instituting a close comparison of the limbs in Nebalia with those in the Malacostraca, and has thereby been led to the result, that the connecting link between Nebalia and the great bulk of the Podophthalmia is represented by the Euphausiidae, from which again all the other forms of that division are supposed to have descended. It wovdd seem that the chief reason that has led Dr. Boas to this view as to the supposed close relationship between Nebalia and the Euphausiidae, is the apparent agreement in the number of segments composing the anterior division of the body (cephalothorax), and the uniformity in structure of the eight pairs of limbs succeeding the oral parts in both. For in all other points the difference is in reality so very great, as in most cases only to admit of the statement of a very general homology, such as could also be made by comparing almost any forms of Crustacea. It has been stated above that the resemblance of the branchial legs to the legs in the Euphausiidae is in the genus Paranebalia considerably more pronounced than in Nebalia. But notwithstanding this agreement, I still believe that there is no true relationship between the Nebaliidae and the Euphausiidas, and that the above mentioned apparent conformity in the con- struction of some of the limbs in both is merely accidental, a view that seems also to be confirmed by the fact that in an otherwise closely related genus, Nebaliopsis, these limbs exhibit a form showing no similarity whatever to the corresponding limbs in the Euphausiidae. If any true consanguinity had existed between the Nebaliidas and the Euphausiidas, the agreement between the two would certainly not be restricted to the above named limbs, but might have been traced also in the rest of the organisation. But this is by no means the case. On the contrary, it may easily be found on closer com- parison, that the Euphausiidae in so far differ from the Nebaliidae even considerably more than is the case with a great number of other Malacostraca. As to my own view on the probable phylogenetical relation of the Nebaliidae to other 1 On Cyclestheria hislopi (Baird), a New Generic Type of Biralved Phyllopoda, Forhandl. Vidensk. Selsk. Clirist, 1887 (to be shortly published). 38 THE VOYAGE OF H.M.S. CHALLENGER. Crustacea, I am inclined to agree on the whole with the suggestion set forth by Dr. A. Packard in his valuable treatise on the Phyllopod Crustacea of North America, believing that the Nebaliidae may have descended from some Copepod-like ancestors, whereas they do not show any relation whatever to the Podophthalmia, the latter having in all probability developed independently by a separate line from some Nauplius- or Zoea-like form. From the same stem that gave origin to the Nebaliidas, I think that also the other Branchiopoda may be derived, the latter having apparently become rather con- siderably modified in various ways to adapt them to the somewhat exceptional conditions under which they live, whereas the Nebaliidte have still preserved much of the external appearance which may have distinguished the progenitors of the order, while their internal organisation has become much more modified. TABLE OF CONTENTS. Introduction, . Principal Works on Eecent Phtllocarida, . General Eemarks on the Morphology of the Phyllocarida Definition of the Family Nebaliid^:, Description of Genera and Species, . Nebaliid^:, . Paranebalia, Claus, . longipes (Willemoes Suhm), Nebaliopsis, n. gen., . hjpica, n. sp., . Homology of the Eecent Pyllocarida, Homology of the Body-Divisions, . n „ Carapace, „ Eyes, . )( „ Antennute, ,, Antennae, „ Oral Parts, „ Branchial Legs, . „ Pleopoda and Caudal Limbs, „ „ Internal Organs, Development, . Concluding Eemarks on the Phylogeny of the Nebaliidas, page 1 2 3 6 9 9 9 10 21 22 29 30 31 31 31 32 33 33 34 35 36 37 PLATE I, (ZOOL. CHALL. EXP. PART LVI. — 1887.) — Kkk. PLATE 1. Paranebalia longipes (Willemoes Suhm). Fig. 1. Adult female, seen from the left side; magnified about thirty- four diameters. The sjjecimen has been rendered pellucid by mounting in Canada balsam, so as to exhibit in sitii the several limbs covered by the carapace, as also some of the internal organs. R, rostral plate ; O, eye ; a1, antennula ; a2, antenna ; M, mandible ; Mp, mandibular palp ; m1, maxilla of first pair ; m2, maxilla of second pair ; brp, branchial legs ; pi, pleopoda ; ep1, ep~, rudimentary caudal limbs of first and second pairs ; ur, caudal rami ; Cm, adductor muscle of carapace ; Ov, ovary. Fig. 2. Rostral plate, seen from below ; magnified fifty diameters. Fig. 3. Left eye, lateral view. Fig. 4. Left antennula. Fig. 5. One of the olfactory filaments from the flagellum of same ; highly magnified. Fig. 6. Serrate projection of last segment of peduncle of same. Fig. 7. Left antenna. Fig. 8. Anterior lip, seen from below. Fig. 9. Mandibles, together with the palps and adductor muscle ; front view. Fig. 10. Masticatory part of a mandible ; more highly magnified. Fig. 11. Mandibular palp, lateral view. Fig. 1 2. Maxilla of first pair. Fig. 13. Maxilla of second pair. TheVcvaSe of H.M.S "Challenger* Phyliocarida,Pl I Lift W Schlachter, Stockholm GO. Sars.del. PARANEBALIA LONGIPES,(Will.Suhm). PLATE II. PLATE II. Paranebalia longipes (Willemoes Sukm) — continued. Fig. 1. Adult female, viewed from above ; magnified about thirty-four diameters. Fig. 2. Branchial leg of first pair ; magnified fifty diameters. Fig. 3. Branchial leg of fifth pair. Fig. 4. Branchial leg of last pair. Fig. 5. Pleopod of first pair. Fig. 6. Pleopod of second pair. Fig. 6a. Lateral lobe of inner branch of same ; more highly magnified. Fig. 6b. One of the chitinous hooks from the same lobe ; magnified three hundred and eighty diameters. Fig. 7. Pleopod of last pair. Fig. 8. First pair of rudimentary caudal limbs. Fig. 9. Rudimentary caudal limb of second pair. Fig. 10. Last caudal segment, together with right caudal ramus, seen from below. Fig. 11. Embryo taken from the incubatory cavity of an adult female specimen, seen from below; magnified seventy-eight diameters. R, rostral plate; O, eyes; a1, antennulae ; a2, antennae ; L, anterior lip ; M, mandibles ; m1, maxillae of first pair; m2, maxillae of second pair ; brp, branchial legs ; pi, pleopoda; ur, caudal rami. Fig. 12. Same embryo, viewed from the right side. C, carapace ; remaining letters as in fig-. 1 1 . The Voyage of H M .""Challenger" PhyUocarida/Pl.II. G.O.Sars del Lift"W ScHachter, Stockholm PARANEBALIA LONGIPES,(Will.Suhm). PLATE III. (ZOOL. CIIALL. EXP. FART LVI. 1887.) — Kkk. PLATE III. Nebaliojjsis typica, n. gen. et sp. Fig. 1. Young specimen, viewed from the left side; magnified about five diameters. Fig. 2. Same specimen, ventral aspect. Fig. 3. Right branchial leg of seventh pair from same specimen ; magnified twenty-four diameters. Fig. 4. Pleopod of third pair from same specimen. Fig. 5. Carapace of another, apparently adult specimen, seen from the left side ; very slightly magnified. Fig. 6. Same carapace, viewed from above. Fig. 7. Rostral plate of same specimen, seen from above ; somewhat more magnified. Fig. 8. Fragment of the cephalic part of same specimen together with its several appendages, viewed from the left side ; magnified about five diameters. 0, eye ; a1, antennula ; a2, antenna ; L, anterior lip ; M, mandible ; Mp, mandibular palp. Fig. 9. Same fragment, front view. The terminal part of the antennulse as also the antennae and mandibles are omitted. 0, eyes; a1, bases of the antennuhe; L, anterior lip. Fig. 10. Maxilla of second pair from same specimen ; magnified fourteen diameters. The Voyage of H MSThallenger" PhyllocancLa,PlIII. G.O.Sarsflel InfttW ScHachter Stockholm NEBALIOPSISTYPICA.n.gen.&sp. THE VOYAGE OF H.M.S. CHALLENGER ZOOLOGY. REPORT on the Pteropoda collected by H.M.S. Challenger during the Years 1873-76. By Paul Pelseneer, D.Sc. (Brussels). PART I.-The GYMNOSOMATA. INTRODUCTION. 1. History of the Group. It is now more than two hundred years since a Gymnosomatous Pteropod was first described and figured;1 but it was only at the beginning of this century that the. organisation of these animals was investigated, and the group Pteropoda established.2 As to the subdivision Gymnosomata, its establishment is of a still later date. Peron and Lesueur, who considered the Pteropoda to form a much larger group than they really do,3 divided these animals into two subgroups : — (l) those without shells; (2) those with shells4 — and it has been found that this division, established on the very empirical character of the presence or the absence of a shell, is quite justified by the anatomical differences, as will be seen further on. Nevertheless, the majority of zoologists continued for a long time to divide the Pteropoda into non-natural groups. Thus, Cuvier5 distinguished those forms with 1 Clime limacina, in 1676, by Martens, Spitzbergisehe oder gronlandische Reisebeschreibung, p. 169, pi. P. fig./. 2 By Cuvier, in 1804 (Memoire sur l'Hyale et le Pneumoderme ; Ann. Mus. Hist. Nat. Paris, t. iv. p. 232), and not in 1798, as noticed by Bronn (Die Klassen unci Ordnungen des Thierreichs, Bd. iii. p. 584). 3 They embraced, in fact, Opisthobranchia (Phyllirhoe, Glaucus), Heteropoda {Carinaria, Firola), and even Ctenophora {Callianira). 4 Histoire de la famille des Mollusques Ptiropodes; Ann. Mus. Hist. Nat. Paris, t. xv. p. 57, 1810. 6 Le Regne animal, ed. 1, 1817. (zool. chaix. exp. — part lviii. — 1887.) Mmm 1 2 THE VOYAGE OF H.M.S. CHALLENGES. distinct heads from those without distinct heads, and, badly applying this character, placed some without shells and some with shells in the same subdivision. Oken1 did the same, placing Pneumonoderma and Clio in one class, and Clione and Cavolinia in another class. It was not until 1824 that de Blainville, relying on very judicious anatomical con- siderations, arrived at the same systematic result as Peron and Lesueur, and divided the Pteropoda (which he named Aporobranchia) into Gymnosomata and Thecosomata.2 He rejected the strange forms that Peron and Lesueur had introduced into the Pteropod group, but he left Phyttirhoe,* which formed in his classification a third group, Psilosomata, of the same value as the two others. Other groupings based on characters drawn from one organ alone, led to very bad results. Thus, Gray4 divided the Pteropoda into Dactylobranchia and Pterobranchia, the genus Cavolinia alone forming the first division and all the other genera being united in the second. Rang,5 who followed the Cuvierian subdivisions, through insufficient study of characters placed some Pteropoda without shells and some with shells in the family with a distinct head, and tried to do the same in the case of the group without a distinct head.6 This was, however, rather exaggerating the love of symmetry, although not so much as Oken,7 who desired that each order should contain four families, each famdy four genera, &c. Latreille,8 on the other hand, divided the Pteropoda according to the size of their fins, into " Macropterygiens " and " Microptdrygiens." Pneumonoderma alone formed the second group, while all the other forms were included in the first. Cuvier9 and his school did not adopt the classification of de Blainville, but followed the divisions established in the first edition of Le Regne animal. Since then, the non-natural arrangements have been gradually abandoned ; and the division of the Pteropoda into two distinct groups, Thecosomata and Gymnosomata, is now generally adopted. There is, however, a slight misunderstanding amongst some zoologists as to the relative extent of these two subdivisions ; but we shall see further on that this disagreement is really without foundation, and that the separation of the two groups is very clear. When Cuvier established his division of Pteropoda, only two species (forming two 1 Lehrbuoh der Zoologie, Bd. i., 1815. 2 Diet, d. Sci. Nat., t. xxxii. p. 271. 3 This animal remained here for a long time, until it was put in its right place, among the Nudibranchia. 4 London Medical Repository, p. 235, 1821. 6 Manuel de l'histoire naturelle des Mollusques, 1829. c Description d'un genre nouveau de la classe des Pteropodes, Ann. d. Sci. Nat., ser. 1, t. v. p. 284, 1825. 7 Loc. cit. 8 Les families naturelles du Regne animal, 1825. 9 Le Regne animal, ed. 2. REPORT ON THE PTEROPODA. 3 genera) of Gymnosoinata were known. But since then, the number of forms, both generic and specific, has increased in a very remarkable manner, chiefly owing to the results of several scientific expeditions. During the first half of this century, it was, above all, French naturalists who, as a result of their travels, described new Gymnosomatous Pteropoda. It is sufficient to name Quoy and Gaimard, Rang, d'Orbigny, and Souleyet. But, more recently, explorations of the great oceanic basins have been carried on in a more scientific and methodical manner and with much more perfect means of investigation. One may say that this new era was inaugurated by the Challenger Expedition, and several nations have already followed the example of England in this department. But, notwithstanding the great progress that these recent scientific expeditions have been the means of effecting in zoology, the Gymnosomatous Pteropoda have remained but little studied and a comparatively little known group. The Gymnosomatous Pteropoda include but few species, they are difficult to gather, and very few naturalists have busied themselves with the task. The published descrip- tions of a great number of these species are very often incomplete and based on insufficient characters ; and the figures which represent them are unfortunately nearly always imperfect.1 The types from which these descriptions have been elaborated have now been lost, so that it is excessively difficult to distinguish between the series of forms hitherto named. In these circumstances, it appeared useful to attempt to bring a little order and light into this chaos. Accordingly, while engaged with the description of the Gymnosomatous Pteropoda collected by H.M.S. Challenger (which description will form the essential part of this Report), I have made, at the request of Mr. John Murray, a critical examination of all the known genera and species of this interesting group of Mollusca. Besides the animals collected by the Challenger Expedition, I have had the advantage of being able to study specimens from several museums and scientific expeditions, viz. : — 1 . The Pteropoda belonging to the Museum of Brussels. 2 . The valuable collection of Gymnosomatous Pteropoda in the Museum of Hamburg. 3. The Gymnosomatous Pteropoda of the U.S. National Museum at Washington (a collection of great interest). 4. The Pteropoda of the " Triton " expedition. 5. The Pteropoda collected by the Italian ship " Vettor Pisani," during its scientific voyage round the world (1882-85). 6. The Pteropoda collected by Mr. John Rattray, F.R.S.E., during the cruise of the S.S. " Buccaneer," off the west coast of tropical Africa (1885-86). 7. The Pteropoda for sale at the Zoological Station at Naples. I take this opportunity to thank especially Professor Spencer F. Baird, and Mr. Wm, H. Dall, who obligingly sent to me the Gymnosomatous Pteropoda in the U.S. 1 The zoologists who deserve the most of these reproaches are indisputably Quoy and Gaimard. 4 THE VOYAGE OF H.M.S. CHALLENGER. National Museum ; Professor Pagenstecher and Dr. George Pfeffer, by whose kindness I was enabled to study those in the Hamburg Museum ; and the Zoological Commission of the " Vettor Pisani," who sent to me for examination the Pteropods collected by this ship. I have also to thank, for information and other assistance, Dr. John D. M'Donald, R.N., Dr. J. E. V. Boas of Copenhagen,1 Dr. W. Giesbrecht of Naples, and Professor Ed. Perrier of Paris. In short, my best thanks are due to all those who, in any way, have assisted me. I have described with much care all the species that I have been able to study, and I have figured them whenever there was not a sufficiently exact representation of them in existence. But in spite of all the material employed for the elaboration of this Report, it is impos- sible to give here a complete monograph of all the living Gymnosomata. For, besides the forms which are now well known, there are many others which are insufficiently described by the naturalists who made them known, of which the types have been lost, and which have not again been found since their discovery. Thus, we have only a few indications of the existence of some species still nearly unknown. I have drawn attention to these forms, hoping that it may facilitate their recognition by future explorers. After the description of the different species I shall examine : — 1. Their geographical distribution. 2. Their mutual phylogenetic relations. 2. The Gymnosomata and their Distinction from the Thecosomata. The adult Gymnosomata are chiefly characterised as follows : — By the absence of a mantle-skirt, pallia! cavity and shell ; by the presence of a well-developed head, bearing two pairs of tentacles, of which the two posterior bear rudimentary eyes ; by two fins of which the anterior edges axe not joined together backwards, above the mouth ; and by the anus being situated at the right side of the body. Every Pteropod possessing these characters belongs to the Gymnosomatous group, and every Pteropod which does not possesses these characters is a Thecosomatous form, so clearly and well-marked is the separation between these two divisions. As I before said, the anatomical characters of the Pteropoda entirely agree with their division into Gymnosomata and Thecosomata. If these names are not strictly exact (because 1 The text and the plates of this Report were already far advanced when Boas' work appeared (Spolia atlantica. Bidrag til Pteropodernes Morphologi og Systematik, K. dansk. Vidensh Selsh Skriv., 1886). As Boas and I are agreed on many points relative to the Gymnosomata, it must he said that our results have been quite ndependently obtained. But I have taken advantage of Boas' special discoveries to render this monograph more complete. REPORT ON THE PTEROPODA. 5 the former possess a shell at the beginning of their larval life), the groups that they represent are quite natural ones. The following table shows the principal differences between the adult Gymnosomata and Thecosomata (the two groups also present considerable differences in their embryonic development) : — Gymnosomata. Thecosomata. No mantle-skirt nor shell. A mantle-skirt and a shell. No pallial cavity. A pallial cavity. A well-developed head. No distinct head. Two pairs of tentacles.1 Only one pair of tentacles. Fins not joining in front, above the mouth. Fins joined at the anterior edge, above the mouth. Stomach without horny plates. Stomach provided with horny plates. Anus on the right side. Anus on the left side. Cerebral ganglia connected together above Cerebral ganglia separated by a long commis- the oesophagus. sure and situated laterally to the oesophagus. Among the other differences which exist between the two groups, the following may be noticed. 1 . The foot in the Gymnosomatous Pteropods is quite distinct from the fins ; it is formed of a posterior lobe and two antero-lateral ones, joined in front, in the form of two longitudinal lips. Between these two lobes a small tubercle is generally found behind, formed by folds of the skin. The foot and fins in the Thecosomatous Pteropods form a single and continuous mass. 2. The penis of the Gymnosomata is latero-ventral and situated on the right side of the foot. The penis of the Thecosomata is anterior and cephalo-dorsal. 3. In the Gymnosomata there is an evaginable proboscis (the anterior protrusible portion of the digestive tract) of the acrembolic type,2 generally bearing buccal append- ages, which may be completely protruded. Tn the Thecosomata there is no proboscis at all. 4. Besides the jaw and the radula, the buccal cavity of the Gymnosomata contains hook-sacs,3 a pair of evaginable sacs opening at each side of the radula. No Thecoso- matous Pteropod is provided with such hook-sacs. 5. The jaw of the Gymnosomata, composed of a great number of horny plates or spines, forms only one piece, in which two halves are, indeed, visible, but these 1 Paul Pelseneer, The Cephalic Appendages of the' Gymnosomatous Pteropoda, Quart. Joum. Micr. Sci., 1885, p. 505. 2 Ray-Lankester, Mollusca, Encyclopaedia Britannica, ed. 9, vol. xvi. p. 653. 3 Coecums buccaux, van Beneden ; spinose tentacular organs, Huxley ; Hakensiicken, Gegenbaur ; cheek pouches, buccal pouches, MacDonald ; cylindres retractiles, Fischer; hook-bearing processes of the mouth, Ray Lankester. THE VOYAGE OF H.M.S, CHALLENGER. are always united in the median line. The jaw of the Thecosomata is always formed of two quite distinct pieces. 6. The radula of the Gymnosomata is provided with long, lateral teeth, with large basal pieces ; the number of the teeth in a transverse row generally increases with age, until the adult state is reached, so that a know- ledge of the radula of young specimens is of little use for specific identification. The radula of the Thecosomata has but one lateral tooth on each side, with a small basal piece. In spite of the well-marked differences between the two groups, there are, however, some forms about the systematic position of which all zoologists do not agree, some placing them among the Gymnosomata, and others among the Thecosomata. Halopsyche [Eurybia), for instance, has been placed by many among the Thecosomata, because its organisation was not well known ; but we shall see further on tbat it is indis- putably a Gymnosomatous Pteropod. As for the Cymbuliidse, even Fol ' cpiestions whether they have not more affinities with the Gymnosomatous than with the Thecosomatous Pteropoda. But they cannot be placed among the former, because they possess all the essential characters of the Thecosomata as given in the foregoing table, and do not exhibit any of those of the Gymnosomata. There is, therefore, nothing to justify Tiberi's opinion that Tiedemannia (Gleba), should be placed among the latter.2 PlG. 1. — Diagram of a Pneunwnoderma, seen from the right side, showing the different external parts of a Gymnosomatous Pteropod. 1, right hook-sac; 2, position of the jaw; 3, ventral papilla of the proboscis ; 4, pro- boscis ; 5, right acetabuliferous buccal ap- pendage ; 6, right buccal tentacle ; 7, head ; 8, position of the right nuchal tentacle; 9, right anterior lobe of the foot; 10, orifice of the penis ; 11, posterior lobe of the foot ; 12, anus ; 13, right fin ; 14, genital orifice and spermatic groove, hidden beneath the fin ; 15, lateral gill ; 16, lateral somatic crest of the lateral gill ; 17, posterior gill (quad- rangular crest) ; 18, radiating crest of the posterior gill ; 19, fringes of the radiating branchial crest. 3. The Habits of the Gymnosomata. The Gymnosomatous Pteropoda do not form a numerous group of species, but they are found in all the seas of the world, and sometimes in great abundance. They are small animals, not more than an inch and a half long, and are carnivorous, often feeding on their Thecosomatous allies. Naturalists are not agreed as to the position assumed by the Gymnosomata when 1 Sur le diiveloppement des Pteropodes, Archives d. Zool. exper., ser. 1, t. iv. p. 173. - Mollusques inarms d'ltalie, Ann. Soc. Malawi. Belg., t. siii. p. 77. REPORT ON THE PTEROPODA. 