DISCOVERY REPORTS H, 0 / WOODS HOLE, MASS. VOLUME XXVII DISCOVERY REPORTS Issued by the National Institute of Oceanography VOLUME XXVII //r CAMBRIDGE AT THE UNIVERSITY PRESS 1956 PUBLISHED BY THE SYNDICS OF THE CAMBRIDGE UNIVERSITY PRESS London Office : Bentley House, N.w. i American Branch : New York Agents for Canada, India, and Pakistan: Macmillan Printed in Great Britain at the University Press, Cambridge (Brooke Crutchley, University Printer) CONTENTS SIPHONOPHORA OF THE INDIAN OCEAN. Together with systematic and biological notes on v related specimens from other oceans (published 28th April, 1954) By A. K. Totton Introduction Pa£e 7 Material and Methods .... Siphonophore Fauna of the Red Sea . Morphology and Relations of the Siphonophora List of Indian Ocean Species Classification Explanation of Terms and Abbreviations . Systematic and Biological Account . List of References Index to Genera and Species Index to Subjects of General Interest, briefly mentioned 8 11 H 15-16 25 31 33 i57 161 162 THE PELAGIC MOLLUSCA OF THE BENGUELA CURRENT. Part I. First Survey, R.R.S. 'William Scoresby', March 1950. With an account of the Reproductive System and Sexual Succession of Limacina Bulimoides (published 30th July, 1954) By J. E. Morton, Ph.D. Introduction PaSe l65 Acknowledgements I. Systematic Ecology-Depth Distribution II. Reproductive System and Sexual Succession of Limacina bulimoides Summary References 165 166 172 183 197 198 THE CIRCUMPOLAR CONTINUITY OF (published 30th July, 1954) By A. de C. Baker Introduction Material and Methods Phytoplankton Zooplankton . Conclusions . Summary Bibliography . ANTARCTIC PLANKTON SPECIES page 203 203 204 208 214 216 217 )890 vj CONTENTS THE PLANKTONIC DECAPOD CRUSTACEA AND STOMATOPODA OF THE BENGUELA CURRENT. Part I. First Survey, R.R.S. 'William Scoresby', March 1950 (published 8th November, 1954) By Marie V. Lebour, D.Sc. Introduction Page 221 Acknowledgements 221 Systematic 221 Literature 232 THE DISTRIBUTION OF SAGITTA GAZELLAE RITTER-ZAHONY (Published nth May, 1955) By P. M. David Introduction page 237 Acknowledgements 237 Methods 23^ Stages of Maturity 24J Synonymy 244 Racial Characteristics 2S4 Horizontal Distribution 26o Vertical Distribution 2°5 Growth Rate and Life History 27! Food 275 Summary 27° References 277 Appendix: Tables 18 and 19 27& CUMACEA OF THE BENGUELA CURRENT (published 6th July, 1955) By N. S. Jones, Ph.D. Introduction Systematic Account References page 281 . 282 290 THE WAX PLUG IN THE EXTERNAL AUDITORY MEATUS OF THE MYSTICETI (published 20th July, 1955) By P. E. Purves Introduction page 295 Location of the Wax Plug Structure and Composition Discussion Acknowledgements Bibliography . Plates XIV-XVIII 296 297 299 302 302 following page 302 CONTENTS ALEPISAUROID FISHES (published 29th July, 1955) By N. B. Marshall Introduction Acknowledgements Part I. Classification of the Iniomi Part II. Comparison of the Alepisauroid and Myctophoid Fishes Summary Bibliography Plate XIX page 305 3°5 306 322 334 334 following page 336 EUPHAUSIACEA OF THE BENGUELA CURRENT. First Survey, R.R.S. 'William Scoresby', March, 1950 (published 30th August, 1955) By Brian Boden Acknowledgements Introduction Distribution of the Euphausiacea in the Benguela Current Developmental Phases Bibliography page 339 • 339 • 34° • 346 • 375 CESTODES OF WHALES AND DOLPHINS FROM THE DISCOVERY COL- LECTIONS (published 17th October, 1955) By S. Markowski Material and Methods page 379 Abbreviations used in the Figures General Discussion List of Cestodes Examined Summary of Discussion Synonymy Systematic Notes Relationships between the Host and the Parasite References Plates XX, XXI 379 380 380 382 382 382 394 394 follozcing page 395 [Discovery Reports. Vol. XXVII, pp. 1-162, Plates I-XII, April 1954.] WOODS HOLE, MASS. SIPHONOPHORA OF THE INDIAN OCEAN TOGETHER WITH SYSTEMATIC AND BIOLOGICAL NOTES ON RELATED SPECIMENS FROM OTHER OCEANS By A. K. TOTTON British Museum (Natural History) CONTENTS Introduction PaSe 7 Material and Methods 8 Siphonophore Fauna of the Red Sea 1 1 Morphology and Relations of the Siphonophora 14 List of Indian Ocean Species iS-1^ Classification 25 Explanation of Terms and Abbreviations 3 J Systematic and Biological Account 33 Chondrophora 33 Siphonophora 3° Physonectae 3^ Calycophorae 73 List of References 157 Index to Genera and Species 161 falDEK TO SUDJEOTC OF GDMPIIAL In I'LIU!!)1! , U1UL1L1 -ftgNTIQNEn ■ 1-fo Diagram of a young Physonect Colony. Legend to Diagram of a young Physonect Colony A. The whole colony, showing the nectosome and the siphosome with five cormidia. A'. This shows that the siphosome has been deliberately twisted to bring the ventral meridian to which the buds are attached from the underside (opposite to that of the nectosome) over to the left for the convenience of the artist. B. Enlarged diagram of the nectosome to show (i) the budding zone of the nectophores, and (ii) how the muscular lamellae of the nectophores all arise in one meridian (dorsal) but bend alternately to left and right. C. The stem in the region of a blastostyle to show the budding of the male gonophores, a palpon and two of the bracteal lamellae. D. Enlarged diagram of the primary terminal gastrozooid. The pneumatophore, Pn., stem, St., and terminal gastrozooid, Gz.1, represent the original polypoid oozooid. The nectosome gives rise to asexual medusa-buds, and the siphosome to both polyp- and medusa-buds, so that the whole animal may be called a floating bud-colony, as suggested by Garstang (1946). The bracts, Br., are probably neither polyp- nor medusa-buds, but are possibly homologous either with larval actinuloid tentacles, or with larval or adult appendages which have arisen de novo. The nectosome with its apical organ — not a medusoid bud, but an air-secreting and air-releasing float (pneumatophore, Pn.) — carries only the propelling nectophores, N. (asexual medusae). The pneumatophore together with the budding-zone for nectophores, which in the larval stage originated on one side of it (Fig. 3), grows upwards (aborally) from the nodal point of minimum growth X. The siphosome grows downwards (orally) from the same nodal point X. It consists in the juvenile stage illustrated, of five cormidia or groups of buds, which originated in the larval stage on the side of the pneumatophore opposite to that which gave rise to the budding zone of the nectophores (Fig. 3). The first formed gastrozooid (protozooid) Gz.1, with its ring of palpons bearing simple tentacles (palpacles, Pal.) together with a reduced secondary gastrozooid, Gz.(rd.), which bears a less highly elaborate 'larval tentacle', forms the terminal part of the oldest cormidium, and is surrounded by a whorl of larger bracts of larval origin. Four successively younger cormidia can be seen between this and the siphosomal budding-zone or blastocrene, Blcr. Each cormidium consists of: (1) A distal gastrozooid with branched tentacle, surrounded by (2) A ring of bracts and palpons, the latter provided each with a simple tentacle and increasing in number with age. (3) Several ventro-lateral rows of bracts. (4) A number of intermediate palpons without palpacles, which increase in number with age. (5) A number of male medusae (gonophores, androphores), Go. 3, which become separated with age from their budding points (blastostyles), and lastly (6) A grape-like bunch of female medusae (gonophores, gynophores) Go. ?, which arises from the base of a gonopalpon at the proximal end of the cormidium. A tentative analysis of the arrangement of palpons and gonophores in successive cormidia suggests that there is a separate growth-gradient in each cormidium and that new palpons and male gonophores are budded off at several loci, characteristic for each species. The order of the development of successive cormidia is indicated in the diagram by numerals added to the symbols, the explanation of which is as follows : Symbol Explanation ^41,2,3,4 The first series 0f gastric palpons, associated with Gz.1, Gz.2, etc., the first palpons to appear. 5 1,2,3,4 Intermediate palpons, the second series of palpons to appear. (71,2,3 Intermediate palpons, 3rd series of palpons to appear. D.lt2'3 Intermediate palpons, 4th series of palpons to appear. E.1-2 Intermediate palpons, 5th series of palpons to appear. F.1,2 Intermediate palpons, 6th series of palpons to appear. G.1 Intermediate palpons, 7th series of palpons to appear. B.br. Bracteal bud. Bl.cr. Blastocrene (budding-zone of siphosome). Bl.st. Blastostyle (gonodendron), a bud from the oozooid from which the gonophores bud. Br. Bract. Br. lam. Bracteal lamella. C.br. Bracteal canal. 6 DISCOVERY REPORTS Symbol Explanation Gz.1 Protozooid (oldest gastrozooid or siphon), part of the oozooid. Gz? Second oldest gastrozooid (siphon), a bud from the oozooid. Gz.3 Third oldest gastrozooid (siphon). Gz* Fourth oldest gastrozooid (siphon). Gz.s Fifth oldest gastrozooid (siphon). Gz.(rd.) Reduced gastrozooid, a bud from the oozooid. Go.cJ Male gonophore (androphore), ? a bud from the oozooid, secondarily acquiring direct connexion with the oozooid. Go.$ Female gonophore (gynophore) ? a bud from the oozooid bearing medusoid-buds. N. Nectophore. N.B. Nectophore bud. Pal. Palpacle. Pn. Pneumatophore, apical organ of the oozooid. St. Stem, part of the oozooid. Tl. Tentacle, part of the gastrozooid. Tl.larv. Larval tentacle. X Nodal point of minimum growth. See also PI. IV. SIPHONOPHORA OF THE INDIAN OCEAN TOGETHER WITH SYSTEMATIC AND BIOLOGICAL NOTES ON RELATED SPECIMENS FROM OTHER OCEANS By A. K. Totton British Museum (Natural History) With a frontispiece, chart, 12 plates, and 83 text-figures INTRODUCTION The original object of this report, based primarily on the very rich collections made by R.R.S. 'Discovery II ', was to investigate the Siphonophore fauna of the Indian Ocean, which, by com- parison with that of the Atlantic and Pacific, was little known. It soon became apparent that, as Bigelow had long suspected, members of the group have very wide distributions, and that it was, indeed, doubtful if any species is either confined to or excluded from some part of the Indian Ocean, except those few species that are confined to antarctic waters. The chief result then of the study of the collections has been to advance our knowledge of the morphology of many of the species con- cerned; and accordingly use has been made, in the systematic account, of some new and critical specimens taken in other waters, not only by ships of the Discovery Committee, but by other expedition ships as well. Our knowledge of the Siphonophora does not increase as rapidly as it might, because the systematist can only, with few exceptions, obtain the material needed for study by laboriously picking it out of plankton hauls. Unfortunately man-power is insufficient for sorting. For this reason there are still some thousands of plankton net hauls among the Discovery Collections alone that have not been examined for Siphonophora, all of which are holoplanktonic animals ; no one can tell what treasures these jars of plankton may contain. One may search for a very long time before being lucky enough to find a really well-preserved specimen, or a larval or other much desired stage, of some particular species, especially if it is a deep-water one. Chuniphyes moserae may be cited as an example. The rarely seen Nectopyramis spinosa, N. [Archisoma] natans and the Rhodaliids, of which I have never found a specimen, are others. There is also the inherent difficulty of dealing with fragmentary specimens. Again, many of the species occur in two phases, since the stems of most Calycophores eventually break up into free-swimming groups of buds called eudoxids, the successive gonophores of which in their turn may break away and swim as free medusae. I have been lucky enough to secure, anaesthetize and preserve for comparison with the fragments secured by nets, fairly complete specimens of Forskalia edwardsii, Apolemia uvaria, Agalma elegans, Nanomia bijuga and Stephanomia rubra — all of them available at times at Villefranche, some in large numbers. The polygastric stages of Diphyiinae and most of their eudoxids, especially those which occur near the surface, as well as the polygastric and eudoxid stages of the Abylopsinae are now well known. The Prayidae are less well known; also the Physonectae, which are the most difficult to determine because they fall to pieces in the nets or in the bottles of plankton, and no one except Bigelow has had the patience to describe and figure accurately more than a few of the pieces. This is regrettable because the nectophores, bracts, gastrozooids and tentacles, siphons and gonophores, both individually 8 DISCOVERY REPORTS and in their arrangement on the oozooid, all have very definite characters. As a rule it is far more worth while to figure accurately the details of these buds than to attempt to give a sketchy figure or description of the whole animal. Indeed, it is a most exacting task to make a figure of a live Siphonophore such as a Forskalia or Apolemia. Most of them are transparent, extremely complex in shape, and very active in movement as a whole, and in expansion and contraction of the stem, tentacles and nectophores. In addition to these difficulties, specimens as a rule quickly become moribund in captivity. Much of the literature is full of names that only with grave doubt can be referred to zoological concepts. For many years now I have been occupied with the laborious task of picking out hundreds of thousands of fragmentary specimens from plankton samples and of trying to form some idea of how many species are represented, and then attempting to apply old specific names and, where none are available, giving new ones. Gradually specialists are building on the foundations laid by their illustrious predecessors and on their own early work, and are beginning to recognize the various species. The greatest stimulus to progress is given by field work, but unfortunately the opportunities afforded to me for this have been very limited. The best known and most easily identifiable Siphonophores are perhaps those of the Mediterranean, a region where so much of the pioneer work on the group was done, and a region where many are often easily accessible at the surface, as long as the sea is calm and sea-temperature does not exceed 2i° C. The least well known are those of the Indian Ocean, for which reason this report has been prepared with no little labour. Many species are difficult to take with tow-nets, except in fragments, and it is probable that manv species of a particular fauna of Siphonophores will always be missed by the nets. As an illustration of this fact, in intensive tow-netting campaigns from 1908 to 19 10 in the Mediterranean, the 'Thor' missed not only all the common Physonects like Agalma, ' Stephanomia ', Forskalia and Apolemia, but also such abundant species as Muggiaea kochii and Lensia subtilis. Another explanation for the absence from ' Thor ' catches of the last two species is that they may be neritic forms. MATERIAL AND METHODS Our knowledge of the Siphonophora of the Indian Ocean was summarized by E. T. Browne (1926) in his report on thirty-six species taken on the ' Sealark ' Expedition by Stanley Gardiner. ' Sealark ' ranged over the banks and slopes of the Chagos Archipelago, the Seychelles and the Mascarene Islands, in the tropical zone during the months May to October. The Siphonophores collected by 'Dana II' in the 1928-30 cruise are being worked out by Dr Mary Sears. Jesperson (1935) gives a summarized list of stations showing that a large number of plankton hauls were made in the Indian Ocean from the north point of Sumatra across to Ceylon and thence via the Seychelles to the north point of Madagascar and to Mombasa. Changing southwards, the route then passed through the Mozambique channel to Durban, thereafter following the coast-line down to Cape Town. The macroplankton on the line Ceylon-Madagascar-Mombasa appears to have been much richer than that on the stretch through the Mozambique channel down to Cape Town. Dr Sears's ' revision of the Abylinae ', published since this report went to press, contains records of Indian Ocean specimens of Abyla bicarinata Moser, A. brownia Sears (sp.n.), A. carina Haeckel [=A. trigona Q. & G.], A. haeckeli L. & van R., A. ingeborgae Sears (sp.n.), A. schmidti Sears (sp.n.), A. trigona Q. & G., Ceratocymba dentata Bigelow and C. leuckartii Huxley. From September 1933 to May 1934, the John Murray Expedition with the ' Mabahiss ' was working principally in the open West Tropical Indian Ocean, in the Somali and Arabian Basins, and brought back thirty-nine species of Siphonophores. MATERIAL AND METHODS 9 Since the return of the ' Mabahiss ', our knowledge of the Indian Ocean siphonophore fauna has been supplemented by a study of eight more collections. The first was a rich collection made by R.R.S. 'Discovery II' which, in 1935, ran a line of stations up from Marion Island to the Gulf of Aden. From the plankton taken at these 'Discovery' Stations I have picked out and identified about 14,000 specimens. They fall into some 170 categories which represent the parts of polygastric and eudoxid colonies of seventy-six species. They were taken in forty-five tow-net hauls (Stations 1566-89), fished at various depths between the surface and 1900 m., all being closing-net hauls. These 'Discovery' Stations provide records of twenty-one known Siphonophores not previously recorded from the Indian Ocean, and not taken by the 'Mabahiss'. In addition a number of new species were taken at these ' Discovery ' Stations : three of these have been known to me for a long time, though descrip- tions have not yet been published, and some others are new to me. There are a few other novelties which will not be mentioned until more is known about them. 'Discovery's' N70V series of nets in particular gave valuable results, and I have separated out no less than 11,000 Siphonophores, including larvae, from these catches. The next collection was made in the Red Sea in 1935-6 by Cdr. J. H. Bowen, R.N., in H.M.S. ' Weston '. Another line of stations across the Indian Ocean was made in 1936 by R.R.S. 'Discovery II' in lat. 320 S., at which thirty-five species of Siphonophora were taken in twenty-nine closing-nets. A fourth collection was made between November 1948 and February 1949 in the Gulf of Aden by Mr A. Fraser-Brunner, while investigating the fisheries of the Aden Protectorate. Off Aden he fished a metre net both by day and night, by anchoring the boat and streaming the net out with the tide, allowing it to fish for half an hour and hauling it. Off Bulhar, Mukalla and Alayu he also fished the same net for half an hour just below the surface at a towing speed of about three knots (see Chart, p. 10). 'Discovery II' made four stations, Nos. 2679-82, in the Gulf of Aden in 1950. I have examined the oblique haul from 200 m. to the surface made at Station 2681. In 195 1 she made another station there, No. 2900. A fifth collection of 925 specimens was made in the Gulf of Aqaba by the British Museum (Nat. Hist.) Expedition in M.Y. 'Manihine'. Thirty-two stations were made between 31 December 1948 and 3 February 1949. Thirteen species of Siphonophora in all were present at thirty of the stations. The fact that not many species were present at any one station in the Gulf of Aqaba has had a useful systematic result, in that an Agalmid larva could be almost certainly identified as that of the species Agalma okenii; the eudoxid of Chelophyes contorta could be definitely identified for the first time owing to the absence of C. appendiculata ; and the eudoxid of Lensia hotspur and its posterior necto- phore could be identified, because, apart from L. subtilis whose eudoxid is known, it was the only other common species of Lensia present at the depth explored. The surface temperatures (21-220 C.) and salinities (40-6-40-8 %0) were high, and no doubt the Siphonophores were at the lower limit of the vertical hauls rather than at the surface. Unfortunately, hauls could not be made from depths below 180 m. or more specimens might have been taken. A sixth collection was made in the Red Sea off Port Sudan by M.Y. 'Manihine' in the winter of 1 950- 1. Three thousand eight hundred and thirty-seven specimens were removed from ten catches and identified, providing almost the first records of Siphonophores from that area. A seventh collection, consisting of six bottles of plankton, was made in November 1951 with oblique hauls from 200 m. or so to the surface in the Red Sea by 'Discovery II ' and was received in January 1952. This collection adds no species that were not taken there earlier by H.M.S. 'Weston' and by M.Y. 'Manihine', and in fact does not contain specimens of seven species taken by them, but it gives valuable corroborative evidence about the fauna of the area. Moie important, it has enabled 10 DISCOVERY REPORTS me to conclude that Sulculeolaria quadridentata is identical with S. quadrivalvis (see Systematic Notes, p. 109). Finally, I have picked out thirty-three species of Siphonophores from ten of the stations made in 1950-1 by ' Discovery II ' along a line in longitude 900 E. The finding at Station 2895 in surface waters of Sulculeolaria biloba led to a long-needed revision of the old genus Galetta and its submergence in Sulculeolaria. It is comparatively easy to preserve Siphonophora in good condition for morphological work, if not for exhibition. They can be engulfed at the surface of a calm sea in a bottle, or picked out with a dip- net, and they should be isolated unless very small. Two or three or more live specimens of a species such as Rosacea cymbiformis, if kept in ajar, soon become inextricably entangled. The use of an isotonic Chart. Stations in the Gulf of Aden occupied by 'Mabahiss', 'Discovery II' and Mr Fraser-Brunner. solution of magnesium chloride makes it possible to achieve a good deal of relaxation of muscles and consequent expansion before killing in formalin, which should contain at least 5 % formaldehyde. It is advisable to add 2 lb. of hexamine (hexamethylenetetramine) to each gallon of 38-40 % formalde- hyde solution to buffer it. Before subsequent examination it is a good plan to stain for a day or two with dilute Grenacher's Borax-carmine. A little of this stain can be added to the jar in which the specimens are preserved. For examination of the external form of nectophores and bracts, brief staining in dilute Delafield's haematoxylin is effective, but if hexamine has been used the specimens must first be thoroughly washed or the stain will be precipitated. To photograph them alive is exceedingly difficult because of the lack of contrast, the depth of the field that it is necessary to cover, and their constant change of form and position. The text-figures, with the exception of a few diagrammatic ones, were all drawn by the author with the aid of a camera lucida. The inking in of the figures was done by Mrs Walker (ne'e E. C. Humphreys) after demonstration and examination of the specimens. A subsequent check with the specimens was made by myself, or by both of us. In order to be able to make drawings of large specimens I removed the foot of a Leitz binocular and made the stage slide on a track, fixed at such a height above the MATERIAL AND METHODS n bench that the drawing could be moved about underneath it. Thus by movements of the specimen on the stage, of the stage on the track, and of the drawing, it was possible to make large drawings, 1 8 in. or more across. Preserved specimens were stained lightly with Delafield's haematoxyl n and photographed in formalin. To obtain the view of the ostium of the nectophore of Abyla tottoni Sears (PL IX, fig. 4) the specimen was held in a vertical glass tube in a vessel of formalin. As well as light from the top, some was reflected upwards through the specimen. Maresearsia praeclara (Pis. VI and VII) was photographed with a Leitz 400 mm. Milar lens. The photomicrographs 4 and 5 of PI. IV were made by Mr M. G. Sawyers, principal photographer, with a 3 in. Holos lens, and the others (unless otherwise stated) were made by Mr J. V. Brown of the Museum Staff, with a Leitz Sumnar 12 cm. lens. Acknowledgements. I am grateful to Dr W. J. Rees for having read the draft of this report, and to Dr Helene Bargmann of the National Institute of Oceanography for a great deal of help. SIPHONOPHORE FAUNA OF THE RED SEA This report gives, for the first time,1 a record of the Siphonophores which live in the Red Sea and in the Gulf of Aqaba. These are areas about whose Siphonophores we previously knew practically nothing. It is therefore satisfactory to be able to include in this work on the species from the Indian Ocean, material collected from these two localities. This fauna, to judge from the material so far available, does not include the following species found elsewhere in the Indian Ocean: Porpita umbella Chnniphyes multidentata A. leuckartii (?) Physalia physalis Clausophyes ovata Chelophyes appendiculata Physophora hydrostatica Sulculeolaria monoica Eudoxoides mitra Rosacea cymbiformis [Galetta] S. biloba E. spiralis R. plicata [G.] S. turgida Lensia achilles Hippopodius hippopus Dimophyes arctica L. campanella Vogtia glabra Diphyes bojani L. conoidea V. pentacantha Abyla haeckeli L. cossack V. serrata A. trigona (sp. aff.) L. multicristata Bigelow & Sears (1937) showed that there was a paucity of siphonophore species in the Medi- terranean as compared with the Atlantic beyond, and noted that two of the reasons for this probably were (a) the excluding action of the deep outflow over the sill upon species which lived below that level, (b) the rather high minimum temperature (12-130 C.) of the Mediterranean deep water. There appears to be an even greater paucity of siphonophore species in the Red Sea as compared with the Indian Ocean outside. Here again the deep outflow over the sill at the southern entrance of the Red Sea probably has an excluding action on deeper-water species, and the even higher (21-5-22° C.) minimum temperature of the Red Sea deep water would be unbearable for such species, even if they entered. Sverdrup, Johnson & Fleming (1946), summarizing the work of Thompson and others on the water masses of the Indian Ocean, say that the entire basin of the Red Sea below sill depth (100 m.) is filled in winter with water which has a salinity of from 40-5 to 4i-o%0, and a temperature between 21-5° and 22° C. : and that it has a very low oxygen content, less than 1 ml./L. in summer, a little more than 2 ml./L. at the end of the winter. The maximum surface temperature in summer is about 30° C. and 1 Schneider (1898, p. 120) mentioned that Steindachner, scientific leader of the 'Pola' Expedition had handed over to him two Siphonophores from the Red Sea, now known as Agalma okenii and (possibly) Melophysa melo. 12 DISCOVERY REPORTS the minimum (February) about i8° C. A species like Stephanomia rubra, which leaves the surface waters of the Mediterranean when the temperature rises above about 21 ° C. and descends into deeper cooler water, appears to be living in the Red Sea in an environment that it can only just manage to tolerate. It is therefore interesting to find that the new records given in Table 1, show at least two dozen species of Siphonophores from the Red Sea. Table 1 . List of Chondrophora and Siphonophora taken in the Red Sea. (New records) R.R.S. 