7 swimming. On this subject Souleyet x contradicts the statements of Peron,2 Eang, 3 and d'Orbigny.4 It appears, however, that they generally swim in a nearly vertical position with the head uppermost, or else slightly sloping, so that the ventral (pedal) side is turned upwards. They swim by moving the ends of the fins successively backwards and forwards. The Gymnosomata, like the other Pteropoda, are nocturnal in their habits, ascending to the surface during the night and sinking to a lower level in the daytime, being driven down by the brightness of the light ; and thus one can scarcely gather Pteropoda on the surface during the day. Mr Murray, however, informs me that he has frequently taken Gymnosomata at the surface in the Arctic seas during the daytime with a hand net. It is impossible to distinguish, as d'Orbigny supposed,5 twilight from nocturnal species. That writer thought that each species inhabits a definite zone of depth, since the different species come to the surface at different times, but it has been experimentally shown by Alexander Agassiz,6 that no species of the pelagic- fauna descends to more than 100 fathoms. It is then in this bathymetrical zone that the Pteropoda perform their daily oscillations, accompanied by some other pelagic- animals, several of which serve as their prey. The anatomical structure of the Gymnosomata will be described in the anatomical part of this Report. 4. The Gymnosomata of the Challenger Expedition. The number of Gymnosomatous Pteropoda collected by the Challenger Expedition is comparatively not very extensive. This can be readdy explained by the fact that these animals are not very numerous in species, and since they are pelagic and nocturnal in habit, they can only be captured by the tow-net during the night, when trawlings were not frequently made, or by sinking the nets to slight depths during the daytime, and also because many common surface forms were not always preserved.7 The number of the Challenger Gymnosomata is nevertheless greater than that of any other scientific expedition. The thirteen8 stations at which they were captured yielded one new genus and specimens of all the other genera, except Clione, furnish - 1 Voyage de la Bonite, Zoologie, t. ii. p. 273. 2 Histoire de la famille des Mollusques Pteropodes, Ann. Mas. Hist. Nat. Paris, t. xv. p. 69. 3 Description d'un genre nouveau de la elasse des Pteropodes, Ann. d. Sci. Nat., ser. 1, t. v. pp. 286, 287. 4 Voyage dans l'Amerique muridionale, t. v. p. 128. 6 Loc. cit., p. 67. 6 Bull. Mus. Com-p. Zo'61., vol. vi. p. 153. 7 Narr. Chall. Exp., vol. i. p. 270, note by Dr. John Murray. 8 I do not include in this number the station (not specified) between Cape Verde and Bahia, at which the Pneu- monoderma (violaceum, I think), recorded in the Narrative of the Cruise, vol. i. p. 219, was taken. This station is unknown to me, because I did not find, in the collection sent to me, specimens of Pnevmonoderma from any station between these two localities. 8 THE VOYAGE OF H.M.S. CHALLENGER. ing altogether eleven different species (comprising nearly half of the known species), of which four are new and several others imperfectly known, or not yet figured. In one case a single station yielded three species. Chierchia was therefore wrong, when he cited the Pteropoda as an example to show that several special collections made by the Italian ship "Vettor Pisani," are more complete than the corresponding series of the Challenger Expedition,1 because during the cruise of the " Vettor Pisani " Gymnosomata were captured at only three stations, and these belong to only two genera and four species. The Challenger specimens are divided as follows :— Genera. Species previously known. New Species. Dexiobranehaea, .... SpongiobrancJisea, .... Pneunwnoderma, .... Clionopsis, ..... Notobranchxa, .... Halopsyche, ..... 22 1 3 1 1 1 1 1 6 7 4 In the following descriptive part of this Report, those species of Gymnosomata procured by the Challenger Expedition are marked with an asterisk. 1 Collezioni per studi di scienze naturali fatte nel viaggio intorno al mondo della K. Corvetta Vettor Pisani, 1882-85, p. 115. 2 The larvae of a Gymnosoniatous Pteropod, described by H. N. Moseley (On Stylocbus pelagicus, &c, Quart. Journ. Mier. Sci., p. 32, pi. iii. fig. 14, 1877), belong to this genus, and very probably to a species previously known. DESCRIPTION OF GENERA AND SPECIES. PTEROPODA, Cuvier. GYMNOSOMATA, de Blainville. Deutocephcda, Wagner, 1885. * Pterota, Boas, 1886.2 THE GENERA AND FAMILIES OF GYMNOSOMATA. Many genera of Gymnosoniatous Pteropods have been described, but several of them cannot be retained. I. Because they are synonyms : — ■ 1. jEgle, Oken,3 = Pneumonoderma, Cuvier. 2. Cirrifer, Pfeffer,4= Pneumonoderma, Cuvier. 3. Cliodita, Quoy and Gaimard,5 = Clione, Pallas. 4. Euribia, Rang,6 non Hiibner = Halopsijche, Bronn. 5. Pneumodermo'psis, Bronn,7 = Dexiobranclisea, Boas. II. Because they only represent the larval state of genera of which the adult is already known : — 1. Trichocyclus, Eschscholtz,8 in which the larvae of different genera have been placed.9 2. Trizonius, Busch,10 = larva of Pneumonoderma. 1 Die Wirbellosen des weissen Meeres, Bel. i. p. 119. 2 Spolia Atlantica, p. 179. 3 Lehrbuch der Zoologie, Bd. i. p. 326. 4 Uebersicht der auf S. M. Sehiff " Gazelle " und von Dr. Jagor gesammelt Pteropoden, Monatsber. d. k. preuss. Akad. d. TViss. Berlin, p. 249, 1879. 6 Description de cinq genres de Mollusques, Ann. d. Sci. Nat., ser. 1, t. vi. p. 74, 1825. 6 Description de deux genres nouveaux appartenant a la classe des Pteropodes, Ann. d. Sci. Nat., ser. 1, t. xii. p. 320, 1827. ' Die Klassen und Ordnungen des Thierreichs, Bd. iii. pp. 584, 645. 8 Bericht iiber die zoologische Ausbeute wahrend der Reise von Cronstadt bis St. Peter und Paul, Oken, Isis, 1825. p. 735. 9 Trichocyclus dumerili, Esch. = larva of Spongidbramchsea australis, d'Orb. Trichocyclus mediterranem, Costa = larva of Clionojjsis krohni, Troschel. 10 Beobachtungen iiber Anatomie und Entwicklung einiger wirbellosen Seethiere, p. 112. (ZOOL. CHALL. EXP. PART. LVIII. 1887.) Mmni 2 10 THE VOYAGE OF H.M.S. CHALLENGER. III. Because they have been insufficiently characterised, never again gathered, and the type specimens have not been preserved ; they are therefore too doubtful to be introduced into systematic nomenclature : — 1. Pelagia, Quoy and Gaimard.1 2. Cymodocea, d'Orbigny.2 3. Homoderma, van Beneden.3 Among the genera hitherto described I only admit six and the new genus Notobranchsea (these six genera are also those that were accepted by Boas,4 before the description of Notobranchsea). They are the following : — 1. Dexiobranchsea, Boas. 2. Spongiobranchsea, d'Orbigny. 3. Pneumonoderma, Cuvier. 4. Clionopsis, Troschel. 5. Clione, Pallas. 6. Halopsyche, Bronn. These six genera, and the new one Notobranchsea, are well known, well charac- terised and distinct. The following table gives their principal distinctive characters :— Key to the Genera. I. A specialised branchial apparatus. 1. A lateral gill. A. No posterior gill, . B. A posterior gill. a. A circular posterior gill, b. A tetraradiate posterior gill, 2. No lateral gill. A. No buccal appendages B. Buccal appendages, II. No specialised branchial apparatus. 1. Body lengthened, posteriorly pointed, 2. Body ovoid, posteriorly rounded, Dexiobranchsea. Spongiobranchxa. Pneumonoderma. Clionopsis. Notobranchxa. Clione. Halopsyche. How must we subdivide these seven genera of the Gymnosomatous group \ Fischer, who considers the Pteropoda to be a class of the same value as the Gastro- poda (although the organisation of the Pteropoda shows, in the clearest manner, that they only form a subdivision of the Euthyneurous Gastropoda), distinguishes among the Gymnosomata two suborders :* — Scleroderma, containing only the genus Halopsyche, and Malacoderma, embracing all the other genera. But the character on which this subdivision is based is very insufficient, because the teguments of Halopsyche are not 1 Voyage de decouvertes de 1' Astrolabe, Zoologie, t. ii. p. 192= Pteropelagia, Bronn. 2 Voyage dans l'Amerique meridiqnale, t. v. p. 133 = Pterocymodocea, Bronn. 3 Exercices zootomiques, p 54, pi. iii. figs. 4, 5. 4 Vorlaufige Mittheilung iiber einige Gymnosomen Pteropoden, Zool. An;:eiger, p. 690, 1885. 5 Manuel de Conchyliologie, p. 422. REPORT ON THE PTEROPODA. 11 more tough than those of some Pneumonoderma. Moreover, the Gymnosomata form a group of too low a status, and contain forms differing too little from one another to enable one to establish within it more important divisions than families. Three of the seven genera I have admitted — Pneumonoderma, Spongiobranch&a, and Dexiobranchsea — show a very close resemblance to one another, and clearly differ from the other genera by the presence of acetabuhferous buccal appendages and of a right lateral gill ; they constitute the family Pneumonoderrnatidae. Clione and Halopsyche are quite separated from the other genera by the complete want of branchial apparatus ; they differ as much by the form of the body as by that of the fins, by the buccal appendages, &c, and respectively represent the families of the Clionidae and Halopsychidae. Clionopsis, till now ranked near Clione, because its organisation was imperfectly known, cannot remain among the Clionidae. In several characters (the presence of a tetraradiate posterior gill and of a dorsal spot) this genus resembles some Pneumono- dermatidae more than Clione; but other more important characters (the absence of a lateral gill and of acetabuhferous buccal appendages, the presence of a retractile proboscis of extraordinary length) exclude it from so natural and so well-characterised a group, which includes Pneumonoderma, Spongiobranchsea, and Dexiobranchsea. There- fore Clionopsis represents a special family — Clionopsidae. As to the new genus Notobranchsea, it must be excluded from all the above cited families : from the Clionidae and Halopsychidag, by the presence of a gill ; from the Pneumonoderrnatidae, by the want of the lateral gill and suckers ; and from the Clionop- sidae, by the presence of buccal appendages and by the form of the gill and the foot. A new family must therefore be established, Notobranchasidae, for this new genus. Hence, the Gymnosomata are divisible into five families, of which the first (Pneumonodermatidae) includes three genera. Family I. Pneumonodermatidae. 1840. Pneumodermidx, Gray (pars), Synopsis of the Contents of the British Museum, p. 86. 1842. Pneumodermidx, d'Orbigny, Paleontologie franeaise, terrains cr(3taces, t. ii. p. 4. 1846. Pneumodermoidx, Agassiz, Nomenclator zoologicus, Index, p. 299. 1852. Clios, Souleyet (piars), Histoire naturelle des Mollusqnes Pteropodes, p. 74. 1855. Clioidea, Gegenbaur {pan), Uiitersuchungen iiber Pteropoden und Heteropoden, p. 212. 1858. Pneinnodermonidx, H. and A. Adams (pars), The Genera of recent Mollusca, vol. L p. 62. 1862. Pneumodermidx, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 645.- 1871. Pneumodermatidx, Dall, Description of Sixty New Forms of Molluscs from the West Coast of North America, Amer. Journ. of Couchology, vol. vi. p. 139. 1881. Cliidx, Fischer (pars), Manuel de Conchyliologie, p. 243. Characters. — Suckers on the ventral side of the protrusible anterior portion of the buccal cavity. A lateral gill on the right side. A jaw. Pigmented skin. 12 THE VOYAGE OF H.M.S. CHALLENGER. Such are the distinctive characters of this family, which only Bronn understood in the sense in which we now treat it, including the three genera Dexiobranclisea, Spongio- branchasa and Pneumonoderma. Description. — The body of the Pneumonodermatidae is more or less elongated, and rounded at its posterior extremity. The length of the cephalic region and the form of the lateral gill vary in the different genera. The foot shows a posterior, generally long lobe, and two rather narrow antero-lateral lobes, united throughout their entire length to the cephalic region, between which posteriorly is a small tubercle. The fins, tapering at their free end, are nearly triangular in shape, with bent margins; they are bluntly contracted at the posterior side of the base. The colour in the individuals of this family is nearly uniform, being violet-brown. A dorsal glandular spot must be noticed as another character common to the three genera of Pneumonodermatidse ; this also exists in Clionopsis,1 but it is here smaller and longer (see Spongiobranchwa, PL I. fig. 7). The suckers of the buccal cavity are generally inserted on appendages ; three in Dexiobranclisea (two symmetrical and lateral and one median, anterior during the evagination of the front part of the digestive tract), and two (lateral and symmetrical) in Spongiobrancliwa and Pneumonoderma. There is on the ventral wall of the buccal cavity in the two last genera, in front of the lateral appendages (during the evagination of the proboscis) a small median papilla. I found this papilla in Spongiobranch&a australis (fig. 2, p. 19), Pneumonoderma violaceum, Pneumonoderma pacificum (PL II. fig. 5), and Pneumononoderma souleyeti (PL II. fig. 6). Boas, who previously noticed this in some specimens of Pneumonoderma, regarded it as the rudiment of the median acetabuliferous appendage of Dexiobranchasa.2 The acetabuliferous appendages, which are inserted on the ventral wall of the buccal cavity, are flattened parallel to the body-wall, and show many different forms, being often short, broad and triangular, or else long, narrow, and with nearly parallel margins. In Spongiobranchsea and Pneumonoderma the lateral appendages, which are typically developed, fold backwards when retracted, so that their base is then their most anterior part. They bear suckers only on the side which is turned against the wall of the buccal cavity during retraction. During the extension of the appendages, the two acetabuliferous surfaces are turned towards one another (since the appendages are inserted on the ventro-lateral sides of the cylinder formed by the exserted part of the digestive tract), in order to be applied together, on opposite sides of any prey which may pass in front of the mouth. 1 Paul Pelseneer, Recherches sur le systeme nerveux ars), Gegenbaur, Untersuchungen iiber Pteropoden und Heteropoden, p. 74. 18G2. Pneumodermopsis, Bronn, Die Klassen und Ordmmgen des Thierreichs, Bd. iii. pp. 584, 645. 1885. Dexiobranchxa, Boas, Vorlaufige Mittheilung iiber einige Gymnosomen Pteropoden, Zool. Anzeiger, No. 210, p. 688. Body rather long, pointed posteriorly. Visceral mass not extending to the posterior end of the body. Head moderately lengthened ; anterior (labial or oral) tentacles strong. Foot ; posterior lobe pointed at its free extremity. Lateral Gill; a simple, more or less long appendage, of a triangular shape, without longitudinal crests or fringes. Proboscis moderately long. The radilla in the adult has always a median tooth. The hook-sacs are but little developed, and consist of small deep depressions which con- tain short hooks. Buccal Appendages. — There are morphologically three acetabuliferous appendages ; one median, which becomes anterior during the evagination of the proboscis, and always bears five suckers, and two lateral appendages. The two latter, or the median appendage alone, or all three appendages, may be atrophied, and their suckers are then directly inserted on the ventral wall of the buccal cavity, forming groups corresponding to the atrophied appendages. The five suckers of the median appendage, or of the corresponding- group, are always disposed in the following manner — a median distal sucker on the free extremity of the appendage, and two lateral ones on each side. The number of the suckers of the lateral appendage varies according to the species. The skin is pigmented, and the two posterior ciliated rings generally remain a very long time. Boas chiefly contributed to make this genus 2 well known ; he was the first to describe with precision the acetabuliferous appendages, and of five species now known he discovered three. I think the name Dexiobranchsea, that he gave to this group, ought to be preserved, although Bronn had previously formed the genus Pneumodermopsis for the same type from one of its principal characters (viz., the absence of a posterior gill), because the latter genus was founded on the insufficient description, without a figure, which Gegenbaur gave of Pneumodermon ciliatum. But Boas was the first to make known the true characters of this genus, and he has figured specimens under the name of Dexiobranckwa, and it is therefore desirable to preserve this last name, rather than that of Pneumodermopsis, which accords with no good characters nor figures. As I have said, this genus includes five species, that may be distinguished as follows : — 1 From h%ios, on the right side, and li^ayxia, gill. 2 Von Ihering regarded it as a larva (Vergleichende Anatomie des Nervensystems und Phylogenie der Mollusken p. 24G). REPORT ON THE PTEROPODA. 15 Key to the Species. I. The two distal lateral suckers of the median appendage larger than the two. proximal ones. 1. The lateral acetabuliferous appendages still distinct, . . Dexiobranchxa ciliata. 2. The lateral acetabuliferous appendages atrophied, . . . Bexiobranchsca minuta. II. The four lateral suckers of the median appendage of the same size. 1. The median appendage atrophied, .... Bcxiobranchsea simplex. 2. The median appendage well developed. A. The suckers corresponding to the lateral appendages disposed in one row, ....... Bexiobranch&a paucidens. B. The suckers corresponding to the lateral appendages disposed in two rows, ....... Bexiobvanchsea polyeotyla. *1. Dexlobrancliaea ciliata (Gegenbaur) (PI. I. fig. 1). 1839. (?) Pneumodermon mediterraneum, van Beneden, Exercices zootorniques, pi. iii. fig. 3. 1855. Pneumodermon ciliatum, Gegenbaur, Untersuchungen iiber Pteropoden und Heteropoden, p. 74. 1862. Pneumodermopsis ciliata, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. G19. 1886. Bexiobranchxa ciliata, Boas, Spolia atlantica, K. dansk. Vidensk. Selsk. Skriv., 6 Rajkke, Bd. iv. p. 159, pi. vii. fig. 104. 1 886. Bexiobranchxa improvisa, Boas, Spolia atlantica, p. 1 60. Foot. — Posterior lobe very long and pointed. Lateral Gill very long. Buccal Appendages. — The lateral appendages are still present, but very sbort ; they have the form of a horizontal projection on each side of the ventral wall of the buccal cavity, and each projection bears on its free edge from six to nine suckers of moderate size. The five suckers corresponding to the median appendage are as follows : — - the median distal sucker is larger than those of the lateral appendages ; of the four other suckers, the anterior (distal) pair is formed of two very large suckers (much larger than the last one), placed on long thin peduncles, and with a broad ledge having a pointed beak on the side opposite to the insertion of the peduncle.1 The suckers of the posterior (proximal) pair are of about the same size as the suckers of the lateral appendages. Raclula. — Formula 7:1 : 7 ; the median tooth tricuspid. Hook-sacs with from thirty to forty hooks. Colour. — Greyish- violet. Length. — Not larger than 1 cm.2 The specimen figured is a little contracted at its posterior part. 1 Boas, Spolia atlantica, pl.viii. fig. 113. 2 The measurements given in this Report are always from the buccal aperture to the posterior extremity of the body, excluding therefore the evaginable parts of the mouth, namely, the proboscis, buccal appendages, and hook-sacs. 16 THE VOYAGE OF H.M.S. CHALLENGER. Challenger Specimen. — An old larva; April 26, 1876; St. Vincent, Cape Verde Islands, lat. 16° 49' N., long. 25° 14' W. Habitat.— Fseroe Channel, lat. 60° 15' N., long. 7° 30' W. ("Triton" Expedition, August 20, 1882 ; surface temperature about 54° F.), and neighbouring localities (Copenhagen Museum) ; west coast of North Africa, lat. 7° 1' N, long. 15° 54' W. (Mr. John Kattray, January 2, 1886 ; surface temperature 82°'9 F.) ; and lat. 16° 49' N., long. 25° 14' W. (Challenger Expedition) ; Messina. *2. Dexiobranchsea minuta, n. sp. (PI. I. fig. 2). Foot. — Posterior lobe very long. Lateral Gill rather long. Cephalic Region long. Buccal Appendages. — The suckers corresponding to the lateral appendages are arranged in two rows, about sixteen in number ; the median appendage is very long and well developed ; the median sucker and the two distal lateral ones are nearly of the same size as those of the lateral groups, and are inserted all three together on the distal extremity of the appendage, on short peduncles ; the two proximal lateral suckers are very small and inserted near the base of the appendage, on very thin peduncles. Radula and Hook-sacs not investigated in the single specimen collected. Colour. — Scarcely any ; a few small spots of brown pigment on the right side of the body. Length, about 3 mm. The evaginated part of the penis resembles that of Clione. Habitat.— Near Station 247, Japan to Honolulu; July 4, 1875; lat. 36° 42' N, long. 179° 50' W. 3. Dexiobranchsea simplex, Boas (PI. I. fig. 3). 1886. Dexiobrancheea simplex, Boas, Spolia atlantica, K. dansk. Vidensk. Selsk. Skriv., 6 E»kke, Bd. iv. p. 160. Foot. — Posterior lobe moderately long. Lateral Gill rather short. Buccal Appendages. — Suckers directly inserted on the wall of the buccal cavity, forming three groups corresponding to the three appendages. Each lateral group includes two small suckers, scarcely pedunculated. The terminal median sucker of the median group is of very large size (broad and high), and measures nearly 1 mm. in diameter;1 it shows a distinct ledge and is placed on a short peduncle ; the two lateral suckers of each side 1 Boas, Spolia atlantica, pi. viii. fig. 114. REPORT ON THE PTEROPODA. 17 are very small and placed on long and thick peduncles, the anterior (distal) pair being- inserted at the same height as the large median sucker. Radula. — Formula 6:1:6; median tooth tricuspid. Hook-sacs. — Depressions containing from forty to fifty hooks. Colour. — Brownish. Length, 4"5 mm. The two posterior ciliated rings remain until a late period. Habitat. — South-East Pacific, off Caldera (Chili), about lat. 27° S. ; " Vettor Pisani" Expedition, February 18, 1883. 4. Dexiobranchasa paucidens, Boas.1 1886. Dexiobranchasa paucidens, Boas, Spolia atlantica, K. dansk. Vidensk. Selsk. Skriv., 6 Raekke, Bd. iv. p. 160, pi. vii. figs. 105, 106. Foot. — Posterior lobe rather short. Lateral Gill very short. Buccal Appendages. — The suckers corresponding to the lateral appendages are inserted on the ventral wall of the buccal cavity, by very short peduncles, from eleven to thirteen in number; they are not arranged in two distinct groups, but are placed in a single row, forming a continuous semicircle. The median appendage is well developed ; its four lateral suckers are larger than those corresponding to the lateral appendages, and are set on long and thick peduncles ; the terminal sucker is larger than these four. Radula. — Formula 2:1:2; the median tooth is bicuspid. Hook-sacs. — Each rudiment of a hook-sac bears from four to six very short, thick, and slightly bent hooks. Length, 5 mm. Habitat. — Atlantic Ocean, on the route from Brazil to Europe (Copenhagen Museum). *5. Dexiobranchasa polycotyla, Boas (PI. I. figs. 4, 5). 1886. Dexiobranchasa polycotyla, Boas, Spolia atlantica, K. dansk. Vidensk. Selsk. Skriv., 6 Raekke, Bd. iv. p. 161. Foot. — Posterior lobe long. Lateral Gill long. Buccal Appendages. — The suckers corresponding to the lateral appendages are arranged alternately in two rows, and form two distinct groups (right and left), each including from ten to twelve suckers with very short peduncles. Median appendage well developed, bearing a terminal sucker larger than those of the lateral groups ; the four lateral suckers of the median appendage are nearly of the same size as these last. 1 I have not been ahle to examine this species ; the following characters are compiled from Boas' description and figures. (zool. chall. exp. — part Lvin. — 1887.) Mmm 3 18 THE VOYAGE OF H.M.S. CHALLENGER. Radula. — Formula 3:1:3; the median tooth is tricuspid. Hook-sacs. — Each with about ten rather long hooks. Colour, brownish. Length, 5 mm. The two posterior ciliated rings remain until a late period. The three rings were still visible in a specimen 3 mm. long. Habitat. — The type-specimens, described by Boas, were gathered by the Italian ship " Vettor Pisani," in the South-East Pacific, off Caldera (Chili), about lat. 27° S., February 18, 1882. In a part of the collection of this scientific expedition, which had not been seen by Boas, I have found larvae of a species of Gymnosomatous Pteropod, caught off Panama about lat. 8° N, December 1883 (PL I. fig. 5), among which the oldest shows the peculiar acetabuliferous appendages of Dexiobranchsea polycotyla. Thus the geographical distribution of this species embraces the whole South-West Pacific, along almost the whole west coast of South America. Challenger Specimens. — The Challenger specimens which belong to this species, are the very young larvae of Gymnosomata described by H. N. Moseley,1 from Station 297, November 11, 1875, Tahiti to Valparaiso, lat. 37° 29' S., long. 83° 7' W. These larvae differ from those of Pneumonoderma, and resemble much that of Dexiobranchsea ciliata, described by Gegenbaur.2 Besides, the abundance of Dexiobranchtea in this region en- ables me to place these larvae with great certainty in this last genus. As these specimens have not been preserved, and since the formula of their radula has not been recorded, it may seem difficult to determine specifically larvae that are so young. Nevertheless, it is nearly certain that they belong to the present species, not only on account of its extensive geographical distribution, of its comparative abundance, and perhaps of the elongated form of the larva, but chiefly on account of the length of the posterior lobe of the foot. Spo7igiobranchsea,z d'Orbigny. 1825. " Cliodita," Quoy and Gaimard (pars), Description de cinq genres de Mollusques, Ann. d. Sci. Nat., s£r. 1, t. vi. p. 74. 1840. Spongiubranchea, d'Orbigny, Voyage dans TAme'rique meVidionale, t. v. p. 130. 1842. Spongiobranchia, d'Orbigny, Pal^ontologie frangaise, terrains cretact5s, t. ii. p. 4. 1856. Spongiobranchxa, Woodward, A Manual of the Mollusca, p. 209. 1862. Spongobranchia, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 645. 1885. Spongobranchus, Wagner, Die Wirbellosen des weissen Meeres, Bd. i. p. 120. Body ovate, rather long, round posteriorly ; visceral mass extending to the posterior extremity of the body ; skin rjigmented. 1 On Styloclms pelagieus, &c, Quart. Joum. Micr. Sci., 1877, p. 32, pi. iii. fig. 14. - Untersuclnmgen iiber Pteropoden und Heteropoden, pi. iv. fig. 11. 3 From tmo'/yiU, sponge, and /^ay^/a, gill. REPORT ON THE PTEROPODA. 19 Foot. — Posterior lobe lengthened, tapering greatly, and pointed at its posterior end. Head rather long ; labial tentacles long. Gills. — (l) Lateral gill little developed and very short, forming a simple prominence on the right side of the body ; (2) posterior or caudal gill as a simple membranous ring, without exterior radii and fringes or folds at its base. Acetabuliferous Appendages. — Two symmetrical appen- dages inserted on the ventral wall of the front portion of the buccal cavity, or protrusible part of the digestive tract ; each appendage bears, on the side which is median during the extension, suckers arranged in a single row, and of which the size increases from the base of the appendage to its distal extremity. The protrusible front portion of the digestive tract con- sists of a rather long proboscis (fig. 2, 1), on the ventral side of which, as in Pneumonoderma, is a papilla (fig. 2, 3) anterior (during the evagination of the proboscis) to the acetabuliferous appendages. Hook-sacs rather short, like those of Clione, and also containing teeth of different length, slightly bent. Radula always with a median tooth in the adult. Jaw rather small, with slender spines. Since d'Orbigny (1840), whose species australis is alone a true Spongiobranchsea,1 this genus has never been studied till lately by Boas and by me. Souleyet,2 Gegenbaur," Fischer,4 &c, have considered it as a Pneumonoderma badly described, and Krohn5 as an immature Pneumonoderma. It is, however, a very good, distinct, and well-characterised genus, as may be seen from the foregoing description and from the figures. Fig. 2. — The proboscis of Spongin- branchcea, expanded ; ventral aspect. Magnified about 8 diameters. 1, proboscis ; 2, hook-sacs ; 3, ventral papilla ; 4, acetabuliferous appen- dages ; 5, labial tentacles ; 6, head. * Spongiobranchsea australis, d'Orbigny (PI. I. figs. 6, 7). 1825. " Cliodita caduceus," Quoy and Gaimard, Description de cinq genres de Mollusques, Ann. d. Sci. Nat., ser. 1, t. vi. p. 74, pi. ii. fig. 2. 1840. Spongiobranchea australis, d'Orbigny, Voyage dans l'Amerique meridionale, t. v. p. 131, pi. xix. figs. 1-7. Acetabuliferous Appendages bearing on their median side from eight to ten suckers, with short peduncles. The three or four more proximal are very small, the largest (distal) are inserted on the ventral margin of the appendages, as in Pneumonoderma mediterraneum. Radula. — Formula 8:1:8. 1 Spongiobranchsea elongata, cl'Orb. = Clione longicaudata, Soul. 3 Untersuchitngen iiber Pteropoden und Htteropoden, p. 173. 2 Voyage de la Bonite, Zoologie, t. ii. p. 272. 4 Manuel de Concbyliologie, p. 424. Beitrage zur Entwickelungsgeschichte der Pteropoden und Heteropoden, p. 17. 20 THE VOYAGE OF H.M.S. CHALLENGER. Hook-sacs bearing about twenty long books, placed in tbe bottom of the sac, and several shorter on the lateral walls, of which the extremities come to the same height as those of the bottom hooks when the sac is evaginated. The other characters are the same as for the genus. The specimens of this species always show a constriction in the posterior half of the body, behind the dorsal spot, and at the height of the lateral gill. Length. — Maximum, 22 mm. Colour. — Grey-brown; a thin longitudinal white line, in the middle of the back, extending backwards to the above constriction. Challenger Specimens. — Station 158, March 7, 1874 ; Termination Land to Melbourne; lat. 50° l'S.,long. 123° 4' E. Station 314a, January 22, 1876 ; Sandy Point to Falkland Islands; lat. 51° 24' S., long. 61° 46' W. Between Stations 332 and 333, March 11, 1876; Rio de la Plata to Tristan da Cunha; lat. 36° 22' S., long. 26° 1' W. Habitat. — Antarctic Seas, within the isothermal line of 50° F. for August (it often does not even pass beyond the isotherm of 40°); from long. 60° W. to long. 123° E., and probably all around the Antarctic Pole. The extreme observed latitudes are lat. 36° 22' S. (Challenger Expedition, March 11, 1876), and lat. 54° 30' S. (d'Orbigny). The localities, " New Britannia, Marquesas," in the Hamburg Museum, are certainly erroneous. Observations. — It is evident from one of the original drawings of Pteropoda by Dr. Hooker (sent to me with the Challenger Pteropoda) that the English Antarctic Expeditin (1840), had obtained Spongiobranchsea australis. But this drawing does not bear any indication of locality. D'Orbigny only knew a single very small specimen of this species (7 to 8 mm. long) ; and it is thus easy to understand why he only saw six suckers on each buccal appendage. He says nothing about the lateral gill ; but he ought to have recognised it, as it may be seen in his figure.1 It is extraordinary that he says : anus " a, gauche,"2 the anus being on the right side as in all the Gymnosomata. I have already 3 expressed the opinion that Trichocyclus dumerili, Esch.,4 is the larva of Spongiobranchsea. I support this interpretation on the length of the labial tentacles of this larva.5 Johannes Midler had already said6 that, for this reason, Trichocyclus appeared to him to be more the larva of Clione than of Pneumonoderma ; 1 Voyage dans l'Amerique meridionale, t. v., pi. ix. fig. 2. 2 Loc. cit., p. 1.31. 3 Description d'un nouveau genre de Pteropode Gymnosome, Bull. Sci. Depart. Nord, 1886, p. 218. * Bericht iiber die zoologische Ausbeute wahrend der Keise von Cronstadt bis St. Peter nnd Paul ; Oken, Isis, 1825, p. 735, pi. v. rig. 4. 6 I niay here notice that Esehseholtz's figure is not exact, because it shows the first ciliated ring anterior to these tentacles. 6 Uebet verschiedene Formen von Seethieren, Archivf. Anat. u. Physiol, 1854, p. 70. REPORT ON THE PTEROPODA. 21 but the larvae of Clione, as well as the adults of this genus, have always the posterior extremity tapering and pointed, whereas that of Trichocyclus is round, as in Spongio- branchsea. The locality where Trichocyclus dumerili was obtained (" Stidsee "),1 also agrees well with my interpretation. I have identified Cliodita caduceus2 with Sp>ongiobranchsBa australis; because, in spite of the imperfect figure given by Quoy and Gaimard, one can there recognise the posterior gill of Spoiigiobranchsea, and the same constriction in the posterior half of the body as in the specimens of Spongiobrarichtea austrcdis. The locality of Cliodita caduceus (" entre l'ile de France et le Cap de Bonne Esperance") also agrees with this opinion. Pneumonoderma3 Cuvier. 1804. Pneumoderme, Cuvier, Meruoire concemant 1'animal de FHyale, &c, Ann. Mus. Hist. Nat. Paris, t. iv. p. 232. 1810. Pneumoderma, Peron and Lesueur, Histoire de la famille des Mollusques Pte'ropodes, &c, Ann. Mus. Hist. Nat. Paris, t. xv. p. 65. 1815. Pneumodermis, Oken, Lehrbuch der Zoologie, Bd. i. p. 326. 1815. JEgle, Oken, Lehrbuch der Zoologie, Bd. i. p. 326. 1819. Pneumodennon, Lamarck, Histoire naturelle des animaux sans vertebres, t. vi. p. 293. 1846. Pneumonoderma, Agassiz, Nomenclator Zoologicus, Index, p. 299. 1855. Pneumouodermt/m, Herrmannsen, Indicis generum Malacozoorum primordia, vol. i. p. 309. Body ovate, rather long, a little shrunk behind, and rounded at its posterior extremity ; visceral mass extending to that extremity. Skin pigmented, more coloured on the right than on the left side (especially at the head). Head very long, or else united to the body by a long neck (this lengthening is a consequence of the great length of the hook-sacs). Labial tentacles short, and wide at their bases. Foot. — Posterior lobe long, always rounded at its free extremity ; anterior lobes attached to the head in their entire length ; between them there is a small plicated tubercle, as in the Clionopsidse. Tlie Lateral Gill placed on the right side, is com- posed of (1) three longitudinal, parallel crests, thin and transparent, placed on the body wall; (2) a free ^^S^^mmS: part, posterior to these crests, of a trianglar shape, L^IX^SVi's^^ crests; 3' more or less elongate, of which the summit points backwards; the two free edges of this triangle bear crests which continue the lateral somatic ones (see fig. 1, p. 6). The median somatic crest continues on the triangular lobe (where it is very projecting) up to its summit. The two crests on the free sides 1 Loo. cit, Oken, Isis, 1825, p. 735. 2 Description de cinq genres de Mollusques, Ann. d. Sci. Nat., ser. 1, t. vi. p. 74, pi. ii. fig. 2. 3 From 7rnvuau, lung, and h^», skin. 22 THE VOYAGE OF H.M.S. CHALLENGER. of the triangular lobe bear fringes at their bases ; these fringes generally are continued on the lateral somatic crests. The median crest of the gill never bears fringes. The Posterior Gill is formed of a quadrangular projecting crest, thin, membranous and transparent, which surrounds the posterior end of the body. There arises from each corner of this principal crest, a longitudinal one directed forwards (the two crests arising on the right side, that is, on the side of the lateral gill, are always shorter than the two crests of the left side). The four radiating crests and the two anterior and posterior sides of the quadrangular one have fringes or foldings at their bases. The posterior gill of Pneumonoderma is an adaptive one (whereas the lateral gill morphologically corresponds with that of the Tectibranchiate Opisthobranchs, such as Aplysia, for example) ; it is but a modification of a part of the cutaneous envelope, which must assist the lateral gill in its breathing function. The formation of crests, the development of fringes at the base of these, and the subdivision of these fringes, are successive adaptations, favourable to the aeration of the blood by increasing the surface in contact with the water. It is difficult to understand how such a good observer as Gegenbaur,1 followed in this respect by Bronn,2 thinking to correct Cuvier, van Beneden, and Souleyet, has stated that the posterior gill of Pneumonoderma only shows three radiating crests. Is his observation based on monstrous or rather on specimens deformed from bad preservation ? Buccal Appendages. — Two symmetrical appendages, latero-ventrally inserted on the proboscis (PI. II. fig. 5) and having the form of flattened ribbons, more or less wide and long. They bear pedunculated suckers,3 on one side only, the median one during the extension of the appendages. Proboscis shorter than in Spongiobranchsea, and showing, in all (?) the species,4 the same median abdominal papilla as in Spongiobranchsea australis, placed in front of the acetabuliferous appendages when the proboscis is protruded. Radula, never with a median tooth in the adult (this tooth exists in the larvse) ; the number of the lateral teeth never varies much in the different species from four to six. Hook-sacs cylindrical and very long; they contain a great many slightly bent hooks, scattered over the whole wall of the sac, the longest at the extremity and the shortest at the base (during the evagination of the sac). The gills of this genus have nearly always been insufficiently described (as in the defective descriptions of Gegenbaur and Bronn), and still more often badly figured. In 1 Untersuchungen iiber Pteropoden, &c, p. 85. 2 Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 607. 3 The structure of these suckers was studied by Nieraiec (Recherches morphologiques sur les ventouses dans le regne animal, Recueil Zool. Suisse, t. ii., 1885), and by me (The Cephalic Appendages of the Pteropoda, Quart. Journ. Micr. Sci., 1885). 4 I found this papilla on Pneumonoderma violaceum, Pneumonoderma pacificum (PI. II. fig. 5) and Pneumonoderma souleyeti (PI. II. fig. 6). REPORT ON THE PTEROPODA. 23 regard to the posterior gill, de Blainville ' gives it as only formed of three sides (the right side of the quadrangular crest wanting). Several writers have figured fringes on the four sides of this crest; in this respect the figure of Rang 2 is much more exact, but the one given by d'Orbigny,3 and republished by so many modern authors,4 is quite erroneous, and makes this gill resemble fins. As for the lateral gill, it is generally forgotten or figured in a shapeless manner. Lastly, in regard to the structure of these gills, there is no figure, not even those' of Souleyet,6 which are by far the best and most exact, that shows the true conformation of the fringes. These fringes are differently formed in the different species. Those of Pneumonoderma souleyeti are very little developed, and only form simple undulations at the base of the crests. Those of Pneumonoderma pacificum form well-marked and regular undulations, those of one side alternating with those of the other, and resembling the fringes of the gill of Clionop>sis grandis. In Pneumonoderma boasi there are a few fringes more distinct, shrunk at their base, but markedly separate. Lastly, in the group including Pueumonoderma violaccum, Pneumonoderma peroni, and Pneumonoderma mediter- raneum, the fringes of the gills show their greatest specialisation; they are numerous, nearly contiguous, and subdivided into parallel ramifications, very close to one another, directed from the base of the fringe to its distal extremity (fig. 3, 3). The conformation of the gills is very useful for the distinction of the species. Huxley,6 and several others after him, as Woodward 7 and Fischer,8 have said that Pneumonoderma possesses a "small anomalous shell." I have examined a great many specimens of Pneumonoderma of all sizes, and even old larvae, and I can assert that like all the other Gymnosomata, Pneumonoderma is absolutely devoid of mantle skin and shell. It is quite probable on the other hand, whatever Lankester may say,9 that the larvae of Pneumonoderma bear a larval shell in the beginning of their development, as in Clione.10 The discrimination of the species of Pneumonoderma is certainly the most laborious part of the systematic treatment of the Gymnosomata. A considerable number of species have been described, but for the most part the descriptions are imperfect and insufficient, and the type specimens have been lost. A comparative and synthetical study of all the specific forms of this genus has not yet been made, and with the materials now available, it might not even yet give 1 Manuel de Malacologie et de Conchyliologie, pi. xliii. figs. 4, 46. 2 Histoire Naturelle des Mollusques Pteropodes, pi. ix. fig. 5. 3 Voyage dans l'Ame'rique me>idionale, t. v., pi. ix. fig. 11. 4 Bronn's Thierreieh, Claus' Text-book, Lankester's Mollusca, in Encyclopaedia Britannica. 5 Voyage de la Bonite, Zoologie, pi. xv. figs. 40-42. 0 On the Morphology of the Cephalous Mollusca, Phil. Trans., 1853, p. 40. 7 A Manual of the Mollusca, p. 208. 8 Manuel de Conchyliologie, p. 423. 9 Mollusca; Encyclopaedia Britannica, ed. 9, vol. xvi. p. 666. 10 Fol, Sur le developpement des Pteropodes, Archives d. Zool. exper., sir. 1, t. iv., pi. x. figs. 6, 7. 24 THE VOYAGE OF H.M.S. CHALLENGER. definitive results. The first, and a very commendable, attempt to bring light into this chaos, was made by Boas,1 and in the following systematic account of the genus Pneumonoderma I have availed myself of the results he has obtained ; this account is thus more perfect, although it cannot be considered as conclusive. I have found in the literature of the Pteropoda eleven different specific names attributed to Pneumonoderma. Among these eleven names there is one that belongs to a Dexiobranchsea — namely, Pneumodermon ciliatum, Gegenbaur, and among the ten others, there are — First, names which are synonyms, as will be shown further on : — 1. Pneumodermon audebardi, Eang,2= Pneumonoderma violaceum, d'Orbigny. 2. AZgle cucullata, Oken,3 = Pneumonoderma per oni, Lamarck. 3. Pneumodermis atlanticum, Oken,4 = Pneumonoderma violaceum, d'Orbigny. 4. Pneumodermon macrocotylum, Boas,s = Pneumonoderma mediterraneum, van Beneden. Secondly, several names applied to species of which the characterisation is quite insufficient, and of which the type specimens have been lost. One is therefore obliged to eliminate them from the actual systematic nomenclature of Pneumonoderma : — 1. Pneumodermon pellucidum, Quoy and Gaimard.0 2. Pneumodermon ruber, Quoy and Gaimard.7 I shall, however, refer again to these two species. The obscurity which surrounds the nomenclature of the genus Pneumonoderma results from the difficulty of finding distinctive characters between the different species, so that the descriptions which have been given of some forms might have been applied to others, which are nevertheless distinct. It may therefore be said that the real cause of this confusion is the fact that the zoologists who have undertaken the systematic classification of the Mollusca possess but little knowledge of the organisation of the animals that they describe, and therefore cannot appreciate which organs maintain constant characters throughout the whole genus, and which have characters that vary from one species to another. To convince oneself of this truth, one must compare the descriptions of Quoy and Gaimard, of Rang, and even of d'Orbigny and Dall, with those of Boas, and he will then understand the accuracy of this observation. The first well-known species of the present genus was Pneumonoderma peroni, Lamarck, from the Indian Ocean. 1 Spolia atlantica, pp. 152-154. 2 Histoire naturelle des Mollusques Pteropodes, pi. x. fig. 13. 3 Lehrbuch der Zoologie, Bd. i. p. 327 (err. typ., 273). 4 Loc, cit., p. 327. 6 Spolia atlantica, p. 152. 6 Voyage de decouvertes de l'Astrolabe, Zoologie, t. ii. p. 390, pi. xxviii. fig. 29. 7 Loc. cit., t. ii. p. 389, pi. xxviii. figs. 19-24. REPORT ON THE PTEROPODA. 25 Up to the time of d'Orbigny (1840), the Pneumonoderma of tlie North Atlantic (Pneumonoderma violaceum), was described under the same name ; and under this name also, Souleyet * reunites some different Pneumonoderma, of which not one was the true Pneumonoderma peroni. It is true that in 1815 Oken2 had distinguished the two forms (Pneumono- derma peroni and Pneumonoderma violaceum), and had respectively named them JEgle cucullata and Pneumodermis atlantica; but the diagnoses he gave of them are so brief and incorrect that these names cannot be retained, and must be regarded as synonyms. The two other names to retain with Pneumonoderma peroni and Pneumonoderma violaceum, are Pneumonoderma mediterraneum, van Beneden, and Pneumonoderma pacificum, Dall. Thus, among the ten different specific names above mentioned, there are only four which can be considered as definitive in the actual nomenclature. Besides these four species I shall describe two new ones, Pneumonoderma souleyeti and Pneumonoderma boasi, of which the former was collected by the Challenger Expedition. As I have already said, the different species of Pneumonoderma are very difficult to distinguish from one another. The distinctive character proposed by Boas is founded on the number and form of the suckers of the buccal appendages ; but this character is not sufficient for all the species. By adding to it the character drawn from the conformation of the gills, one can much more easily distinguish the six species above mentioned. By the conformation of the posterior gill, these six forms are divided into two groups, each including three species : — I. In the first group the radiating crests of the posterior gill are one-fifth of the length of the body, and bear numerous fringes (from seven to twelve), which are long and subdivided. The species of this group generally have the free part of the lateral gill rather short ; one of them does not possess fringes on the lateral somatic crest of this gill. II. In the second group, the radiating crests of the posterior gill are short (one-tenth of the length of the body), and bear a few fringes (about three), little projecting, and simple. Two species of this group (Pneumonoderma pacificum and Pneumonoderma souleyeti) have the free part of the lateral gill long, and the fringes of the posterior gill little developed. The following table, combining the characters taken from the gills with those of the acetabuliferous appendages, allows of the easy discrimination of the six species of Pneu- monoderma described in this Beport. 