'Discovery II', 1951, St. 2905, N 70 B R.R.S. 'Discovery II', 1951, St. 2905, TYF B H.M.S. 'Weston', 1935-6, Kamaran Islands, Apr. 1936 M.Y. 'Manihine', 1950-1, Port Sudan, Jan. 1951 H.M.S. 'Weston', 1935-6, Port Sudan, Nov.-Dec. 1935 R.R.S. 'Discovery II', 195 1, St. 2906, N 70 B R.R.S. 'Discovery II', 1951, St. 2906, TYF B R.R.S. 'Discovery II', 1951, St. 2907, N 70 B R.R.S. 'Discovery II', 1951, St. 2907, TYF B M.Y. 'Manihine', 1950-1, Koseir and Jaffatin Islands, Nov. 1950 M.Y. 'Manihine', 1948-9, Gulf of Aqaba, Jan. 1949 H.M.S. 'Weston', 1935-6, Safajin Islands, 260 40' N. 340 0' E., Jan. 1936 H.M.S. 'Weston', 1935-6, Aqaba, Feb. 1936 Winter or summer South, central or north part of Red Sea W s W S W/S s W C W C W C W C W N W N W N W N W N W N L P1 X1 Velella velella Forskalia sp. or spp. Nanomia bijuga Stephanomia rubra Agalma elegans A. okenii Athorybia rosacea Amphicaryon sp. Cordagalma cordiformis Sulculeolaria chuni S. quadrivalvis Diphyes dispar D. chamissonis Abylopsis eschscholtzii A. tetragona Bassia bassensis Enneagonum hyalinum Lensia campanella L. fozvleri L. hotspur L. tneteori L. subtilis L. subtiloides Chelophyes contorta Muggiaea atlantica Sphaeronectes sp. X X X X X 2 2 X X X X X X xc 2 X 2C X X X X 2C 2 2C X xc X X X X X 2C 2c 2C 2° P1 p X P1 P1 xc 2 X 2 E P1 X P1 X X X X X X 2 E X 2 X X X X X X 2 X P1 2 X X X X 2 2 2 X 2 2 P2 X X X X X p p X X X X X L X X1 xc X X 2 P 2 P 2C 2 X Explanation of symbols : x = present, stage and numbers not recorded. — = absent. P=polygastric stage. E = eudoxid stage. x \ P\ P2, P3 = number of specimens. c = common. 2 = both stages. L = larval stage. The almost cosmopolitan Chelophyes appendiculata has not, so far, been taken there, though its close ally C. contorta is common. The evidence for the occurrence of Agalma elegans depends on the SIPHONOPHORE FAUNA OF RED SEA i3 identification of young nectophores and a few larval bracts, and needs corroboration. The rather uncommon, associated pair Diphyes chamissonis and Lensia subtiloides are abundant. Comparison of the records from the northern, central and southern areas of the Red Sea, for what they are worth, show that there are eleven species in the southern, twenty-one in the central and seventeen in the northern. All the records are winter ones. Red Sea records by the ' Manihine ' for the ten following species show that they were living at the upper limit of their known temperature range : Diphyes dispar, D. bojani, D. chamissonis, Lensia subtiloides, L. subtilis, Chelophyes appendiculata, Eudoxoides mitra, Abylopsis tetragona, A. eschscholtzii, Bassia bassensis. SIPHONOPHORES FROM THE GULF OF ADEN From the Gulf of Aden, where forty-five 'Mabahiss ' Stations were made, five hauls were made with plankton nets but only nine Siphonophores were captured. However, Mr Fraser-Brunner, who made some very rapid oblique hauls off Aden, Mukalla, Alayu and Bulhar on 10 and 26 November 1948, on 5 and 6 December 1948, and on 23 February 1949, respectively, showed that the water was teeming with Siphonophores. He took more than 9000 specimens, yet Vogtia glabra was caught by the ' Mabahiss ' and missed by Fraser-Brunner, probably because it is not a surface species. One station, No. 2681, was also made by R.R.S. 'Discovery II ' in 1950 in the Gulf of Aden. Table 2. Siphonophore colonies from the Gulf of Aden. Analysis of catches made by 'Mabahiss' (1933-4) and by Fraser-Brunner (1948-9) Fraser- Brunner, Alayu 26. xi. 48 \ m. net 0 m. Night 'Mabahiss', St. 186 5- v- 34 2 m. net 952-0 m. Day Fraser- Brunner, Mukalla 10. xi. 48 1 m. net 0 m. Day 'Mabahiss', St. 18 21. ix. 33 2 m. net 900-0 m. Day Fraser- Brunner, Aden 6. xii. 48 1 m. net 0 m. Day Fraser- Brunner, Aden 5. xii. 48 1 m. net 0 m. Night Fraser- Brunner, Bulhar 23. ii. 49 1 m. net 0 m. Day 'Mabahiss' St. 14 J9- ix- 33 Agassiz trawl 988 m. Nanomia bijuga Agalma oketiii Vogtia glabra Sulculeolaria chuni Sulculeolaria sp. indet. S. quadrivalvis Diphyes dispar D. bojani D. chamissonis Lensia subtiloides Muggiaea atlantica Chelophyes appendiculata Ch. contorta Eudoxoides mitra Abylopsis tetragona A. eschscholtzii Bassia bassensis Enneagonum hyalinum Abyla trigona sp. aff. ) i P 27 12 26 14 99 3 3 2 20 < 1 1 7 E 188 2 4 3 4 3 33 3 P 1 [ E P 506 226 0 1024 E 748 i55 10 4820 P 1 E X P 36 7 59 15 0 10 IS K X X [ E 6 1 5 2 ? 4 3 0 47 X P 20 2 16 6 402 46 3 1 260 X X 2 E 16 1 61 0 ? 70 5 1 0 888 P 5 3 88 10 22 31 X E 2 1 18 27 5 P 3 E P = polygastric stage. £ = eudoxid stage. H DISCOVERY REPORTS SALINITY TOLERANCE Siphonophores evidently have a wide range of salinity tolerance as well as of temperature. Although Bigelow (191 1 b) found that in the Eastern Tropical Pacific they are an almost negligible factor in the plankton from waters with a salinity less than 35 %0, and that they are entirely absent where the salinity is below about 30%0, in the Gulf of Aqaba with surface salinities as high as 40-8 %0 (Deacon, 1952), the records are far from scanty. SIPHONOPHORES FROM THE INDIAN OCEAN Table 3 gives a list of species from the Indian Ocean. It will be seen from the brief summary of the broad hydrographic conditions in these waters, given below, that in these regions they are living in surroundings very different from those obtaining in the Red Sea and the Gulf of Aqaba. The penetration of water of high salinity from the Red Sea can be detected in the Indian Ocean at depths of about 500 m. in latitude 8° N. to depths of 1250 m. in latitude 200 S. According to Sverdrup et al. (1946), the southern limit of the Indian Ocean is marked by the sub-tropical convergence in approximately latitude 400 S. To the north it is closed by the land-mass of Asia. The temperatures of the surface layers are uniformly high during the greater part of the year, varying between 25-290 C. In August the south-west monsoon causes upwelling of cold water along the East African and south-east Arabian coasts as far south as the Equator, and in February the north-west monsoon brings about the same phenomenon in the Bay of Bengal. Consequently, lower surface temperatures are met with in these regions in these months. Salinity values also show an annual variation because of these influences, but otherwise have a sub-tropical maximum of 35'4%o. Below the surface-layers are three main water masses : (1) Indian Ocean Central water, probably formed at the subtropical convergence by sinking, and having temperatures of 8-150 C. and salinities of 34-6-3 5 -5 %0. (2) Indian Ocean Equatorial water formed by sub-surface mixing, and having temperatures of 4-170 C. and salinities of 34-9-35,25%0. (3) Deep-water mass below depths of 2000 m. approximately, formed partly by antarctic inter- mediate water (temperature 2-2° C. and salinity 33'8%0) and partly by antarctic bottom water (tem- perature 2-0° C. and salinity 34-8 %0). Their mixing produces deep water in the Indian Ocean of relatively high salinity 3476 %0 and temperatures about 2-5° C. MORPHOLOGY AND RELATIONS OF SIPHONOPHORA An outstanding event in the history of siphonophore literature was the publication in 1946 in the Quarterly Journal of Microscopical Science by Walter Garstang of his ' Morphology and Relations of the Siphonophora '. It is indeed, as John Colman has said, an admirable exercise in classical comparative morphology.1 Try as one will it is difficult to escape from Garstang's main conclusions. ■Written in less than two years as the last of his great papers, it has its roots deep in his personal history. He told me that he could never have tackled the larval development of the big groups Tunicata, Mollusca, Crustacea and Siphonophora had it not been his duty in the early years of the Plymouth Laboratory to make himself familiar with all branches of the Plymouth fauna, both as regards structure and habitat, an experience which taught him to turn easily from one group to another. No doubt Garstang was not familiar with more than a few Siphonophores, alive or preserved, and the paper itself is really a masterly review of the literature. It is very satisfactory for one who has spent more than twenty years making an intensive and fresh study of the Siphonophora, both preserved and alive, to find himself so much in agreement with Garstang's main conclusions. 1 Personal communication. MORPHOLOGY AND RELATIONS OF SIPHONOPHORA IS Table 3. List of species identified by the author in the Indian Ocean - 0- e 0 -a u -a >-. 0 0 cd u Pk c CU -a < O (4-1 O Red Sea, 'Discovery II'; 'Manihine' c 'JB 'a a § a Xi a cr < O O CHONDROPHORA Porpita umbella 33 X X — — X — — — Velella velella 34 ? — — X — — X — SIPHONOPHORA: PHYSO- NECTAE A. Cystonectae Physalia physalis X X — — X — — — B. Physonectae Athorvbia rosacea 38 X X X Melophysa melo 4i ? X Physophora hydrostatica 42 X X — — X — — — Stephanomia rubra 47 0 X X — X — X — Nanomia bijuga 52 0 X — — — X X — Marrus orthocannoides gen.n., sp.n. 59 0 X Agalma elegans 61 X X X A. okenii 64 X X — X X X X X Cordagalma cordiformis 69 0 X — — — — X X Bargmannia elongata gen.n., sp.n. 69 0 X Forskalia sp. 73 X X — — — — X X SIPHONOPHORA: CALYCO- PHORAE Sphaeronectidae Sphaeronectes sp. X X — — — — X — Prayidae Praia ? dubia 0 X Nectopyramis thetis 78 0 X X N. natans 82 0 X — X — — — — N. diomedeae 83 0 X X X — — — — Rosacea cymbiformis 89 ? X — — X — — — R. plicata 89 0 X — — X — — — Amphicaryon acaule 92 X X X — — — X ? A. ernesli sp.n. 94 0 X — — — — X — A. pelt if era 96 0 X Maresearsia praeclara gen.n., sp.n. 97 0 X HlPPOPODIIDAE Hippopodius hippopus 99 X X X X X — — — Vogtia glabra 100 0 X X X X X — — V. pentacantha X X X X X — — — V. spinosa 0 X — X — — — — V. serrata X X [6 Table 3 (continue i) • B 0 -a V T3 u 0 CJ -. u > 0 CJ cfl Q - u - c a -g «l 0 > CD g IS ■3 s a XI c« a" < 0 <^H 0 SIPHONOPHORA: CALYCO- PHORAE (continued) Diphyidae: Sulculeolariinae Sulculeolaria biloba 104 O X X — X X — — S. turgida 107 O X X — ? — — — S. chuni X X X — X X X X S. angusta sp.n. 108 0 X ■S. quadrivalvis 109 X X X X X X X X S. monoica X X X X X — — — DlPHYINAE hernia achilles 0 X X — X — — — L. ajax X X L. campaneUa X X — — X — X — L. conoidea 114 0 X X X X — — — L. cossack 112 X X X X X — — — L. exeter x x X L. fowleri X X X X — — X X L. grimaldii 0 X X L. havock X X X L. hostile X X — X — — — — L. hotspur 110 X X X — X — X X L. hunter X X X L. lelouveteau 0 X L. meteori II7 0 X X — — — X X L. multicristata "3 X X X X X — — — L. reticulata sp.n. 118 0 X L. subtilis 114 X X X — X — X X L. subtiloides 112 0 X — — X X X — Eudoxia macra sp.n. 118 0 X X Muggiaea atlantica 120 0 — — — — X X — Dimophyes arctica 123 X X X Chelophyes appendiculata 127 X X — X X X — — Ch. contorta 130 X X X — X X X X Eudoxoides mitra X X X X X X — — E. spiralis X X X X X — — — Diphyes dispar X X X X X X X — D. bojani X X X — X X — — D. chamissonis 130 X X — — X X X — Chuniphyinae Chuniphyes multidentata J31 0 X X X X — — — Ch. moserae sp.n. 131 0 X X Clausophyes ovata J33 0 X X X X — — — Crystallophyes amygdalina 134 0 X — ' Heteropyramis macidata 137 0 X X Thalassophyes crystallina 141 0 X X Abylidae: Abylopsinae Abylopsis tetragona 155 X X X X X X X X A. eschscholtzii X X X X X X — — Enneagonum hyalinum X X X — X X X — Bassia bassensis X X — X X X X — Abylinae Abyla schmidti 147 X X X X X X — — A. haeckeli 151 X X X X X — — — Ceratocymba sagittata X X X X — — — — C. leuckartii X X X — X MORPHOLOGY AND RELATIONS OF SIPHONOPHORA 17 GARSTANG ON PHYLOGENY Reduced to its elements Garstang's paper consists of some corrections of interpretation which permit a more satisfactory alignment of the orders, and support the idea that the pelagic habit of the Siphonophores results from a prolongation of a free-swimming larval phase, coupled with precocious development and sundry adaptations of the ancestral characters, including budding. Garstang attractively rounded off his critical discussion of siphonophore phylogeny by suggesting that a myriotheline relationship of Siphonanths is indicated by the resemblance of their lateral, paddling (sic) bracts to the larval tentacles of the actinula of the hydroid Myriothela, which are aboral in position, locomotive in function and precociously developed. But he thought that this relationship of Siphonanths to Myriothela had to be reconciled with the corymorphine affinities of Disconanths before the common origin which he suspected could be firmly assured. He suggested that Pelagohydra, Siphonanths and Disconanths may in fact be independent pelagic offshoots from three different types of gymnoblastic hydroid, the larvae of which failed to attach themselves but succeeded in keeping afloat until their gonophores ripened. Pelagohydra absolutely, and Disconanths to some extent, may be said to have survived, he continued, by adaptive aggrandizement of their oozooids at the expense of their colonies, while Siphonanths, by precocious budding, elaborated the colony at the expense of the oozooid. Rather surprisingly, Garstang did not mention another floating tubularid hydroid, Marge- lopsis haeckeli; and had he been familiar with the little-known Siphonophore, Apolemia uvaria, I feel sure that he would have commented on the fact that the unique feature it presents, namely the presence of tufts of tentacles at various points on the nectosome, recalls the arrangement in Pelagohydra.1 RELATIONSHIPS BETWEEN SIPHONOPHORA AND HYDROIDA Having noted Garstang's conclusions about the phylogeny of Siphonophora and their relationship with corymorphine and myriotheline hydroids, let us now take a brief look at this group of hydroids and the rest of the Hydroida (Anthomedusae) Capitata. Whenever these hydroids bud off medusae they belong to the old Haeckelian family Codonidae, considered by medusologists to be a simple and primitive group of Anthomedusae. The Capitata include hydroids which have two contrasting habits. In one the oozooid secretes a permanent tube, simple or branched, attached to the substratum at its basal end by a branching stolon. In the other we find either planktonic forms, or insecurely attached solitary forms anchored by hydrocauline processes to the sand or mud. It is perhaps natural to ask which of these two contrasting types appears to be primitive and which derivative. Kramp (1949) gave a most interesting review of the corymorphine hydroids, including several of the less sedentary forms of which very little is known. He said that actinula larvae are unknown in this family, but both Hartlaub (1907) and Rees (1937) have shown that actinuloid larvae or polyp buds are indeed known. Kramp, who, in his important paper, nowhere mentioned either Garstang's (1946) work nor that of Gronberg (1898) on the canals of Tubularia, presented an outline of the course which he thinks has been taken in the phylogeny of a Tubularia line of hydroids and a Corymorpha line, and he derived the solitary, anchored forms like Corymorpha from those with fixed stolons like the Corynidae. From general considerations this seems to be most improbable, but I shall not attempt myself to speculate on the detailed phylogenies of genera of hydroids within the Capitata. Garstang's argument for the derivation of the sedentary forms like Tubularia and Pennaria with fixed stolons, from the free forms like Corymorpha, was based on good evidence from comparative morphology. I have re-examined the type specimen of the latter and have made efforts to find detached medusae referable to it amongst plankton hauls from the New Zealand region, but without success. I was interested to hear recently from Dr Elizabeth J. Batham that Canterbury Museum, Christchurch, New Zealand possessed a second specimen of Pelagohydra. D XXVII , 18 DISCOVERY REPORTS Kramp does not appear to have produced arguments of similar worth for concluding that evolution has proceeded in the reverse direction, nor did he anywhere refer to Garstang's work, as I have already mentioned. ACTINULOID ANCESTORS The question whether it is theoretically impossible for the ancestral Metazoa to have been pelagic is of critical importance. A morphological character possessed by so many of the Corymorphidae is the presence on various parts of the stem of papillae and root-processes, which bear suggestive signs of homology with the larval tentacles of actinuloid animals. What are we to conclude from the fact that so many groups of Hydrozoa have actinuloid larvae — Trachymedusae, Narcomedusae, Chondrophora, (?) Siphono- phora, Margelopsidae (Margelopsis haeckeli, Pelagohydra, Climacocodon), Corymorphidae, Myriothe- lidae, Tubulariidae, and Corynidae and allies? It must mean that the common ancestor of all these animals had an actinuloid larva. Garstang appears to have believed that, as in the Tunicata, this larva was a purely locomotive one, that it was followed by fixation, and then by an adult stage of feeding and growth, with budding last of all, but that in some forms fixation was postponed and finally abandoned altogether. I should like to think that the adult ancestor had always been planktonic, like the ancestral larva, and that fixation did not take place until much later in phylogeny. According to Carter (1949) who briefly reviewed Hadzi's (1944) Turbellarian theory of the Cnidaria, Hadzi believes that ' only the smaller Protista can float without either an elaborate structure adapted to enable them to do so or considerable expenditure of energy. We must think, then ', he says, ' of the ancestral metazoan as a bottom-living form, and not as a floating spherical colony. ... It would develop an intermediate cell-layer, a parenchyma, between the ectoderm and the endoderm, to give solidity and support to the body.' But if animals like the trachymedusan Liriope, whose youngest stages show neither elaborate structure, except a mesogloea, nor appear to be obliged or able to expend con- siderable energy, can live at the surface, it would seem that this argument is not valid. The Trachylina are thought to be the most primitive Cnidaria. It is true that most Trachymedusae are well-marked bathypelagic forms, as are the Narcomedusae from northern latitudes, but in the warm seas the latter suborder are found in the surface layers. POLYP AND MEDUSA The argument for thinking of the medusa as the original adult form of Cnidaria rather than as a dispersal device was stated by Brooks (1886) following Bohm (1878) and Claus (1878). He advanced a theory that the polypoid generation in Hydromedusae is a persistent larval form, and that the adult ancestral hydrozoan was some sort of medusoid, perhaps (see Brooks, 1886, pi. 41, figs. 6-7) rather like the earliest medusa-stage of development of the trachymedusan Liriope, the adult stage of which has made only a slight advance beyond its larval polypoid stage (Brooks, fig. 8). The ancestral poly- poids, as suggested by Brooks and supported by Libbie Hyman, developed the habit of budding before some of them took up a prolonged attached existence as polypoid oozooids or colonies. After this, instead of becoming metamorphosed directly into a medusoid, the actinuloid produced the medusoid by budding. Libbie Hyman (1940) pointed out that this view harmonizes with the develop- ment of Siphonophores, where the planula develops into the polypoid and buds off the medusoid. LIBBIE HYMAN ON PHYLOGENY Libbie Hyman, though repeating some outmoded theories about the interpretation of the morphology of Siphonophora, has given a more satisfactory account (pp. 636-40) of the probable phylogeny of the various groups of Cnidaria and Ctenophora than is found elsewhere. I find it hard to believe in MORPHOLOGY AND RELATIONS OF SIPHONOPHORA 19 that part of Hadzi's (1944) Turbellarian theory of Cnidaria which derives Anthozoa from Turbellaria. But then I have only a text-book knowledge of the Turbellaria, and if text-books on Cnidaria are any criterion, they can be grossly misleading. SOMATOCYSTS AND BRACTS Garstang (p. 143) boldly revived Schneider's contention that the Calycophoran somatocyst is homo- logous with the cavity of the cap-shaped bract of a larval Physonect. But whereas, as Garstang says, Schneider, Woltereck and Moser all realized this, they thought that the Calycophorae were primitive, and that in the Physonect (1) the primitive Calycophoran bract had secondarily become separated from the nectophore, (2) this separated nectophore had become the pneumatophore, and (3) the separated somatocyst had become the Physonect bract. The true significance of the homology according to Garstang is this : the Physonects are primitive and the independent primary bract of the Physonect (e.g. the larval bract of Physophora) has become secondarily fused with the precocious nectophore in the Calycophorae; and this shows that originally Calycophorae as well as Physonectae possessed larval, locomotive bracts, reminiscent of the larval tentacles of hydroid actinulae. So Garstang put forward, and himself adopted what he called the illuminating view that the original larva of the Siphonophore ancestors was not a planula but an actinula, in which, before the typical actinuloid characters could be developed, precocious budding became established. It is quite impossible to be sure that the somatocyst of Calycophorae represents a larval actinuloid tentacle at all and, in addition, that it represents a larval tentacle of the ancestor of a particular group of gymnoblast hydroids within the Capitata. In ontogeny it always arises, in my view, as an integral part of the precocious larval nectophore-bud, and may in reality have no more connexion with the larva than that. It may, in fact, be an oil-storing organ developed late in phylogeny which now appears early in ontogeny. I think that Garstang was probably right in believing that bracts were developed from actinuloid tentacles, though another possible explanation is that larval bracts are precociously developed adult bracts, developed late in phylogeny. If a close study is made of the bracts of Athorybia and Melophysa, and they are compared with the larval bracts of Agalma elegans, and particularly of the larval and young stem bracts of A. okentt, it will be seen that the stem bracts of Physonectae in general are probably all homologous with larval bracts (see Text-figs. 8, 9 and 26). ORIGIN OF NECTOPHORES One of the most difficult questions to answer about siphonophore phylogeny has always been how nectophores were evolved. Judging by their simple nematocysts and tentacles, lack of nectophores, and structure of the pneumatophore, Cystonects are the most primitive group of Siphonophora. Long ago, Kawamura (1910) in an inaccessible * Japanese work, suggested a way in which the three-branched tentilla of Rhizophysa filiformis might have evolved from a simple unbranched tentacle. The process can be watched in the course of ontogeny. The great puzzle was how, if Cystonects are primitive, did the nectophores of Physonects originate ? Garstang linked his explanation that they represented pre-siphonophore (i.e. hydroid) gonophores with an argument that does not appear to be very sound, namely that hydroid gonophores were originally produced on the hydrorhiza of the ancestral hydroids, and that only later did they invade the circumoral region via the hydrocaulus with which he homo- logizes the siphonophore stem. Before proceeding with the discussion of Garstang's theory of the origin of nectophores it should be noted that Garstang made it somewhat difficult for a non-specialist to follow some of his arguments 1 Through the good offices of Dr Mary Sears most of Kawamura's work on the Siphonophora has been translated from Japanese into English by Rodney Notomi. 