1 Voyage de la Bonite, Zoologie, pi. xiv. figs. 7-16. 2 Lehrbuch der Zoologie, Bd. i. p. 327. (ZOOL. CHALL. EXP.— PART lviii.— 1887.) Mmm 4 26 THE VOYAGE OF H.M.S. CHALLENGER. Key to the Species. I. Radiating crests of the posterior gill long, with subdivided and numerous fringes. 1. Each acetabuliferous appendage bearing seven large suckers inserted on the margins of the appendage, . . Pneurnonoderma mediterraneum. 2. Each acetabuliferous appendage bearing numerous small suckers on their whole median face. A. Each appendage with about forty suckers, . . Pneurnonoderma violaceum. B. Each appendage with about one hundred suckers, . Pneurnonoderma peroni. II. Radiating crests of the posterior gill short, with simple and not numerous fringes. 1. Fringes well marked, strongly separated, and shrunk at their tase, ....... Pneurnonoderma boasi. 2. Fringes slightly developed and forming merely simple, continuous and slightly projecting undulations. A. Triangular, rather short acetabuliferous appendages, each bearing about fifty suckers, . . . Pneurnonoderma pacificum. B. Very long acetabuliferous appendages with nearly parallel margins, each bearing about seventy-five suckers, ...... Pneurnonoderma souleyeti. 1. Pneurnonoderma mediterraneum, van Beneden (PI. I. fig. 8 ; PI. II. fig. 1). 1838. Pneumoderrnon mediterraneum, van Beneden, Note sur une nouvelle espece de Pneumo- derme, Mem. Acad. Sci. Bruxelles, t. xi. p. 13, pi. hi. figs. 1-2. 1852. Pneumoderrnon peronii, Souleyet (pars), Voyage de la Bonite, Zoologie, t. ii. p. 274, pi. xiv. figs. 13, 1 4, 16. 1855. Pneumoderrnon mediterraneum, Gegenbaur, Untersuchungen fiber Pteropoden und Hetero- poden, p. 213. 1859. Spongiobranchea australis, Chenu (pars), Manuel de Conchyliologie, t. i. p. 116, fig. 509. 1873. Pneumoderrnon peronii, Costa, Pteropodi della Fauna di Napoli, p. 22, pi. v. figs. 1-3. 1886. Pneumoderrnon macroeotylum, Boas, Spolia atlantica, K. dansk. Vidensk. Selsk., Skriv., 6 Raekke, Bd. iv. p. 152. 1886. Pneumoderrnon audebardi, Locard, Prodrome de Malacologie franchise, Mollusques marins, p. 19. Body rather long. Foot- — Posterior lobe very long (no other known Pneurnonoderma has it so long). Lateral Gill rather short, with fringes on the lateral somatic crests. Posterior Gill. — The four radiating crests are long, and possess on each side of their base about seven ramified fringes ; such fringes are also to be found on the anterior and posterior sides of the quadrangular crest. Acetabidiferous A-ppcndages rather long and not very wide at their base ; each bearing generally seven (sometimes six) suckers inserted on the median side of the appendage, the four largest on the abdominal margin, and the three smallest at the distal extremity and on the dorsal margin ; the latter are sometimes very small. The shape of the EEPOET ON THE PTEEOPODA. 27 suckers is nearly that of a fiat porringer with a rather narrow edge ; the largest are about one line in diameter. Radula. — Formula 6:0 : 6. Hook-sacs very long. Colour. — -Violet brown, more or less dark (Costa1 describes this species as red coloured). Length. — The adult specimens measure 17 mm. Habitat. — Mediterranean Sea (Naples, Messina, Nice) ; according to Boas,2 this species is also to be found in the Atlantic Ocean, and in the China Sea. Observations. — Contrary to the opinion of Boas, I consider the species with seven suckers, obtained frequently at Naples (and which he designates Pneumodermon macro- cotylum), as identical with the species from Nice, formerly described by van Beneden as Pneumonoderma rnediterraneum ; I therefore think that the latter name must be retained for it. Boas identifies his Pneumodermon macrocotylum with " Pneumodermon rnediter- raneum" Gegenbaur, which he consecpiently distinguishes from van Beneden's species. But Gegenbaur3 says : — "Acetabula five or six"; and van Beneden's figures4 also show the same number of suckers (five or six) . The suckers in van Beneden's figures really appear smaller than those of a well- preserved Pneumonoderma from Naples ; but in very contracted specimens (as those of van Beneden), the shape of the suckers is altered and they seem smaller. I have already said that the Mediterranean Gymnosomata are not yet well known. Gegenbaur observed a " Pneumodermon violaceum " " more common at Messina, during autumn, than Pneumodermon rnediterraneum" and bearing from ten to fourteen suckers on each buccal appendage. This form, of course, differs from Pneumonoderma violaceum, d'Orbigny, and has not since been seen in the adult state. Boas 5 has found some larvae from Messina bearing ten or eleven suckers on each appendage, and these undoubtedly belong to the species Pneumonoderma violaceum of Gegenbaur. Up to the present time adult Pneumonodermata have only been found in the Mediterranean, at Naples, Messina and Nice. But they must also live in other parts of this sea, since larvge have been found at Malaga (Trizonius csecus), by Busch,6 and in the Adriatic sea, at Trieste, by Johannes Midler.7 1 Pteropodi della fauna di Napoli, p. 22. 2 Spolia atlantica, p. 153. 3 Untersuclmngen uber Pteropodeii und Heteropoden, p. 213. 4 Loc. cit, p. 213. 6 Spoiia atlantica, p. 151. 6 Beobaebtungen iiber Anatomie und Entwicklung einiger wirbellosen Seetbiere, p. 112. 7 Ueber verschiedene Formen von Seetbiere, Archivf. Anat. u. PhysioL, p. 69, 1854. 28 THE VOYAGE OF H.M.S. CHALLENGER. *2. Pneumonoderma violaceum, d'Orbigny (PI. I. fig. 9). 1815. Pneumoderrnis atlantica, Oken, Lehrbuch der Zoologie, Bd. i. p. 327. 1832. Pneumodermon peronii, Quoy and Gairnard, Voyage de decouvertes de r Astrolabe, Zoologie, t. ii. p. 388, pi. xxviii. figs. 1-6. 1840. Pneumodermon violaceum, d'Orbigny, Voyage dans lAme>kpue meridionale, t. v., pi. ix. figs. 10-15. 1850. Pneumodermon cucutlatum, Gray, Catalogue of the Mollusca in the Collection of the British Museum, p. 39. 1852. Pneumodermon peronii, Souleyet (jmis), Voyage de la Bonite, Zoologie, t. ii. p. 274, pi. xiv. figs. 12, 15. 1852. Pneumodermon audebardi, Bang, Histoire naturelle des Mollusques Pt6ropodes, pi. x. fig. 13. 1859. Spongiobranchea australis, Chenu (pars), Manuel de Conchyliologie, t. i. p. 116, fig. 508. 1879. Cirrifer paradoxus, Pfeffer, Uebersicht der auf S. M. Schiff " Gazelle " und von Dr. Jagor gesammelten Pteropoden, Monatsber. d. k. preuss. Akad. d. Wiss. Berlin, p. 249, fig. 20. Body and head resembling those of Pneumonoderma mediterraneum. Foot. — Posterior lobe not so long as in the foregoing species. Lateral Gill short ; lateral somatic crests without fringes. Posterior Gill with long radiating crests, each of which bears about seven ramified fringes on each side of the base. Acetabidiferous Appendages triangular in shape, of moderate length, and rather large at their base. Each appendage bears about forty suckers covering their whole median face ; these suckers are smaller than those of Pneumonoderma mediterraneum, but have nearly the same shape. Colour. — Brownish ; the pigment being chiefly distributed on the right side. I have seen specimens in the Hamburg Museum which are nearly colourless, or albinos, and have but very little pigment on the right side of the head. Length. — About 13 mm. Souleyet1 gives the length of the Atlantic Pneumonoderma that he studied (and which belongs to the present species), as 25 mm.; but this is certainly with the proboscis and hook-sacs everted. Challenger Specimens. — Near Station 354, May 7, 1876; Atlantic, off the Azores; lat. 34° 22' N., long. 34° 23' W. In the Narrative of the Cruise (vol. i. p. 219), there is also mentioned a "Pneumo- dermon " (without indication of species) collected between Cape Verde and Bahia, Stations 93 to 128, July 27 to September 14, 1875. 2 I did not, however, find, in the collections sent to me, any specimens of Pneumonoderma from any station between the 1 Voyage de la Bonite, Zoologie, t. ii. p. 274. 2 In this trip there were only four trawlings (Stations 101, 104, 106 and 107, from St. Vincent to St. Paul's Rocks), between lat. 5° 48' N. and lat. 1° 22' N., and from long. 14° 20' W. to long. 26° 36' W.; this zone quite agrees with the area of distribution of Pneumonoderma violaceum. REPORT ON THE PTEROPODA. 29 two above cited. But since all the stations between these latter are north of lat. 15° S., the specimens noticed undoubtedly belong to Pneumonoderma violaceum. Habitat. — This species lives only in the Atlantic Ocean, and especially in that part adjoining the African coast, and is distributed between lat. 45° N. and lat. 15° S., the extreme western localities (long. 45° W.) are in the northern hemisphere, the most eastern (long. 6° W.) in the southern hemisphere. Observations. — Boas,1 with some doubt, ascribes to the present species specimens from the Indian Ocean, but they are certainly distinct, because each of their buccal appendages only bears from twenty-five to thirty suckers. In this species Boas 2 also includes specimens from the south-east Pacific (off Caldera, Chili), but they are the types of a new species, Pneumonoderma boasi, described further on. #3. Pneumonoderma peroni, Lamarck (PL II. fig. 2). 1815. sEgle cucullata, Oken, Lelirbuch der Zoologie, Bd. i. p. 327. 1819. Pneumodermon peronii, Lamarck, Histoire naturelle des Animaux sans Vertebres, t. vi. p. 294. Body. — Head and foot as in the foregoing species. Lateral Gill short, with fringed lateral somatic crests. Posterior Gill with long radiating crests, each of which bears, on each side of the base, as many as twelve contiguous fringes, very much developed and subdivided. Acetabuliferous Appendages triangular, with a large base, bearing about a hundred suckers a little smaller than those of Pneumonoderma violaceum, more convex and inserted on very extensible peduncles. Radida. — In the largest specimens I was able to examine (about 15 mm.), the formula was 4:0:4. Colour. — Bluish-grey ; pigment divided into numerous little spots, visible under a magnifying glass. Length. — A little greater than that of Pneumonoderma violaceum; the adult specimens measure 15 mm. Cliallenger Specimens. — Station 222, March 16, 1875; off the Admiralty Islands; lat. 2° 15' N., long. 146° 16' E. Habitat. — Indian Ocean, "from the equator to lat. 40° S., and the seas of the Malay Archipelago ; this species also lives in the south Atlantic Ocean, but only south of the Tropic of Capricorn, while Pneumonoderma violaceum is only to be found north of this tropic and has its southern limit at lat. 15° S. ; Pneumonoderma peroni, on the contrary, is only to be found from lat. 25° S. to lat. 37° S., and between long. 0° and 34° W. 1 Spolia atlantica, p. 154, Nos. 42 and 43. 2 Spolia atlantica, p. 154. 30 THE VOYAGE OF H.M.S. CHALLENGER. 4. Pneumonoderma boasi, u. sp. (PI. II. fig. 3). 188G. Pneumodermon violaceum, Boas (pars), Spolia atlantica, K. dansk, Yidensk. Selsk. Skriv., 6 Kaekke, Bd. iv. p. 154. Body, head, and foot as in the two foregoing species. Lateral Gill short, with fringes not subdivided, and fringed lateral somatic crests. Posterior Gill with short radiating crests, of which the fringes are not subdivided, but distinct, markedly separated, projecting and narrow at their base ; they number three on each side of the radiating crests. Acetabuliferous Appendages bearing about forty small suckers, of the same shape as those of Pneumonoderma mediterraneum. Hook-sacs rather short. Colour. — Brownish ; pigment existing as small spots visible to the naked eye. Length, 7 mm. (in a contracted state). Habitat. — South-East Pacific, off Caldera (Chili), about lat. 27° S. ; " Vettor Pisani" Expedition, February 18, 1883. Observation. — A young specimen, about 4 mm. long, still showed well-marked traces of the second ciliated ring of the larva ; the fringes on the branchial crests were not yet formed. *5. Pneumonoderma pacificum, Dall (PL II. figs. 4, 5). 1871. Pneumodermon pacificum, Dall, Descriptions of Sixty New Forms of Molluscs from the West Coast of North America, Amer. Journ. Conch., vol. vii. p. 139. Body. — Head and foot like those of the other species of the genus. Lateral Gill long, with fringed lateral somatic crests. Posterior Gill with short radiating crests, of which the fringes are simple well-marked foldings on each side of the crests, those of one side alternating with those of the other. These fringes are close to one another, and the radiating crests being short, they are not numerous. Acetabuliferous Appendages triangular, bearing about fifty small suckers of the same shape as those of Pneumonoderma mediterraneum. Radula. — Formula 4 : 0 : 4, in specimens 1 cm. long. Hook-sacs long. Colour. — Purplish-brown ; pigment chiefly abundant on the right side, and distributed in small spots, visible with a magnifying glass. Length. — Dall 1 gives 1 inch as the maximum size of this species ; the specimens from the Challenger Expedition are about 1 cm. long. Challenger Specimens. — Station 240, June 21, 1875; Yokohama to Honolulu; lat. 35° 20' N., long. 153° 39' E. 1 Amer. Journ. Conch., vol. vii. p. 140. REPORT ON THE PTEROPODA. 31 Station 254, July 17, 1875; Yokohama to Honolulu; lat. 35° 13' N., long. 154° 43' W. Habitat. — The type-specimens of this species were obtained by Dall in lat. 37° 8' N., long. 136° 10' W. The following localities have also been furnished to me by Mr. W. H. Dall:— Off Santa Barbara Islands (California), about lat. 33° 30' N., long. 118° 45' W.; off Sitka (Alaska), and near the Kodiak Island, about lat. 58° N., long. 135° W. Including the two Challenger localities, this species is therefore distributed throughout the whole north-east Pacific Ocean, and along the west coast of North America, from lat. 58° N. to lat. 33° N., and between long. 160° E. and long. 118° E. Observation. — This species has not hitherto been figured, but as it was found among the Challenger Pteropoda, I have been able to give good figures of it in this Report. The description of Dall was not only not accompanied by figures, but like the descrip- tions of Quoy and Gaimard, and of Rang, it did not indicate any character which might specifically distinguish it. Happily, Mr. Dall had the kindness to send to me some specimens of this species, and after comparison I recognised that the specimens of Pneumonoderma from Stations 240 and 254 were identical with them. Having thus identified these specimens with certainty, I have been able to give a descrip- tion sufficient to distinguish Pneumonoderma pacificum from the other species of the same genus, by means of characters drawn from the acetabuliferous appendages (e.g., their form and the form and number of the suckers), from the gills, radula, &c. *6. Pneumonoderma soideyeti, n. sp. (PL II. fig. 6). Body rather short, swollen in its anterior half ; neck of moderate length. Foot as in the foregoing species. Lateral Gill long. Posterior Gill with very wide quadrangular crest and very short radiating ones ; undulations of the base of the crests hardly developed and indeed almost absent. Acetabuliferous Appendages as long as the body, with a narrow base, and nearly parallel margins, bearing about seventy-five small suckers of the same shape as those of Pneumonoderma mediterraneum, and the size of which regularly decreases from the base to the extremity of the appendage. Colour. — The single specimen collected by the Challenger Expedition is but little coloured ; it shows brown pigment chiefly on the right side. Length, 4 mm. Cliallenger Specimen. — Station 254, July 17, 1875; Yokohama to Honolulu; lat. 35° 13' N., long. 154° 43' W. 32 THE VOYAGE OF H.M.S. CHALLENGER. Observation. — The appearance of this small Pneumonoclerma, when its fins are a little contracted, is very peculiar, and resembles very much that of some Cephalopods, as, for example, Cranchia scabra, Leach. I have dedicated this species to Souleyet, whose remarkable works will always remain the basis of our knowledge of the Pteropoda. After the death of this courageous worker, his exceedingly able works on the anatomy of the Mollusca were nearly forgotten in France, especially among the zoologists of Paris, where, however, the comparative anatomy of Molluscs is held in esteem, and it is English, German, and Danish zoologists (Huxley,1 Bronn,2 and Boas3), that have given to Souleyet the position to which in my opinion he is entitled. And since in that group of Pteropoda that Souleyet studied so much and contributed so much to make known there is not one species bearing his name, I have thought that the best way of showing my respect for his memory, was to give his name to that new species which has the most remarkable appearance. As remarked previously, the other specific names of Pneumonoclerma are synonyms of other already known species, with the exception of two names of Quoy and Gaimard, already referred to, " Pneumodermon pellucidum"* and " Pneumodermon ruber."5 Boas thinks6 that the first may be a Clionopsis, because it possesses a foot like that in this genus. But according to the figure of Quoy and Gaimard, it has a well-developed lateral gill like that in Pneumonoderma ; and Clionopsis does not possess such a gill. This is therefore a point which throws doubt upon the exactness of the drawing of the above mentioned writers, and which will at present prevent any decision in regard to the systematic position of this species. As to the " Pneumodermon ruber" of the same writers, it certainly includes different forms. First, a species (loc. cit., pi. xxvii. figs. 19, 20) which also possesses a foot as in Clionopsis and a lateral gill as in Pneumonoderma, and in regard to which one must make the same remark as regarding " Pneumodermon pellucidum." Second, a species (loc. cit, pi. xxvii. figs. 21-24) which possesses a foot and gill as in Pneumonoderma, but which is so badly described and figured that there is no character to distinguish it from the other species of this genus, or sufficient to identify it with one of these species. Nothing can therefore be decided, either in regard to "Pneumodermon ruber" or " Pneumodermon pellucidum," untd the forms which Quoy and Gaimard described under these names are again obtained. 1 On the Morphology of the Cephalous Mollusca, Phil. Trans., 1853, p. 52. 2 Die Klassen und Ordnungen des Thierreichs, Bd. iii. pp. 585, 873. 3 Spolia atlantica, pp. 9, 10. 4 Voyage de decouvertes de l'Astyolabe, Zoologie, t. ii. p. 390, pi. xxvii. fig. 25. 6 Better rubrum. Loc. cit., t. ii. p. 389, pi. xxvii. figs. 19-24. 6 Spolia atlantica, p. 171. REPORT ON THE PTEROPODA. 33 Family II. Clionopsi d.e. 1855. Clioidea, Gegenbaur (pars), Untersuchungen iiber Pteropoden und Heteropoden, p. 212. 1859. Pneujywdermonidse, H. and A. Adams (pars), The Genera of Recent Mollusca, vol. ii. p. 613. 1862. Clionidx, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 645. 1873. Clionopsidei, Costa, Pteropodi della Fauna di Napoli, p. 24. Characters. — Proboscis very long. Buccal appendages absent. A jaw. A quadri- radiate posterior gill. No lateral gill. Skin not pigmented. Description. — Body voluminous, rather transparent, barrel-shaped, a little contracted at both extremities ; posterior extremity rounded ; visceral mass extending to this extremity, forming an axial mass. Skin without pigment, showing generally a great many small spots due to fatty glands ; dorsal spot rounder than in Pneumonoderma. Head small, neck rather short ; labial tentacles thick, but short and shrunk at their extremity. Foot lying far forward, and without posterior lobe ; anterior lobes like those of Pneumonoderma, connected in all their length ; between these two lobes, on the posterior part of the foot, there is a plicated tubercle. Fins small in proportion to the body. Posterior Gill. — Membranous posterior crest rather short, forming a hexagon elongated in the dorso-ventral diameter, and the four ventral and dorsal angles of which each give rise to a small membranous crest which is turned forwards. These crests (the hexagonal and the four radiating ones) show different degrees of specialisation in the different species, and may bear at their base fringes or foldings like those of Pneumonoderma and Spongiobranchtea. Proboscis or evaginable anterior portion of the digestive tract (penis, Troschel) excessively long ; it is about three times the length of the body, while in Spongiobranchasa (which of all the other Gymnosomata possesses the longest proboscis), it is not a third of this length. When the proboscis of Clionopsis is invaginated, the horny parts, the radula, jaw, and hook sacs are placed very far from the buccal aperture. Radula always with median tooth (tricuspid) in the adult. Hook-sacs (whose true signification Krohn was the first to make known ') very short, possessing strong hooks, rather numerous, short, of nearly uniform length, slightly bent and mostly arranged in a bunch. 1 Beitriige ziir Entwickelungsgeschichte der Pteropoden und Heteropoden, p. 40. (ZOOL. CHALL. EXP. — PART LVIII. — 1887.) Mm m 5 ;54 THE VOYAGE OF H.M.S. CHALLENGER. Clionopsis,1 Troschel. 1854. Cliopsis, Troschel, Beitriige zur Kenntniss der Pteropoden, Archiv f. Naturgesch., Jahrg. xx. p. 222. 1855. Clio, Gegenbaur (jiars), Untersuckungen iiber Pteropoden und Heteropoden, p. 212. 1862. Clionopsis, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 645. Characters and Description as for the family Clionopsidse. The animals of this genus have been wrongly placed next to Clione by nearly all zoologists. Since the organisation of Clionopsis has become better known (Boas and I have independently worked at the subject, but Boas' researches have priority to my own), it is evident that in some respects this genus comes nearer to Pneumonoderma ; but it possesses some peculiar characters in which it differs as much from the Pneumono- dermatidse as from the Clionidse. It must therefore be placed in another family, the Clionopsidse, already established by Costa. The three species of this genus may be distinguished as follows : — Key to the Species. I. Animal of large size ; the posterior gill possessing four very distinct radiating crests. 1. Radiating crests of the posterior gill without fringes, . . . Clionopsis krolini. 2. Radiating crests with fringes, ...... Clionopsis grandis. II. Animal of small size, with the radiating crests on the posterior gill nearly obsolete, Clionopsis modesta. 