3-2 20 DISCOVERY REPORTS by using what seems to be a faulty nomenclature — one that leads to a confusion of identities. He said much about two not very well-known Physonects which, like Bigelow, he called ' Anthophysa' and ' Athorybia '. The proper names of these two forms are Athorybia Forskal (for Garstang's ' Anthophysa ') and Melophysa (Haeckel) (for Garstang's 'Athorybia '). The synonymies are explained in my systematic notes (p. 37). It is most unfortunate that these corrections must be made, since Garstang's paper will become a classic. But in places Garstang uses both names Athorybia and ' Anthophysa ' for one and the same animal, Athorybia rosacea Forskal: at other times he uses one of these names 'Athorybia' for a distinct species Melophysa melo (Haeckel). The confusion in siphonophore synonymy is notorious. According to Garstang's theory the rudimentary medusoid buds, discovered by Chun in ' Athorybia ' {^Melophysa Haeckel) and hitherto regarded as vestigial nectocalyces, are (bearing Pelagohydra in mind) tentatively claimed as obsolescent hydroid gonophores ; in other families they are said to have been converted into nectocalyces; and in 'Anthophysa' ( = Athorybia Forskal) they are thought to have disappeared altogether. I have now collected and examined quite a number of these Melophysa nectophores (see Text-fig. 7). They may be more primitive in many ways than other nectophores, but show no obvious sign of being obsolescent gonophores, nor of being vestigial. In fact ' Anthophysa ' ( = Athorybia Forskal) may never have possessed nectophores at all, being descended from an 'Athorybia larva '-like ancestor comparable with the stage in the ontogeny of an 'Athorybia larva'1 before nectophores arise. In that case Athorybia would be neotenous. Garstang's 'Athorybia' ( = Melophysa Haeckel) is in my opinion another neotenous form descended from a similar 'Athorybia larva '-like ancestor, only it is comparable with a rather later stage in the ontogeny of an ' Athorybia larva' when nectophores have begun to develop. But of course we cannot be sure whether or not an 'Athorybia larva' represents an actinuloid larva with precocious gonophore (nectophore) buds. Even if it does it is probable that Garstang's ' Anthophysa ' ( = Athorybia Forskal) arose at a stage in phylogeny before the appearance of the precocious buds which, in Garstang's theory, represent the obsolescent hydroid gonophores. Garstang tried to clinch his argument by saying that the nectocalyces of the nectosome of Physonects must represent the gonophores of pre-siphonophore (i.e. hydroid) ancestors because they are produced from the oozooid, which in Siphonophores remains sterile, whilst sexual siphonophore medusa buds are borne on secondarily budded gonopalpons. The explanation of the fact that nectophores (sterile swimming medusae) arise from the oozooid is, I believe, that they were evolved later than the bud-borne, sexual medusae distributed along the stem of all Physonects, that they were developed precociously in the larvae of Physonects purely for swimming purposes, and were then further developed and carried into adult life — a curious case of adult retention of an organ that was originally transformed and developed precociously for larval use. That a medusoid gonophore bearing a sexual manubrium can lose this manubrium and be specially developed for swimming is proved by the transformation of gonophores into special swimming nectophores in the eudoxids of forms like Diphyes dispar, D. bojani and D. chamissonis. LARVAL AND DEFINITIVE NECTOPHORES From a study of the ontogenies of many calycophore species of the genera Rosacea, Hippopodius, Vogtia, Chelophyes, Muggiaea and Galetta, we can deduce that as successive nectophores are budded off, so in the course of evolution they have become more and more specialized. For instance, the second definitive or adult (posterior or inferior) nectophore of Chelophyes appendiculata is very unlike the first definitive one, which replaces the original, caducous larval one. This type of ontogeny is typical 1 The expression 'Athorybia larva' is used, not for the larva of Athorybia rosacea, but for the larvae of Physonects like Agalma spp., at the stage when they look rather like miniature specimens of A. rosacea. MORPHOLOGY AND RELATIONS OF SIPHONOPHORA 21 of diphyine Diphyidae, but we know little of the ontogeny of any species of the Diphyidae Tribus- Intermedia of Moser, or of the Abylidae. Text-fig. i. Phylogeny of Abylids. Anterior nectophores (in position for horizontal progression) arranged in sequence to suggest evolutionary trend. Fig. 1 H will be seen, on comparison with Text-fig. 2, to resemble larval forms. A, Chuniphyes multidentata ; the tooth T and ridge XT are missing in subsequent forms ; B, Ceratocymba sagittata ; C, Ceratocymba leuckartii; D, Ceratocymba dentata; E, Abyla trigona; F, Abyla haeckeli; G, Abylopsis tetragona; H, Bassia bassensis. Ridges XY and ab respectively are homologous. Ridge abl is an extension of ridge ab. Ridge XA in C, E and F is homologous. B Text-fig. 2. A, The larval 'calyconula' (Muggiaea atlantica, after F. S. Russell) representative of many Calyconects, compared with B, Sphaeronectes sp., and C, the anterior nectophore of Bassia bassensis. N, larval nectophore; N2, first definitive necto- phore; S, somatocyst, G, gastrozooid. In tentatively reconstructing a phylogenetic sequence of forms, and in crudely using present-day forms as some sort of guide, it is often difficult to decide in which direction evolution has proceeded. So, for instance, the present-day species of the last two groups mentioned above may be arranged pro- visionally as a sequence, which suggests that evolution either started or ended with a Bassia-like form, 22 DISCOVERY REPORTS as the first diagram (Text-fig. i) shows. But there are to be found in the plankton very young stages of the anterior nectophore of Abylines and Abylopsines with a single gastrozooid, which suggests that they are larvae. I have accordingly arrived at the tentative conclusion that the anterior nectophore of Abylids is the persistent larval nectophore, and that the anterior nectophore of Bassia represents an initial rather than a late stage in evolution. It may be compared in the second diagram (Text-fig. 2) with larval nectophores of other Calycophorae, and with the nectophore of Sphaeronectes. I have bred one species of Sphaeronectes at Villefranche, and it is evident that the adult nectophore is the persistent larval one. It is interesting then, to see that although, as Garstang pointed out, the pioneers and the writer of one modern text-book, Moser, misconstrued the structure of Sphaeronectes as representing a medusa with an exumbral manubrium, they were not far wrong in thinking it was the nearest approach to an archetype of the Siphonophora, since the adult is at any rate most like what we imagine the ancestor of the Calycophorae to have been. Perhaps it is a neotenous form, like Nectopyramis diomedeae. BUDDING ZONES It is evident from a fresh study of living and preserved post-larvae that in such forms as Nanomia bijuga, Agalma okenii (Text-fig. 27), and to some extent also in Physophora hydrostatica, the budding zones for the nectosome and for the siphosome, where true gonophores are now budded, arise on opposite sides of the larval air-sac (Text-figs. 3-5), although the significance of this is not yet understood.1 In the adult of Forskalia edwardsii the nectophores, siphons (gastrozooids) and gonophores appear all to be budded from the ventral side, as far as I have been able to ascertain. But even in an anaesthetized Forskalia there is so much crowding together of the siphosomal buds that it makes observation very difficult. In many Physonects whose siphosomal budding zone I have examined, there are indications that the successive gastrozooid-buds arise in a curved line that lies obliquely to the main axis or stem (Text-fig. 4) as in the nascent lobe of Physophora hydrostatica and in Athorybia rosacea (Text-fig. 5). Further work is needed on this point, for in both species of Rhizophysa, a contracted specimen shows that, though the budding zone of cormidia has always been thought to be single and ventral, it is only very close to the start of the budding zone that this can be seen to be true. As successive gastrozooids move distad they shift alternately to the right and left sides, leaving the gonophores between in the mid-ventral meridian. It is only further down the stem that they secondarily appear to arrange themselves in a single meridian again. Although, as I shall show below, it is possible theoretically to derive the calycophoran budding arrangement from that of the Physonectae, no one, as far as I am aware, has drawn attention to the true nature of these budding arrangements and to the striking difference between the relationships of siphosome to nectosome in the two groups. In a mature, long-stemmed Physonect such as Nanomia (Stephanomia) bijuga and Stephanomia rubra (originally known as Agalma rubra Vogt), the youngest proximal part of the siphosome is distal to the oldest part of the nectosome, so that the two parts of the stem must have increased in length in the same direction ; and the budding zone of the siphosome must have been carried distad (in an oral direction) in the course of growth from the larval stage, with the result that nectosome and siphosome form a single pendant stem, and the nectosome separates the two budding zones.2 Garstang's description of the arrangement (1946, pp. 106, 147) is misleading and gives the impression that the two budding zones lie close together between nectosome and siphosome in Physonectae — a statement which is true of Calycophorae but not of Physonectae. 1 Metchnikoff (1874) figured the two budding zones on opposite sides of the stem in the larvae of Sulculeolaria quadrivalvis (' Epibulia aurantiaca'), pi. 7, figs. 12-14, an^ in the post-larva of Agalma elegans ('A. sarsii'), pi. XI, fig. 2. Chun (1897a, fig. 6c) figured the two zones in a larva of Hippopodius hippopus. 2 See frontispiece. 23 yOAmm Gz-^ n Text-fig. 3. Larva of Nanomia bijuga captured at the surface at Villefranche, 17 June 1949, to show budding zones for stem-groups and nectophores (N.B.).Gz.\\ rotosiphon ; Gz.2, reduced secondary siphon budded from region near protosiphon ; X103. BrB . Blcr to s mm Text-fig. 4. Upper part of young specimen of Nanomia bijuga taken at Villefranche, o m., 23 June 1949, to show budding zones of nectophores and of stem groups on opposite sides of stem, x 69. 24 DISCOVERY REPORTS The calycophoran condition is best seen in the Prayids and in the Hippopodiids. Here the budding zones arise and remain close to each other, and the increase in length of nectosome and siphosome takes place in opposite directions. This is particularly noticeable in Hippopodiids, where the region adjoining the joint budding zones forms an upper, progressively twisting loop, and the nectosome forms a spiral-shaped, pendant housing (hydroecium) for the siphosome, which hangs down inside and is retractable within it. In the early stages of Hippopodiid growth the larval nectophore is attached above the joint growing-zone, but is later cast off. Bz.Gz.-_ Text-fig. 5. A, larva of Physalia physalis (after Okada, 1932) for comparison with B, larva of Nanomia bijuga, C, young Physophora hydrostatica (diagrammatic) after elongation of nectosome, and D, adult Athorybia rosacea (diagrammatic), E, diagrammatic ventral view of D to show the succession of gastrozooids. In all four species there is a terminal protosiphon Gz1, one or more secondary gastrozooids in a nearby group and a separate budding zone Bz.Gz. for the main succession of gastrozooids. Leaving out of account the larval Hippopodiid nectophore, which is specialized precociously to support the bud-colony, it would be possible theoretically to derive the Hippopodiid condition from that of the long-stemmed Physonect by imagining that with the disappearance of the pneumatophore, the pedicel or foot-stalk of the first nectophore grew continuously in length like that of the protosiphon (first gastrozooid), and that each successive nectophore was budded from the foot-stalk of the one before. Thus the nectosomal budding zone would remain near the siphosomal one, instead of being carried upwards further and further away from it as it is in Physonectae. In this way the lengthening of nectosome and siphosome in Calycophorae would take place in opposite directions instead of in the same one (cf. frontispiece). In both Prayids and Diphyids the definitive nectophores are reduced to a number which varies between four and one, but the arrangement of budding zones is difficult to see because it is not easy to anaesthetize the stem in that region, so that details are obscured by overcrowded buds. PHYSONECTAE: PRIMITIVE OR DERIVATIVE Once having been evolved, it is evident that the nectophores of Calycophorae and Physonectae are homologous structures, although the situation in Calycophorae is complicated by what Garstang supposes to be the addition of a bract, now represented by the somatocyst. The relationship of the budding zone of the nectophores with that of the siphosome in the two groups was discussed above. We now have to decide which group, Physonectae or Calycophorae, is primitive and which derivative. Garstang published the view which I myself have always held, namely, that the Physonect ancestors appeared first. In text-books the order in which the groups are dealt with implies that they appeared last. As Garstang said, it is the Calycophore which has undergone the most radical change, and it may MORPHOLOGY AND RELATIONS OF SIPHONOPHORA 25 even be that calycophoran larvae have discarded, in favour of a precocious nectocalyx, a previously possessed aboral float. He quite justifiably trounced the systematists for not having noticed the significance of the atrophy of the anterior end of the larval planula in Calycophorae. Moreover, although the early development of Hippopodiids is still not well known,1 probably because it takes place in deep water, Metchnikoff 's figures of early stages seem to indicate that the larval nectophore of Hippopodius is in fact produced by the aboral (anterior) part of the planula. I reminded Garstang of this some time after the publication of his paper, but by then he was approaching the end of his life's work and felt unable to go into the matter again. The reduction in number and kind of appendages in the Calycophorae and their detachment in the form of highly specialized, free-swimming eudoxid groups is almost certainly an evolutionary advance, and their significance lies in the fact that this arrangement tends to increase the chances of cross- fertilization with all its attendant advantages. Also, the fact that calycophoran species are so much more numerous than physonect ones would suggest that the evolution of eudoxids had been followed by deployment of the group. And judging by the long tubular nectosac of the posterior nectophore of Ceratocymba dentata and the complex structure of the gonophores of its eudoxid, one would certainly think that evolutionary progress had gone furthest in the Calycophorae. In considering which group is derivative, however, it has to be borne in mind that in Physonects the polyps are of two kinds, as Vogt realized long ago, each with a specialized function. First, there is the sterile gastrozooid with a large nematocyst factory (basigaster) for the highly evolved, branched tentacle. Secondly, there are the sexual palpons (with minute basigaster and simple tentacle), some of which bear at their base a prominence of varying length that carries the sexual medusae. This prominence ('stalk of the gonophore ') is obviously not itself a polyp or gonopalpon because it has no tentacle : it is part of the base of a palpon, and is quite short as a rule in the male gonophore. In the Calycophorae, on the other hand, the gonophores generally arise from the base of polyps of the highly evolved, gastrozooid type. Does the absence of the palpon-type of sexual polyp in the Calycophorae mean that reduction has taken place, or does the presence of the gastrozooid type of sexual polyp, with a highly evolved kind of tentacle mean that in this respect Calycophorae are an advance on the Physonectae? LARVAE Since Garstang wrote his valuable thesis I have seen post-larvae of three forms related to Hippopodius hippopus, namely Vogtia glabra, V. serrata— also figured by Moser (1925)— and Rosacea plicata (see PP- 73-77)- A study of these has convinced me that the adult Nectopyramis diomedeae is a neotenous form closely resembling in basic plan all these post-larvae. The Hippopodiids seem to have some characters more primitive than those of Rosacea. CLASSIFICATION The classification given by Moser (19246) in Kukenthal's Handbuch almost completely obscures the natural relationships of the Siphonophora and should not be followed. Libbie Hyman's text-book (1940) will be very widely used. Her classification of Siphonophora (PP- 370, 475) is: Qrder SIPHONOPHORA Suborder CALYCOPHORAE Suborder PHYSOPHORIDA Group PHYSONECTAE Group RHIZOPHYSALIAE Group CHONDROPHORAE 1 I failed in my attempts at Villefranche to breed Hippopodiid larvae (see Systematic Notes, p. 77). 26 DISCOVERY REPORTS The tentative scheme outlined by Garstang (1946, p. 190) treats the Chondrophora and the Siphonantha as two suborders: I DISCONANTHA— Order CHONDROPHORA II SIPHONANTHA A PHYSOPHORIDA (1) Cystonecta (2) Amphinecta ( = Physonectae Haeckel) B CALYCOPHORIDA On grounds of priority Eschscholtz's old name Physophorae should perhaps be retained in preference to Garstang's new one Physophorida, Haeckel's old name Physonectae for Amphinecta, and Calyco- phorae for Calycophoridae. I doubt whether Garstang's Cystonecta ( = Rhizophysaliae Chun) and Amphinecta ( = Physonectae Haeckel) are more closely related to each other than are his Amphinecta and Calycophorida ( = Calycophorae Leuckart). And I believe that the Chondrophora are as distantly related to the Siphonophora as the Margelopsid genera Pelagohydra, etc. So the outline of the revised classification I have adopted is: Order CHONDROPHORA (Chamisso & Eysenhardt, 1821) =DISCONANTHAE Haeckel, 1888 Genera Velella, P or pita, Porpema Order SIPHONOPHORA (Eschscholtz, 1829, part) = SIPHON AN THAE Haeckel, 1888 Suborder CYSTONECTAE Haeckel, 1888 = 'PNEUMATOPHORIDEN' Chun, 1882 = RHIZOPHYSALIAE Chun, 1897 Suborder PHYSONECTAE Haeckel, 1888 = PHYSOPHORAE Eschscholtz, 1829 (part) Suborder CALYCOPHORAE Leuckart, 1854 [Order HYDROIDA Suborder ANTHOMEDUSAE Family MARGELLOPSIDAE Genera Pelagohydra Margelopsis (haeckeli) Climacocodo?i] Much more work and consideration is needed before a sound and detailed classification of Siphonophora can be arrived at, but for the sake of planktonologists and others who handle these animals, the following very tentative arrangement of Calycophoran genera is suggested. The first genus contains what are probably the most simply organized siphonophores, and the last the most highly evolved Suborder CALYCOPHORAE Family 1. SPHAERONECTIDAE Sphaeronectes Family 2. PRAYIDAE Nectopyramis Rosacea CLASSIFICATION 27 Family 2. PRAYIDAE (co?it.) Praia Desmophyes \ Lilyopsis \ (see p. 88) Stephanophyes) Amphicaryon Maresearsia gen.n. (see p. 97) Family 3. HIPPOPODIIDAE Hippopodius Vogtia Family 4. DIPHYIDAE SULCULEOLARIINAE Sulculeolaria ( = Galeolaria, Galetta) DlPHYINAE Lensia Muggiaea Chelophyes Eudoxoides Dimophyes Diphyes (dispar, bojani, chamissonis, antarctica. These are probably proto-abylids) Chuniphyinae (these are probably proto-abylicls) Ckuniphyes Thalassophyes Crystallophyes Heteropyramis Clausophyes Family 5. ABYLIDAE Abylopsinae Abylopsis Bassia Enneagonam Abylinae Ceratocymba Abyla The relationships of the Cystonects are not at all certain. The gonophores of Rhizophysa arise in a longitudinal meridian on the lower part of the air-sac, and are flanked on either side by a meridional row of gastrozooid buds. Distally they all come to occupy a single meridional row. Each branchlet of the gonophore has a subterminal medusa-bud in which, presumably, the eggs are developed. But the early development of cystonects is quite unknown, except for Physalia. Hitherto they have been placed with the Physonectae because in both groups there is an air-sac or pneumatocyst. But it seems possible that the Physonectae have not passed through a Cystonect stage in their phylogeny, and that the Calycophorae may be more closely related to the Physonectae than either group is to the Cystonectae. The larva of Physalia would certainly seem to indicate a much earlier origin of 4-2 28 DISCOVERY REPORTS the Cystonectae than that suggested by the larvae of Physonectae or of Calycophorae, namely at an ancestral stage in which nectophores had not developed (Text-fig. 5 A). For the purpose of this report I have placed Athorybia, Melophysa and Physophora at the head of the list of Physonectae because they are probably neotenous forms, but the relationships between the others are not understood. In the list of Calycophorae, about the relationships of which, one with another, Garstang was silent, Sphaeronectes, as the ancestral type, comes first. It is followed by the Prayidae, which appear to be primitive in that the adults resemble the larvae fairly closely in their organization. The Hippopodiidae develop heteromorphous definitive nectophores, but their larvae are very similar to those of the Prayidae. These are followed in my list by Moser's group, the Tribus Diphyidae Intermediae, which appears to lead on to the genus Diphyes (sensu stricto) and to the Abylidae. The relationships inter se of the remaining Diphyidae are unknown, except that the species of Galettinae seem to be related to one another. It still remains to be ascertained whether, as I suspect, there is a fundamental difference between those species — included in the three groups Tribus Diphyidae intermediae, Diphyes {sensu stricto) and Abylidae on the one hand, and the remaining Diphyidae on the other, the suspected difference being, that in the former three groups, the larval nectophore is retained as the anterior, definitive one, whilst in the last group it is caducous and is followed by one or more heteromorphous ones. NUMBER OF SPECIES In his famous ' Challenger ' Report Haeckel dealt with about sixteen species taken by ' Challenger ', eleven of them very common, and two not seen since. In that report he described and figured altogether about thirty distinct species of Siphonophora. Haeckel listed 240 so-called species, of which not more than fifty are recognized as distinct to-day. Twenty-six of these 240 he called new species, which he promised to describe later. Four of them prove to be synonyms for known species and the rest are nomina nuda, since Haeckel never did describe them. No one, as far as I know, has commented on the fact that there are so few species — something in the order of 150 — of Siphonophores, which occur in great abundance in most seas. Is the small number of species due to the fact that so many of the species have very wide distributions and so consist of virtually single, interbreeding populations?1 There is a second remarkable phenomenon, namely the existence of many genera with 'pairs' of species such as Abylopsis tetragona and A. eschscholtzii; Diphyes dispar and D. bojani, Rosacea cymbiformis and R. plicata, Chelophyes appendiculata and Ch. contorta, Agalma okenii and A. elegans, Eadoxoides tnitra and E. spiralis, Chuniphyes multidentata and Ch. moserae, Nanomia bijuga and N. car a, Rhizophysa filiformis and Rh. eysenhardtii. The number of monotypic genera is also large, Apolemia, Lychnagalma, Erenna, Nectalia, Physophora, Athorybia, Melophysa, Dromalia, Rhodalia, Physalia, Porpema, Porpita, Velella, Enneagonum and Bassia. These two categories account for thirty-five out of the total of 150 species, but the analysis cannot be completed at present. There has appeared recently a note by C. B. Williams (195 1) on the relative sizes of genera in the classification of animals and plants. ' One general principle ', he said, ' of a mathematical nature, has several times been pointed out and discussed. In almost every classification the number of genera with one species is greater than with two, the number with two greater than with three, and so on ; so that if we plot the classification in the form of a frequency curve, we get a " hollow curve " somewhat resembling a hyperbola.' He then proceeded to show that we can only recognize as real the scheme of relative specific and generic relationships commonly used by good systematists if we bear in mind that the point in time-past, at which to 'draw the line' or make 1 If it can be shown that waste products play some part in causing mutation, it seems possible that there may be a connexion between the small number of species and the fact that waste products must quickly diffuse out into the surrounding sea-water. CLASSIFICATION 29 a cross-section to demonstrate these relationships, is a matter of opinion, and that there must be consistency in the choice. Even if the phylogeny of a number of species were known, it might be quite correct to adopt any one of a number of generic groupings, according as to where in the past the cross- section was taken. Judge, then, of the systematists' difficulty when the phylogeny is not known. THE PIONEER WORK OF HENRY B. BIGELOW Many additions have been made to our knowledge of the Siphonophora since Bigelow, forty years ago, published his famous 'Albatross' Report, which has been a key-work on the group ever since. It was a study of this delightful work that first lured me to examine these animals and to build up an enormous collection at the British Museum (Nat. Hist.). Hardly a day passes when it is not necessary to refer to the work again. If in my systematic notes I often appear to criticize this classic, it is only with the desire to build firmly on sure foundations. Bigelow's figures of siphonophore structure in that, and subsequent reports, are unrivalled. Only he and Dr Mary Sears know how painstakingly he worked upon them. To orientate these animals, whose density is so nearly that of the media in which they have to be examined, and to hold them in position without movement for drawing is a major difficulty in the study of the group. GENERAL CONCLUSIONS To sum up my tentative conclusions on the phylogeny of Siphonophora — reached, it has to be admitted, without full consideration of either Rhodaliids or Apolemids — there is evidence that leads to the deduction that siphonophore ancestors, like the primitive Trachylines, had an adult medusoid stage (medusoid gonophores, nectophores), as well as a larval polypoid one (present-day larvae, gastrozooids, palpons). In the course of evolution, and after the adoption of the budding process, the two phases instead of following one another, as in scyphomedusan and trachyline metamorphosis, appeared simultaneously in one and the same bud-colony (gastrozooids and gonophores) as, for example, in Agalma. After the development of an aboral organ as a gas-secreting gland (pneumatophore), this gland was once again lost (Calycophorae), when medusoid gonophore-buds were developed precociously in the larval stage as swimming organs. The precociously developed and functionally altered (gonophore -> nectophore) larval organs were then carried into adult life and improved upon as adult swimming organs (nectophores). But some of the larvae of these bud-colonies never adopted the device of using a precociously developed gonophore as a swimming organ and so the adults (e.g. Cystonects, and Athorybia rosacea) do not swim to-day by jet propulsion. It should here be emphasized that this interpretation of a very complicated evolutionary story bears no resemblance to the old classical ' medusome ' theory of Haeckel's, which was exploded by Garstang. It follows that acceptance of the conclusions set forth above is compatible with finding in Siphonophora much evidence of neoteny (Physalia, Nectopyramis diomedeae, Sphaeronectes, Athorybia). This phe- nomenon probably affords the explanation of the existence of both short-stemmed forms (brachysteles) and long-stemmed forms (macrosteles) among the Physonectae, the macrosteles having increased in length and the brachysteles having failed to lengthen in one or two directions. There seems to be every chance that in the future other unspecialized larvae will evolve on new neotenous lines. We may, perhaps, in some ways compare the story of the siphonophore nectophore with that of the tail of the Ascidian tadpole, which appears to have been developed as a larval organ, is lost at meta- morphosis by Ascidians, but is retained and developed in the adult stage of other Chordata. In the siphonophore story, of course, the organ (gonophore) adapted specially for larval use (as a nectophore) was originally an adult ' organ ', or rather the adult (medusa) phase itself, produced by budding, and 3o DISCOVERY REPORTS developed in the larva precociously (by heterochrony). It was then developed still further in the adult for the same purpose (swimming). Some adult Physonects took to swimming by jet propulsion, it might almost be said, because of an ' afterthought ' on the part of an ancestral larva. The old suborder Disconanthae contains three forms Velella, Porpita and Porpema that are really the floating Tubulariid hydroid-phases of inconspicuous adult Codonid medusae known as Chryso- mitra. There is nothing in their morphology in the least like the Siphonanthae, and it is possible that their pneumatophore has been acquired independently. They form a separate Order Chondrophora. Garstang (1946) interpreted the Conaria larva, but omitted reference to important papers by Delsman (1923) on the Conaria larva of Porpita and its development, and by Leloup (1929), who, after studying Velella at Villefranche, pointed out the homology between the Conaria larva and the actinula of Tubularia larynx, showing that the peculiar ' crimson cone ' of Conaria corresponded with the ' gastric diaphragm' of Tubularia. This work of Leloup appears not to have been known to Garstang. Flrud -Tlarv. Text-fig. 6. A, Actinula of Tubularia. After Leloup (1929). B, Conaria of Porpita. Based on Delsman (1923, figs. 16, 17, 18). X Y= homologous planes. Text-fig. 6 will make this point clear. There is no need to endeavour, as Garstang did, to 'close the gap ' between Disconanths and Siphonanths. The old order Siphonophora without doubt had a double origin, as Haeckel realized. As to the position of Pelagohydra, one is struck by the origin of the tentacles between the axial parenchyma in the meshes between the endoderm canals, from which the gonophores (gonozooids now reduced) spring. In Chondrophora Bigelow's sections show that the tentacles arise from the canals themselves. The difficulty of exactly homologizing the structure of the two forms was stressed by Garstang (1946, p. 126), and of course Pelagohydra has no pneumatophore. Garstang maintained that the ' float ' of Pelagohydra represents the hydrocaulus of Corymorpha, on account of the position of the gastric diaphragm, and that the gonophores are cauline. In the Chondrophora the gonophores are on the oral cone. Possession by hydroids of cauline gonophores is probably a later evolutionary stage than that of oral ones and in the opinion of my colleague Dr W. J. Rees leads on to a hydrorhizal position. The Chondrophora appear to be aberrant, pelagic, tubularian Anthomedusae. They have lateral (radial) extensions of the walls of the aboral chamber, the edge of the extension being provided with TERMS USED IN THIS REPORT 31 mucus glands. They have radial extensions as well of the oral cone, which bears a large number of gonozooids. They possess too, rings of aboral tentacles, and large floats invaginated into the aboral end. They differ widely from Siphonophora, though the two groups, no doubt, had a common ancestor. The Chondrophora are so much more aberrant than the somewhat similar pelagic tubularian Margellopsinae, that they should be treated, perhaps, as a separate order, neither as Hydroida, with which Vogt, Kolliker and Agassiz were inclined to associate them, nor as Siphonophora, as has for long been customary. TERMS USED IN THIS REPORT I have used the same terms as those used by Bigelow (191 ib) in his 'Albatross' report. A long- stemmed Agalmid is divided into the nectosome above, and the siphosome below. In the early larva or oozooid these two parts occupy opposite sides of the float or pneumatophore.1 As growth proceeds the growing point of the nectosome, lying as it does just below the pneumatophore, is carried upwards from the level of the growing point or blastocrene of the siphosome, whose oldest gastrozooid (protosiphon) is thus carried downwards.2 The outer wall of the pneumatophore is called the pnenmatocodon, and the inner wall the pneumatosaccus. Libbie Hyman regards these as the exumbrella and subumbrella respectively of a medusa, a point of view that I hope will be abandoned after reading Garstang (1946) on the subject. The rhythmical contractions of the muscular lining (nectosac) of the nectophores provides jet-propulsion. The nectophores of the Calycophorae include a somatocyst which has vacuolated walls and usually contains a quantity of fat. It is in direct communication with the stem and all the buds (blastozooids) of the original larva (including the protosiphon) or oozooid. The whole forms a floating bud-colony of larval-type polyps or zooids, to which are added precociously the adult-type medusae or nectophores — for swimming only — without manubrium or mouth, and later the gonophores with manubrium but without mouth. At the base of the feeding zooid or gastrozooid is a swollen part, the basigaster, filled with growing nematocysts, from which arises the single tentacle. Its lateral branches or tentilla are provided with complicated nematocyst batteries or cnidosacs. The cnidosac is attached to a pedicel of the tentillum by two structureless, much folded, elastic bands {angle-bands). From this point a broad band (cnidoband) of probably adhesive nematocysts, often spirally coiled, is doubled back towards the pedicel and partially or wholly covered by an involucre. Distad to the attachment of the angle-bands may be a central ampulla and a pair of lateral horns. In the Calycophorae there is usually a distal patch of pear-shaped nematoblasts, and a (sub-terminal) terminal filament bearing numerous nematocysts. Flanking the inner end of the cnidoband of an unexploded cnidosac is a series of large penetrant nematocysts. Upon appropriate stimulation, whose nature, in spite of experiment, is still not understood, the inner end of the cnidoband is flung forward, its outer end anchored to the pedicel by the elastic bands. The Calycophorae appear to have lost the aboral end of the larva after the precocious formation of nectophores. They now have no pneumatophore. Their nectophores differ from those of the Physo- nectae in that they include a somatocyst, about the significance of which there has been much specula- tion. Old authors sometimes referred to the other type of nectophore, found in Physonectae, as pure nectophores. There is only one other type of zooid commonly called the palpon. In Physonectae it is generally much less developed than the gastrozooid. It does not feed, though its terminal end may occasionally open. Its basigaster is very reduced, and its tentacle or palpacle is simple. In the Calycophorae there is no such distinction, or in other words there are no palpons. So that either Physonectae possess reduced or undeveloped zooids (palpons) or in Calycophorae the palpons are upgraded to gastrozooids or else have disappeared. 1 Text-fig. 3. 2 See frontispiece. 32 DISCOVERY REPORTS Calycophorae have made progress from an evolutionary point of view by freeing the segmentally repeated groups of gastrozooids, bracts and gonophores, in the form of eudoxids. In some cases one of the gonophores loses its manubrium and becomes an organ for jet-propulsion, the special nectophore. Bracts are found in larval Physonectae, and in the adult Physonectae they reach the peak of their development, being present in several ventro-lateral rows on the stem. They are not found in larval Calycophorae, unless their canals are represented by the somatocyst. In adult Calycophorae, as stated above, there is only one bract to each gastrozooid, and it eventually becomes the protective part of the free eudoxid. Explanation of abbreviations used in figures Abor. aboral Go. gonophore Ang.ap. apical angle Gz. gastrozooid Ang.d.-l. dorso-lateral angle Gz.rd. gastrozooid (reduced) Apoph. apophysis Gz.m. mouth of gastrozooid Bg. basigaster H. hydroecium Blcr. blastocrene H.op. hydroecial opening Br. bract L.musc. muscular lamella Br.B. bract-bud Lob. ant. -1. r. right antero-lateral lobe Br. lam. muscular lamella of bract Man. manubrium B.z.Gz. budding zone of gastrozooids M.b.-l. baso-lateral margin B.z.N. budding zone of nectophores N. nectophore C.br. bracteal canal N.ant. anterior nectophore C.br.d. dorsal bracteal canal N.B. bud of nectophore C.br.s. semicircular bracteal canal Nek. notch C.br.spur bracteal spur-canal Nee. nectostyle C.br.v. ventral bracteal canal N.def. definitive nectophore C.circ. circular canal Nem. nematocysts C.com. commissural canal Nem.p.ect. ectodermal patch of nematocysts C.d. dorsal canal N.larv. larval nectophore C.d.-l. dorso-lateral canal N.post. posterior nectophore C.desc. descending canal Ns. nectosac C.h.l. left hydroecial canal Ns.def. nectosac of definitive nectophore C.h.l.l. left lateral hydroecial canal Ns.larv. nectosac of larval nectophore C.h.l.r. right lateral hydroecial canal Ns.obs. nectosac of obsolescent nectophore C.h.r. right hydroecial canal N.spec. special nectophore Clot. lateral radial canal Ns.v. ventral wall of nectosac C.long. longitudinal canal Nuc. nuclei Cmb. comb N.vest. vestigial nectophore C.med.long. median longitudinal canal O.c. central organ Con.cr. crimson cone 01. oil C.pa. pallial canal Ost. ostium C.pa.{som.) pallial canal Ov. ovum C.ped. pedicular canal P. palpon C.rad. radial canal Pal. palpacle C.rad.d. dorsal radial canal Phc. phyllocyst C.rad.lat. lateral radial canal Pl.M. mouth-plate C.rad.v. ventral radial canal Pn. pneumatophore d. dorsal Pr.dist. distal process D.g. gastric diaphragm Pr.lat. lateral process Eud.Br. eudoxid bract Prom. prominence F.con. convex facet Proj.ang. projecting angle F.d. dorsal facet Prot.ch. chin-shaped protuberance F.d.-I. dorso-lateral furrow R.ap. apical ridge F.h. hydroecial fold R.ap.-lat. apico-lateral ridge Fl.rud. rudiment of float R.buc. buccal ridge SYSTEMATIC AND R.d. dorsal ridge R.d.-l. dorso-lateral ridge Ret. rete R.inf.lat. inferior lateral ridge R.l. left ridge R.lat. lateral ridge R.lat.as. asymmetrical lateral ridge R.lat./. left (asymmetrical) lateral ridge R.long. longitudinal ridge R.or.lat. oral lateral ridge R.post. posterior ridge R.r. right ridge R.vert.lat. lateral vertical ridge Som. somatocyst Som.N.def. somatocyst of definitive nectophore Som.N.larv. somatocyst of larval nectophore spx. spadix Str.m. mamma-like structure IOLOGICAL ACCOUNT Surf .art. articulating surface T. tentacle T.abor. aboral tentacle 77. tentillum T.larv. larval tentacle To. tooth To.d. dorsal (median) tooth To.int. internal tooth To.l. left dorso-lateral tooth To.r. right dorso-lateral tooth To.v.l. left ventral tooth To.v.r. right ventral tooth V. ventral Vet. velum W.hyd.l. left hydroecial wing W.hyd.r. right hydroecial wing W.lat. lateral wing XB growing apex of pallial canal 33 SYSTEMATIC AND BIOLOGICAL ACCOUNT CHONDROPHORA Porpita Lamarck, 1801 Forty years ago Bigelow (191 16) reviewed the question of generic and specific limits in Porpitidae. I do not think further progress will be made by any survey of literature, but only by the fresh examina- tion of all growth stages from both Atlantic and Pacific. On a priori grounds it seems unlikely that species of these surface animals from tropical and sub-tropical waters would be isolated in one or other ocean — that is, if our knowledge of distribution of other siphonophores, and if Sewell's (1948) views on transportation of planktonic copepods are any guide. Haeckel recognized no less than six genera and twenty species of Porpitids. Bigelow regarded many of Haeckel's forms as growth stages and provisionally recognized two genera and five species. Bigelow considered five points in the anatomy of Porpita: (1) the tubercles on the upper surface of the disc, (2) the arrangement of the stigmata, (3) comparative width of the limbus, (4) radial or branching arrangement of the canals of the limbus, and (5) the number of stalked tentacular nematocyst clusters on the tentacles, and from a study of these anatomical points, he came to the conclusion that an Atlantic form Porpita umbella could be distinguished from a Pacific form P. pacifica; but he doubted if a Pacific and Atlantic species of Porpema — the only other genus he recognized — could be distinguished. I am inclined to think that there are only two Porpitids, Porpita porpita (L.) and Porpema prunella (Haeckel, 18886), and that one genus Porpita is enough to contain them. Porpita porpita (Linne), 1758. (Synonymy in Bigelow, 191 ib, p. 353.) Porpita umbella (O. F. Miiller), 1776. (Synonymy in Bigelow, 191 ib, p. 352.) Porpita pacifica Lesson, 1826. (Synonymy in Bigelow, 19116, p. 333.) 34 DISCOVERY REPORTS Seven specimens were taken at the surface at 'Mabahiss' Stations 39 (see Chart, p. 10) and 99. It was on Indian Ocean material that Linnaeus based his Medusa porpita, to which the present material probably belongs. So far, a critical comparison of Indian Ocean, with Atlantic or Pacific material appears not to have been made, although Bigelow (1904) described and figured some from the Maldive Islands. Leloup (1934ft) identified some taken off the Madras coast, stating that this species is common in the Indian Ocean, but without giving further details. I am unable, without good material of all growth stages from all three oceans, to make the desired critical com- parison. Examination of such material and critical descriptions as are available, leads me to suspect that one species only is to be found in all three oceans. If so, it should bear the name porpita (Linne.) A few counts of the stalked tentacular nematocyst clusters of the long rows gave the figure 21-24, and of the short rows 10-13. The numerous tubercles are obvious and have no connexion with the stigmata which are few and become covered not far from the inner edge of the limbus. The radial plications of the surface of the disc are slight. The limbus is comparatively narrow as in Porpita pacifica. In a specimen ('Mabahiss', 23. xii. 33, near St. 99) 3-9 cm. in diameter, the limbus measured 3 mm. In another 2-7 cm. in diameter, it measured 2 mm. The canals of the limbus are irregularly arranged and not radial. In all these respects the Indian Ocean and Pacific forms agree. Velella velella (Linne), 1758. Velella spirans Forskal, 1775; Chun, 18976. The development of the Conaria-larva, first described by Woltereck (1904, 1905) was interpreted by Leloup (1929) and Garstang (1946); and the further development of the Rataria- or Ratarula- post-larval stages are now fairly well known. Likewise the young medusae {Discomitra Haeckel, 1888 ft, p. 39, pi. 50, fig. 8) from the time when they arise as buds on the walls of the gonozooids to the time when they are shed, are well known. They are of the Tubulariid type, the tentacles, usually undeveloped, bearing four radial canals and four radial, exumbrellar rows of nematocysts. In the youngest stages these rows extend only half-way to the apex, but later reach all the way up. Such Discomitra medusae may be compared with the well-known Tubulariid medusae, Hybocodon prolifer and Ectopleura dumortieri, the earliest stage of the latter having only one tentacle, three and four being developed later on. The possession by Velella and Porpita of this type of medusa appears to confirm Garstang's view about the Tubulariid ancestry of the Disconanthae (Chondrophora). Not so well known as Rataria and Discomitra is an apparently quite different type of medusa Chrysomitra striata (Gegenbaur), the genus founded by him in 1857 for a little free-swimming medusa Phorcynia striata (not stricta) described by Kolliker (1853 a) with from thirteen to sixteen radial canals and one or two tentacles. This Chrysomitra medusa (not larva) is commonly said by writers of text- books to come from Velella, but I am not yet satisfied that this has been proved. It possesses exum- brellar nematocyst tracts, but in appearance differs from Discomitra. Its direct connexion with Velella was not observed by Gegenbaur, but on account of the nematocyst tracts and the ' yellow-cells ' in the sub-umbrella in both forms, and the appearance of rudiments of further radial canals growing out from the stomach in some specimens of Discomitra, Gegenbaur thought it reasonable to assume their identity. That this is very probable has been shown by the micro-photograph by Ankel (195 1) of living specimens of Discomitra recently freed by Velella. One of the medusae has a single small tentacle of the Chrysomitra type, but the rest of the medusae are in the Discomitra stage. Metchnikoff (1886) recorded that on two occasions in January and February 1883, he fished at Messina some sexually mature medusae of Velella and he gave good figures of the single amoeboid egg, the four testes, the sperm and the medusa itself. The medusa had only four radial canals and a single tentacle. He likened SYSTEMATIC AND BIOLOGICAL ACCOUNT 35 them to Keferstein & Ehlers' (1861) Rhabdoon singular e, but left any decision about their identity with Gegenbaur's Chrysomitra for future research to decide. Bigelow (191 1 b) in his review of the genus mentioned that eleven so-called species have been described from the Indo-Pacific. Most recent authors have agreed that there is only one Indo-Pacific species of Velella and that it is distinct from the Atlantic one, but Schneider (1898), and Bigelow & Sears (1937) with whom I am inclined to agree, thought it was the same as the Atlantic one. Bigelow (191 1 b) found no features to differentiate more than one Pacific species, but was not prepared to unite it with an Indian Ocean one, because a critical review of Indian Ocean specimens was not then possible ; but he believed that the separation of Pacific and Atlantic species was justified, the breadth in proportion to the length being greater in Pacific than in Atlantic specimens : in other respects he could find no constant difference. Bigelow gave measurements of sixteen shells from the Atlantic and of the same number from the Pacific. The ratio of breadth to length of the Atlantic ones was from 32 to 37%, and of the Pacific ones 44 to 51%. For Mediterranean specimens, more than 8 mm. in length, Bigelow & Sears gave a mean ratio of 38-5 %■ It is convenient to call the long axis of a specimen of Velella the 'N./S.' one, and a specimen in which the sail runs ' N.W./S.E.' is called a ' N.W.' one and the type in which it runs 'N.E./S.W.' a'S.W.' one. Station 2669 (W. Medite rranean) Stations WS 1057, WS 1058 (S. Atlantic) Height of sail Height of sail Length Breadth from apex Length Breadth from apex (mm.) (mm.) of liver (mm.) (mm.) (mm.) of liver (mm.) 20 9 7 13 6 4 22 10 8 17 10 5 22 14* 10 18 10 6 23 9 9 27 16 8 25 9 9 — — — 26 13 9 — — — 27 x3 8 — — — 28 !3 10 28 17 9 28 12 10 28 H 11 29 12 8 28 16 10 29 12* 11 — — — 30 14 10 — — — 30 14 9 — — — 31* 14 8 — — — 32 15 12 — — — 32 15 10 — — — 33 15 12 — — — 33 12* 12 33 20 12 34 J7 12 34 21 J3 34 16 11 — — — 35 19 10 35 18 9 35 !3 10 35 22 12 35 H 10 37 23 12 37 !7 H 37 J9 12 40 17 14 38 22 !3 44 22 J3 40 22* 10 45 21 H 41 24 15 Ratio average Br./L. 45% Ratit 3 average Br./L. 55% An asterisk indicates that the measurement is an estimate of an imperfect specimen. 5-2 36 DISCOVERY REPORTS The 'Discovery II' brought back a 'S.W.' specimen from Station 1738 in the South Indian Ocean that measured: length 33 mm., breadth 18 mm., height of sail from apex of 'liver' 9 mm. This gives a ratio of 54-5 % for breadth and length. The 'Discovery' collections also contain a series of twenty-seven 'N.W.' specimens from Station 2669 in the Western Mediterranean, as well as an interesting series of ' S.W.' specimens from Stations WS 1057 and WS 1058 made in the South Atlantic between St Helena and Tristan da Cunha. The ratio L/Br for these two groups of specimens is given in the table on p. 35. Only those specimens fit for measurement are listed. Bigelow showed that the 'N.W.' condition of the sail was much more common in the Pacific Velella lata than in the Atlantic V. velella. More data are needed on proportions and direction of sail. Distribution. In the Atlantic Velella extends from Cape Cod to at least 280 S. Feeding. The specimen from Station 1738 was 'brilliant indigo-blue' in colour when fresh; 'the sail pale purple almost transparent '. Many of its gonozooids had been feeding on Crustacea, which I have not attempted to examine or identify, and in many cases the mouths are still open. One gonozooid had extended its mouth to a width of 2 mm. in an attempt to engulf a Calanoid copepod (family Euchaetidae) which measured 2-5 mm. in length (excluding terminal setae). SIPHONOPHORA PHYSONECTAE At least eleven Physonect species are represented in these Indian Ocean collections, an increase of 50% on numbers last reported by Browne (1926). So far, very few Physonects at all have been described and figured to the satisfaction of specialists, and though the systematic work on them is always going on, it is necessary to send this report to press before it has been possible to complete it. Quite apart from the morphological work entailed, the synonymy is very involved, as shown by the classical example of Agalmopsis elegans Sars, 1846. Sars himself realized later that he had based his name on two species. A discussion of the matter was published by Bigelow (191 1 ft). The two species occur together commonly not only in Norwegian waters but off South-west Ireland and in the 'Celtic' Sea. Unfortunately each component, generally referred to as ' Agalma'' (trifid tentilla) and ' Stephanomia' (filiform tentilla), has been renamed sarsii, the first by Kolliker (1853ft); ana the second by Haeckel (1888 ft) as Cupalita sarsii. And finally Bigelow (191 1 ft) chose to use the name elegans for the minor 'Agalma' component thinking that the major one was 'Stephanomia' bijuga. My contribution to the discussion is to affirm, on the evidence of the nectophores (Sars, 1846, Tab. 6, figs. 3-4) and of the palpon (Tab. 6, fig. 11), that fig. 1 of both of Sars's plates 5 and 6 repre- sented at the time a new genus and species, a species later redescribed from the Gulf of Maine by Agassiz (1865) and Fewkes (1888a) as Nanomia cara (see Text-fig. 19). The main concept that Sars had in mind when describing his novelty was the species represented in twenty-four out of twenty-six of his figures. Agalmopsis elegans obviously has priority as the name of the species to which they belong, but Haeckel (1888ft, p. 234) chose to- restrict Agalmopsis to the other species. Two short visits to Villefranche have enabled me to study alive in the field six of eleven Physonects found in the Indian Ocean and here reported upon, viz. Physophora hydrostatica, Stephanomia rubra, S. bijuga, Agalma elegans, Athorybia rosacea Forskal (non Bigelow) and Cordagalma cordiformis, together with a species of Forskalia ; and also to study alive the larvae of Physophora hydrostatica, Stephanomia bijuga, S. rubra, Agalma elegans, and Forskalia sp. One of these species, Cordagalma cordiformis Totton, 1932, must be the smallest of all the Physonects. At Villefranche I found loose nectophores frequently, and some minute, colourless colonies that SYSTEMATIC AND BIOLOGICAL ACCOUNT 37 I believe to belong to them. A similar stem was associated with many nectophores at one ' Manihine ' Station in the Gulf of Aqaba. One of the eleven Indian Ocean Physonects, a new species (see p. 59) is of special interest because of its apparent relationship with the one and only known high-arctic Siphonophore Stephanomia orthocanna described by Kramp (1942). Kramp's species is characterized by the straight lateral canals of the nectophores. In the Antarctic there is a third Physonect of this group, also new, and the type species of a new genus which has been named (p. 55) after Mr J. W. S. Marr. It has very characteristic bracts with truncated distal ends and straight lateral canals in the nectophore. To this new genus both S. orthocanna and the two new species belong. Athorybiadae Huxley, 1859 Syn. Anthophysidae Brandt, 1835 Eschscholtz (1829) included in his genus Athorybia: (1) Rhizophysa heliantha Quoy & Gaimard; (2) R. melo Quoy & Gaimard; and (3) Physsophora (sic) rosacea Forskal. Haeckel (1888a) took Rhizophysa melo Quoy & Gaimard as the type of a new genus Melophysa. I now select Physsophora (sic) rosacea Forskal as the type of Athorybia1 Eschscholtz. I believe that Rhizophysa heliantha Quoy & Gaimard is a synonym of Athorybia rosacea (Forskal). The two genera Athorybia and Melophysa are quite distinct. In the latter a distinct nectosome is found. Schneider (1896, p. 590) suggested that Athorybia was really a sexually mature, larval (neotenous) form, with which I agree. I go further in suggesting that Melophysa melo represents a form that has similarly become sexually mature but at an early post-larval stage. This would account for the absence of nectophores in Athorybia and the presence of a single functional nectophore at any one time (?) in Melophysa. Furthermore, I consider an important point in this argument, to be the shape of the nectophore in Melophysa. In Bigelow's (1931) fig. 217, I suggest that the nectophore has become twisted through 1800, as so often happens, about one axis or the other,2 even in a species like Diphyes dispar. I believe that in Melophysa melo the pedicular canal ascends and does not descend to the nectophore. Now the important point is that in young nectophores of Agalma okenii there is a very distinct inferior pedicel, a condition which I believe to be homologous with that in Melophysa melo. If Bigelow's nectophore were to be reversed, the loops of the lateral canal could also be homologized with those of Agalmids. The external features of the nectophore are unlike those of other Agalmids as my figures show (Text-fig. 7), a fact that once more suggests that Melophysa1?, nectophore is a survival of an earlier stage of development of these organs. Schneider (1896) gave no clue as to the identity of the form he was discussing except that it was evidently the larval form described by Haeckel (1869a). The identity of Haeckel's larvae can only be surmised. He took a unique specimen of Athorybia ocellata (syn. rosacea) at Lanzarote in 1876, and the larvae, which he was observing and describing at the same time, probably came from this same specimen. Athorybia Eschscholtz, 1829 Anthophysa Brandt, 1835. Angela Lesson, 1843. Athorybia Kolliker, 18536; Huxley, 1859; Gegenbaur, 1859; Fewkes, 1882; Haeckel, 18886; Fewkes, 18886. Ploeophysa Fewkes, 18886. Bigelow (193 1, p. 579) discussing Melophysa melo said, 'it will never be possible to settle conclusively whether or not it was this same form Athorybia rosacea [Bigelow, syn. Melophysa melo Q. & G.] that Forskal described and pictured as Physsophora (sic) rosacea, for his figures and account would apply equally to any Athorybia'. But, as Haeckel says, Forskal was describing a typical Mediterranean form, which Melophysa melo is not. I have even witnessed in Agalma elegans the loss of one of a series of nectophores, shot off when the stem contracted, followed by the twisting round of a neighbouring nectophore to take its place. 