1. Clionopsis Jcrohni, Troschel (PI. II. fig. 9 ; PL III. fig. 1). 1854. Cliopsis krohnii, Troschel, Beitriige zur Kenntniss der Pteropoden, Archiv f. Naturgesch., Jahrg. xx. p. 222, pi. x. figs. 2-4. 1855. Clio mediterranea, Gegenbaur, Untersuchungen iiber Pteropoden und Heteropoden, p. 212, pi. iv. fig. 14. 1869. Triehoeyelus inediterraneus (larva), Costa, Osservazioni su taluni Pteropodi del Mediterraneo, Annuario del Mus. Zool. d. R. Univ. di Napoli, Anno v. p. 46, pi. i. fig. 3. 1873. Clionopsis krohnii, Costa, Pteropodi della Fauna del Regno di Napoli, p. 25, pi. v. figs. 7-9. Body long and thick. Head small ; nuchal tentacles much projecting (this is the Gymnosomatous Pteropod on which these appendages are the most developed). Foot. — Anterior lobes with posterior right angle, so that their posterior extremity is neither free nor very projecting ; the folded tubercle on the posterior end of the foot is not divided by a median longitudinal wrinkle. Fins rounded at their distal extremity. 1 From Clione, a Gymnosomatous Pteropod, and o\pi;, appearance. REPORT ON THE PTEROPODA. 35 Posterior Gill a simple hexagonal crest, from the four ventral and dorsal angles of which there arise four short crests, without foldings or fringes, directed forwards. (In one specimen I only found three radiating crests, two dorsal and one ventral ; perhaps the two ventral crests were fused together, or the left one was absent.) This gill was unknown to Troschel, who founded his genus Cliopsis on young specimens not more than 8*5 mm. long,1 in which the place of the gill was still hidden by the posterior ciliated ring ;2 but it was figured by Costa in 1873. 3 Proboscis very long; in the figured specimen (PL III. fig. l), it is not yet quite everted, since the radula, jaw, and hook-sacs do not appear at its extremity. Radula. — The formula is 6 : 1 : 6; the formula 4:1:4 given by Troschel4 is that of young specimens ; and even for them it is probably inexact, because Macdonald has found that larvae measuring but 3 mm. (PL II. fig. 9) already possessed five lateral teeth on each side, the most distal being exceedingly small. Hook-sacs. — Each containing about sixty hooks. Colour. — The animal is colourless, rather transparent, with numerous small brown spots (in the specimens preserved in alcohol) visible to the naked eye, and due to cutaneous fatty glands. Length. — Up to 24 mm. Habitat. — Central parts of the Mediterranean Sea ; Naples, Messina, and south of Sicily, 20 miles east of Malta, lat. 36° 1' N., long. 15° 5' E., where Macdonald found larvae. Observations. — I have already said that the genus Trichocyclus is not a good one, and that the forms described under this name are larvae of different genera of Gymno- somata. All the Gymnosomata indeed, in the second part of their embryonic develop- ment, exist as naked larva?, with three ciliated rings. I may state that Trichocyclus mediterraneus, Costa,5 is the larva of Clionopsis krohni; a fact that I was able to ascertain by original drawings, much more perfect than Costa's figure, which Dr. J. D. Macdonald kindly sent to me. One of these drawings (PL II. fig. 9) represents a larva, 3 mm. long, with the two posterior ciliated rings still attached, and one may say that its foot is similarly formed to that of the adult Clionopsis krohni, and that there is no longer a posterior lobe. The axial visceral nucleus is another proof which shows that this larva belongs to the genus Clionopsis, as do also the radula (of which the formula is then 5:1:5), and the form of 1 Archivf. Naturgesch., Jahrg. xx. pi. x. fig. 1. 2 Von Iheriug therefore thought that Clionopsis was the " Jungstadium" of Clione, instead of a good genus (Vergleichende Auatomie des Nervensystems und Phylogenie der Mollusken, pp. 245, 24(i). 3 Pteropodi della Fauna di Napoli, pi. v. fig. 8. 4 Beitrage zur Kenntniss der Pteropoden, Archiv f. Naturgesch., Jahrg. xx. p. 231. 6 Anniuvrio del Mus. Zool. d. R. Univ. d. Napoli, t. iii. p. 46, pi. i. fig. 3. 36 THE VOYAGE OF H.M.S. CHALLENGER. the hook-sacs, which are very short and contain numerous hooks disposed in the form of a bunch. This larva was gathered by Macdonald, in lat. 3G° 1' N., and long. 15° 5' E. 2. Clionopsis grandis, Boas (PI. II. figs, 7, 8). 1885. Pneumodermon peronii, Verrill, Third Catalogue of the Mollusca recently added to the Fauna of the New England Coast, Trans. Conn. Acad. Arts and Sci., vol. vi. p. 431. 1886. Cliupsis grandis, Boas, Spolia atlantica, K. dansk. Vidensk. Selsk. Skriv., 6 Eaekke, Bd. iv. p. 170. Body a little more truncated at its posterior extremity than in the foregoing species. Nuchal tentacles less developed. Foot. — The lateral lobes are a little larger and have the posterior extremity terminating in an acute angle, so that a small part of this extremity is free; the folded tubercle, situated behind and between these two lobes, is divided into two by a longitudinal median wrinkle. Fins short and wide, rounded at their free extremity. Posterior Gill. — Hexagonal crest of the same shape as in Clionopsis krohni ; radiating crests longer. All these crests both radiating and hexagonal, and the latter on its six sides, lateral, ventral and dorsal, bear fringes on each side at the base, those of one side alternating with those of the other, whilst in Pneumonoderma fringes are developed only on the dorsal and ventral sides of the quadrangular crest and on the radial crests ; these fringes are simple rounded foldings, short, not contracted at the base or subdivided, and resembling those of Pneumonoderma pacijicum. Radida. — The formula is 5 : 1 : 5 in the adult. Hook-sacs with about the same number of hooks as in Clionopsis krohni. Colour. — The animal is colourless ; the small brown spots are not so large as in the foregoing species, and are chiefly numerous on the posterior part of the body and inside the hexagonal crest of the gill. Length. — The specimen I have studied (from the North Atlantic) measured 22 mm.; but one specimen observed by Boas (from the China Sea), reached a length of 30 cm. Habitat. — Australasian locality — China Sea (Copenhagen Museum). North Atlantic localities — (l) Twenty miles south-east of Cape Hatteras, lat. 35° 0' N., long. 75° 0' W. (U.S. National Museum and Brussels Museum); (2) lat. 39° 37' N., long. 71° 18' W., collected by the steamer "Albatross" of the U.S. Fish Commission, August 1883 (Verrill, as " Pneumodermon peronii"). I have not seen the type specimens of Boas in the Copenhagen Museum. The above description and the figures are taken from a specimen from the North Atlantic, sent to the Brussels Museum by the U.S. National Museum, under the name of "Pneumodermon, n. sp. (?)" REPORT ON THE PTEROPODA. 37 I did not find any specific differences between the characters of this specimen and those enumerated in Boas' description. If the comparison of living specimens from the two localities confirms the identity of the two forms, as I believe it will, it will prove that this species possesses an extensive geographical distribution. *3. Clionopsis modesta, n. sp. (PI III. fig. 2). Body a little more oval and less truncated at the extremities than in the two fore- going species. Head small ; labial tentacles rather short. Foot rather large, with lateral lobes terminating behind in an acute angle ; folded tubercle rather long. Fins contracted towards their distal extremity. Posterior Gill with scarcely distinct crests, and without foldings or fringes ; radiating crests nearly obsolete. Radula and Hook-sacs not investigated in the single specimen collected. Colour. — Very pale rose, and without the small spots which are seen in the foregoing species ; rectum visible as a dark brown line. Length, 3 mm. Notwithstanding the small size of this species it is not a larva, but an adult specimen, the three cUiated rings having disappeared. Challenger Specimen. — Station 254, July 17, 1875 ; Yokohama to Honolulu ; lat. 35° 13' N., long. 154° 43' W. In regard to the " two species of Pncumodermon " of Quoy and Gaimard, which Boas thinks may possibly belong to Clionopsis, see p. 32. Pelagia alba of the same zoologists1 (from Amboina ; length, 23 mm.) seems also to have some resemblance to Clionopsis, in the general form of the body and the absence of a lateral gill. The absence of a posterior gill in Quoy and Gaimard's figure is explained by the slight development of this organ in most species of Clionopsis, and the absence of the foot by the fact, that when its two anterior lobes are removed, the foot, which possesses no posterior lobe, becomes almost indistinct. The enlargement of the cephalic region in the above mentioned figure, quite agrees with the supposition that the two anterior lobes of the foot were removed from one another in the figured specimen. Family III. Notobranc h^e i dm. 188G. Notobranchxidx, Pelseneer, Description d'un nouveau genre de Pteropode Gyrunosome, Bull. Sci. depart. Nord, p. 224. Characters. — Buccal appendages conical. Lateral gill absent. Posterior gill ; three radiating crests (the dorsal one fringed) meeting posteriorly. Pigmented skin. 1 Vnyage de decouvertes de 1' Astrolabe, Zoologie, t. ii. p. 392, pi. xxviii. figs. 7-9. 38 THE VOYAGE OF H.M.S. CHALLENGER. Description. — Body ovate, contracted and pointed at the posterior extremity. Visceral mass not extending to this extremity. No dorsal spot as in the two foregoing families. Head short and swollen ; neck short. Foot. — Posterior lobe long, pointed at its free extremity ; anterior lobes long, free for a great part of their length ; between them, on the foot, a small tubercle. Fins large, broad and rounded at their free extremity. Posterior Gill. — Three longitudinal membranous crests which meet together at the posterior extremity of the body, and spread over about the posterior third of the body ; of these crests, one is median and dorsal, the two others lateral (right and left) and symmetrical, the ventral side being naked. The dorsal crest alone is fringed, the fringes of the right side alternating with those of the left side ; the two lateral crests are simple, without fringes or foldings. There is no lateral gill. Buccal Appendages. — One pair of thick, short (perhaps shortened by shrinking in the alcoholic specimens), symmetrical cones, inserted on the lateral sides of the wall of the buccal cavity, flattened on their median side, and with a smooth surface. Radula, Jaw, and Hook-sacs unknown. I had only a single specimen for examina- tion from the Challenger collection (type of NotobrancJisea inopinata), and one specimen in the Brussels Museum (type of NotobrancJisea macdonaldi); thus I was not able to extract the horny buccal organs of these specimens without damaging them. But from the shortness of the neck it may be inferred that the hook-sacs are also rather short ; and from the general resemblance of Notobranchasa to Clione, I think it probable that the radula of the adult, in this faindy, always possesses a median tooth, and that the jaw is perhaps absent. As these horny pieces are important, from a systematic point of view, for the com- parison of genera and species, they ought to be examined in the first specimen which is gathered in the future. Notobranchaea,1 Pelseneer. 1825. Clio, Rang (pars), Description d'un nouveau genre de la classe des Pt^ropodes et de deux especes nouvelles du genre Clio, Ann. d. Sci. Nat., ser. 1, t. v. p. 286. 1863. Clio, Macdonald (pars), On the Zoological Characters of the living Clio caudata, Trans. Roy. Soc. Edin., vol. xxiii. p. 186. 1886. Notobranchsa, Pelseneer, Description d'un nouveau genre de Pteropode Gynmosonie, Bull. Sci. depart. Nord, p. 224. Characters and Description as for the family Notobranchseidas. I made this genus known in June 1886, founding it on a Gymnosomatous Pteropod from the North Atlantic, which was sent to the Brussels Museum by the U.S. National Museum. I have previously said that many specimens of Pteropods were prepared for micro- 1 From vinos, back, and /3g«yx'«. gill. REPORT ON THE PTEROPODA. 39 scopical examination (stained and mounted in Canada balsam) by R. von Willemoes Suhm, without any previous examination. It was in this collection of preparations, which was sent to me along with the alcoholic specimens of Pteropods, that I found the following specimens of Gymnosomata from four Stations : — Station 170, ..... Halopsyche gaudichaudi. Station 240, . . . . . Pneumonoderma pacificum. Between Stations 332 and 333, March 11, 1876, Spongiobranchaea australis. Near Station 354, May 7, 1876, . . Pneumonoderma violaceum. A little after the publication of my Description d'un nouveau genre de Pteropode Gymnosome, and when the text of this Report was already far advanced, I re-examined this series of preparations to make certain that no Gymnosomatous Pteropod had escaped me. It was then that I found a specimen of a naked Pteropod from the North Atlantic, which was in such a state, from the colouring and the compression between the slide and cover-glass, that I did not recognise it on the first examination. When this specimen was freed from the balsam, by dissolving the latter in chloroform, I was able to study it more easily, and found that it belonged to the new genus I had just established. As the form that I described in June 1886 was only obtained in 1883, and the Challenger specimen from the Pacific was caught on June 24, 1875, the priority of the discovery of the genus N otobranchsea belongs to the Challenger Expedition, and to it also would have belonged the nominal priority, if the single specimen from Station 242 had not unfortunately been stained and compressed on the glass slide instead of being preserved in alcohol. The two known species of the present genus are distinguished as follows : — I. Dorsal crest of the gill with large and long fringes ; anterior lobes of the foot free for two-thirds of their length, ..... Notobranchxa macdonahli. II. Dorsal crest of the gill with short and small fringes ; anterior lobes of the foot free for half their length, ...... Notobranchxa inoyinata. 1. Notobranchsea macdonaldi, Pelseneer (PI. III. figs. 3, 4). 1884. Clione longicaudatus, Verrill, Second Catalogue of the Mollusca recently added to the Fauna of the New England Coast, Trans. Connect. Acad. Arts and Sci., vol. vi. p. 215. 1886. Notobranchxa macdonaldi, Pelseneer, Description d'un nouveau genre de Pteropode Gymnosome, Bull. Sci. depart. Nord, p. 225. Head round, neck contracted. Foot with anterior and posterior lobes long and pointed at their free extremity ; anterior lobes free for the two posterior thirds of their length ; between these two lobes 40 THE VOYAGE OF H.M.S. CHALLENGER. there is, on the foot, a small tubercle, at the same place as in the Pneumonoderrua- tidae and Clionopsidae. Posterior Gill with dorsal crest bearing on each side from eight to ten fringes, those of one side alternating with those of the other. Buccal Appendages. — Two thick symmetrical cones, occupying the whole width of the buccal cavity. Length. — About 1 cm. (the specimen, contracted in its posterior part, measures 8 mm. in the contracted state). Colour. — Brownish-grey. Habitat.— North Atlantic, lat. 38° 10' N., long. 74° 15' W.; off Carolina (U.S. National Museum and Brussels Museum) ; lat. 39° 22' N., long. 68° 34' W., October 3, 1883, steamer "Albatross," of the U.S. Fish Commission (Verrill, as Clione longi- caudatus). Observation. — The specimens collected by Professor Verrill are certainly identical witli the one sent to the Brussels Museum, because the latter was also identified by Professor Verrill as Clione longicaudatus, and because they were collected near the same place. The larvae noticed by Professor Verrill under the name of Trichocyclus dumereillii (sic) are certainly not identical with the Trichocyclus dumerilii of Eschscholtz. Verrill's identification is due to the fact that all the Gymnosomatous larvae with ciliated rings resemble each other very much. Although I have not seen these " Trichocycli," it seems to me that they are in reality the larvae of Notobranchasa macdonaldi, because they were collected on the same day and at the same place as that species. *2. Notobranchzea inopinata, n. sp. (PI. III. figs. 5, 6). Body and general aspect as in the foregoing species. Neck a little less constricted. Foot. — The anterior and posterior lobes are shorter than in Notobranch&a mac- donaldi ; the two anterior ones are less narrow and free only in their posterior half. Posterior Gill. — The dorsal crest bears small fringes, somewhat spreading. Buccal Appendages. — So far as I have been able to make out the single specimen I have studied, they are like those of the foregoing species. Length, 4 mm. Colour. — It cannot be estimated, the specimen having been stained with carmine by von Willemoes Suhm. The state of the specimen does not allow of a more complete description being given. Challenger Specimen. — Station 242, June 24, 1875 ; east of Japan; lat. 35° 29' N, long. 179° 50' W. REPORT ON THE PTEROPODA. 41 To the genus Notobranchsea, I refer Clio capensis, Rang,1 from the Cape of Good Hope, which was considered by Gray,2 d'Orbigny,3 and Boas,4 as a Pneumonoderma. The foot of " Clio capensis " shows the same disposition indeed as that of Notobranchsea macdonaldi ; the neck and the head are also short, and finally, one can distinguish5 the three crests of the posterior gill. But the figures given by Rang are not accurate in regard to the foot and still less so in regard to the gill ; besides, the buccal appendages are not at all known. It is therefore impossible to identify the two forms. On the other hand, Macdonald6 has noticed, as "trigonal tailed Clio," a little Gymnosomatous Pteropod in which he has observed the same conformation of gill as in Notobranchsea. But the foot is not described, and there should be two (?) pairs of buccal cones, like those of Clione. Macdonald wrote me that he had obtained this Pteropod off Sydney Harbour. It follows then that the genus Notobranchsea is distributed in both the northern and southern hemispheres, in the Atlantic as well as in the Pacific Ocean. Family IV. Clionid^:. 1840. Clionidx, Gray, Synopsis of the Contents of the British Museum, p. 86. 1842. Pneumodermidx, d'Orbigny {pars), Paleontologie francaise, terrains cretac^s, t. ii. p. 4. 1846. Glionoidx, Agassiz (pars), Nonienclator Zoologicus, Index, p. 90. 1850. Clioneidx, Gray, Catalogue of the Mollusca in the Collection of the British Museum, pt. ii., Pteropoda, p. 35. 1852. Clios, Souleyet (jiars), Histoire naturelle des Mollusques Pteropodes, p. 74. 1855. Clioidea, Gegenhaur (pars), Untersuchungen liber Pteropoden und Heteropoden, p. 212. 1856. Cliidx, Woodward {pars), A Manual of the Mollusca, p. 208. Characters. — Two or three pairs of conical buccal appendages. Jaw absent. Gill absent. Skin not pigmented. Description. — Body long, contracted behind and pointed at its posterior extremity, visceral mass not extending to that extremity, and in the adult occupying only the anterior half of the body. Skin generally not pigmented ; no dorsal spot. Head round and short ; long labial tentacles ; neck short. 1 Description d'un nouveau genre de la classe des Pteropodes et de deux especes nouvelles du genre Clio, Ann. d. Sci. Nat., ser. 1, t. v. p. 286, pi. vii. figs. 3, 4. 2 Catalogue of the Mollusca in the Collection of the British Museum, pt. ii., Pteropoda, p. 40. 3 Voyage dans l'Amerique meridionale, t. v. p. 129. 4 Spolia atlantica, p. 167. 6 Rang, loc. cit., pi. vii. figs. 3, 4. 6 On the Zoological Characters of the living Clio caudata, Trans. Roy. Soc. Edin., vol. xxxiii. p. 186, pi. ix. fig. 3, 2. (zool. chali. exp. — paut Lviu. — 1887.) Mmm 6 42 THE VOYAGE OF H.M.S. CHALLENGER. Foot. — Anterior lobes large, broad in front, narrow and pointed behind, attached in nearly their whole length ; posterior lobe short and pointed. Gills. — Quite absent. Proboscis short. Buccal Appendages. — Two or three pairs of retractile, glandular and sensory cones, symmetrically inserted on the two sides of the buccal cavity. Jaio. — Absent. Radula always possessing in the adult a rather broad median tooth, without long projecting spines, and rather resembling that of Halopsyche than of any other Gynmoso- matous Pteropod. Hook-sacs of moderate size, with hooks of various lengths, so that when the sac is evaginated the extremities of all the hooks reach to about the same level. The species of this family are generally described by zoologists under the name of Clio. But this name was established by Browne in 1756,1 and preserved by Linne for a Thecosomatous Pteropod which most naturalists now name Cleodora pyramidata. As the description of the genus Cleodora by Peron and Lesueur was only published in 1810,2 the name Clio, on account of its priority, must be preserved instead of Cleodora, which is more generally used. 0. F. Muller was therefore wrong when in 1776 3 he applied the generic name Clio to a naked Pteropod, for this name was originally used for a Thecosomatous form, and those zoologists who have followed him to this day are equally in error. The Gymnosomatous Pteropods that are generally described under the name of Clio must be placed in the genus Clione which Pallas established as far back as 1774 4 for the type species, Clione limacina (his Clione borealis). For the various Gymnosomata which certainly belong to the genus Clione, it was proposed to form new generic sections. Thus for some species whose buccal appendages were not described, Quoy and Gaimard in 1825 established 5 the genus Cliodita, based upon inexact differential characters. But these naturalists recognised later that this group could not stand and they abandoned their genus altogether.6 It is therefore rather surprising to find that some more recent writers nevertheless preserve this genus.7 On the other hand, several zoologists, without giving new generic names, have shown 1 The Civil and Natural History of Jamaica, p. 386. 2 Histoire de la famille des Mollusques Pteropodes, Ann. Mus. Hist. Nat. Paris, t. xv. p. 66. 3 Faunae Danicse prodromus, p. 226. 4 Spicilegia zoologica, fasc. x. p. 28. 6 Description de cinq genres de Mollusques, Ann. d. Sci. Nat., st'r. 1, t. vi. p. 74. 6 Voyage de decouvertes de l'Astrolabe, Zoologie, t. ii. p. 371, 1832. 7 Gray, Catalogue of the Mollusca in the Collection of the British Museum, pt. ii., Pteropoda, p. 37; H. and A. Adams, The Genera of Recent Mollusca, vol. i. p. 62 ; Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii, p. 645 ; Fischer, Manuel de Conchyliologie, p. 424. REPOET ON THE PTEEOPODA. 