38 DISCOVERY REPORTS Diplorybia Fewkes, 18886. Anthophysa Chun, 18976; Bedot, 1904; Bigelow, 191 \b (excl. synonymy); Totton, 1936; Leloup, 1941. Athorybia Garstang, 1946, pp. 135, 140, 141, 190; non i^rj, 167, 175 ( = Melophysa melo). Anthophysa Garstang, 1946, pp. 141, fig. 45; 142, 147, 148, 167, 168, 170, 171, 172, 175; non 190 ( = Melophysa melo). Brandt (1835) established a genus Anthophysa based on unpublished figures and a manuscript description by Mertens. Haeckel (18886) said that he had seen these and made a few comments on them, but his observations are not sufficiently critical to enable me to come to a certain conclusion about the generic identity of A. rosea Brandt. As far as I can judge it is congeneric with Kolliker's Athorybia rosacea, but not with Bigelow's (193 1) A. rosacea. The species described and figured by Bigelow (191 1 b) as Anthophysa rosea is of the same genus as Forskal's Physsophora (sic) rosacea and Kolliker's Athorybia rosacea. I hope that the selection of the lectotype of Athorybia will end this confusion of generic names. Athorybia rosacea (Forskal), 1775. (Plates I— III.) This Mediterranean species was figured in colour by Quoy & Gaimard in 1827 as Rhizophysa heliantha, by Kolliker in 1853 b as Athorybia rosacea and again, probably, by Haeckel from the Canaries in 18886 under the name of Athorybia ocellata. Bigelow (191 lb) gave a detailed description and figures of it under the name Anthophysa rosea. Finally, Leloup (1941) gave us the latest summary of our knowledge of this genus under the name Anthophysa Brandt, 1835, together with some notes on a specimen from the Gulf of Guinea. Leloup also plotted the distribution of the species. His figs. A-C (194 1, pi. 1) are of interest for they tend to question the significance of Bigelow's similar photographs of a section of an alleged superior, gas-secreting, secondary ectoderm of the pneumatochone. In Bigelow's fig. 1 (191 1 b, pi. 23) the thickened secondary ectoderm reaches right across the vault of the pneumatosaccus, a condition unparalleled, as far as I know, in any Physonect. But the pneumato- phore lies obliquely, and the section may have missed the apex of the pneumatosaccus. In Leloup's figures the condition is more normal. Further work on this point is needed. I recently studied Athorybia alive at Villefranche, where Sabine Baur made some excellent coloured sketches1 of it. The only species of this genus Athorybia is characterized by having a large air-sac that forms the bulk of the specimen. It is half covered by what Fewkes called the hood. Haeckel called it the necto- style or attachment of the bracts, which he claimed were used in swimming. There are no nectocalyces. Behaviour. Whilst I was aboard the launch belonging to the Station Zoologique at Villefranche in May 1950 a living specimen, 12-14 mm. in diameter, was secured at about 9 a.m. after being detected by the keen eyes of the mechanic, M. Raibaud, as an inconspicuous floating object ' a peine rouge '. No specimens had been seen there in the last forty years. Observation soon showed that the animal could ascend and descend at intervals in the jar of sea-water, which I protected from the sun with a duster. Before sinking, with closed bracts, a number of small bubbles would appear amongst the bracts, and be removed by their movement. It was not possible to locate the pore from which the bubbles must have come, although my notes twice mention that they appeared to be released from the lower or basal part of the pneumatocyst, but I now believe that they must have come from an apical pore. After an interval of ten minutes the animal would begin slowly to leave the bottom of the jar again, after intermittent closure of the bracts that caused a slight rise followed by a falling back. The partially raised bracts were held motionless during ascent and not flapped in the manner described by Kolliker, except that when about an inch from the surface they opened somewhat and closed with a kick. This action had the effect of sending the animal downwards for an inch before rising again 1 Mile Baur has generously allowed me, with Professor Portman's permission, to publish reproductions of these unfinished drawings. I consider them of greater value as records than finished ones. SYSTEMATIC AND BIOLOGICAL ACCOUNT 39 to rest at the surface. My desire to obtain sketches of the living animal, and then to anaesthetize and preserve it for detailed examination, as well as my fear that it might become moribund and lose all its bracts, forbade me to keep it under observation alive for more than two hours or so, but the ascent and descent with contracted tentacles were repeated some half-dozen times after capture. I watched the descent three times. When, finally, 7% magnesium chloride was added in three stages, after drawing off most of the sea-water, a slow beating of the bracts, which opened out at once, was seen for a long time. Cook's first voyage. Whilst examining the third volume of Sydney Parkinson's unpublished water- colour drawings made during Captain Cook's first voyage, I came across a page of sketches of this animal which is reproduced here (Pis. II— III) since so few figures exist. It was 'painted from life in October 1768 between the tropicks', and appears to be the earliest record of the species. Perusal of a photographic copy of Banks's MS. Journal shows that the specimen was probably taken, with other invertebrates on 4 or 7 October, at a position which I estimate to have been about 9-1 20 N. and 22-23 ° W. Haeckel's ocellus. Haeckel described near the tip of each palpon a red ocellus and figured it as if it had a small lens. I was interested therefore to observe in some of the living palpons small cercariae. The palpons had a pink spot, but none had a lens. It may be that Haeckel was deceived by the appearance of cercariae into thinking that there was a lens associated with the pink spot. Anatomy. Haeckel claimed to have discovered, what he said previous observers had missed, the true nature of Athorybia. However, it seems that even Haeckel did not make the correct interpretation of its anatomy. My suspicion, gained from examining one or two better preserved specimens in the British Museum (Nat. Hist.) Collection, that the line of budding gastrozooids was not strictly median- ventral, as Haeckel had described it, but took a curved course from a 'kink' or 'hilum' at the upper right-hand side, down to the lower left-hand one (Text-fig. 5E), was recently confirmed by the examination of several more specimens taken by 'Discovery', 'Carnegie VII ' and by Beebe, as well as of the Villefranche specimen, which is probably the most complete one to be found anywhere. In young specimens the ovoid pneumatophore rises with long axis vertically above the protosiphon, and its ventral side is covered by bracteal lamellae. There are, in a young specimen, in addition to the large protosiphon, three or four secondary gastrozooids on the right-ventral side, the youngest bud placed on a higher level. This arrangement is comparable with the stage at which the nectophores first appear in a larval Nanomia bijuga (Text-fig. 3). There is a corona of bracts attached to the 'nectostyle', which projects above the ventral side of the pneumatophore. Below the bracts there is, in an early stage, another corona of about six comparatively large palpons provided with palpacles, although these simple tentacles cannot always be seen. Later, more buds of secondary gastrozooids appear on the upper right-hand side, just in front of the nectostyle. In one full-grown specimen from Bermuda, 13 mm. in length, I counted eighteen gastrozooids, including buds. They form a staggered row, descending in a curved line running from the budding zone towards and round the left-hand side of the basal area (Text-fig. 5 E), as in Physophora (Text-fig. 5 C). By the time that they have all appeared the animal has ' toppled over ' to the ventral side (or end) so that the original vertical axis is now at an angle, and the apex of the pneumatophore is nearly covered by the nectostyle. The gonopalpons are produced on both right and left sides, in an order that has not yet been ascertained, and from each common, gonodendral stalk arise five or six palpons. In addition to the secondary gastrozooids already mentioned there is usually, I think, another gastrozooid budded near the protosiphon. The arrangement of the gastrozooids that I have described has not been noticed before. It is extraordinarily interesting because it links Athorybia not only with other Brachystelia (short-stemmed forms), especially Physophora but also with such Macrostelia as Nanomia bijuga. 4o DISCOVERY REPORTS It seems, then, that in Athorybia the pneumatophore is never carried up by a growing zone above the level of the young gastrozooids, and that an adult Athorybia is homologous in its general organization with a larva of Nanomia bijuga turned on its side. It is homologous also with a young Physalia in which the same turning has occurred, for the gastrozooids are in the same relative position to the pneumatophore as they are in a larval Nanomia bijuga (Text-fig. 5 A, B). Bracts. The basal part, to which the muscle is attached, is flattened from side to side. The main part of the bract is flattened dorso-ventrally. There are seven slight longitudinal ridges of nematocysts on the dorsal surface. When the ectoderm is stripped off no ridges remain. Only the median row reaches the proximal end. New Indian Ocean Records. 'Manihine' Aqaba Station 18, 1 example (juv.), bracts; Red Sea Station 10, 1 bract. 'Discovery II' Station 1589, 1 example (juv.), 1 bract. Other new records. 'Terra Nova' Station 57 (Tropical Atlantic, 13 May 1913), 1 example. ' Discovery '. Four specimens were taken by dip-net on 1 6 October 1 925 at a position 290 56' 50" N . , 150 03' 10" W. and fixed in Schaudin. A note on the label by Professor A. C. Hardy reads 'small fish [14 mm. long] was caught clasped by one of the Athorybia, but separated on being transferred from net to fixative '. 'Discovery' Stations 281, 282, 284, 299. 'Discovery II' Stations 689, 697, 698, 2069, 2648. Bermuda, Beebe Collection. Four specimens were recorded in a preliminary note under my name in 1936 as Anthophysa formosa Fewkes. Beebe took another specimen number 29453, net 267. H.M.S. 'Rodney'; Kingstown, St Vincent, B.W.I. , 20 February 193 1 (Totton Coll.); 1 example. Melophysa Haeckel, 1888 Type species Rhizophysa melo Q. & G., 1827 Eschscholtz (1829) included three species in his Athorybia. Of these, two — no doubt the same species — Physsophora (sic) rosacea Forskal and Rhizophysa heliantha Quoy & Gaimard belong to one genus Athorybia Eschscholtz, and the other Rhizophysa melo Q. & G. belongs to the quite distinct genus which Haeckel (18880) named Melophysa. Bigelow (191 1 b) rejected this name Melophysa on the grounds that Haeckel in his 'Challenger' Report (18886) referred to, but did not describe under this generic name, a single specimen taken in the Strait of Gibraltar in 1867. It was lost before Haeckel could draw it, but the form of its bracts was said to be similar to that in Athorybia melo, which Quoy & Gaimard had observed in the Strait of Gibraltar forty years before and figured in their 'Astrolabe' Atlas (pi. II, figs. 7-12). However, prior to publication of this 'Challenger' Report in 1888, Haeckel had published in December 1887 and again in May 1888 in his System der Siphonophoren the new name Melophysa, citing Rhizophysa melo Quoy & Gaimard as the only species. Perhaps this was overlooked by Bigelow, but, bearing it in mind, we can hardly agree that Haeckel set up a new genus for a problematical, undescribed and unfigured specimen that was soon lost. In fact I see no reason why Haeckel's name, Melophysa should not be used for Quoy & Gaimard's species, Rhizophysa melo. The only good account of this species is by Bigelow (193 1) under the name Athorybia rosacea. The probable origin of the confusion, both zoological and nomenclatural, is the fact that till figures were published by Chun (18976) it was not realized that Rhizophysa melo had a distinct nectosome with nectocalyces and that there was a characteristic difference between the heavy, ribbed bract of melo and the light, striated bract of rosacea. SYSTEMATIC AND BIOLOGICAL ACCOUNT 4i Melophysa melo (Q. & G.), 1827. The nectophores are shown in Text-fig. 7; and the bracts are like that figured by Bigelow (1931, fig. 217). New Indian Ocean records. 'Discovery II' Stations 1375, 1580, 1581, 1586, 1755, 1758, 1763. Other new records. 'Discovery' Stations 282, 284, 288; 'Discovery II' 692, 695, 701, 702, 2068, 2635, 2638. Beebe Collection, Bermuda. Nos. 29,266; 29,270; 29,349; 29>400; 301,072; 301,427; 311,008; 312,067; 312,071. Corrigendum. Items 27, 28 in my (1936) list of Wm Beebe's siphonophore material taken at Bermuda in his 193 1 Oceanographic Expedition refer to M. melo and not Athorybia rosacea. The name was used sensu Bigelow. Text-fig. 7. Nectophores of Melophysa melo. x 7. A, C, viewed from upper side; B, D, from abaxial end; E, from left side. A, from Bermuda, Beebe Collection, 1463m. B, 'Discovery II' St. 692, 350-0 m. C, D, E, 'Discovery II' St. 1580, 450-0 m. A post-larval specimen (Text-figs. 8, 9) taken by H. B. Moore at Bermuda in 1938-9 bears two gastrozooids. It is particularly interesting because (1) it still bears an attached larval bract with smooth outer surface, (2) the larval bract is attached near the tip of a long, conical nectostyle whose axis is at right angles to that of the pneumatophore, and not overlapping it as in A. rosacea. Only about half a dozen bracts were attached to the upper parts of the nectostyle, and buds for younger ones are attached much lower down on the ' body ' of the specimen. There appears to be a young nectophore bud at the base of the pneumatophore, but it is doubtful whether buds of different types can certainly be distinguished. The nectosome has not started to elongate. There is a ring of nine palpons. The length, from apex of nectostyle to mouth of protosiphon is 5 mm. ; the length of the attached bract is 7 mm. The tentilla are like those figured by Bigelow (193 1). 42 DISCOVERY REPORTS P Oz Text-fig. 8. Larva of Melophysa melo. A, dorsal view of the attached larval bract. B, C, lateral views of larva. A, B X4; C x 10. Coll. H. B. Moore, Bermuda. Text-fig. 9. Larval bracts of Melophysa melo. B, side-view of A. Coll. H. B. Moore, Bermuda, x 3-4. Physophora hydrostatica Forskal, 1775. This almost cosmopolitan species, which is especially common in the North Atlantic Gulf Stream area and the Mediterranean, has been shown by Kramp (1942) not to occur in the cold tracts of the Polar Currents. So far the only Indian Ocean record is that of Haeckel (1888 b) from off Ceylon. There are records from the Malaysian region by Lens & van Riemsdijk (1908); off South Australia (43 ° 26' S., 1260 37' E.), by Huxley (1859); the North-eastern Pacific, Bigelow (191 ib, 193 1) — I have examined young specimens from La Jolla, California — and North-western Pacific, Bigelow (1913). Bigelow (1931) pointed out that it had not been recorded from as high latitudes in the Pacific as in the Atlantic. From the South Pacific (Station 967) 'Discovery II ' took this species midway between New Zealand and South America at a temperature of 9-5° C, at the junction between the sub-antarctic South Pacific upper water and the eastern South Pacific central water. It was taken also at Station 943 off the east coast of New Zealand, at a temperature between 7-37° and 7-40° C. It was taken in the Pacific equatorial water by 'William Scoresby' in the Peru current (Sts. WS 687, 766) in 070 20' S. and 070 42' S. lat. SYSTEMATIC AND BIOLOGICAL ACCOUNT 43 In the South Atlantic it was taken by ' William Scoresby ' at several stations on the Patagonian Shelf between Magellan Straits and the Falkland Islands at temperatures of from 50 to 6° or 70 C. On the eastern side of the South Atlantic 'Discovery' found the species at Station 104, about seven degrees south of the Cape, and south of the sub-tropical convergence in lat. 410 33' 30" S. and long, if 58' W. Bigelow gave its known temperature range as between 7-2° and 2110 C. Kramp (1942) gave records from South-west Greenland at temperatures between about 30 and 470 C. ' Mabahiss ' took a small specimen at the surface, temperature 26-6° C, salinity 35-82 %0 at Station 39, about 80 miles north-west of Sokotra on 25 November 1933. The fact that Physophora hydrostatica was at the surface at such a high temperature is in itself of interest. No specimen has ever been recorded from this part of the Indian Ocean before. The only other Indian Ocean records are those of Haeckel (Discolabe quadrigata), but there may be some doubt whether all his Ceylon specimens belonged to this species (Garstang, 1946). The so-called record by Quoy & Gaimard (1833) of Physophora australis is based on figures which do not for certain, in my opinion, represent Physophora. The ' Mabahiss ' specimen is small, about 15 mm. overall, and has five attached buds of nectophores and six detached, crushed and deformed ones. All the palpons are missing, but one half-grown gastrozooid with a half-grown tentacle is attached, as are three younger ones, each with only a rudiment of a tentacle. Some of the tentilla on the former are unmistakably characteristic of similar stages in Physophora. The nectophores and the course of their radial canals have been compared and correspond with those of good Mediterranean specimens. The nectophore was well figured in Bigelow & Sears's 'Thor' Report (fig. 49 — a little masterpiece). The 'Mabahiss ' specimen is somewhat younger than that figured by Haeckel (18886; pi. xx, fig. 13). Like it, the small discoidal siphosome exhibits a one-sided dextral proliferation of buds. There are one or two small buds proximal to the youngest gastrozooids which may be young gonophores. Though the specimen is poor, incomplete and difficult to study, I have not the slightest doubt about its identity. A great deal more work is needed on the development and structure of this species. On 18 and 19 June 195 1, 'Scotia' took three young specimens to the north-west of the British Isles in the region of 190 N., 120 W. It was these that first made me doubt whether there is such a regular, dextral sequence of cormidia arranged in vertical, trimorphic series as maintained by Garstang (1946). My own tentative view is that the peripheral margin of the disc to which the cormidia are attached does indeed correspond with the ventral median line of a twisted macrostele stem. This does not mean that that is its phylogenetic history. I made a large collection of larvae of the species at Villefranche, and studied alive, anaesthetized, and preserved two full-grown specimens. There is no time to complete the morphological study of the material now in the British Museum collection before submitting this report, but a cursory examination of the ' Scotia ' specimens shows that the lower one of the two rows of palpons on the nascent lobe is a secondary one, and has not been caused by deformation, through squeezing, of a single row, as suggested by Garstang. It also suggests that the upper row itself is not a single sequence. This view is borne out by a study of the series of larvae. Garstang could not understand how the spiral winding of a uniserial trimorphic stem could, in theory, produce the parallel coronas ' as we see them ' in Physophora, but I think that Claus's (i860, 1878) interpretation of homology is quite correct, though the Physophora condition is probably primitive and the macrostele is derivative. Let us start with Text-fig. 10 [1], the generalized, uniserial, macrostele arrangement: theoretically this could give rise to, or be derived from [2] and [3]. In Physophora, omitting secondary palpons, we find the arrangement [4] . If the budding area is not studied closely we can easily get the impression, which Garstang seems to give, of an arrangement such as [5]. 6-2 44 DISCOVERY REPORTS Arrangements [i] and [5] are at first sight incompatible. Further study of available pickled material should clear up this point, though it could be done best in the field where musculature can be relaxed. A study of old figures and diagrams will not help much. Haeckel's figures of larvae are not reliable. He would have us believe, for instance, that the pneumatophore arises as the basal part of the bracteal canal. Physophora is an interesting animal to study alive: it is very active and responds readily to stimuli. If the air-sac is touched when the animal is at rest all the palpons are at once thrown up. Even a larva can be recognized by the irritability and activity of the palpons. In the adult they very often poke 0 0 X X • • 0 0 X X o O X o * o oc*>\+' XX ** cP,+ • 00000 X X X X X [I] [2] [3] W H Text-fig. 10. Diagrammatic representation of the possible steps in derivation of the 'Physophora' arrangement of stem appendages [5] from that of Agalma [1] or vice versa, x = gastrozooid ; O = palpon; # = gonophore. Text-fig. 1 1 . Physophora hydrostatica. Sketches from life of A, resting position on bottom of glass vessel ; B, swimming posture. Nat. size. about amongst the nectophores, and their tips even enter the nectosacs. When swimming (Text- fig. 11B) all the palpons are trained aft, their tips coming to a point outside the retracted tentacles. The air-sac on its relatively long neck turns vertically upwards, the main axis being horizontal. It is an active swimmer. A specimen measuring 6-5 cm. in length and having four pairs of fully developed nectophores moved about 5-0 cm./sec. The rate of contraction of the nectophores in a closed vessel at Villefranche on 28 April 1950 was about i/sec. In horizontal progression half of the nectophores had their orifices facing upwards, and only these nectophores contracted in unison. The initial con- traction when moving from the vertical resting position was made by the youngest nectophore. The simple tentacles of the palpons extended to about the same length as the palpons themselves, namely an inch or more. When the animal sank, the tentacles of the palpons trailed upwards. At rest they hung down inside the ring of palpons, which were of a rose-pink colour. The float was tipped with a plum-coloured pigment. SYSTEMATIC AND BIOLOGICAL ACCOUNT 45 Garstang tried to analyse the structure and development of Physophora, one of five types of brachy- stele Physonects, Rhodalia, Physophora, Athorybia, Epibula and Nectalia. He had, perforce, to rely on very unsatisfactory, old and partly restored figures for developmental stages. Thanks to a generous grant from the Browne Fund which I received in 1949 from the Royal Society, I was able to study alive at Villefranche two mature and some dozens of developmental stages of this species. There was a great deal of other work to be done at Villefranche and it was only recently, after staining the anaesthetized and fixed larvae, and after examining young specimens from La Jolla and others taken in 1951 by 'Scotia' off North-west Ireland, that I was able to gain a more correct idea of the mor- phology and ontogeny. It would not be profitable to mention all the mistakes that have been made in the interpretation of the morphology and development of this species. Haeckel was not, I think, able to distinguish the different types of bud from one another, so that his figures of developmental stages do not make it clear which are palpon-, gastrozooid- and nectophore-buds. His ' Challenger ' figures of larvae give the erroneous impression that the pneumatophore is contained in the mesogloea of the bract. How to orientate the larva is a puzzle. In one place Haeckel calls the bracteal cleft dorsal, as does Garstang, and in another ventral. Perhaps the meridian in which the second gastrozooid is budded, close to that in which the first palpon is budded, and near the bracteal cleft, might be called ventral, and the region of attachment of the nectophore buds dorsal. The meridian, in which the tentacle of the protosiphon (first gastrozooid) and bract lie, would then be close to the ventral. But there are no two well-marked and opposite meridians. The function of the bract seems to be to provide cover for the retracted tentacle. The picture Garstang gave of a young stage of development was that of a protosiphon hanging down from the middle of a basal stem surface. This consisted of a wide space surrounded by a ring of palpons. Round this open space he pictured the buds of siphons, gonodendra, and other palpons circulating, with much jostling and squeezing, as they advanced from a marginal budding zone in three parallel whorls, grouped into cormidia along the meridians. My own observations show that this is not a good picture of what takes place. There is no open space until after most of the appendages have been formed and have grown apart. The first palpons are not, I believe, budded regularly in one meridian, though the last formed ones may be. For instance, palpon II appears commonly to be budded to the right of palpon I, palpon III to its left and palpon IV between I and II. Later on, palpon-buds appear in other meridians, between and distal to earlier formed ones. The second gastrozooid is budded quite early on, when there are only two or three palpons. When the bract drops off, its position is taken by palpons, which by that time form a complete ring. The first palpons are budded from a region on the apical side of the constriction that separates the pneumatochone from pneumatosaccus, at a time when the pneumatochone extends down to the level of the distal end of the basigaster. As the growth of the budding zone of the nectophores carries the pneumatophore upwards and away from the ring of the palpons, the external wall, to which they are attached, increases in diameter, but not to an equal extent in all meridians. The radius of curvature increases steadily as we pass from the growing point. At first sight, it might be thought that the marginal area, to which palpons are attached, should be regarded as part of the nectosome, because of their relative larval position. But this area is not muscular like the nectosome proper. Since there is virtually no special growing zone between the level of attachment of the palpons and the base of the protosiphon, there is little to compare with the siphosome of Macrostelia. Any argument, therefore, as to whether the siphosome in Physophora forms a twisted stem round an imaginary central axis is unnecessary. But if the basal part of the proto- siphon was extended by a growing zone, we should get something very like a macrostele stem. Examina- tion of hundreds of larvae of Nanomia bijuga1, a macrostele, shows that the budding zone of the 1 Commonly known as Stephanomia bijuga. 46 DISCOVERY REPORTS siphosome bears a great resemblance to that in Physophora, and forms a pointed outgrowth (Text- fig- 3)- As Haeckel truly stated, the gonodendra develop on meridians between those on which the gastrozooids are found, but on their apical side. Haeckel's figure, on the contrary, would give the impression that successive groups of palpon, gonodendron and gastrozooid were each borne on one meridian. The whole arrangement of the cormidia is very like that found in Rhodaliids, but whereas the nectophores, in that family, form a bilateral corona, in Physophora the budding zone of the nectophores is carried up in an apical direction. Whether their muscular lamellae are all attached to a single meridian or not needs further, careful determination. Successive lamellae overlap one another, but the pedicular canals may lie in one meridian. Physophora is a highly specialized form, and it is unlikely that its larva would tell one much about its phylogeny. I agree with Garstang in thinking that there is no reason why it should ever have had a macrostele ancestor. On the other hand the elongation of the proximal part of the protosiphon of one of its ancestors conceivably might have given rise to a macrostele type of Physonect. ■ 'Stephanomia' No one can be sure at present what is the identity of a siphosome originally figured by Lesueur & Petit (1807-1 1, Atlas) which they called Stephanomia amphytridis . Peron (1807-16) did not mention it by name in his text (Tome 1, p. 45) but gave a short description of the living siphosome, its feeding habits, its phosphorescence, and a reference to Lesueur's and Petit's figure. Strictly speaking, then, the name was published by Lesueur & Petit. I have found in Peron's text no clue as to where it was taken — Bigelow (191 1 b, p. 288) said, 'in the Atlantic' — but in Peron's narrative it was mentioned after that part of the text which dealt with arrival at He de France. It might, therefore, have been taken anywhere between Le Havre and Mauritius. There are two courses open to systematists when dealing with this name and with long-stemmed Physonects that have single terminal filaments to their tentilla. The first is to use the name ' Stephanomia ' as a temporary convenience and to acknowledge that S. amphytridis of Lesueur & Petit is at present unidentifiable ; and therefore cannot be the name of the type of a genus Stephanomia. The second course is to abandon the generic name Stephanomia and to take instead one that can be used for a species of which topotype specimens, or a type specimen, can be re-examined — specimens, that is, of a long-stemmed form with single terminal filaments. For the present I am taking the first course, but I am excluding and placing in a new genus named after Mr J. W. S. Marr three species found chiefly in the Arctic and Antarctic. In the present state of our knowledge, I think it would be unwise to attempt to identify specimens of 'Stephanomia' (sensu Bigelow) unless they bear nectophores, for nectophores do, I believe, have constant specific characters which establish their identity. Lesueur's, Huxley's and Bigelow's famous specimens of Stephanomia lacked the nectosome, and no nectophores have even been described under the name of amphytridis or amphitridis. Indeed, there are few proper descriptions of the nectophores of any species of ' Stephanomia '. Different species of Physonects may have nectophores that are superficially alike, so great care must be taken with their identification. It is generally necessary to stain1 both the radial canals and the ridges. Figures of nectophores published long ago are not, as a rule, to be relied upon. I recommend the use of very dilute borax-carmine for the canals, and of Delafield's haematoxylin for ridges. (See p. 10.) SYSTEMATIC AND BIOLOGICAL ACCOUNT 47 Stephanomia rubra (Vogt), 1852. (Text-figs. 12-18.) Agalmoptis punctata Kolliker, 18536. This well known, but not well described or figured Mediterranean species, which I have examined alive at Villefranche, and of which there is a good deal of material preserved in the British Museum (Nat. Hist.) collections, was taken by 'Manihine' in the Gulf of Aqaba; and in the Red Sea and west Indian Ocean by 'Discovery II'. Although specimens living at Villefranche are easy to identify from the sum of the characters, and because of the restricted fauna, it is not easy to establish the presence of this species in another fauna by deduction from the identification of loose nectophores taken in the tow-net. But this, as a rule, is the only sort of evidence that is available, and for many years past I have been working to achieve such an aim. There is a species of Physonect whose nectophores can easily be confused with those of Stephanomia rubra by the less expert, and that is Agalma elegans. Of course the living or complete animal cannot be confused, on account of their very different tentacles. If we compare the nectophores of the two species, and bear in mind that fully grown nectophores look very different from young ones, we find the following distinctions: 1. General shape of nectophores. If nectophores of A. elegans and Stephanomia rubra are examined together as they lie, abaxial sides uppermost on the bottom of a watch-glass of formalin, it will be seen that the paired lateral wings or wedges of Agalma elegans are proportionately longer than those of Stephanomia rubra. It will also be noticed that on each side, on a level with the ventral (adaxial) wall of the nectosac of Agalma elegans, there is a vertical ridge that runs from the upper (apical abaxial) to the lower (oral adaxial) lateral ridge. In other words, the lateral facets, in A. elegans, have the ' latero-ventral (adaxial) triangle ' cut off by a complete vertical ridge to form a separate facet. It will be seen, too, that in A. elegans the upper (apical, abaxial) lateral ridge forms a prominent tooth half- way between the ostium and the ventral (adaxial) wall of the nectosac. In Stephanomia rubra there is sometimes a slight prominence at this point, but usually not a well-marked tooth. In a strictly lateral view of Agalma elegans (Text-fig. 24 A) the vertical ridge, described above, can be seen, with appropriate staining, running down in line with the adaxial (ventral) wall of the nectosac : and it will be noticed that the upper part of the adaxial, median ' thrust-block ' (in which runs the pedicular canal), does not project so far in A. elegans (Text-fig. 24 B) as it does in Stephanomia rubra (Text-fig. 12 C,B,D). The much greater length of the long (abaxial-adaxial) axis of the nectophore when compared with that of the vertical (oro-aboral) axis is another distinguishing feature of the nectophore of. Agalma elegans. 2. Lateral radial canals. Another useful distinguishing feature of Stephanomia rubra is the way in which the lateral radial canal dips down as it runs on to the lateral face of the nectosac, before making its ascending, abaxial run to the semicircular dorso-lateral curve, and descent to the circular velar canal. In a strictly lateral view this dipping down of the lateral radial canal as it runs on to the lateral face of the nectosac takes it well below the level of the inner end of the pedicular canal, and below the horizontal part of the lateral canal that runs on the adaxial wall of the nectosac (Text-fig. 12D). In Agalma elegans the lateral canal does not dip down noticeably as it runs on to the lateral wall of the nectosac, but runs horizontally (Text-fig. 24 A). It should be pointed out that in both A. elegans and Stephanomia rubra and many other Physonects, immature nectophores, when viewed from above (apically), have the upper lateral ridges somewhat folded in towards the sagittal plane, so that care must be taken, when comparing nectophores, to allow for immaturity. 48 o / 2 j/nrn - -R vert lot -W/Ot -Wlat Crodd ,-C.rad.lat Text-fig. 12. Stephanomia rubra. A, C, side and upper view of nectophores from 'Manihine' Aqaba St. i. A x 7; C x 12; B, D, side views of Mediterranean specimen from Villefranche, x 8 ; E, underside of nectophore of Mediterranean specimen, x8. 49 Text-fig. 13. Stephanomia rubra. Bract of a Mediterranean specimen from Villefranche, x 12. A shows the tip of B enlarged, x 47. Text-fig. 14. Stephanomia rubra. Bracts of a Mediterranean specimen from Villefranche. B, C, young bracts, x 20, to show apical position of pad of nematocysts. A, E x6;D xn. D XXVII - 5° Text-fig. 15. Stephanomia rubra. Bracts from 'Discovery II' St. 1587. A, B, D, Ex 7. C, tip of bract D, x 38. The pecked area of the bracteal canal shows the area of attachment of the muscular lamella. Text-fig. 16. Stephanomia rubra. A, B, E, upper, lower, and side views of nectophores of type ' e\ from 'Discovery II' St. 1585; C, D, F, similar views of specimen from 'Discovery II' St. 1586 (type '/'). x 6-5. Si Text-fig. 17. Stephanomia rubra. A, C, upper and side views of nectophore from 'Mabahiss' St. 145; B, upper view of nectophore from 'Discovery II' St. 1586. xo. Text-fig. 18. Stephanomia rubra. Bracts from 'Discovery II ' St. 1586. x 6. The pecked area of the bracteal canal shows the area of attachment of the muscular lamella. 7-2 52 DISCOVERY REPORTS It is interesting to find such a close similarity between nectophores of Agalma elegans and Stephanomia rubra because their tentilla are so very different. Bracts. In Mediterranean specimens the bracteal canal rises towards the dorsal (abaxial) surface near the tip and forms an enlarged subterminal, elongated, cylindrical vessel provided with about 200 nematocysts (see Kolliker, 18536, Tab. iv). This subterminal section of the bracteal canal lies under a short ridge. In the youngest bracts the subterminal section is bent at right-angles to the main axis, and so is terminal (Text-fig. 14 B, C). In Indian Ocean collections, as indeed in those from the Atlantic, there are many small nectophores and bracts that closely resemble those of the Mediterranean species S. rubra, and also some more complete specimens that bear parts of both nectosome and siphosome. There are differences in shape between many of these nectophores and some well-preserved ones of S. rubra that I was able to anaesthetize at Villefranche. But in the present state of our knowledge it would be premature to establish a new species for the Indian Ocean forms. For a long time it seemed possible to divide the Indian Ocean nectophores into two categories, one, type '/', appearing to be somewhat more compressed or shortened in the abaxial-adaxial axis than the other type 'e\ The appearance of the two types of nectophore as they lie flat on the bottom of a watch-glass is quite different (cf. Text-figs. 16 A, B, E — type ' e\ and C, D, F — type '/'). But if a different view-point is taken of the type '/' nectophore shown in Text-fig. 16C, we get the appearance shown in Text-fig. 17B, which is that of type le\ Although in the Mediterranean form of S. rubra there is only a very small mouth-plate in the nectophore, if any at all, in these Indian Ocean nectophores the mouth-plate is often conspicuous. Their canal system is indistinguishable from that of S. rubra. Their bracts are very similar to those of the Mediterranean form but the bracteal canal does not rise up abruptly towards the terminal ridge as it does in the Mediterranean form. Tentatively I regard all these specimens as belonging to S. rubra. New Indian Ocean records. ' Manihine ' Aqaba Station 1 ; ten small but well-preserved nectophores. 'Discovery II' Stations 1567, 1568, 1585, 1586, 1587, 1589, 2688, 2907 (Red Sea). Much more work must be done on the large amount of available Atlantic material before the size-range of specimens and the variation in form of nectophores and bracts can be determined. It is still uncertain whether there are other closely related species. The nectophores described and figured in my (1932) Great Barrier Reef report as Agalma sp. indet. belong to Stephanomia rubra. Nanomia bijuga (Delle Chiaje), 1841. This species has been mentioned a good deal in the literature under many names : Physsophora bijuga Delle Chiaje, 1841. Anthemodes canariensis Haeckel, 18696. Halistemma pictum Metchnikoff, 1870. H. tergestinum Claus, 1878. Agalmopsis picta Fewkes, 1880. Cupulita picta Chun, 1897a; Kawamura, 191 ib. Stephanomia bijuga Bigelow, 191 ib (see synonymy). This species was not taken in the Indian Ocean either by 'Mabahiss' or by 'Manihine'. 'Discovery II' took only a very few, small nectophores as follows: St. 1 587 45°-° m- TYF B One nectophore. St. 1587 450-0 m. N 70 B Two small nectophores. St. 1589 600-0 m. N 70 B One small nectophore. SYSTEMATIC AND BIOLOGICAL ACCOUNT 53 Attention may be called to Kawamura's excellent figures (191 16; pi. 7, figs. 1-10) of CupuKta picta, which have escaped mention in the synonymy lists. He took many specimens at Misaki in the months of January and February. He noted that they carried from ten to fifty nectophores and more than ten cormidia. The bracts and their distribution were described by Kawamura. He said they were generally in four longitudinal rows round the stem, each siphon having an extremely large and long bract close to it. The only colour he noted was that of the reddish cnidoband of the tentillum. He agreed that Stephanomia bijuga is not identical with Agassiz's Nanomia cara, but he had not resolved the confusion of at least two species under Sars's (1846) name Agalmopsis elegans, and in error said in his generic diagnosis that Cupulita had trifid tentilla. His description of C. picta, however, was good, and his figure of the whole animal is perhaps the best in existence. His figures of the nectophores are excellent and characteristic, and show that they are quite different from those of the species figured by Sars (1846; Tab. 6, fig. 3) and renamed by Haeckel (18886) C. sarsii. The Siphonophore whose abundance at Salcombe, Devon, was noted by Berrill (1930) but not studied morphologically, was probably C. sarsii and not Stephanomia bijuga. The oil-filled diverticulum at the base of the palpon appears to be characteristic of Cupulita sarsii. Forty years ago Bigelow raised the question whether the southern Stephanomia bijuga (Cupulita picta Metchnikoff) was distinct from a northern form Nanomia cara Agassiz. I have examined specimens labelled cara by Bigelow and taken by ' Albatross II ' (St. 213 1 5) just north of Cape Cod, and other specimens that agree with them but differ from those of the Mediterranean species bijuga. These other specimens were collected (1) by Miss Delap at Valentia Island, Ireland, where I myself collected a mature specimen on 3 May 195 1 ; (2) in the Celtic Sea between Ireland and Cornwall, and sent to me by the Marine Biological Association, Plymouth; (3) by ' Scotia' in 195 1 to the north-west of the British Isles; (4) in Norwegian waters, a single necto- phore. The nectophores of all these specimens agree with each other, and with the figure made by Sars, in being flattened at right-angles to the long axis of the whole animal. They differ from those of Nanomia bijuga ( = Cupulita picta) which are flattened in a plane parallel with the long axis of the whole animal. I have therefore come to the conclusion that there are two closely allied species (a) Cupulita sarsii Haeckel, 18886, found in Norwegian and British waters, and probably identical with the North-west Atlantic form Nanomia cara Agassiz, and (b) a Tropical Atlantic, Mediterranean, West Indian, Indian Ocean and warm-water Pacific species, referred to by Bigelow (191 lb) as Stephanomia bijuga. As to applying to these two related species the prior generic name valid for either of them, my opinion is as follows: Stephanomia (1807) is the name of a monotypic genus, the identity of whose only species amphytridis Les. & Petit is exceedingly doubtful. Cupulita (1824) is also monotypic, and the identity of its only species, bowditchii Q. & G., is very doubtful. Agalmopsis Sars, 1846, was set up for two species; and was subsequently used by Kolliker (18536) for the component with trifid tentilla. But most of Sars's account and all the figures, except for two of the tentilla of his form b (an Agalma) refer to my species (a). Agalmopsis was restricted by Haeckel (18886, p. 234) to the other component with trifid tentilla. There seems to be little doubt that Agalmopsis is the prior and correct name of the genus in question. Halistemma (1859) was published by Huxley for a species from East Australian waters, represented only by the siphosome of a specimen. The identity of Huxley's species has not been established. And so we come to the name Nanomia Agassiz, 1863. The figures of nectophores and palpons given by Agassiz (1865) for N. cara show the relationship of this species to, and its differences from bijuga ; and the differences of both species from species of other genera. I give three views of a nectophore of N. cara (Cupulita sarsii) taken by me in May 195 1 on the ebb tide, off the landing stage of Valentia Island, Co. Kerry, Ireland; and a view of a nectophore of Nanomia bijuga taken 54 DISCOVERY REPORTS at Villefranche (Text-fig. 19). In order to avoid confusion at the moment I shall use the name Nanomia. A a I— 3/nm 1 Rap .-/at. C. radio t ^R.vert /at Nemp ect ■C rod lot 3 mm 1 Text-fig. 19. A, B, C, upper, abaxial and side views of nectophore of Nanomia cara from Valentia Harbour, Kerry, Ireland, x 12. D, upper view of nectophore of Nanomia bijuga from Villefranche for comparison with A. There were colour differences between N. bijuga, as observed alive at Villefranche on 23 March 1949, and Cupulita sarsii as observed alive at Valentia Island, Co. Kerry, Ireland on 3 May 1951. N. bijuga N. cara (Cupulita sarsii) 1. Tip of air sac Plum coloured Light red 2. Nectophores (a) Young (b) Oldest five pairs (c) Elder dozen or more Red patch on each side. Red patch in middle of upper side, with sometimes one or two red patches on either side of the dorsal patch Red pigment all the way round circular, velar canal Colour round velar canal lost or absent entirely Some or all of colour round velar canal lost or absent No pigment 3. Stem of nectosome Numerous irregular flecks (half bands) of red Irregular flecks 4. Tentilla Red Red 5. Gonads Red pigment round bases 6. Gastrozooids Irregular spots of red, six round mouth, two circlets above and below basal bulb Small dark plum-coloured spot on side next to ten- tacle, in basal half only 7. Siphosome Numerous irregular patches of red on stem Irregular flecks 8. Palpons Red pigment round base : band or patch of red round the terminal part Red pigment round base of oil globule SYSTEMATIC AND BIOLOGICAL ACCOUNT 55 The gastrozooids of C. sarsii appeared pink as the animals were carried along in the ebb tide off the pier at Valentia Island. Further observations on the pigmentation of this species are desirable. At times it is abundant at Valentia Island. Marrus gen.n. Type species Marrus antarcticus sp.n. Holotype (in fragments) from 'Discovery II' Station 943 Species of this genus are known from fragments only. The lateral radial canals are unlooped, the tentilla are unicornuate. Marrus antarcticus sp.n. I have named the new genus to which this large Physonect from antarctic and southern sub- antarctic waters belongs after Mr J. W. S. Marr, who has had more than thirty years active association with the Antarctic and has always treated siphonophore material with loving care. He also made useful colour notes on this Siphonophore when it was alive. It appears to be closely related to a species ' Stephanomia' orthocanna from deep cold water in Baffin Bay described by Kramp in 1942, and since collected by 'Scotia' at Station 688 in 1951 at 1100 m., some 50 W. of the south-west end of the Faroe-Shetland Channel.1 Through the kindness of Kramp I have been able to compare nectophores, stems and siphons. The nectophores of both have straight lateral canals and their other details are very similar. The siphons of both have a long, cylindrical basigaster, and are in other respects similar. Unfortunately, the northern specimens, whose stems show numerous attachments of bracts, were not associated with either attached or unattached bracts except for one specimen which had one small attached one. As described by Kramp, the bract is of a different shape from those so characteristic of Marrus antarcticus. The specimens of the northern form had very few tentilla, so that a satisfactory description of them has not yet been made. Description. Nectosome at present known only from loose nectophores. Nectophores measuring uptoi7Xi5X9 mm., lateral facets triangular but not cut off by an oblique cross ridge; two broad lateral wedges, and broad median lappet (thrust block) ; dorso-lateral ridge bifurcated distally. These ridges are present, though not easily seen except after staining, also in 'Stephanomia' orthocanna, but not figured by Kramp. Nectosac lacks musculature on its broad basal (adaxial) face as in' S.' orthocanna. The blind end extends in the form of two widely rounded lateral arms; lateral-radial canals unlooped. Siphosome long, coiled in tight turns when preserved, apparently unsegmented, about 6 mm. in diameter when contracted and preserved. The appendages are all budded from the ventral, non- muscular groove. Gastrozooids of a specimen from 'Discovery II' Station 2010, reached a length of 30 mm., of which about half is the cylindrical basigaster. Another gastrozooid (St. WS 552 E) measured 27 mm., the basigaster accounting for 5 mm. It was dilated with food at the distal end which measured 5 mm. in diameter. A third gastrozooid, 15-5 mm. in length, from the same 'William Scoresby' Station 552 E consists of three parts: (1) the basal, cylindrical basigaster, which measures 6-omm. in length and i-omm. in diameter; (2) the stomach proper with 'liver streaks', which measures 7-25 mm. in length and 1-25 mm. in diameter; and (3) the oral section (closed) which measures 2-25 mm. in length and 0-5 mm. in diameter. Mature tentilla are unicornuate, have no involucrum and a simple, straight canal. The pedicel of a retracted tentillum measures about o-8 mm., the cnidoband, in three coils, 0-9 mm. and the filament 1 mm. in length. A more extended tentillum 1 Dr J. H. Fraser of the Scottish Fisheries Department has been kind enough to inform me that evidence from hydrographic data as well as from planktonic data would suggest that there is very little outflow, if any, into the Atlantic through the Faroe-Shetland Channel even in deep water. He thinks that it is much more likely that 'S.' orthocanna and other cold-water plankton species taken in haul S 51/688 had penetrated southwards via the area west of the Faroes than through the channel. S6 DISCOVERY REPORTS measures: pedicel i mm., cnidoband 2-5 mm., filament 5 mm. in length. In young stages the tentilla are finger-shaped, with a patch of large, pavement-epithelial cells over the developing cnidoband: the three sections show histological differences at an early stage. The filament contains numerous smaller ovoid nematocysts (? isorhizas) measuring 18/x in length and 5/x in diameter, as well as larger ovoid nematocysts, type unidentified, measuring 18 x 13/x. The cnidoband has on each side a row of about 125 large nematocysts (? mastigophores) measuring 50x20/^. In between these are about 5000 finger-shaped nematocysts (? anisorhizas) regularly arranged in a dozen rows measuring jn in diameter and 40^ in length. Their outer ends, which form a pavement, are overlain by large C.rad.d. Crad/at C.po C.ped "Crodv. Crad.v Cro'dlot C.ped c)pa Text-fig. 20. Marrus antarcticus gen.n., sp.n., 'Discovery II' St. 943, 1000-750111. Nectophore; A, underside; B, upper side; C, lateral view; D, adaxial view, x 4-6. The muscular lamella is attached along the pecked continuation of the pedicular canal. polygonal, epithelial cells measuring 0-04 x 0-02 mm. Often the exploded cnidoband may be found on bracts. Palpons are of the same length as siphons, but more slender. No filament has been observed at the base. Some of them, probably all, arise from the gonodendra. Gonophores observed have all been female and monovon (Text-fig. 21 E). The largest eggs measure about 1-25 mm. in diameter. The position of the gonophores relative to the gastrozooids has not been determined. Dissection of female gonophores with a micro-manipulator showed that the endoderm of the outer thin coat (umbrella) forms four or more radial canals, and that there is no terminal opening. Inside this outer tunic is another, formed by the irregularly branched spadix (manubrium) (cf. Vogt, 1854, Tab. 10, fig. 25; and Huxley, 1859, pi. vi, figs. 13, 14; pi. ix, fig. 14) that is flattened out by the large ovum. When the ovum is partly grown the spadix grows up to one side of it and wraps branches SYSTEMATIC AND BIOLOGICAL ACCOUNT 57 round it, but leaves an oval area on the other side free. Under the spadix lies the nucleus. A similar condition was figured by Weismann (1883, Taf. xxn, and p. 203) for Agalma rubrum {Stephanomia rubra). Bracts are very characteristic in shape (Text-fig. 21). The bracteal canal terminates on a papilla on the concave terminal facet. The presence of young bracteal buds of this characteristic form is the criterion for identification of large pieces of the siphosome, and at once distinguish such pieces from those of another antarctic physonect Stephanomia convolnta, whose young bracts have a terminal sausage-shaped mass of nematocysts lying in the sagittal plane of the bracts, separating the two oblique, dorso-lateral, terminal facets. The muscular lamellae of bracts and buds of bracts are closely crowded together in at least six rows on either side of the gonophores in the contracted and preserved material, and most of the gastrozooids are detached. Text-fig. 21. Marrus antarcticus gen.n., sp.n., WS St. 552E. A, B, two views of adult bract; C, D, side of young bracts; E, gonophore, x 13. A, B, C, D x 12. Colour. There are colour notes on Antarctic physonects by members of ' Discovery ' staff made on four occasions on board while sorting the plankton. They related to specimens of two physonect species that turn out to be other than Pyrostephos vanhoeffeni. It has been a difficult matter to apply these notes because on each of the four occasions (Sts. 890, 2001, 2006 and 2010) all three Antarctic physonects were present in the catch; but, bearing in mind that it is possible that I have made some mistake, the following collectors' notes appear to apply to Marrus antarcticus: (1) Station 890, N 100 B. A large part of the siphosome (no nectosome) labelled 'Pyrostephos vanhoeffeni ' ; ' typical deep brown colouring [probably applies to gastrozooids] central stalk portion bright pink'. (2) Station 2001. 