43 the necessity of forming new genera for some Gymnosoniatous forms which I consider undoubtedly to belong to the genus Clione. Gegenbaur, indeed, unfortunately confounding the tentacles and buccal appendages under the same name, divided Clione into two groups1: — (1) those with more than two "tentacles" — " Clio borealis" and " Clio australis"; (2) those with two tentacles — " Clio cajoensis, Clio longicaudata, Clio limacella, Clio jlavescens, and Clio mediterranea " ; and he says that if he had treated the question from the systematic point of view, he should have created a new genus for these last species. But there are two species among these which do not belong to the family Clionidse, namely, Clio capensis, Rang ( = Notobranchiea sp.), and Clio mediterranea, Gegenbaur ( = Clionopsis hrolvni, Troschel). "Clio" longicaudata, Souleyet, besides its true tentacles, which number two pairs as in Clione limacina (=Clio borealis), possesses two pairs of buccal cones of the same nature as the three pairs in the latter species. Clio limacella, Rang (which has never been described, but only figured), is a species very closely allied to Clione longicaudata, if not identical with it, and it very probably also possesses two pairs of buccal cones. Finally, respecting " Clio " Jlavescens, Gegenbaur, we shall see further on that our knowledge of its larva shows that it also possesses two pairs of buccal cones, as in Clione longicaudata. Thus, these latter species do not differ from the type of the genus Clione, except by having two pairs of buccal cones instead of three ; and I think that this difference is not at all a generic one. Macdonald also thought that one might generically separate the forms with three pairs of buccal cones from those which only possess two.2 In his group with two pairs of buccal appendages, Macdonald includes two forms ; one very close to Clione longicaudata, the other with a posterior gill, which thus does not belong to the family Clionidas, and which must be removed from the Clione with two pairs of buccal appendages as well as from those which possess three pairs, and will be placed in the genus Notour anchsea. It is quite as inadmissible for me to generically separate forms of Gymnosomata so nearly resembling one another, because they have two or three pairs of buccal cones, as to place in the same genus a Gymnosomatous Pteropod with a gill (Notobranchsea) and another without a gill [Clione). All the Gymnosomata with an elongated body, without a gill and with buccal cones (two or three pairs), must be placed in a single genus Clione. Lastly, Fol, on account of the species he has called " Clio " aurantiaca, also thought that he should establish a new genus.3 As we shall show further on, " Clio " aurantiaca must be considered as the old larva of a previously known species of the genus Clione. 1 Untersuchungen iiber Pteropoden und Heteropoden, p. 2] 2. 2 On the Zoological Characters of the living Clio caudata, Trans. Roy. Soc. Edin., vol. xxiii. p. 187. 3 Sur le developpement des Pteropodes, Archives d. Zool. exp^r., ser. 1, t. iv. p. 172. 44 THE VOYAGE OF H.M.S. CHALLENGEK. Clione,1 Pallas. 1774. Clione, Pallas, Spicilegia zoologica, fasc. x. p. 28. 1776. Clio, 0. F. Miiller (non Browne), Faunae danicse prodromus, p. 226. 1825. Cliodita, Quoy and Gaimard (pars), Description de cinq genres de Mollusques, Ann. d. Sci. Nat., ser. 1, t. v. p. 74. 1840. Spongiobranchea, d'Orbigny (pars), Voyage dans lAmerique meridionale, t. v. p. 132. Characters and Description. — As in the family Clionidse, great confusion has existed for a long time in regard to the cephalic appendages of Clione. In a paper on these organs 2 I have shown that there may be distinguished : — (1) The buccal appendages or cephaloconi, inserted on the wall of the buccal cavity. (2) The true tentacles, situated on the external wall of the head, and numbering two pairs. First, the anterior, labial or oral pair, and second, the posterior, nuchal or cervical pair. A great number of Gymnosoniatous Pteropods are described under the name Clione (or Clio) ; but a considerable proportion of these forms are so little known, so imperfectly and insufficiently described and figured, that it is impossible since the type specimens have been lost to give a diagnosis of them, and to admit them into the nomenclature. I have found in the literature (after eliminating, of course, the Thecosomata bearing the name Clio, which are more often called Cleodora) nineteen different names of Clione (or Clio). Among these nineteen names there are : — 1 . One given to a Mollusc of another group — (1) Clio amati, Delle Chiaje3 = Gastropteron meckeli, Kosse. 2. Names given to forms belonging to other genera of Gymnosomata — (2) Clio capensis, Eang = Notobranchasa sp. (3) Cliodita caduceus, Rang = Spongiobranchsea australis, d'Orbigny. (4) Clio mediterranea, Gegenbaur =Clionopsis krohni, Troschel. 3. Names which are synonyms of other specific names of the genus Clione, as we shall see further on — (5) Clio aurantiaca, Fol = Clione flavescens, Gegenbaur. (6) Clione borealis, Pallas = Clione limacina, Phipps. (7) Clio caudata, Gray (non Macdonald) = Cfr'ewe longicaudata, Souleyet. (8) Clione dalli, Krause= Clione limacina, Phipps. 1 From Clio, a mythological name. 2 The Cephalic Appendages of the Gymnosomatous Pteropoda, Quart. Journ. Micr. Sci., 1885, p. 493. 3 Memorie sulla storia e notomia degli animali senza vertebre del Regno di Napoli, t. i. p. 52. REPORT ON THE PTEROPODA. 45 (9) Clione elegantissima, Dall = Clione limacina, Phipps. (10) Clio miquelonensis, Eang= Clione limacina, Phipps. (11) Clione papilionacea, auctorum = Clione limacina, Phipps. (12) Clio retusa, 0. F. Miiller (non L.) = Clione limacina, Phipps. 4. Names applied to species not well enough known and described to be distinguished from one another — (13) Clio australis, Bruguiere. (14) Cliodita fusiformis, Eang. (15) Clio limacella, Eang (without description). (16) Clio pyramidata, Eang. I shall speak afterwards of these last four species. There are, then, only three forms well enough known and characterised to be distinguished from one another, as in the following table : — I. With three pairs of buccal cones, ....... Clione limacina. II. With two pairs of buccal cones. 1. Animal small, of a brownish colour, with the posterior lobe of the foot very short, Clione longicaudata. 2. Animal of moderate size, of a yellowish colour, with the posterior lobe of the foot rather long, ........ Clione flavescens. Of these three species only one is very abundant in the Arctic Seas, and the two others are very localised, and thus the Challenger, which did not go to the Arctic Seas, did not bring back specimens of the genus Clione. It is the only genus of Gymnosomata that was not collected during the cruise. ■e 1. Clione limacina (Phipps). 1773. Clio limacina, Phipps, A Voyage towards the North Pole, p. 195. 1774. Clione borealis, Pallas, Spicilegia zoologica, fasc. x. p. 28, pi. i. figs. 18, 19. 1776. Clio retusa, 0. P. Miiller, Faunae danicee prodromus, p. 226. 1825. Clio miquelonensis, Rang, Description d'un nouveau genre de la classe des Pt<5ropodes et de deux nouvelles especes du genre Clio, Ann. d. Sci. Nat., st5r. 1, t. v. p. 285, pi. vii. fig. 2. 1871. Clione elegantissima, Dall, Description of Sixty New Forms of Molluscs from the West Coast of North America, Amer. Journ. of Conchology, t. vi. p. 139. 1885. Clione dalli, Krause, Ein Beitrag zur Kenntniss der Mollusken Fauna des Berings Meeres, Archiv f. Naturgesch., Jahrg. li. p. 299, pi. xviii. fig. 19a. Clione papilionacea, auctorum. Body long, tapering to a point behind. Head round and swollen ; neck short. Skin not pigmented over all its surface, but of a bright red colour at the posterior extremity and on the head and buccal cones l of the living animal (the alcoholic specimens are of an orange colour). Foot. — Large anterior lobes, broad in front, pointed behind, attached in nearly their 1 See the figure of Eschriclit, Anatomische Untersuchungen iiber die Clione borealis, pi. i. figs. 1-4. 46 THE VOYAGE OF H.M.S. CHALLENGER. whole length ; posterior lobe short, pointed behind and nearly concealed when the anterior lobes fall back against one another. In some specimens from Behring Strait, the posterior lobe is so short that it is only distinguished by a small undulation from the anterior lobes. Fins nearly triangular in shape, pointed at their free extremity. Penis very long, bifid towards its base, and formed of a short lateral branch and a long posterior one. Buccal Appendages. — Three pairs of contractile cones, symmetrically inserted on the wall of the anterior part of the digestive tract, so that they occupy the whole circumference of this part. At the median base of these cones the digestive tract is contracted into a pad, so that it forms a pair of false " lips." 1 The dorsal pair of cones are the longest and the ventral the shortest. When they are expanded, these organs are very long ; their surface is covered with granulations visible with a magnifying glass. (These appendages contain special nervous terminations and glandular follicles.2) Behind the false lips is the short evaginable portion of the digestive tract or proboscis. Radula. — Its formula varies with the size of the specimen. Therefore Krause gives 6:1:6 or 7 : 1 : 7/ Sars 8:1:8/ Loven 12:1:12/ and Boas 14:1:14/ &c. It is easy to find all these formulas and the intermediate ones when examining specimens o various lengths. In specimens of large size which I have examined, I have found the same formulae as Loven and Boas, or the intermediate one (13:1:13). Hook-sacs of moderate length, with hooks of various sizes, regularly decreasing from the bottom of the sac to the edge, slightly bent (the larger less than the small) and assuming a fan-like arrangement when the sac is evaeinated. Colour. — Transparent ; in the living animal the extremities are coloured bright red ; the visceral mass (liver) is brownish-violet. Length. — The largest specimens measure 35 to 40 mm. Habitat. — Clione limacina is found in all the seas around the North Pole ; Kara Sea ; Nova Zembla ; Waigatz Straits ; White Sea ; Spitzbergen ; Norwegian coast — Finmark and Lofoten Islands to Karmoe Island (lat. 59° N.) ; Kattegat, Bohuslan ; Island of Mull ; west of Hebrides (lat. 59° N, long. 10° W.); Jan May en Island; Iceland; Coasts of Greenland ; Baffin Bay ; Davis Strait ; Hudson Strait ; Labrador ; Newfoundland ; Arctic Ocean, Alaska, Cape Lisburne ; Behring Strait and Sea ; Aleutian Islands (Akutan Pass) ; North Pacific, lat. 81° 80' N, long. 161° 26' W.). The most northern point where Clione limacina was observed is lat. 81° 40' N. (Ross) ; the southern limit of its geographical distribution nearly corresponds to the isothermal 1 Pelseneer, The Cephalic Appendages of the Gymnosomatous Pteropoda, loc. cit., pi. xxsv. fig. 4, d. 2 See my paper on this subject, loc. cit., pp. 495-500. 3 Ein Beitrag zur Kenntniss der Mollusken Fauna des Berings Meeres, Archiv f. Naturgesch., Jahrg. li. p. 299. 4 Bidrag til Kundskaben om Norges arktiske Fauna, I., Mollusca Regionis Arcticas Norvegias, pi. xvi. fig. 21,/. 6 Oin tungans bevapning hos Mollusker, Ofversigt h. Vetensk. Akad. Handl., 1847, p. 188, pi. iii. 0 Spolia atlantica, p. 162. REPORT ON THE PTEROPODA. 47 line of 60° F. (for August) or with the isocryme of 44° F. When the temperature sinks very low Clione limacina may reach more to the south ; thus, from April 7th to the beginning of May 1868, it was observed at Portland, Maine, U.S.A. (lat. 43° 30' N) 1 and in 1833 at New York (lat. about 41° N, the latitude of Madrid and Naples). Notwithstanding that these two localities are comparatively southern from a geographical point of view (isothermally New York corresponds with Scotland), they are less remarkable than the European locality of Falmouth, Cornwall (about lat. 50° 9' N),2 for, isothermally, Falmouth nearly corresponds with the localities in lat. 35° N. on the east coast of North America. Observation. — The larvae of this species had not been described until quite recently, and having been able to examine a great many larva? of all sizes, I prepared drawings of young Clione ; but since Boas has lately figured it,3 I do not think it necessary to introduce these into the plates of this Report. The naked larvae of Clione limacina, in the first stage of their development, are certainly testaceous like that of Clione Jlavescens, Gegenbaur (Clio aurantiaca, Fol).4 They differ from the adult by the length of the posterior lobe of the foot (which is often longer than figured by Boas), and by the fact that the visceral mass extends more posteriorly in the body. The fins of the very small larvae are but little developed. The ciliated rings often remain to a very late stage, but in a rather variable manner ; the anterior ring has nearly always disappeared from specimens measuring from 2 to 3 mm., and the second ring from specimens of 5 or 6 mm., but old larvae of 15 mm. occasionally retain the three ciliated rings. Traces of the posterior ring are often found on specimens of rather large size. The hook-sacs of the larvae with ciliated rings are identical with those of adults, while the radula differs by the small number of lateral teeth. Clione limacina is the only Gymnosomatous Pteropod of which the copulation has been observed. Copulating specimens, preserved in alcohol, are disposed as follows : the two animals are in contact ventrally, in symmetrical positions, with their axes parallel ; the union is reciprocal, the two penes crossing each other, the short branch being placed at the genital orifice of the other animal, and the long branch turned behind and applied by its extremity to the ventral side of the other branch. With Clione limacina I identify Clione elegantissima, Dall. This form was found by Dall in the North- West Pacific ; the collected specimens measured about 7 mm. One of these specimens which Dall kindly sent to me does not differ at all from Clione limacina 1 Gould, Report on the Invertebrata of Massachusetts, p. 507. 2 There is in the collection of the Museum d'Histoire Naturelle of Paris a specimen from Falmouth, presented by Leach. 3 Spolia atlantica, pi. vii. fig. 103. In the same plate (fig. 101) there is a good figure of the adult. Therefore a new figure of the adult Clione limacina in the plates of this Report is quite unnecessary. 4 Sur le developpement des Pteropodes, Archives d. Zool. exper., ser. 1, t. iv. pi. x. figs. 6, 7. 48 THE VOYAGE OF H.M.S. CHALLENGER. of the same length ; the hook-sacs, the radula, the buccal cones, the foot, &c, are quite similar. In the living animal a red line exists on the fins, which, however, is not now visible in the alcoholic specimens; but in other respects Clione elegantissima is exactly the same as the young of Clione limacina. With respect to Clione dalli, Krause, this author says that it only differs from Clione elegantissima by having two ciliated rings. The collected specimens measured from 5 to 6 mm., and had the radula similar to that in specimens of Clione limacina of the same size. They are thus really the young of this species. 2. Clione longicaudata (Souleyet). 1840. Spongiobranchea elongata, d'Orbigny, Voyage dans rAmeiiqiie nieridioDale, t. v. p. 132, pi. ix. figs. 8, 9. 1850. Clione caudata, Gray, Catalogue of the Mollusca in the Collection of the British Museum, pt. ii., Pteropoda, p. 37. 1852. Clio longicaudatus, Souleyet, Voyage de la Bonite, Zoologie, t. ii. p. 28G, pi. xiv. figs. 17-21. Body very long, with posterior part very pointed. Foot. — Anterolateral lobes large, wide in front, pointed posteriorly ; posterior lobe probably very short ;* Souleyet says2 that the foot of Clio longicaudatus is like that of Pneumonoderma, without the posterior lobe. But the anterior lobes being very large, they probably hid the posterior lobe, as often happens in Clione limacina. Besides, we have seen that, in this last species, there are some specimens with the posterior lobe of the foot very small ; but however minute it may be, this lobe is always distinct, and exists in all the Clione {Clione limacina, Clione jlavescens) and in all the other Gymno- somata, except Clionopsis. It is therefore very improbable (as Ihering3 following Souleyet 's figure states) that Clione longicaudata has no posterior lobe to the foot. Proboscis rather short. Buccal Appendages. — Two pairs of cones, like those of Clione limacina. Colour. — Dark brown. Length. — Does not measure 10 mm. (Souleyet). Habitat. — Atlantic Ocean, lat. 10° N., long. 21° W.; " Spongiobranchea elongata," which I identify with the present species, was gathered in lat. 4° N., long. 25° W. Observation. — " Spongiobranchea elongata," d'Orbigny, is in reality not a Spongio- brancheea; in the shape of the body, head, and foot, it quite resembles Clione longi- caudata. The habitats of the two forms make their identity still more certain, the localities where they were gathered being very close to one another. 1 The type-specimen of Souleyet, in the Museum of Paris, being quite destroyed, and d'Orhigny's specimens not having been preserved, it was impossible to definitively decide on this point until new specimens had been obtained. 2 Voyage de la Bonite, Zoologie, t. ii. p. 286. 3 Vergleichende Anatomie des Nervensystem und Phylogenie der Mollusken, p. 248. REPORT ON THE PTEROPODA. 49 3. Clione flavescens, Gegenbaur. 1855. Clio flavescens, Gegenbaur, Untersuchungen uber PLeropoden und Heteropoden, p. 73, pi. iv. fig. 15, A, B. 1875. Clio aurantiaca (larva), Fol, Sur le developpement des Ptdropodes, Archives d. Zool. exper., s<5r. 1, t. iv. p. 178, pi. x. fig. 10. Body of moderate length, and terminating behind in a point (quadrangular, ac- cording to Gegenbaur). The visceral mass only extends to the middle of the body. Head rather wide (in Gegenbaur's figure the nuchal tentacles are distinguishable). Foot. — Anterior lobes rather wide, pointed behind ; posterior lobe rather long and pointed. Fins wide at the base, but like those of Clione limacina. Buccal Appendages. — They have not been described, and Professor Gegenbaur wrote me that the type specimen of this species no longer exists. But from the larva, it is known that Clione flavescens possesses two pairs of cephaloconi, as in the foregoing species. Radula. — Unknown in the adult ; in the larva, there are 2 : 1 : 2 teeth (Fol)1 ; (in the adult the radula probably possesses twice or three times as many lateral teeth). The median teeth resemble those of Clione limacina, except that they possess a small median denticle. Hook-sacs rather short in the larva. Colour. — The skin is transparent with a yellowish reflection ; the visceral mass has a yellow tint. Length. — About 22 mm. Habitat. — Messina (Sicily). The first " Preisverzeichniss" of the zoological station 2 mentions " Clio flavescens" from the Gulf of Naples, but the form there obtained was Clionopsis krohni. Clione flavescens has only been found at Messina, once in the adult, and twice in the larval state.3 It is a true Clione, as much by the shape of the body and foot as by the absence of gills. Bronn was therefore wrong in saying that only Clionopsis was found in the Mediterranean Sea.4 Observation. — I consider " Clio aurantiaca," which Fol thought " probably adult," as the young of Clione flavescens. Notwithstanding that it has been seen to lay eggs, "Clio aurantiaca" is not an adult. It hardly reaches 2 mm. in length, and has still all the three ciliated rings ; besides its radula has a larval formula. 1 Sur le developpement des Ptdropodes, Archives d, Zool. exper., sdr. 1, t. iv. p. 179. 2 Mitth. axis d. zool. Stat, zu Neapel, Bd. i. 3 Besides Fol, J. Miiller, Ueber verschiedene formen von Seethieren, Archivf. Anat. u. Physiol, p. 70, 1854. 4 Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 639. (zool. chall, exp. — part lviii.— 1887.) Mmm 7 50 THE VOYAGE OF H.M.S. CHALLENGER. The following considerations clearly prove that this form is the young state of Clione Jlavescens : — 1. The presence of but one true Clione in the Mediterranean has been so far ascertained, viz., Clione Jlavescens. On the other hand, only one larva of a true Clione, " Clio aurantiaca," is known to exist in the Mediterranean. This form is indeed a true Clione, a fact made evident by the shape of the foot, as well as by that of the buccal appendages. As to the larvae of Clione described by Johannes Midler,1 they are identical with " Clio aurantiaca," as may be seen from the buccal appendages and the radula. 2. Clione jlavescens has never been collected except at Messina. At Messina also, and there alone, the larva called " Clio aurantiaca" has been twice caught. 3. Clione Jlavescens is yellow ; this also is the colour of " Clio aurantiaca." I therefore think there can be no doubt at all as to the identity of these two forms, and hence one may safely conclude that Clione Jlavescens possesses two pairs of buccal cones or cephaloconi, as exist in Clio aurantiaca. I cannot figure this species nor the foregoing one, because the only known specimens do not now exist. They and Dexiohvanclisea •paucidens are the only species that I have not myself seen. Among the species of the genus Clione which are imperfectly known, there are two which in their general aspect rather closely resemble Clione longicaudata ; they are : — 1. " Cliodita fusiformis," Quoy and Gaimard,2 collected between the Cape of Good Hope and Mauritius. 2. " Clio limacella" Rang,3 without description or recorded habitat ; only known from figures. A Clione from the South-West Pacific, of which Dr. Macdonald gave me an original drawing, very much resembles " Clio" limacella, and is of the same length (6 mm.) ; they are both adults, because they no longer show any trace of ciliated rings. Lastly, " Clio " caudata, Macdonald,4 which this writer identifies with Clione longi- caudata,5 also resembles very much " Clio " limacella. " Clio caudata," Macdonald, which was collected off Sydney Harbour, possesses a radula of which the formula is 6 : 1 : G, and bears two pairs of buccal cones. Probably " Clio" fusiformis and " Clio" limacella are also provided with two pairs of buccal appendages, and belong to the group of Clione longicaudata. Our knowledge of these species shows that the genus Clione is distributed south of 1 Archivf. Anat. u. Physiol, p. 70, 1854. 2 Description de cinq genres de Mollusques, Ann. d. Sci. Nat., sir. 1 , t. vi. p. 74, pi. ii. figs. 3, 4. 3 Histoire naturelle des Mollusques Pteropodes, pi. x. figs. 10-12. 4 On the Zoological Characters of the living Clio caudata, Trans. Roy. Soc. Edin., vol. xxiii. p. 185, pi. ix. fig. 3, 1. 6 I cannot confirm this identification, because Macdonald gives no description of this form, and especially no indication of the conformation of the foot. REPORT ON THE PTEROPODA. 