'Stolon deep orange crimson [this appears to refer to a specimen of the necto- some, without nectophores, of Stephanomia convoluta]. Coloured part of bells [Marrus antarcticus] very rich deep orange red.' 58 DISCOVERY REPORTS (3) Station 2006. ' Stolon [this appears to refer to two specimens of the nectosome, without necto- phores, of Stephanomia convoluta] and coloured part of swimming bells [Marrus antarcticus] very deep rich scarlet orange.' (4) Station 2010. Part of the siphosome and numerous appendages of Marrus antarcticus: ' Stolon milk white, tinged with a lovely rose pink, large tasters [? gastrozooids] deep purple brown: faint purple-red streak [canal] in bracts, also in swimming bells.' Distribution: a a rr A. Antarctic Zone (1) Boundary depths between the cold bottom water (which must have a small northward creep as well as eastward movement) and the warm deep layer (in which the water must be assumed to have a small southward creep as well as eastward movement) : (2) Warm deep water: 'Discovery II' Stations Depth (m.) 661 1500-1000 663 1 500- 1 000 2001 I7S0-1 3°° 2006 1750-1400 2010 700-400 ' Discovery II ' Stations Depth (m.) 1728 800-350 WS552 290-190 B.A.N.Z.A.R. Exp. Station Depth (m.) 3i 1 000-0 the warm deep layei : and the c 'Discovery II' Station Depth (m.) 890 240-110 B. Sub-antarctic Zone (4) Water layer into which the antarctic surface water sinks when it sinks below the surface at the antarctic convergence: Discovery II' Stations Depth (m.) 671 943 1000-750 1000-750 ing antarctic water Discovery II' Station Depth (m.) 2023 1500-1000 I am obliged to Dr G. E. R. Deacon for identifying the water-masses concerned, and for suggesting the generalization that Marrus antarcticus thus appears to be associated with the antarctic zone and with water that has recently left the zone, or to live in close proximity to such water. It appears to be a deep water or mid-water form; but a scarcity of very deep net hauls prevents the formation of conclusions about its distribution at great depths. The figured holotype (a nectophore) came, with other fragments, from 'Discovery II ' Station 943, and bears the Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 26 (Text-fig. 20). SYSTEMATIC AND BIOLOGICAL ACCOUNT 59 Marrus orthocannoides sp.n. A haul of the young fish trawl from 1400 to 700 m. at 'Discovery II' Station 1585 in the West Tropical Indian Ocean contained loose nectophores (Text-figs. 22, 23), bracts and denuded stems (nectosome and part of siphosome) of a new Physonect that appears to be closely akin to the high-arctic species ' Stephanomia' orthocanna Kramp. Dr H. Bargmann of 'Discovery Investigations ' has kindly made an approximate estimate of what would have been the temperature and salinity reading at the time and place of capture. Her figures are : Depth (m.) T. °C. S.%o 1400 4-85 34-84 700 8-35 34-98 _ A-C.ped C.radd Crad.v. -Cradlat C.ped Text-fig. 22. Marrus orthocannoides sp.n. Upper and side views of a nectophore from 'Discovery II' St. 1585, 1400-700 m. x 6. I have already described (p. 55) a new antarctic Physonect which appears, in the present state of our knowledge, to be congeneric with 'S.' orthocanna, and owing to the fact that, since the name ' Stephanomia ' is used in a loose way for a number of macrostele Physonects, whose tentilla have single terminal filaments, these three species have been placed in the new genus Marrus. The study of each of these three related species has been difficult because of the fragmentary condition of the material. The nectophores of the three related species are characterized by the absence of any oblique ridge dividing the lateral facets. On the dorsal surface there is a pair of ridges dividing it from the lateral surface. Each divides again before reaching the ostium of the nectosac. These ridges are not shown in the figures by Kramp (1942) of ' Stepha?iomia' orthocanna, but, after staining a nectophore kindly 8-2 60 DISCOVERY REPORTS supplied by Kramp, I am satisfied that these ridges are present, though not well marked. In all three species the place of attachment to the muscular lamella is deeply sunk, as Kramp says, in a narrow longitudinal furrow. The lateral radial canals follow a straight course to the ring canal. There is a ventral, medianly cleft mouth-plate. Whilst in 'S.' orthocanna, and in its antarctic counterpart men- tioned on p. 59, the proximal side of the nectosac, bearing the junction of the canals, is free of muscular tissue, in orthocannoid.es it is not so, nor is the wall of the nectosac of Marrus orthocannoides sunken in as it is in those two species. Cbr -rS.mm Text-fig. 23. Alarms orthocannoides sp.n. A, fully grown; B, C, young bracts, x 10. The pecked portion of the bracteal canal indicates the area of attachment of the bracteal muscle. Bracts. In a row ( ? rows) flanking the palpons, thin and leaf-like, the largest about 1 5 x 8 mm., ovate, without marginal teeth. There are signs of thickened oblique truncate distal ends. The bracts' attachment to the muscular lamella takes place along the proximal two-thirds of its length, so that the terminal canal measures about one-third of the bract length. This canal appears to terminate in the middle of one margin of the end-facet, without any sign of a pit or other surface marking. All the bracts are detached. Palpons. Very numerous, slender, 9-2 mm. long x 0-2 mm., on short pedicels, close set in four to five longitudinal rows, each with a fine tentacle. Gonodendra. Fine, about 4-6 mm. long, two or three per segment of stem, bearing a dozen or more male gonophores on very short pedicels. The gonophores are globular, the largest measures i-i mm. in diameter. As the gonophores of Marrus antarcticus were all female, it appears that Marrus may be dioecious. SYSTEMATIC AND BIOLOGICAL ACCOUNT 61 Stem (? contracted). Air-sac oblong, 6x2 mm.; nectosome 14x2 mm.; siphosome 23 cm. long, 4-6 mm. measured dorso-ventrally, 3-7 mm. side to side, very muscular, segmented. Gastrozooids and tentilla missing. Other appendages (not well preserved) in a ventral band 2-7-3-7 mm- wide. I take this opportunity to publish a second record of the capture of Marrus (Stephanomia) orthocanna Kramp by ' Scotia ' on 31 August 1951 in haul 688. Locality: 590 50' N., 1 1° 24' W., two loose necto- phores, which I have compared with type material. The figured holotype of Marrus orthocannoides bears the Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 4. Agalma elegans (Sars), 1846. Agalmopsis sarsii Kolliker, 1S536. I have found no adult specimens, not even loose nectophores, of this species amongst the 20,000 siphonophore specimens that I have examined from the Gulf of Aden, or from the Somali or Arabian basins of the Indian Ocean. But in November 1950 and January 1951 'Manihine' took eighty-four small nectophores at five stations in the Red Sea (16-17 January 195 1) at a distance of from 5 to 35 miles east of Sanganeb Lighthouse, off Port Sudan, where the temperature must have been about 21-5-22° C. Also 'Discovery II ' took a young specimen at Station 2906 in the Red Sea. It had two gastrozooids (no tentilla are visible) and two nectophores 4 mm. in diameter, which I have compared carefully with Mediterranean specimens (Text-fig. 24). The only other specimens that I have found Text-fig. 24. Side and upper views of nectophore of Agalma elegans from Villefranche, x 8. in the whole of the vast amount of ' Discovery ' material from the Indian or Atlantic Oceans were some well-preserved young ones, of typical form taken at Station 273, differing in no way from those that I have examined from Villefranche, the Celtic Sea (M.B.A. Plymouth 'Mackerel Cruises'), Valentia harbour (E. T. Browne collection) and to the north-west of the British Isles ('Scotia' 1951, St. 683, 6o° 18' N., 12° 20' W.). I must mention that Bigelow's figure (191 ib; pi. 19, fig. 2) of a nectophore does not perhaps give a very good idea of the typical form. In fact for years this figure misled me into mis-identifying the juvenile nectophores of Agalma okenii, which have only one ridge on the lateral facet (see p. 64) as those of A. elegans. I give this warning because Bigelow's famous report is a standard reference work, and his figure is the only one likely to be used by planktonologists. Leuckart (1854; pi. 12, fig. 22) illustrated well the characteristic dorsal view, and showed its difference from that of Stephanomia rubra, which he illustrated on the same plate in fig. 12. Leuckart's drawing of the 62 DISCOVERY REPORTS lateral radial canals, however, is not very good. My new drawing of the nectophore (Text-fig. 24) is still not quite satisfactory. From 'Discovery II' (East African line) Stations 1567, 1568, 1581, 1583, 1584, 1586 and 1587 come larvae with bracts typical of larvae thought to be those of this species. These bracts are identical in shape with those of larvae taken with adult specimens in the Celtic Sea, and also with those of a series of post larvae taken by 'Scotia' and kindly sent by Dr J. H. Fraser of Aberdeen. A study of the ' Scotia ' specimens has enabled me to say that the well-known larval Agalmid stage (see Haeckel, 18886; pi. xxi, fig. 12) in which larval bracts cover the air-sac and all buds, is followed by one in which the nectosome elongates and emerges from the whorl of larval bracts. The last of these, however, remain for a time attached by their muscular lamellae to their supporting prominence, and have been found in specimens 3 or 4 mm. in length, after the appearance of a second functional gastrozooid, and even in the terminal cormidium of a specimen 7 cm. in length. The buds of the nectophores now grow and begin to function. There is a second smaller gastrozooid and a ring of from four to eight palpons present at this stage, which is prior to the elongation of the siphosome in the reverse direction to that of the nectosome and to the appearance of a succession of secondary stem groups of buds. The stage before the appearance of secondary groups is reminiscent of the adult Nectalia loligo, and might be referred to as the ' Nectalia-stage ' in the same way in which earlier stages are referred to as ' Athorybia- larvae ' because of their resemblance to miniature specimens of Athorybia rosacea. In the ' Nectalia- stage ' the terminal gastrozooids are surrounded, just as they are in Nectalia, by a number of long, trifid bracts of the general shape figured by Haeckel (18886; pi. xm, figs. 1, 9-12). They are very much like those of Athorybia rosacea and, since they are probably pre-cormidial as Garstang suggested, they may be homologous with them and with the larval bracts of A. okenii. It remains to work out in Agalma elegans the succession of bracteal types and their situation on the stem. A slightly later stage at which four or five nectophores are functional, is represented by a series of beautiful specimens taken at 09.15 hr. at Villefranche on 28 March 1949 near the surface in a i-m. net. In these, the larger of the two terminal gastrozooids described above for the ' Nectalia-stage ' bears an adult-type tentacle (but with only two turns of the cnidoband, and the involucrum either covering it completely or only partially on one and the same tentacle), whilst the smaller one bears a tentacle with kidney-shaped (larval-type) tentilla. If anyone was able to devote the spring of a year like 1949, when Agalma elegans was abundant at Villefranche, to watching the development of this species he would, I think, be able to confirm that in the course of its life-history it passed through successive stages reminiscent of Athorybia, Melophysa and Nectalia1 and would therefore come to the conclusion that species of these genera have become neotenic at successively later evolutionary stages. I take this opportunity of publishing the following data for a live specimen 12 cm. in length, taken at Villefranche on 14 April 1949. Nectosome: 3-5 cm. long, 1 cm. diameter; ten nectophores on one side and nine on the other, and buds. Siphosome: three complete gastrozooids, one young one. Gastrozooids 5 cm. apart. The terminal gastrozooid was surrounded by a ring of palpons. But the tentacle of one ' palpon ' bore larval type tentilla so that it was probably the secondary gastrozooid. The rest had ordinary thread-like tentacles. The terminal gastrozooid bore a typical tentacle 12 cm. 1 Nectalia Haeckel, 18886, should be accepted with some caution, since specimens with the very peculiar tentilla that Haeckel figured have not certainly been seen again, and his figures are often not dependable. Bigelow's (191 1 b, pi. 20, figs. 4, 5) specimen may be a young stage of a long-stemmed form. I have examined many such specimens taken at Bermuda by Beebe, but a great deal more work has still to be done on the large amount of Agalmid material sent me by Beebe before the question can be resolved. I hope eventually to be able to link up these Nectalia-like specimens with a giant form that I provisionally called Stephanomia amphitridis. They have a new type of nectophore. SYSTEMATIC AND BIOLOGICAL ACCOUNT 63 or more in length (expanded) with twenty-seven tentilla, 5 mm. apart when expanded. The expanded horns of the tentilla measured 3 mm. in length. I have examined closely the arrangement of the gonophores in a specimen that I anaesthetized and fixed in formalin at Villefranche on 5 May 1949. The female gonophore is at the proximal end of each segment, just distal to the last gastrozooid. It has a plain stalk which subdivides distally. Following it is a gap, and then male gonophores and palpons are scattered over more than half the length of the segment. Another gap follows and then the gastrozooid and a group of palpons are found. A newly budded gastrozooid is accompanied by a single palpon. More are added in the course of growth in the region of the male gonophores. The general arrangement was shown quite clearly, if a little crudely, by Kolliker (18536, pi. 3). The pneumatophores of live specimens (7-12 cm. in length) varied in shape, presumably according to the state of air-secretion. In jars of sea-water at Villefranche, specimens at rest took up a vertical position, but swam round horizontally at the surface. On 5 May 1949, the calm water off Cap Ferrat, Villefranche, was full of Agalma elegans and Stephanomia rubra, and six dozen of the former were picked out with dip-nets at 8.45 a.m. on this fine sunny morning (wind S.E., light). Specimens had fourteen to twenty necto- phores a side, six gastrozooid groups, and were about 17 cm. in length (contracted). It may be noted that specimens like these can be lifted out of the water on a flat dip-net and jolted into jars of sea-water without damage. Kawamura (191 1 b) studied alive a species which he identified as Agalma elegans. He captured and illustrated one perfect specimen 26 cm. in length at Misaki on 29 January 19 10, and observed, without being able to catch them, a number of much larger specimens that were swimming at some depth from the surface. From a glance at Kawamura's figures (191 ib, pi. 7) it will be seen that there are two important differences between the figure of a Misaki and a Villefranche nectophore. One is the abaxial shift in position of the vertical ridge in Kawamura's specimen, and the other is the way in which the lateral radial canal, in his fig. 13, crosses the lateral face of the nectosac. If these prove to be constant features of a Pacific form it will merit specific rank. But Kawamura's fig. 1 3 does not appear to have been so carefully drawn as others. The upper and lower vertical ' branches ' of the pedicular canal described by Kawamura form the area of attachment of the muscular lamella. Kawamura commented on the length of the paired horns of the tentilla, and figured them as somewhat less than 1 mm. in length. I measured them in an expanded, live specimen and found them to be three times this length. I found also that specimens (from 7 to 12 cm. in length) were very susceptible to vibration and contracted violently. Agalma elegans was hitherto known to occur in the Indian Ocean only from Browne's (1926) record of a post larva with a couple of not very characteristic bracts, and without nectophores. It has a protosiphon, a ring of palpons, and many buds. It is a pity that Haeckel (18886, p. 234) restricted the name Agalmopsis elegans to that one of Sars's two species which had trifid tentilla, because nowhere did Sars describe or figure the rest of a specimen of that species. All his figures and most of his text, based on notes made during the eight years 1835 to 1842, is descriptive of what Agassiz (1865) later called Nanomia cara. Kolliker (18536) had already given the Agalmid a new specific name sarsii. Haeckel seems, mistakenly, to have thought that the whole of Sars's plate 6 referred to the specimens with trifid tentilla. He seems, too, to have overlooked the fact that eight years before his description of Agalmopsis was published, Fewkes (1880, p. 135) restricted Agalmopsis elegans to the species (now Nanomia cara) that had a single terminal filament. But I dare not cause further nomenclatural havoc by putting these matters right now. The best thing to do is to have the name put on the list of nomina conservanda. We can be certain, however, 64 DISCOVERY REPORTS of the identities of Sars's two species because of their frequent association in these northern waters. Agalma okenii Eschscholtz, 1825. Plethosoma cristalloides Lesson 1826; 1838. (For synonymy see Bigelow 191 lb.) Kawamura (1911a) studied specimens at Misaki in the winter. He observed, described and figured specimens of what he considered to be two separate species, Agalma okenii Eschscholtz and Crystallomia polygonata Dana. Specialists for the last forty years have thought that the two names were synonyms of one and the same species, but Kawamura's criteria are clearly set out and illustrated on his plate 16. Much research will be needed before his claim can be either supported or refuted. He considered one of the characters of the smaller C. polygonata to be that in its nectophores there is only one vertical ridge on the lateral facet, so that it has only two parts instead of the three found in Agalma okenii, though reference to Dana's (1858) original figure shows that the nectophores had two vertical ridges. -C ped Text-fig. 25. Three views of a young nectophore of Agalma okenii from the Gulf of Aqaba, 'Manihine' St. 1, x 21. Only one ridge is developed on the lateral facet. Meanwhile, I am able to contribute towards what I believe to be the solution of the problem of the existence of two types of nectophore associated with okenii type bracts, the two differing only in the number of the longitudinal ridges on the lateral facets. The solution is that Kawamura's smaller Crystallomia polygonata is a young stage of Agalma okenii. A long search through accumulated material of A. okenii at last revealed the presence of two young specimens of A. okenii each about 1-5 cm. in length having only two functional gastrozooids. A few larval type bracts were still attached, and the budding nectophores all possessed two longitudinal ridges on the lateral facets. But each specimen was associated with one or two nectophores, 8 mm. in diameter, that possessed only one longitudinal ridge. I tentatively suggest, therefore, that the earliest nectophores, like the larval bracts, may have only one ridge whilst the subsequently budded bracts and nectophores have more. The critical speci- mens upon which this view is based are : 'Terra Nova', 1910-13: Station 32. A series of nine young specimens taken in the tropical Atlantic south of the Cape Verde Islands. The seven youngest specimens possess nectophores and buds of the 1 -ridge type. The two elder ones possess nectophore buds of the adult (2-ridge) type. Station 63. Two young specimens taken a little further south than the last. One of them had still attached to it a nectophore of the juvenile (1 -ridge) type, and was associated with detached nectophores of the same (1 -ridge) type. In the catch also were pieces of a larger specimen and detached necto- phores of the adult (2-ridge) type. SYSTEMATIC AND BIOLOGICAL ACCOUNT 65 Station 107. Three young specimens were taken off Three Kings Islands, North Island, New Zealand. One of these had nine attached nectophores of the juvenile (1 -ridge) type as well as buds of the same type, but not of the adult type. There were present also detached nectophores of the adult (2-ridge) type. Mr Fraser-Brunnei 's collection (Gulf of Aden). On the night of 26 November 1949, Mr Fraser- Brunner took at the surface, ten miles north-east of Alayu one young specimen (only two gastrozooids developed) that had still attached to it one nectophore of the larval (1 -ridge) type, as well as a larval bract. It was associated with one detached nectophore of the adult (2-ridge) type. 'Discovery II' Station 1598, 460-300 m. (tropical Atlantic), 1 juv. ex. 1-5 cm. in length with one attached nectophore possessing a single longitudinal ridge on its lateral facets. It was associated with two more young specimens, one of which still had two bracts of the larval type attached near the Text-fig. 26. Larval bracts of Agalma okenii from the Gulf of Aqaba. A:-J x 5 ; H1, H2 x 9. The numerals indicate that more than one view is given of a particular bract. penultimate gastrozooid. Four gastrozooids are present, a reduced terminal one — the protozooid, but no tentacle is visible — two functional gastrozooids and one well-developed bud. The bracts are similar to those shown in Text-fig. 26 D, E. Dr W. Beebe's collection (Bermuda, where the species appears to be plentiful). A young specimen numbered 312117, net 1314, and 1-5 cm. in length, has some larval type bracts, and only two functional gastrozooids. There is no sign of larval type tentilla and the budding nectophores all possess two longitudinal ridges on the lateral facets. Associated in the phial with this specimen were two unattached nectophores, 8 mm. in diameter, and some unattached larval type bracts. One of these nectophores had only one ridge, the other had traces of a second. H.M.S. 'Challenger' in October 1950 took some larvae and post-larvae about 30 miles off the Pacific coast of Costa Rica. There were in the catch small detached nectophores of the 1 -ridge type. The temperature of the surface water, in which the larvae were probably living, though the tow was made at a depth of 27 m., was 26-6° C. Long ago Bigelow (191 1 b) dealt with Lens & van Riemsdijk's (1908) attempt to split the species Agalma okenii on the form of the tentillum, which is one of Kawamura's criteria. 66 DISCOVERY REPORTS The number of specimens taken by 'Discovery II' or 'Manihine' in the Indian Ocean was small when one remembers that, quantitative considerations apart, at a single station, No. 4600, in the Eastern Tropical Pacific 'Albatross' took fifty-seven complete specimens, as well as innumerable loose cormidia and nectophores. At three stations off the East African Coast, ' Discovery II ' took four small specimens, fifty-six small nectophores and fifty-two bracts, while at six Aqaba stations ' Manihine ' took two very small specimens, three small nectophores and a small bract, as well as seven specimens of what appears to be the little known larvae of this species, and twenty-one loose larval bracts. These bracts are the field-mark of these larvae. They differ considerably in shape from those of A. elegans. Mr Fraser-Brunner, however, took many specimens in the Gulf of Aden as indicated below. Material. The following specimens belong to the stage in which the nectophores have a single ('equatorial') lateral, longitudinal ridge: ' Manihine ' Aqaba Depth Complete ex. and Small loose Small loose Post- Loose larval Station (m.) cormidia nectophores bracts larvae bracts 1 183 1 3 12 13 130-0 2 3 17 180-0 2 4 23 183 1 1 1 • 25 180-0 . . 1 29 180-0 1 2 'Discovery' Stations: SS 43 (1 nectophore); 288 (2 juv. ex.); 299 (nects); ' Discovery II ' 679 (1 ex. and nects); 689 (1 nect.); 690 (nects and bracts); 693 (part of ex. and cormidia); 694 (nects); 695 (2 ex. and nects); 696 (cormidia and nects); 698 (nects); 704 (nects); 705 (nects); 1372 (cormidia and nects); 1373 (1 ex., nects and bracts) ; 1567 (nects and bracts); 1580 (nects); 1581 (1 juv. ex., nects and bracts); 1585 (4 juv. ex., cormidia and nects); 1586 (nects and bracts); 1752 (1 ex., no nects); 2648 (nects). 'Terra Nova' 1910-13 Stations: 45, Atlantic (nectophores); 56, Atlantic (1 nectophore); 311, 350 29' S., 510 21' W. (1 nectophore). The following specimens belong to the typical adult stage in which there is a second lateral, longi- tudinal ridge on the nectophores : 'Discovery II' Stations: 1375 (nects and bract); 1576 (1 ex.); 1580 (nects and bracts); 1585 (6 juv. nects); 1747 (1 ex. and cormidia); 1759 (nects and cormidia); 1760 (2 adult ex.); 2681 (6 adult ex.). The specimens from Station 1760, where the surface temperature was 21-22° C. are from 5 to 6 cm. in length, and the siphosome consists of from ten to twelve segments. The terminal (? primary) gastrozooid is much smaller than those in the middle region, and has a reduced basigaster. The bracts are arranged in five rows on each side in the middle region and in three rows near the end. Numerous specimens with nectophores of the 2-ridge type were taken in Bermuda by Beebe. A fine specimen of Agalma okenii (siphosome only) from ' Discovery II ' Station 1752 is nearly 6 cm. in length. Some of the tentilla have as many as sixteen turns of the cnidoband, and the involucre covers from six to nine of them. The first seven turns are larger in diameter, and are flanked by the larger, elongated type of nematocysts, whose threads can penetrate through at least 6 mm. of bracteal jelly. The terminal eight or nine turns, which are not so flanked, are smaller in diameter. In the absence of nectophores this specimen cannot be placed in either of the above groups of specimens, though it is typical of the species. 