51 the equator, but is rather localised and not very abundant, for no one has ever collected numerous specimens of the different species just mentioned. In regard to two other species, " Clio" pyramiclalis, Quoy and Gaimard,1 and " Clio" austrcdis, Bruguiere,2 they are very imperfectly known. The first was taken in the harbour of Amboina, and measured 10 mm. in length ; it is represented as of so strange a shape, that the accuracy of the figure of Quoy and Gaimard is a little doubtful. The second species is known only from an obscure description and more obscure figure by Bruguiere, who says that it was abundant on the south coast of Madagascar. Since the time of Bruguiere, some naturalists have made collections of the Mollusca of that island, but no one has again found this species. I hope, however, that further researches will be made, for it would be very interesting indeed to obtain this Pteropod, because it would be the largest of all (according to Bruguiere, it measures 2 inches in length), and it is said to have also three pairs of buccal cones, as in Clione limacina. The other species described under the name of Clione or Clio (after deducting the Thecosomata which bear the latter name, and which most zoologists still call Cleoclora), are, as we have already seen, synonymous with other forms previously known. Ross 3 recorded, under the name Clio borealis, a Gymnosomatous Pteropod obtained between lat. 60° and 64° S., along with " Argonauta artica" (the latter is really a Limacina, and was found by the Challenger).4 According to Souleyet,5 the naturalists of the " Astrolabe " also found, during the last voyage, a Clione " among the ice of the South Pole," which must certainly be the same species as the " Clio borealis " of Ross. Unfortunately, in the zoological account of this voyage, there is no Pteropod mentioned, and I do not know what became of the specimens of Clione noticed by Souleyet. However, if one may be allowed to make a hypothesis respecting these examples, it seems to me rather probable that they are only Spongiobranchsea australis, a species which is widely distributed throughout the cold Antarctic Seas, where it was previously observed from long. 60° W. to long. 123° E. ; towards the ecpiator it scarcely passes beyond the isothermal line of 50° F. (for August), and it has up to the present time been collected as far as lat. 51° S. It is therefore a species belonging to the cold regions, and I think it will be found to exist all around the South Pole, as Clione limacina does around the North Pole. Bronn 6 mentions an Australian Clione, according to Lamartiniere. But the species described by this last writer is a Clio (Thecosomatous Pteropod) known to zoologists under the name of Clio (or Cleoclora) pyramidata. 1 Voyage de decouvertes de 1' Astrolabe, Zoologie, t. ii. p. 371, pi. xsvii. fig. 37. 2 Encyclopedie rnethodique, Vers, t. i. p. 507, pi. lxxv. figs. 1, 2. 3 A Voyage of Discovery and Research in the Southern and Antarctic Regions, vol. i. p. 169, 1847. 4 Station 153, lat. 65° 42' S., long. 79° 4C E. 5 Histoire naturelle des Mollusques Pteropodes, p. 86. 6 Die Klasaen und Ordnungen des Thierreichs, Bd. iii. p. 582. 52 THE VOYAGE OF H.M.S. CHALLENGER. Family V. Halopsychid^. 1S50. Cymbuliadse, Gray {pars), Catalogue of the Mollusca in the Collection of the British Museum, pt. iL, Pteropoda, p. 24. 1856. Hyaleidx, Woodward (jiars), A Manual of the Mollusca, p. 204. 1859. Eurybidse, Chenu, Manuel de Conchyliologie, t. i. p. 115. Characters. — One pair of long flattened buccal appendages, not retractile into the front part of the digestive tract. Proboscis absent. Gill absent. A bifid jaw. Unpig- mented skin. Description. — Body ovate, of medium length, wide and thick, rounded behind. Visceral mass extending to the posterior part of the body. The integuments are hard, but do not form a shell at all. The head and fins of the contracted specimens may be retracted into the body-wall, as in Pneumonoderma ; in that state, a dorsal extension of the integuments of the body covers and protects the head. Head very small relatively to the body. Foot. — Anterior lobes free and rounded behind ; posterior lobe of moderate length, contracted posteriorly. The separation of the posterior lobe of the foot from the fins is more difficult to make out than in the other Gymnosomata, although the disposition is the same, and the figures of Souleyet 1 and Huxley 2 are inexact on that point. Fins. — They differ from those of all other Gymnosomata. In the latter, they are always abruptly contracted at the base, and their greatest width is placed immediately behind this contracted part ; but the fins of the Halopsychidse on the contrary are very long, and gradually narrow from the extremity towards the base, so that the greatest width is at the distal end. This extremity is nearly truncated and slightly sinuous in the middle. Proboscis and Hook-sacs absent. Jaw. — The two halves extend dorsally, on eaeh side of the buccal orifice, forming a long double or triple row of small toothed horny plates. Radula always with a median tooth, without denticles ; resembling that of Clione. Cephalic Appendages. — They have been described under various names, and generally in an inexact and incomplete manner, by the zoologists who have studied the organisation of Halopsyche. In this genus, as well as in all the Gymuosomata (with the exception of Clionopsis), there are : — 1. Tentacles properly so called. — As in all the other genera of Gymnosomatous Pteropods, they number two pairs ; (l) the anterior, oral or labial tentacles, placed more dorsally than in the other Gymnosomata ; they were called by most of the writers, " the median branch inserted at the base of the tentacles ; " (2) the posterior, nuchal or cervical tentacles, which have not yet been seen by any naturalist, except Huxley, who calls them " rudiments of eyes upon the outer side of the base of the tentacles,"3 whereas 1 Voyage de la Bonite, Zoologie, pi. xv. fig. 1. s On the Morphology of the Cephalous Mollusca, pi. iv. fig. 3. 3 On the Morphology of the Cephalous Mollusca, Phil. Trans., p. 41, pi. iv. fig. 3, i, 1S53. REPORT ON THE PTEROPODA. 53 these nuchal tentacles are quite separated from the so-called " tentacles." They are terminated by a small swelling, and although rather short, they are more projecting than in all other Gymnosomata, except Clionopsis krohni ; they are more laterally inserted than in the other genera but are innervated in the same manner.1 These tentacles are shown in PI. III. figs. 7, 8. 2. Buccal Appendages. — They number one pair, are very long (nearly as long as the fins), flattened and extensible, and are inserted one on each side of the buccal orifice (3, fig. 4). They are the so-called tentacles of the majority of zoologists, less the " medial branch of their basis," which is, as we have seen, the true anterior tentacles (2, fig. 4). Souleyet calls them gills.2 Ihering3 is the only one who considered them as corresponding to the buccal appendages of Clione, &c; but he did not recognise the tentacles properly so-called. (y^\J"~ Kegarding the union by their bases of the anterior Fl^i/ff0ntvieffr^td^ tentacles and buccal appendages, I explain it as follows. The \° ^rioTS'teutkd^;Ct wli genus Halopsyche is the only one among the Gymnosomata ^n age ' ' antenor which does not possess the so-called cephalic hood, which covers the buccal appendages when these are retracted within the buccal cavity. The buccal orifice of Halopsyche, therefore corresponds to the " false lips " of Clione ; the buccal appendages indeed are inserted on its sides, like the cephaloconi in Clione. The anterior tentacle not being therefore inserted on a cephalic hood in this genus, has been confounded, so far as the base is concerned, with the buccal appendage, because both have a common retractor muscle. The structure of the buccal appendages of Halopsyche is not known, and I had no specimens sufficiently good for histological investigation. Halopsyche? Bronn. 1825. Pysche, Rang (not Linne), Description d'un nouveau genre de la classe des Pte>opodes, Ann. d. Sci. Nat., ser. 1, t. v. p. 284. 1827. Euribia, Rang (not Hiibner, 1806), Description de deux genres nouveaux appartenant a la classe des Pteropodes, Ann. d. Sci. Nat., ser. 1, t. xii. p. 328. 1856. Euryhia, Woodward, A Manual of the Mollusca, p. 206. 1862. Halopsyche, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 645. 1862. Theceurybia, Bronn, Die Klassen und Ordnungen des Thierreichs, Bd. iii. p. 645. Characters and Description the same as for the family Halopsychidse. Rang established the genus Pysche in 1825 for a Pteropod without a shell that he called 1 Pelseueer, Recherches sur le systeme nerveaux des Pteropodes, Archives de Biologie, t. vii. p. 110. 2 Voyage de la Bonite, Zoologie, t. ii. p. 247. 3 Vergleichende Anatomie des Nervensystems und Phylogenie der Mollusken, p. 246. 4 From &*;, sea, and ipixh a mythological name . 54 THE VOYAGE OF H.M.S. CHALLENGES. Pysche globulosa, and which he considered to be one without a distinct head. But since the name Pysche had already been applied by Linne to a genus of butterflies, it had to be eliminated from the nomenclature of the Pteropoda, and Bronn, therefore, substituted the name Halopsyche, which has been adopted by Verrill,1 Lankester 2 and Boas.3 In 1827 Eang formed another new genus, Euribia, for a Pteropod that he thought to be testaceous (a supposition since proved to be incorrect) ; but here again the name Eurybia had been already used by Hiibner (1806), and it could not be retained for a Pteropod, and Bronn therefore changed it to Theceurybia. But, on the other hand, Souleyet 4 showed that the two genera "Psyche " and "Euribia" of Rang are not distinct, and the name Halopsyche must therefore be retained for the whole group. Eang,5 Cuvier,6 Gray,7 Owen,8 Woodward,9 Adams,10 Bronn u and Lankester,12 have all considered these animals as Thecosomatous Pteropods ; Macdonald 13 and Boas 14 alone maintaining that they are Gymnosomata. This last opinion is the true one. Halopsyche indeed does not possess any character of the Thecosomata ; on the contrary, it shows all the characters common to' the Gymnosomata ; it possesses neither mantle skirt, shell, nor pallia! cavity ; it has a distinct head, bearing two pairs of tentacles (the posterior with rudiments of eyes) ; the foot and fins are separated ; there are buccal appendages and a jaw of which the two halves are united in the median line ; there are no stomachal horny plates ; the anus is on the right side (and not on the left, as stated by Huxley 15) ; the cerebral ganglia are united above the oesophagus ; and finally, the larvae resemble those of all the Gymnosomata (PI. III. fig. 9). Halopsyche therefore has no relation to the Thecosomata. If its radula seem to resemble that of the Thecosomata in the number of the teeth, it differs from it by the conformation of the teeth, as may be seen below. Halopsyche can no longer be regarded, as some zoologists believe, as a transitional 1 Notice of Recent Additions to the Marine Invertebrata of the North- East Coast of America, Proc. U.S. Nat. Mus., vol. Hi. p. 393. 2 Mollusca, Encyclopaedia Britannica, ed. 9, vol. xvi. p. 666. 3 Spolia atlantica, p. 171. 4 Voyage de la Bonite, Zoologie, t. ii. p. 252. 6 Description de deux genres nouveaux appartenant a la classe des Pteropodes, Ann. d. Sci. Nat., s6r. 1, t. xii. p. 328. 6 Le Regne animal. 7 Catalogue of the Mollusca in the Collection of the British Museum, pt. ii., Pteropoda, p. 27. 8 Mollusca, Encyclopaedia Britannica, ed. 8, vol. xv. p. 381. 9 A Manual of the Mollusca. 10 The Genera of Recent Mollusca, vol. i. p. 56, 57. 11 Die Klassen und Ordnungen des Thierreichs, Bd. iii. pp. 645. 12 Mollusca, Encyclopedia Britannica, ed. 9, vol. xvi. p. 666. 13 On the Anatomy of Eurybia gaudichaudi, Trans. Linn. Soc. Lond., vol. xxii. p. 248. 14 Spolia atlantica, p. 171. 15 On the Morphology of the Cephalous Mollusca, Phil. Trans., p. 41, 1853. REPORT ON THE PTEROPODA. 55 type, with common characters of the two groups Thecosomata and Gymnosomata, because these groups are very distinctly separated from one another. * Halopsyche gaudichaudl, Souleyet (PI. III. figs. 7-9). 1852. Euribia, gaudichaudil, Souleyet, Voyage de la Eonite, Zoologie, t. ii. p. 253, pi. siv. figs. 1-6. 1883. Thecearybia norfolkensis, Lankester, Mollusca, Encycloptedia Britannica, ed. 9, vol. xvi. p. 666, fig. 83. 1886. Halopsyche gaudichaudil, Boas, Spolia atlantiea, p. 173, pi. viii. fig. 119. Body ovate, nearly globose, barrel-shaped, rounded at the posterior extremity. Head small and rather short. Fins long, wide and truncated at their extremity, of which the margin is sHghtly sinuous. Radula. — Formula 1:1:1; the median tooth entirely differs from that of the Thecosomata, and on the contrary much resembles that of Clione ; the lateral teeth are long and narrow and have a wide basal part, as in all the Gymnosomata. Colour. — Nearly absent ; the body-wall is translucent and slightly yellowish. Length, 4 mm.; very large specimens reach G or 7 mm. (Souleyet). Challenger Specimens. — West Pacific locality — Station 170, July 14, 1874; off the Kermadec Islands; lat. 29° 55' S., long. 178° 14' W. Australasian locality — Station 201, October 26, 1874; off Mindanao, Philippine Islands ; lat. 7° 3' N., lat. 121° 48' E. (numerous specimens). Habitat. — West Pacific Ocean, to long. 174° E.. between lat. 40° N. (Kiel Museum), and lat. 29° S. (Challenger Expedition, Station 170); generally in large shoals. Observation. — According to Macdonald1 this species possesses the curious peculiarity of being ovoviviparous. The larvas are proportionally longer than the adults; they are pos- teriorly a little pointed and thus possess the general form of the larvae of the Gymno- somata (PL III. fig. 9). The two posterior ciliated rings remain visible until a late period. Cymbulia norfolhensis, Quoy and Gaimard,2 must be related to the present genus ; it seems very close to Halopsyche gaudichaudl, from which it only differs (according to the figures of Quoy and Gaimard) by having small tubercles on the wall of the body. Cymbulia norfolhensis was caught near the Norfolk Islands, and measures about 5 mm. in length, but it is not well enough known to definitively decide its systematic position. In regard to Euribia hemispherica, Rang 3 and Psyche globulosa, Rang,4 both from 1 On the Anatomy of Eurybia gaudichaudl, Trans. Linn. Soc. Lond., voL xxii. p. 246, pi. xliii. fig. 9. 2 Voyage de decouvertes de r Astrolabe, Zoologie, t. ii. p. 376, pi. xxvii. figs. 31, 32. 3 Description de deux genres nouveaux appartenant a la classe des Pteropodes, Ann. d. Sci. Nat., ser. 1, t. xii. p. 329, pi. xivB. figs. 9-11. The specimens called Euribia hemispherica in the British Museum are not Pteropods. * Description d'un nouveau genre de la classe des Pteropodes, Ann. d. Sci. Nat, s&\ 1, t. v. p. 284, r>l. vii. fie 1. 56 THE VOYAGE OF H.M.S. CHALLENGER. the Atlantic Ocean (the latter from Newfoundland), it is very difficult thoroughly to appreciate them, so summary are the descriptions and figures of Eang ; but they appear to resemble one another very closely. Verrill says1 that he found the " Halopsyche" globulosa, Rang, on the shores of Newfoundland and Nova Scotia. I have several times asked him for a specimen for examination, but have never obtained one. Souleyet 2 thinks that Argivora parva, Lesueur,3 from La Martinicpie, also resembles Halopsyche ; but the information available in regard to this species is very fragmentary, and its affinities cannot at present be decided. It is described, indeed, as a naked Cym- bulia; it may, therefore, be a Tiedemannia (Gleba), that is, a Thecosomatous Pteropod. In conclusion, I must notice the genus Thliptodon, Boas, that I mentioned when speaking of the larvae of the Pneunionoderniatidae. This genus is known only from its larva, and thus its systematic position remains uncertain among the various families of Gymnosomata. These larvae have a larger head than any other known larva of Gymnosomatous forms. Its radula has a formula of 4:1:4; the median tooth is very short and with- out denticles ; on the contrary the lateral teeth are very long, narrow, and slightly bent at their free extremity. The jaw, the hook-sacs, and the buccal appendages have never been seen ; perhaps some of them are wanting. The largest larvae have a foot, the posterior lobe of which is well developed. The third or posterior ciliated ring has the form of a star with three radii ; it may, therefore, be asked if the posterior gill of the adult is not perhaps like that of Notobranchsea. Two larvae observed by Krohn 4 (one of them at Funchal, Madeira), possess a radula which resembles that of the above larvae, formerly described by Gegenbaur6 (and which Boas called Thliptodon gegenbauri). With respect to the genus Cymodocea (Cymodocea diaphana, d'Orbigny),8 it is a form that is quite indeterminable from d'Orbigny 's description ; it was, moreover, founded upon a mutilated specimen. It is very doubtful whether the organism called Homoderma7 was a Pteropod at all; it has no fins, and its mouth is lateral not axial. 1 Notice of Recent Additions to the Marine Invertebrata of the North-East Coast of America, Proc. U.S. Nat. Mus., vol. iii. p. 393. 2 Histoire naturelle des Mollusques Pteropodes, pp. 71, 72. 3 De Blainville, Manuel de Malacologie, p. 655. 4 Beitrage zur Entwickelungsgeschichte der Pteropoden und Heteropoden, pp. 11, 14, p]. i. figs. 6, 11. 6 Untersuchungen liber Pteropoden und Heteropoden, pi. v. figs. 14-17. 6 Voyage dans l'Amerique meridionale, t. v. p. 133, pi. ix. figs. 16, 17. 7 Van Beneden, Exereices zootomiques, p. 54, pi. iii. figs. 4, 5. REPORT ON THE PTEROPODA. 57 SUMMARY OF THE KNOWN SPECIES OF GYMNOSOMATA. It will be seen from the foregoing descriptive monograph, that we now know fairly well, and in the adult state, twenty-one species of Gymnosomatous Pteropoda, which are divided as follows : — Genera. Dexiobranchxa, . SpongiobrancJi xa, Pneumonoderma, Clionopsis, . Notobranchxa, . CHone, Halopsyche, Species 5 1 6 3 2 3 1 21 Of these twenty-one species, eleven were collected by the Challenger Expedition ; these eleven species belong to six of the seven known genera, the genus CHone being alone wanting (for the arrangement of the species in the different genera, see the table in the Introduction, p. 8). One of these genera is new {Notobranchxa), as, are also more than a third of the species collected (four out of eleven). The eleven Challenger species were obtained at thirteen different Stations, as shown in the following list ; all being collected, like all the living Pteropoda, in surface gatherings. (ZOOL. CHALL. EXP. PART LVIII.— 1887.) Mmru 8 GEOGRAPHICAL DISTRIBUTION. STATION LIST OF THE GYMNOSOMATOUS PTEROPODA OF THE CHALLENGES EXPEDITION. Station 158. March 7, 1874; south of Australia; lat. 50° 1' S., long. 123° 4' E.; surface temperature, 45° F. Spongiobranchsea australis. Also at Station 314a; and on March 11, 1876 Station 170. July]4,1874; off the Kermadec Islands ; lat. 29° 55' S., long. 178° 14' W.; surface temperature, 65° F. Halopsyche gaudichaudi. Also at Station 201. Station 201. October 26, 1874; Philippine Islands; lat. 7° 3' N, long. 121° 48' E.; surface temperature, 83" F. Halopsyche gaudichaudi. Also at Station 170. Station 222. March 16, 1875 ; north of New Guinea; lat. 2° 15' N, long. 146° 16' E.; surface temperature, 8 2° "8 F. Pneumonoderma peroni. Station 240. June 21, 1875; North Pacific, east of Japan; lat. 35° 20' N., long. 153° 39' E.; surface temperature, 64°'8 F. Pneumonoderma pacijicum. Also at Station 254. Station 242. June 24, 1875; North Pacific, east of Japan ; lat. 35° 29' N., long. 161° 52' E.; surface temperature, 68°-5 F. Notobvanchsea inopinata. Between Stations 247 and 248. July 4, 1875; North Pacific; lat. 36° 42' N., long. 179° 50' W.; surface temperature, 73°"2 F. Dexiobranchsea minuta. 60 THE VOYAGE OF H.M.S. CHALLENGER. Station 254. July 17, 1875; North Pacific, north of the Sandwich Islands; lat. 35° 13' N., long. 154° 43' W.; surface tempera- ture, 72° F. Pneumonoderma pacificum. Also at Station 240. Pneumonoderma soideyeti. Clionopsis modesta. Station 297. November 11, 1875; off the western coast of South America; lat. 37° 29' S., long. 83° 7' W.; surface temperature, 57° F. Dexiobranch&a polycotyla (larvae). Station 314a. January 22, 1876; between Cape Virgins and the Falkland Islands; lat. 51° 24' S., long. 61° 46' W.; surface temperature, 49° F. Spongiobrancliiea australis. Also at Station 158 ; and on March 11, 1876. Between Stations 332 and 333. March 11, 1876; west of . Tristan da Cunha; lat. 36° 22' S., long. 26° 1' W.; surface temperature, 64°7 F. Spongiobranchaea australis. Also Stations 158 and 314a. April 26, 1876; St. Vincent, Cape Verde Islands; lat. 16° 49' N. long., 27° 14' W. ; surface temperature, at noon 730,2 F., at midnight 72°-2 F. Dexiobrancheea ciliata. Near Station 354. May 7, 1876 ; Atlantic, south of the Azores; lat. 34° 22' N, long. 34° 23' W.; surface temperature, 67° -8 F. Pneumonoderma violaceum. GENERAL CONCLUSIONS. A. The Pelagic Provinces. The geographical distribution of littoral marine] [organisms has been the subject of many researches, and we possess considerable knowledge of their arrangement and groupings into distinct great natural faunas. But it is very different with regard to pelagic animals, our knowledge of their geographical distribution being much more fragmentary. In order to satisfactorily determine the natural pelagic provinces into which oceans REPORT ON THE PTEROPODA. 61 and seas are divided, the study of all, or nearly all pelagic animals, ought to be com- bined, because by limiting investigation to one single group of organisms, one is erroneously led to multiply the geographical subdivisions. But since up to the present time there are but few groups of pelagic animals (Crustacea, Polyzoa, &c), the geographical distribution of which has been investigated, all attempts at synthesis in this way would as yet be premature, and would give merely provisional results. The distribution of the Pteropoda will add to our knowledge on this subject, and I shall here give the conclusions at which I have arrived after studying the geographical distribution of the Gymnosomatous Pteropoda. The importance of currents in influencing the geographical distribution of pelagic organisms is much greater than in the case of littoral animals ; and it seems to me that each great surface area of water coinciding with an important current or system of currents, forms a distinct pelagic province. Eelying chiefly upon the distribution of the Pteropoda, I think I may at least divide the surface of the ocean into the ten following provinces, many of them resembling those established by Mr. George Busk in his Report on the Polyzoa, and which have been followed in several other Reports. I. Arctic Province. — This province corresponds to the region generally accepted by zoologists under this name. It extends north of both the Old and the New Continents, and between them, as the Behring Sea, the Arctic current, and the current of Hudson's Bay (or the Labrador current). Thus, this province extends all around the North Pole, within the limit of the floating ice, or, more exactly, of the isotherm of 60° F. for August. II. North Atlantic Province.- — This division comprises all the Atlantic Ocean, south of the foregoing province, to the equator ; it includes the Gulf of Guinea and also a small part of the Central Atlantic, south of the equator. This province thus embraces the whole area traversed by the following system of currents ; the Gulf Stream, Rennel's current, North African and Guinea current, and the main equatorial current, besides the whole Mediterranean Sea ; that is to say, this province is nearly equivalent to the following littoral regions of the malacologists : — Celtic, Lusitanian, West African, Car- ribean and Atlantic. It corresponds to Dana's north temperate and torrid zones of the Atlantic, without the subtorrid region of the South Atlantic. The part of the Atlantic Ocean which is traversed by the Brazfl current perhaps constitutes a special province ; but I cannot be certain about this, not having received any Pteropoda from this region. III. South- West African Province. — This division includes the parts of the South Atlantic traversed by the South Atlantic current. On account of the origin of this current this province has some relations with the following one. IV. Indian Ocean Province. — This province comprises the whole Indian Ocean, 62 THE VOYAGE OF H.M.S. CHALLENGER. from the east coast of Africa to Australia, and extending to about lat. 36° S. This province was not visited by H.M.S. Challenger, and requires to be explored. V. Australasian Province. — This division nearly corresponds to the Indian province of Keferstein,1 that is, including the seas of the Malay Archipelago, the China Sea, the seas of the Philippines, of Celebes, of New Guinea, and of the north of Australia. VI. West Pacific Province. — This province includes the part of the Pacific Ocean which is traversed by the equatorial currents — north equatorial drift current, equatorial counter current, south equatorial drift current. It extends no further than from lat. 40° N. to lat. 30° S., and rather corresponds with the Polynesian province of Keferstein.2 VII. East Australian Province. — This division occupies that part of the ocean between Australia and New Zealand. VIII. North Pacific Province. — This division extends, south of the Behring Sea, from one continent to the other, and comprises all that part of the North Pacific Ocean which is traversed by the North Pacific drift current. It extends to the south as far as the foregoing province, and towards the American coast its southern limit is near lat. 30° S. This province is thus limited by the Aleutian and Californian regions of the malacoloa;ists. Until the voyage of H.M.S. Challenger this province was scarcely explored at all, and its exploration, therefore, gave remarkable results. All the four new species of Gymnosomata were obtained in it. IX. South-East Pacific Province. — This province lies along the American continent, and occupies all that part of the Southern Pacific Ocean which is traversed by Humboldt's current, that is to say, from lat. 37° S. to the Bay of Panama. X. Antarctic Province. — This province surrounds the South Pole, and comprises the region corresponding to the Arctic province, namely, nearly the whole part of the sea included within the isothermal line of 50° F., for August, and traversed by the Cape Horn current, the southern connecting current, the South Australian current, and the Antarctic drift ; it is nearly the same as the Magellanic littoral region of the malaco- logists. This province has been hitherto but little explored. As to the Pteropoda, its short exploration by H.M.S. Challenger procured, with other well-known species, three species of Thecosomata, which had never been observed since their first discovery, namely, Limacina antarctica, Woodward (Station 153), Limacina (or Spinalis) australis, Souleyet (Stations 146, 153), and a new species of Clio (or Cleodora) (Station 156), which had probably been previously collected by the naturalists of the " Astrolabe " during her last voyage to the South Pole,3 but which was not described. 1 Bronn, Die Klassen unci Ordnungen cles Thierreiehs, Bd. iii. p. 1135. 2 Loc. cit., p. 1136. 3 Souleyet, Histoire nalurelle cles Mollusques Pteropodes, p. 48. REPORT ON THE PTEROPODA. 63 B. Geographical Distribution of the Genera. 1. The genus Dexiobranchsea is found in the Atlantic (north of the equator) in the Mediterranean Sea, and in the North and South-East Pacific Ocean. Among the five known species, two inhabit the North Atlantic province, near the west coast of the Old Continent ; two inhabit the South-East Pacific province, on the west coast of the New Continent ; the fifth species lives in the North Pacific. 2. The genus Spongiobranchsea inhabits the whole Antarctic province, all round the South Pole, where its single species seems quite to correspond to Clione limacina of the Arctic province. Like this last species, indeed, Spongiobranchsea australis never passes the limit of floating ice, and its area of distribution is enclosed within the isotherm of 50° F. for August. 3. The genus Pneumonoderma, which most abounds in species, is also the genus which has the widest area of distribution. It was observed in the whole Atlantic Ocean between lat. 45° N. and lat. 38° S., in the Mediterranean Sea, in the Indian Ocean, and in the Pacific Ocean from lat. 58° N. to lat. 42° S. In the Atlantic Ocean, Pneumonoderma violaceum is hardly found south of the equator ; Pneumonoderma peroni, in the Atlantic, is never found north of the tropic of Capricorn ; it also lives in the Indian Ocean and Australasian provinces, always south of the equator. In the Pacific Ocean, Pneumonoderma pacificum and Pneumonoderma souleyeti are localised in the North Hemisphere, and Pneumonoderma boasi in the South Hemisphere. Some species, as Pneumonoderma peroni and Pneumonoderma mediterraneum, seem to be very widely dispersed (the former in the South Hemisphere, the latter in the Northern). This genus, however, requires fresh study when new and more abundant material has been accumulated. 4. The genus Clionopsis is only distributed in the North Hemisphere — North- West Atlantic Ocean, Mediterranean Sea, China Sea and North Pacific Ocean. 5. The genus Notobranchsea was observed in both hemispheres, both in the Atlantic and Pacific Oceans. 6. The genus Clione (with the exception of the badly known species that Bruguiere found at Madagascar) is distributed throughout the whole Arctic province (the area of dispersion of Clione limacina exactly corresponds with this province), the North Atlantic Ocean, the Mediterranean Sea and the South- West Pacific Ocean. 7. The genus Halopsyche has been observed in the whole of the West Pacific Ocean, north and south of the equator. The following table summarises the description given above of the geographical distribution of the genera of Gymnosomatous Pteropoda (this table only gives the 64 THE VOYAGE OF H.M.S. CHALLENGER. minimum areas of distribution, because all doubtful indications have been removed from it) : — Arctic. North Atlantic. South- West African. Indian Ocean. Austra- lasian. West Pacific. East Austra- lian. North Pacific. South- East Pacilic. Ant- arctic. Dexiobranchsea, SpongiobrancJisea, . . Pneumonoderma, . , Clionopsis, .... Notohranclieea, . Clione, Halopsyche, .... + + + + + + + + 1 + + + + + 2 + 3 + + + + + With respect to the species of Gymnosomata, the following list shows how they are distributed in the different provinces. C. The Species of Gymnosomata arranged in Provinces. I. Arctic Province. Clione limacina. II. North Atlantic Province. Dexiobranchsea paucidens. * 4 Dexiobranchsea ciliata. Pneumonoderma mediterraneum, also V. # Pneumonoderma violaceum. Clionopsis krohni. Clionopsis grandis, also V. Notobranchsea macdonaldi. Clione longicaudata. Clione Jlavescens. III. South- West African Province. Pneumonoderma peroni, also IV. and V. 1 Clio capensis, Rang. 2 The "trigonal tailed Clio" of Macdonald ; see pp. 41, 43. 3 The " Clio caudata" of Macdonald. 4 Species marked with an asterisk have been collected in the province by the Challenger Expedition. REPOBT ON THE PTEEOPODA. 65 IV. Indian Ocean Province. Pneumonoderma peroni, also III. and V V. Australasian Province. Pneumonoderma mediterranean,1 also II * Pneumonoderma peroni, also III. and IV. Clionojjsis grandis, also II. *Halopsyche gaudichaudi, also VI. VI. West Pacific Province. *Halopsyche gaudichaudi, also V. VIII. North Pacific Province. *Dexiobranchwa minuta. * Pneumonoderma souleyeti. * Pneumonoderma pacificum. *Clionopsis modesta. * Notobrancheea inopinata. IX. South-East Pacific Province. Dexiobrancheea simplex. Dexiobranchsea polycotyla. Pneumonoderma boasi. X. Antarctic Province. *Spo7igiobranchsea australis: " Pneumodermon macrocotyhim," Boas. (zool. chall. exp. — part Lvin. — 1887.) Mmm 9 PHYLOGENETIC RELATIONS OF THE GENERA OF GYMNOSOMATOUS PTEROPODA. In the anatomical part of this Report, I shall show the close affinities that these Mollusca possess to the Tectibranchiate Opisthobranchs. The zoologists who have carefully studied the organisation of the Thecosomata and Gymnosomata have always recognised the great differences that separate the two groups ; and many have thought that they have not had a common origin. Boas, however, was the first who clearly stated this opinion,1 against which no solid argument can be opposed. Here I have only to consider the origin of the Gymnosomata. According to my opinion, their origin ought to be found in the group Aplysiidae, Opisthobranchia which already show a remarkable adaptation to swimming habits. In the anatomical part of this Report I shall show the great and numerous resemblances which exist between the organisation of the Gymnosomata and that of the Aplysiidae. With regard to the relations of the different genera of Gymnosomatous Pteropoda among themselves, we find rather primitive genera which have retained much of their resemblance to the primitive stock, and others which are highly specialised. Between these two extremes, there are a whole series of forms, by means of which the evolution of the group could almost be traced, and which are not difficult to arrange in a phylo- genetic tree, according to their affinities. The stage of evolution of the group, which each genus represents, is chiefly characterised by the development of the respiratory organs. Thus, Dexiobranchsea, which among all the known genera is the most primitive, only possesses the lateral gill corresponding to that of the Aplysiidae. Spongiobranchsea has, moreover, a very simple posterior gill, a specialisation of the posterior ciliated ring which remains very late in the foregoing genus. Pneumonoderma shows a great increase in the complexity of this posterior gill, by the formation of four crests radiating from the posterior quadrangular ring, which corresponds to the membranous ring of Spongiobranchaia. Clionopsis already shows a retrogression, the lateral gill having quite disappeared ; and 1 Spolia atlantica, p. 14. 68 THE VOYAGE OP H.M.S. CHALLENGES. the posterior gill being generally simplified, but formed, however, on the same plan as in Pneumonoderma. Notobranchsea has also lost the lateral gill ; the posterior crest of the terminal gill has also disappeared, and there only remains a gill formed of three radiating crests. Lastly, Clione has lost every vestige of gill, as in Halo-psyche. A form closely resembling some very specialised Dexiobranchsea? with atrophied median buccal appendage, and with the posterior ciliated ring converted into a respiratory ring, might give rise : — 1. To Sjwngiobranchsea, by the lateral gill becoming rudimentary. 2. To Pneumonoderma, by specialisation of the posterior gill, and the development of radiating crests on the annular crest.2 Pneumonoderma, by the disappearance of the lateral gill and buccal appendages, and by the atrophy of the posterior lobe of the foot, gave rise to Clionopsis. A form allied to Pneumonoderma, in which the lateral gill has disappeared as well as the quadrangular crest of the posterior gill, and in which the latter bears only three radiating crests, might lead to Notobranchsea. This last genus certainly represents the forms from which Clione has been derived, by the total disappearance of the branchial apparatus. 1 The evolution of the buccal appendages in Dexiobranchsea illustrates the evolution of the genus. The first specialisation was the production of suckers on the proboscis (ventrally and on the two lateral sides) ; this state is shown in Dexiobranchsea simplex. The ventral suckers then united their peduncles into a single median appendage, as seen in Dexiobranchsea paucidens. The number of the lateral suckers then increased, and they were arranged in two rows, as in Dexiobranchsea rainuta, and Dexiobranchsea polycotyla. Lastly, the ventral appendage became atrophied, and, on the other hand, the lateral suckers became more developed, and they were borne on projecting crests of the proboscis, the first indication of the lateral appendages of the following genera, Spongiobranchsea and Pneumonoderma. This last state is shown in Dexiobranchsea ciliata. Therefore the living species of Dexiobranchsea must be phylo- genetically arranged as follows : — 1. Dexiobranchsea simplex. 2. Dexiobranchsea paucidens 3. Dexiobranchsea minuta. 4. Dexiobranchsea polycotyla. 5. Dexiobranchsea ciliata. - In the Pneumonodermata, the evolution is chiefly shown in the branchial apparatus. In the most primitive species the lateral gill is still long as in some Dexiobranchse, and the fringes on both lateral and posterior gill are hardly developed. Pneumonoderma souleyeti corresponds to this stage. The branchial fringes then became more developed, but the lateral gill continued long — Pneumonoderma pacificum. The lateral gill became shorter, and the fringes more distinct, but not yet subdivided — Pneumonoderma boasi. The fringes became subdivided — Pneumono- derma violaceum. Then, by specialisation of the buccal appendages the other species arose — by increase of the number of suckers, Pneumonoderma peroni ; by decrease, Pneumonoderma mediterraneum. The living species of Pneumono- derma must therefore be phylogenetically arranged as follows : — 1. Pneumonoderma souleyeti. 2. Pneumonoderma pacificum. 3. Pneumonoderma boasi. 4. Pneumonoderma violaceum. 5. Pneumonoderma peroni. 6. Pneumonoderma mediterraneum. REPORT ON THE PTEROPODA. 69 Halopsyche is the most specialised genus of all the Gymnosomatous Pteropoda ; it has lost, besides every vestige of a branchial apparatus, the cephalic hood and the proboscis. These statements may be represented in the following diagram, which shows the mutual relations of the Gymnosomata. This phylogenetic tree only exhibits, of course, the grouping of the living forms according to their reciprocal affinities, and does not at all suggest their mutual descent, which is quite impossible among cotemporary organisms. Olionc Halopsyche Notobranchsm Spongiobranct Pnewrumoderma ClionopsU Demobranchma CONTENTS. Introduction, . 1. History of the Group, . 2. The Gymnosomata and their Distinction from the Thecosomata 3. The Habits of the Gymnosomata, 4. The Gymnosomata of the Challenger Expedition, Description op Genera and Species, . The Genera and Families of Gymnosomata, Family I. Pneumonoderniatidae, Dexiobranchsea, Boas, ciliata, Gegenbaur, minuta, n. sp., simplex, Boas, paucidens, Boas, polycotyla, Boas, Spongiobranchsea, d'Orbigny, australis, d'Orbigny, Pneumonoderma, Cuvier, mediterraneum, van Beneden, violaceum, d'Orbigny peroni, Lamarck, boasi, n. sp., pacificum, Dall, souleyeti, n. sp., Family II. Clionopsidas, Clionopsis, Troschel, hrohni, Troschel, grandis, Boas, modesta, n. sp., Family III. Notobranchaidae, Notobrandima, Pelseneer, macdonaldi, Pelseneer, inopinata, n. sp., Family IV. Clionidse, Clione, Pallas, limaeina, Phipps, longicaudata, Souleyet, flavescens, Gegenbaur, page 1 1 4 6 7 9 9 11 14 15 16 16 17 17 18 19 21 26 28 29 30 30 31 33 34 34 36 37 37 38 39 40 41 44 45 48 49 72 THE VOYAGE OF H.M.S. CHALLENGER. Family V. Halopsychidae, Halqpsyche, Bronn, gaudichaudi, Souleyet, Summary of the known Species of Gymnosomata, Geographical Distribution, Station List, .... General Conclusions, A. The Pelagic Provinces, B. Geographical Distribution of the Genera, C. The Species of Gymnosomata arranged in Provinces, Phylogenetic Relations of the Genera of Gymnosomatous Pteropoda, page 52 53 55 57 59 59 60 60 63 64 INDEX. Synonyms are printed in italics, the pages where species are described are indicated by dark type. JEgle, Oken, 9,21. cucullata, Oken, 24, 25, 29. Argivora parva, Lesueur, 56. Cavolinia, 2. Cirrifer, Pfeffer, 9. paradoxus, Pfeffer, 28. Clio, Browne, 42. Clio, Gegenbaur, 34. Clio, 0. F. Miiller, 2, 34, 38, 41, 42, 44, 51, 62. amati, delle Chiaje, 44. aurantiaca, Fol, 43, 44, 47, 49, 50. australis, Bruguiere, 43, 45, 51. borealis, Gegenbaur, 43. borealis, Eoss, 51. capensis, Kaug, 41, 43, 44, 64. caudata, Gray, 44. caudata, Macdonald, 50, 64. flavescens, Gegenbaur, 43, 49. fusiformis, 50. limacella, Eang, 43, 45, 50. Umacina, Phipps, 45. longicaudata, Souleyet, 43, 4b. mediterranea, Gegenbaur, 34, 43, 44. miquelonensis, Eang, 45. pyramidalis, Quoy and Gaimard, 51. pyramidata, Eang, 45, 51. retusa, 0. F. Miiller, 45. Cliodita, Quoy and Gaimard, 9, 18, 42, 44. caduceus, Quoy and Gaimard, 19, 21, 44. fusiformis, Quoy and Gaimard, 45, 50, 51. Clione, Pallas, 2, 7, 9, 10, 11, 16, 19, 20, 21, 23, 34, 35, 38, 41, 42, 43, 44, 45, 47, 49, 50, 51, 52, 53, 57, 63, 64, 68, 69. borealis, Pallas, 42, 44, 45. caudata, Gray, 48. dalli, Krause, 44, 45, 48. elegantissima, Dall, 45, 47, 48. (ZOOL. CHALL. EXP. — PART LVIII. — 1887.) Clione flavescens, Gegenbaur, 44, 45,^47, 48, 49, 50, 64. limacina, Phipps, 42, 43, 44, 45, 46, 47, 48, 49, 51, 63, 64. longicaudata, Souleyet, 19, 43, 44, 45, 48, 50, 64. longicaudatus, Verrill, 39, 40. pajnlionacea, Auctorum, 45. Clionopsis, Troschel, 8, 10, 11, 12, 32, 33, 34, 35, 37, 48, 49, 52, 57, 63, 64, 67, 68, 69. grandis, Boas, 23, 34, 36, 64, 65 (PI. II. figs. 7, 8). krohni, Troschel, 9, 13, 34, 35, 36, 43, 44, 49, 53, 64 (PI. II. fig. 9 ; PI. III. fig. 1). modesta, n. sp., 34. 37, 60, 65 (PI. III. fig. 2). Cliopsis, Troschel, 34, 35. grandis, Boas, 36. krohnii, Troschel, 34. Cymbulia, 56. norfolkensis, Quoy and Gaimard, 55. Cymodocea, d'Orbigny, 10, 56. diaphana, d'Orbigny, 56. Dexiobrancruea, Boas, 8, 9, 10, 11, 12, 13, 14, 18, 24. 57, 63, 64, 67, 68, 69. ciliata, Gegenbaur, 15, 18, 60, 64, 68 (PL I. fig. 1). improvisa, Boas, 15. minuta, n. sp., 15, 16, 59, 65, 68 (PI. I. fig. 2). paucidens, Boas, 15, 17, 50, 64, 68. polycotyla, Boas, 15, 17, 18, 60, 65, 68 (PI. I. figs. 4, 5). simplex, Boas, 15, 16, 65, 68 (PI. I. fig. 3). Euribia, Eang, 6, 9, 53, 54. guudichaudii, 55. hemispherica, Eang, 55, 56. Halopsyche, Bronn, 6, 8, 9, 10, 11, 42,53, 54, 56, 57, 63, 64, 68, 69. gaudichaudi, Souleyet, 39, 53, 55, 59, 65 (PL III. figs. 7-9). Mmm 10 74 THE VOYAGE OF H.M.S. CHALLENGER. Halopsyche globulosa, Rang, 56. Homoderma, van Beneden, 10, 56. Notobranchsea, Pelseneer, 8. 10, 11,38, 39,41,43, 41, 56, 57, 63, 64, 68, 69. inopinata, n. sp., 38, 39, 40, 59, 65 (PI. 111. figs. 5, 6). macdonaldi, Pelseneer, 38, 39, 40, 41, 64 (PI. 111. 'figs. 3, 4). Pelagia, Quoy and Gaimard, 10. alba, Quoy and Gaimard, 37. Pneumoderma, Peron and Lesueur, 21. Pneumodermis, Oken, 21. attantica, Oken, 24, 25, 28. Pneumodermon, Gegenbaur, 14, 21, 37. audebardi, Rang, 24, 26. dliatum, Gegenbaur, 14, 15, 24. cuSullatum, Gray, 28. mdcrocotylum, Boas, 24, 26, 27, 65. mediterraneum, van Beneden, 15, 26, 27. pamficwn, Dall, 30. pellucidum, Quoy and Gaimard, 24, 32. peronii, Souleyet, 26, 28, 29. peronii, Costa, 26. peronii, Quoy and Gaimard, 28. peronii, Verrill, 36. ruber, Quoy and Gaimard, 24, 32. violaceum, Gegenbaur, 27, 28, 30. Pneumodermopsis, Bronn, 9, 14. eiliata, Bronn, 15. Pneumonoderma, Cuvier, 2, 6, 7,8, 9, 10, 11, 12, 13. 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 31, 32, 33, 34, 36, 41, 48, 52, 57, 63, 64, 67, 68, 69. boasi, 23, 25, 26, 29, 30, 63, 65, 68 (PI. II. fig. 3). mediterraneum, van Beneden, 19, 21, 23, 24, 25, 26, 27, 28, 30, 31, 63, 64, 65, 68 (PI. I. fig. 8; PI. II. tig. 1). Pneumonoderma pacificum, Dall, 12, 22, 23, 25, 26, 30, 31, 36, 39, 59, 60, 63, 65, 68 (PL II. figs. 4, 5). peroni, Lamarck, 23, 24, 25, 26, 29, 36, 59, 63, 64, 65, 68 (PI. II. fig. 2). souleyeti, n. sp., 12, 22, 23, 25, 26, 31, 60, 63, 65, 68 (PI. II. fig. 6). violaceum, d'Orbigny, 7, 12, 22, 23, 24, 25, 26, 27, 28, 29, 39, 60, 63, 64, 68 (PI. I. fig. 9). Pneumonodermum, Herrmannsen, 21. Psyche, Rang, 53, 54. globulosa, Rang, 54, 55, 56. Pterocymodocea, Bronn, 10. Pteropelagia, Bronn, 10. Spongiobrancbtea, d'Orbigny, 8, 10, 11, 12, 18, 19, 20, 21, 22, 33, 44, 48, 57, 63, 64, 67, 68, 69. australis, d'Orbigny, 9, 12, 13, 19, 20, 21, 22, 39, 44, 51, 59, 60, 63, 65 (PI. I. figs. 6, 7). Spongiobranchea, d'Orbigny, 18, 44. elongala, d'Orbigny, 19, 48. australis, Chenu, 26, 28. Spongiobranchia, d'Orbigny, 18. Spongobranchia, Bronn, 18. Spongobranchus, Wagner, 18. Theceuryhio, Bronn, 53, 54. norfolkensis, Lankester, 55. Thliptodon, Boas, 13, 56. gegenbauri, Boas, 56. Trichocydus, Eschscboltz, 9, 13, 20, 21, 35, 40. dumereillii, Verrill, 40. dumerili, Escb.snb.oltz, 9, 13, 20, 21, 40. mediterrcmeus, Costa, 9, 13, 34, 35. Trizonius, Buscb, 9. eeecus, Buscb, 13, 27. PLATE I. (ZOOL. CHALL. EXP. PART LVIII. 1887.) MlMll. PLATE I. Fig. 1. Dexiobranchsea ciliata, Gegenbaur. Ventral aspect ; magnified ten diameters. Fig. 2. Dexiobranchsea minuta, n. sp. Ventral aspect ; magnified twenty diameters. Fig. 3. Dexiobranchiea simplex, Boas. Ventral aspect ; magnified fifteen diameters. Figs. 4, 5. Dexiobranchwa polycotyla, Boas. Fig. 4. Adult, ventral aspect; magnified twelve diameters. Fig. 5. Larva, ventral aspect; magnified forty diameters. Figs. 6, 7. Spongiobranchasa australis, d'Orb. Fig. 6. Ventral aspect ; magnified eight diameters. Fig. 7. Dorsal aspect; magnified eight diameters. Fig. 8. Pneumonoderma mediterraneum, van Ben. Ventral aspect; magnified four diameters. Fig. 9. Pneumonoderma violaceum, d'Orb. Lateral (right) view ; magnified five diameters. The Voyage of H.M.S"CliaTLenge Pterop /•'hi. / Fy.e. Fig. , Fig :,■ Fig. 7. Fig. 5. Fig. 3. P Pelseneer ad na* del FHuth(l.u>.' Editf PLATE II. PLATE II. Fig. 1. Pneumonoderma mediterraneum, van Ben. Lateral (right) view; magnified four diameters. Fig. 2. Pneumonoderma peroni, Lam. Ventral aspect ; magnified five diameters. Fig. 3. Pneumonoderma boasi, n. sp. Ventral aspect ; magnified ten diameters. Figs. 4, 5. Pneumonoderma pacificum, Dall. Fig. 4. Ventral aspect ; magnified six diameters. Fig. 5. Lateral (right) view; magnified six diameters. Fig. 6. Pneumonoderma souleyeti, n. sp. Ventral aspect : magnified ten diameters. Figs. 7, 8. Clionopsis grandis, Boas. Fig. 7. Ventral aspect; magnified three diameters. Fig. 8. Dorsal aspect ; magnified three diameters. Fig. 9. Clionopsis krohni, Troschel. Larva, ventral aspect ; magnified thirty diameters. From an unpublished drawing by Dr. J. I). Macdonald. snger" Pteropoda Fia, I. '-y* Fig. 5 Fig : Fvcf. 6. 7r*\r*siy i >. Fia. 8. PLATE III. (ZOOL. CHALL. EXP. — PART LVIII. — 1887.) — Mmm. PLATE III. Fig. 1. Clionopsis krohni, Troschel. Adult, ventral aspect ; magnified three diameters Fig. 2. Clionopsis modesta, n. sp. Ventral aspect ; magnified eighteen diameters. Figs. 3, 4. Notobranchsea macdonaldi, Pelseneer. Fig. 3. Ventral aspect; magnified nine diameters. Fig. 4. Dorsal aspect ; magnified nine diameters. Figs. 5, 6. Notobranchsea inopinata, n. sp. Fig. 5. Ventral aspect; magnified eighteen diameters. Fig. 6. Dorsal aspect; magnified eighteen diameters. Figs. 7-9. Halopsyche gaudichaudi, Souleyet. Fig. 7. Adult, ventral aspect; magnified ten diameters. Fig. 8. Adult, dorsal aspect ; magnified ten diameters. Fig. 9. Larva, ventral aspect; magnified twenty-nine diameters. The Voyage of H ' ' illenger" ' P.Pelseneer ad.) } ':'