'Terra Nova' 19 10-13 Stations: Station 46, 1 specimen possessing buds of nectophores and detached nectophores of the adult (2-ridge) type : 56 (1 nectophore) ; 59 (nectophores) ; 69 (nectophores). SYSTEMATIC AND BIOLOGICAL ACCOUNT 67 Mr Fraser-Brunner's collection (Gulf of Aden). Off Aden (night, 5. xii. 1948) several adult specimens (and five larvae); off Aden (day, 6. xii. 1948) parts of several adult specimens; 20 miles north-west of Bulhar one small specimen with adult (2-ridge) type nectophores. 'S.V. Cap Pilar' (Seligman collection). At a position in 350 21-5' S., 66° 22-5' E. (to the south of 'Discovery II' Station 1754) at the surface on 30 April 1937 an adult specimen was taken bearing twelve to thirteen gastrozooids. Larvae. There is no record of larvae having been bred since Haeckel (1869 a, Crystallodes rigidum) watched their development in the Canary Island of Lanzerote in the winter of 1866-7, though Lens & van Riemsdijk (1908) recorded what is probably a series of thirty-four post-larvae of this species under the name ' Crystallomia spec, group I '. Twenty-three of these formed a series, specimens increasing gradually in length from 1-5 to 12 mm. Four of them were in the monogastric stage, but bore no larval bracts, except one which evidently had a bud of one. Lens & van Riemsdijk described the succession of types of tentilla. It will therefore be of interest to illustrate some post-larvae taken by ' Manihine ' in the Gulf of Aqaba Station A 1 (Text-fig. 27). Their specific determination depends Text-fig. 27. Larva of Agalma okenii from the Gulf of Aqaba, 'Manihine' St. 1, x 23. on the facts that they have trifid tentilla like Agalma spp. and that the bracts differ from those of the only other Agalma species elegans, but agree with those figured by Haeckel for Crystallodes rigidum ( = Agalma okenii). The three interesting monogastric larvae bearing larval bracts and a larval type of tentillum were taken at 'Manihine's' Aqaba Station Ai made on 31 December 1948 in 15 fathoms. The smallest of the tentilla measured about 0-2 x 0-13 mm., having three tiny processes at the distal end, one turn of the cnidoband, and a number of nematocysts (0-01-0-06 mm. long) set obliquely at the base (cf Lens & van Riemsdijk, 1908; fig. 98, pi. xm, 'type 1 '). Dawydoff (1928) has given a warning that accounts of siphonophore development prior to those of Metchnikoff must be treated with caution. This warning should be extended to Dawydoff 's own short summary, which contains false statements as well as interpretations. The remarkable allegation by Haeckel ( 1 869 a) that the pneumatophore originated as a diverticulum of the common ' Primitivhole ' was by implication denied in 18886, when he said that in all cases the pneumatophore arose as a gland- like invagination of the exoderm. It is very difficult to see the structural details of these siphonophore larvae when they are alive, except in the early stages. I am concerned here only with the larval bracts figured by Haeckel. They have a shape that is characteristic, and is shown in lateral view by Haeckel's (1869 a) plate vn, fig. 51. The upper and outer faces are nearly at right-angles to one another, and the outer face is divided by a longitudinal ridge into two facets (Text-fig. 27). 9.2 68 DISCOVERY REPORTS Material. A. ' Manihine ', Gulf of Aqaba Stations No. of necto- Station Date (Jan. 1949) Hour Depth (m.) T. °C. S-%o phores (when counted) i 15th 1300-1415 0 21-59 40-79 6 137 — 40-78 — 274 — 40-38 — 2 15th iMS^S00 0 21-65 40-78 — 137 21-65 40-71 1 5 15th 1150-1410 0 137 274 21-61 40-72 40-68 40-80 1 6 16th 1010-1115 0 21-56 40-74 1 J37 18+ 40-80 — 274 18+ 40-79 — 8 16th i35°-I44° 0 21-72 40-79 3 9 1 6th HSS-^0 0 21-57 4°-75 10 274 — 40-79 — 548 18 + 40-66 — 1 1 17th 1105-1125 . 0 21-51 40-76 2 12 17th 1015-1045 0 2i-53 40-77 13 13 17th 1240-1330 0 22-0 40-74 2 IS 1 8th 0930-1000 0 21-35 40-71 6 18 1 8th 1310-1415 0 21-57 40-78 16 J9 1 8th 1430-1530 0 21-58 40-78 3 137 — 40-74 — 823 18 + 40-66 — 21 19th 1 120-1200 0 21-56 40-70 1 137 18 + 40-74 — 548 18 + 40-64 — 24 19th 0900-0930 0 21-20 4°-75 1 29 2ISt 1000-1030 0 21-5 40-66 4 3° 2ISt 1100-1150 0 22-38 40-50 3 137 18 + 40-66 — 548 18 + 40-68 — B. R.R.S. 'Discovery II' No. of necto- Station Date (Apr. 1935) Hour Depth (m.) T. °C. S-/io phores (when counted) 1581 28th 1815 50 22-29 35-26 1 100 1721 35-3° — 1583 30th 0930 0 29-53 35-32 1 50 27-18 35-37 — SYSTEMATIC AND BIOLOGICAL ACCOUNT 69 I have re-examined Browne's ' Sealark ' specimens from Mauritius and Farquhar (he did not keep them apart). They are well preserved and show stages of development a little later than those of the Aqaba ones. Two are post-larvae in the monogastric stage, one still bearing a larval bract, and the other one three larval bracts. They are about 3 mm. in length. A third specimen is 10 mm. in length, and two others about 18 mm. These last two have gastrozooids, and one has lost at least seven nectophores. All bear buds of nectophores, and two of them large ones. A sixth specimen is a single cormidium to which is attached three bracts. The largest of these is of adult type, o-8 mm. broad, with four primary facets and two outer secondary ones. The larva of A. okenii differs from that of A. elegans in having none of the simple, larval type of tentilla, so well known in Nanomia bijuga. The explanation of this is that this simple type of tentillum in Agalma elegans is borne by the tentacle of a reduced type of gastrozooid intermediate in type between an ordinary gastrozooid and a palpon, and that there is no gastrozooid of this type accom- panying the normal, terminal gastrozooid in A. okenii. The larval bract of A. okenii is comparable with that of A. elegans, but its proximal end makes an angle with the distal one. In A. elegans six or seven bracts of this larval type, much enlarged, may still be attached to the terminal cormidium when the specimen has reached a length of 7 cm. Cordagalma cordiformis Totton, 1932. This must be the smallest of all Physonects, and it has been overlooked by everyone. I described it first in 1932 from the Great Barrier Reef of Australia. I have found it in daylight close to the surface in the West Indies, and on many occasions at Villefranche, and am now able to report its presence at sixteen of the 'Manihine' Aqaba Stations and at 'Discovery II' Stations 1581 and 1583 (see Tables A and B on p. 68). It was not taken by ' Mabahiss '. My figure, published in 1932, will serve to identify the nectophores of this very obscure species, about which we know so little at present. Measurements. A nectophore was taken at random from the catch made by 'Manihine ' Station 18. Its length is 2-0 mm., breadth 1-2 mm., and thickness i-i mm. measured dorso-ventrally. The other nectophores are of similar size. Bargmannia elongata gen.n., sp.n.1 This species has long been known to me by its very characteristic nectophores, and since they are very unlike those of species of known genera, I have made this species the type of a new genus. Two views of a nectophore of this species from 'Discovery II' Station 681, which I had already prepared, are now published for the identification of this very characteristic nectophore, but no description will be attempted, beyond saying that the lateral canals do not form loops as they do so commonly. In this respect it resembles Kramp's (1942) ' Stephanomia' orthocanna and related species. Distribution: 'Discovery' Stations 107, 'Discovery II' 681, 686, 698, 699, 701, 1554, 1585, 1586, 1587, 1589, 1769, 2061. Holotype (Text-fig. 28 D) and eighteen paratype nectophores Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 7-25 from 'Discovery II' Station 699, 370-0 m. I have found a single nectophore in a tube of material submitted by Dr Leloup from ' Michael Sars ' Expedition Station 58 (12-13 January 1910, 370 37' N., 290 25' W.). 1 Named for Dr Helene Bargmann. 70 DISCOVERY REPORTS Text-fig. 28. Characteristic nectophores of a new Physonect Bargmannia elongata, the other parts of which are unknown. A, B, specimen 'Discovery II' St. 681, x 3-2; C, D, specimen from 'Discovery II' St. 699, 370-0 m., x6'3; E, F, specimen from 'Discovery II' St. 1769, 1000-750 m., x 3-5. The muscular lamella is attached along the pecked line in figs. E and F. Forskalia Kolliker, 18536 Very little advance has been made in our meagre knowledge of species of Forskalia since Bedot (1893), who was familiar with living Mediterranean forms, reviewed the genus. He recognized five species, three of them Mediterranean ones. I am familiar with one living Mediterranean species, whose identification depends, as a field-mark, on a lemon-yellow coloured spot at the junction of the ventral radial and circular canals of the nectophores. This spot is not visible after fixation in formalin. Schneider (1898) maintained that this species should bear the name ophiura Delle Chiaje 1829. Bedot called it edzvardsii Kolliker, 18536. I have also often seen preserved nectophores of the leuckarti type, namely those with a small disc-shaped rete on the pedicular (apical) canal. Very little information has been given in descriptions about the shape and disposition of bracts, but I believe that bracteal characters will prove to have great systematic value. Leuckart (1854) gave the impression that, in F. contorta1, the bracts on the polyp-stalk are very numerous. In F. edwardsii I estimate that there are ten, including the small distal buds. We really know very little about the different species. That quantitative processes of change in animal tissues produce allometric growth-rates is well known. Child and others have done much classical work on the growth gradients in gymnoblast 1 Redescribed by Bedot (1893) as F. leuckarti. SYSTEMATIC AND BIOLOGICAL ACCOUNT 71 hydroids, but I am not aware of any work that may have been done on gradient phenomena in other coelenterate bud-colonies. That such phenomena occur, for instance, in Aglaophenid hydroids becomes evident when the denticulation of the theca-margin from the stem-region to the tip of a hydrocladium is examined; and there are axial gradients, too, from top to base of the colony. Lately I have come across similar growth-gradient phenomena whilst making a tedious study of the bracts of F. edwardsii, the Mediterranean species with the lemon-yellow spot on the velum. In no account of any species of Forskalia can a satisfactory description or figures of the morphology be found, especially of the bracts. Forskalia edwardsii Kolliker, 18536. In the spring of 1949, at Villefranche, I took the opportunity to anaesthetize and carefully preserve parts of thirty-two large colonies of F. edzuardsii, and to obtain some photographic records1 of the living animal (PI. IV, figs. 4, 5, 8) — a very difficult procedure. Examination of a small, random sample of the loose, preserved bracts of a specimen of F. edwardsii showed that it was easy to pick out bracts of certain shape-categories, and that each category appears in two forms, one the mirror-image of the other. Further examination of the specimen after staining — the bracts are very transparent — showed that the stem bore leaf-like bracts, and that the long, muscular peduncles of the siphons, so characteristic of Forskalia spp., carried bracts of at least three other kinds, the proximal pair being Cbr Jmm Text-fig. 29. Forskalia edwardsii. Bracts of a specimen captured at Villefranche, 31 March 1949, x 4. The three bracts, B, are of a kind found only on the stem. C, a bolster-shaped bract found only on the upper part of the pedicels of the gastrozooids. A, a series of eight stem bracts to illustrate constancy of shape. All bracts have enantiomorphs. Homologous points on the bracts are lettered a, b, c. of the most simple form (Text-fig. 29 C). After making a series of camera-lucida drawings of numbers of each category, it was apparent that the shapes of bracts at different relative positions on the peduncle were fairly constant and characteristic. Finally I was able to homologize their structure. Though it appeared in life, while appendages, some of them almost invisible, were writhing about under the binocular, that there was no discernible ordered arrangement, it is now clear that all the bracts of F. edzvardsii have essentially the same structure, but that their proportions in different positions on the stem and peduncles vary markedly. The change in form takes place suddenly in passing from stem to peduncle, but progressively from the proximal to the distal end of each peduncle. It remains to be seen whether there is further progressive change from the anterior to the posterior end of the whole bud-colony. 1 I have to thank Dr W. G. R. Marden for co-operation in this difficult task. 72 DISCOVERY REPORTS The small Forskalia specimens occasionally brought back in plankton-collections are invariably strongly contracted. No doubt the sudden contraction in length of siphon-peduncles caused by formalin has the mechanical effect of tearing the long and delicate membranous, muscular attachments of the bracts. Sometimes a few distal bract-buds remain, but siphons and bracts are nearly always shed. In cross-section the long, peduncular bracts of F. edwardsii are triangular, and have a distal facet set on at varying angles. As a result the distal facet is of varying length, relatively short in the proximal pair, relatively long and oblique at the distal end. The feature of all the bracts, which makes it possible to homologize the various categories, is a shallow pocket between the distal facet and what I will call the dorsal facet. By labelling the two distal teeth of the dorsal facet a and b in the figures (Text-figs. 29, 30), and attaching other letters c, d and e to the distal tooth, the two median teeth of the distal facet and the penultimate tooth respectively, it is hoped to make the homologies clear. \4.mm 2 mm. Text-fig. 30. Forskalia edwardsii, Villefranche. A-E, knee-shaped bracts from pedicels of gastrozooids. F, bolster-shaped bract from base of pedicel. A-C x 6, D, E x 1 1, F x 9. Homologous points on the various bracts are lettered similarly, a, b, c, d, e. It will be seen that in a cross-section of a stem-bract the longer transverse axis is from side to side, and the ' pocket ' is very wide, whilst in the distal peduncular bract the longer transverse axis might be called dorso-ventral, and the 'pocket' is very narrow. In January 1952 I received four bottles of plankton containing Forskalia material, taken in the Red Sea by ' Discovery II ' in November 195 1. The material is in better condition than usual, and although it cannot be worked up for this report it can be said that it corresponds well with specimens of F. edwardsii from the Mediterranean. The material consisted of: 'Discovery II' Net Depth (m.) Material Station 2905 Station 2906 TYFB N70B TYFB N70B 250-0 200-0 Two young colonies Pieces of several colonies and young ones SYSTEMATIC AND BIOLOGICAL ACCOUNT 73 Forskalia sp. or spp. A species was taken at one 'Manihine' Aqaba Station and in the following 'Discovery II' hauls: 1375 (x ex-> no nectophores); 1568 (nectosome and 31 nectophores) ; 1573 (5 nectophores) ; 1581 (4Juv. nectophores) ; 1583, 100-50 m. (1 juv. nectophore); 1585, 1400-700 m. (11 nectophores), 500-0 m. (11 bracts); 1586, 1650-950 m. (1 nectophore); 1587 (1 ex., no appendages). The above nectophores are unmistakably those of a Forskalia species and have the characteristically shaped nectosac, with straight radial canals. The flattened asymmetrical ventral process that carries the pedicular canal is also characteristic. But the pedicular canal itself is noteworthy for lacking the branched expansion in its middle region found in F. leuckarti. Cbr Pr.lat. y-Prdist Text-fig. 31. Knee-shaped bracts of a species of Forskalia from 'Discovery II' St. 1585, 1400-700 m., x 15. Of the above bracts, five are similar (Text-fig. 31) in having a concave oval facet on the dorsal side, a cavity on the ventral side, a convex facet at right angles to the dorsal one, a three-cornered lateral process with a ridge and a canal that has a right-angled bend in the middle of its course. All the bracts are somewhat hollowed on one side and have a distal process, triangular in section, on which the canal ends. The knee-shaped bracts described above can be mirror images of each other, as the figures show. At present we do not know enough about the morphology of species of Forskalia to be able to identify species from loose bracts and nectophores, unless they are of living Mediterranean specimens. CALYCOPHORAE Post-larvae of Prayidae and Hippoppdiidae It has long been recognized that Prayids and Hippopodiids more nearly resemble one another than either resembles the remainder of the Calycophorae. In both of the families there appears to be a similar larval, caducous nectophore, combined possibly with a bract, like the one first figured by Chun (1888 a) for the Canary Islands Hippopodius hippopus. Up the present time we have known very little about the larvae of any Prayid or Hippopodiid species, and as far as I am aware no one has succeeded in breeding any except Metchnikoff (1874). Eighty years ago he reared some larvae of H. hippopus at Villefranche, but could not keep them alive for more than ten days. The rest of our knowledge is based on larvae of unknown parentage captured in the sea. Great caution should be used in accepting identification of such larvae unless there is within the hydroecium a young, definitive nectophore of recognizable shape. 74 DISCOVERY REPORTS Quite recently, in June 195 1, I was able to identify such larvae of a Hippopodiid and a Prayid from the Antarctic ice-edge. But we know only one Antarctic species of each family, so that after seriation of growth stages we can reasonably assume that these ice-edge larvae are those of Vogtia serrata and Rosacea plicata. These last grow to a considerable size, a diameter of 12 mm. The larvae of Vogtia serrata and Rosacea plicata can be separated by the shape of the somatocysts. Without attempting to go into detail it may be stated that in larvae of R. plicata the somatocyst is ovate or globular, whilst in Vogtia serrata it is elongated. There appears to be a much more pronounced angle in the dorsal wall of the hydroecium of the larval nectophore in Rosacea plicata than in Vogtia serrata. Both have comparatively straight lateral subumbral canals, those of Rosacea plicata being somewhat sigmoid in shape. The post-larva of R. plicata was recorded by Bigelow (19110) as Hippopodius hippopus in his report on Biscayan Siphonophores. I have re-examined some of these ' larvae ' from ' Research ' Station 36K. One is an eudoxid, and three are post-larvae of Rosacea plicata (Text-fig. 32). Som Cradlat Som Gz- Text-fig. 32. Rosacea plicata. Larva from 'Research' (1900) St. 36^, Bay of Biscay. A, lateral view of whole larva, x6; B, larval nectophore, x 24 ; C, view of nectophore from aboral end, x 22. Thanks to the generosity of Dr Tregouboff, Director of the Station Zoologique at Villefranche, where so much of the pioneer work on Siphonophores was carried out, I have been able to examine a very well-preserved post-larva of Vogtia glabra (PI. IV, fig. 2 ; Text-fig. 33). It is identifiable because within the hydroecium of the larval nectophore is a young, definitive, heteromorph nectophore characteristic of the species.1 Its identification is important because the larval nectosac shows clearly that only dorsal and ventral radial canals are developed at this stage in V. glabra, if my identification is correct. Both types of shed larval nectophores, i.e. with two and with four radial subumbral canals, have been found from time to time in plankton samples, and would constitute distributional records for the species concerned. In Chun's (18886) figures of Naples larvae he omitted the dorsal canal (inserted by Moser in a copy, 1925). R.R.S. 'Discovery II' has taken many dozens of post-larvae of Rosacea plicata measuring up to 12 mm. in length. There is a clear discontinuity between these and the smallest definitive nectophores. From this fact I deduce that the larval nectophore of R. plicata is caducous, as in Hippopodius. Moser (19240) figured larval nectophores of Rosacea cymbiformis, Vogtia serrata and V.pentacantha. The subjects of her plates (pi. 1, fig. 3 ; pi. 11, fig. 5 ; and pi. in, fig. 4) contained a young definitive nectophore, but the last two cannot be identified with much certainty. 1 Only two prominences are present above the nectosac and not four. In the larva from Orotava, attributed by Chun (i888«) to Hippopodius hippopus, he did not mention the number of prominences, but figured what looked like one of a pair only. But Chun's larva had four radial subumbral canals. SYSTEMATIC AND BIOLOGICAL ACCOUNT 75 We are now in a position to see how closely alike are the larvae of two species of Prayidae and of three species of Hippopodiidae, for Chun (18880, b, 1913) had already figured three larvae of Hippo- podius hippopus. In all except Rosacea cymbiformis there is a deep, pocket-shaped hydroecium, and in none is there an extension of the somatocyst distad to the pedicular canal of the nectosac. All, except Vogtia glabra, where they are missing, have simple straight, lateral subumbral canals such as are found in Desmophyes. Reference to Bigelow's (191 \b) figure of Nectopyramis diomedeae will show that these same features appear in the definitive nectophore of the species, which I consider to be a good instance of a neotenous form which has subsequently evolved on new lines. We do not, of course, know whether the common ancestor had a very different adult stage. Text-fig. 34 will make this discussion plain. CpeoJ N.detl Nsdefi- ■ Nlorv- - Text-fig. 33. Vogtia glabra. Prof. Tregouboff's larva from Villefranche, x 19-5. For photograph, see PI. IV, fig. 2. It may be mentioned at this point that the definitive nectophores of Hippopodiid species resemble the larval type common to Hippopodiidae and Prayidae in that there is no distal prolongation of the somatocyst beyond the pedicular canal of the nectosac. There is present in Prayid and Hippopodiid larval nectophores an ovate or globular structure that lies between the pallial canal (somatocyst) and the gastrozooid. In older stages it is transformed or absorbed into the stalk canal that traverses the muscular lamella. It has been considered that the free end of the pallial canal in Prayids is homologous with the somatocyst in other Calycophorae, but the origin of the rest of the pallial canal is in doubt. Metchnikoff 's figures of Hippopodius hippopus larvae show no sign of a homologue of the somatocyst, which appears so clearly in all other Calycophorae. The stalk canal of the larval nectophore (not the pedicular canal of its nectosac) must lie somewhere near the morphologically aboral end of the post-larva, and the curved pallial canal, so characteristic of Prayids, has probably arisen as an endodermal proliferation at the junction of the yolky larva proper and its enormous mesogloeal nectophore. This pallial canal terminates at the inner 76 DISCOVERY REPORTS PRAYIDAE HIPPOPODIIDAE Nectopyramis diomedeae Rosacea adult [neotenouS] y°ur>9 \ ff-\ | larva ' I (I J| \unknowri\ I \ "^r I I adult [heteromorpn\ full grown larva-not sexually mature [caducous] ancestral ontogenies 3 " 9 tgg Text-fig 34. Phytogeny of Prayids and Hippopodiids, to show the nature and relationship of the nectophore of Nectopyramis diomedeae. In this diagram arrows are used to indicate growth stages, and long links to show how successive somatocysts are budded from their predecessors. A break in these links indicates that the larval nectophore is caducous. SYSTEMATIC AND BIOLOGICAL ACCOUNT 77 end of the pedicular canal of the nectosac. In the definitive nectophore the pallial canal extends for some distance beyond this point, and is usually called the descending branch of the somatocyst. One is tempted to speculate as to why such large nectophores as those of Rosacea spp. appear to need no somatocyst of the type found in other Calycophorae. The somatocyst of Hippopodiids is of an inter- mediate type. The branched somatocyst of species of Praya and Stephanophyes seem to fulfil a circu- latory rather than a flotational function, such as it is presumed to fulfil in other Calycophorae. Garstang (1946) made the interesting suggestion that in Calycophorae the aboral end of the larva is aborted, and that for this reason Calycophorae do not develop the pneumatophore which he assumes that their ancestors possessed. But Metchnikoff 's account of 9- and 10-day old larvae of Hippo- podius hippopus, raised by himself at Villefranche eighty years ago, seems to show that it is from the aboral end of the larva that its united bract and nectophore develops, although the nectosac itself lies to one side. It appears to be a matter of some importance to observe the early stages of development of either H. hippopus or Rosacea cymbiformis, both common Mediterranean species, so that the account of larval development can be completed. In the spring of 1950 I made attempts at Villefranche to rear larvae of Hippopodius hippopus, but specimens were scarce during my short visit, and although I kept the actively swimming and apparently ripe female gonophore alive for ten days, and introduced active sperms to the eggs, I was unable to induce the gonophores to shed their eggs and obtained no larvae. Chun (18880) reported that he, too, had tried in vain. Metchnikoff himself reported that he had great difficulty in getting the species to breed. My experiments were carried out in an improved Harvey ' rocker ', consisting of a battery of sixteen 500 ml. beakers, in which plates of glass were kept rocking by a siphon-emptying water-can and counter weight. The air temperature of the sous sol of the Station Zoologique in which the rocker was installed was about 21 ° C, and subdued daylight entered from a small north window. The gonophores appear to be heavy, and no doubt sink to some depth in the sea, where the temperature would be lower. PRAYIDAE Nectopyramis Bigelow, 191 1 Type species N. thetis The two ' deep-sea species described forty years ago under this generic name, N. thetis Bigelow (1911a) and N. diomedeae Bigelow (191 ib) differ in several respects that are of some importance. First, the dorsal and ventral radial canals of the nectosac of N. thetis do not arise in common with the lateral pair, but are far separated. In this N. thetis agrees with Archisoma natans. Secondly, the eudoxid of Nectopyramis thetis resembles that of Archisoma natans and differs from that of Nectopyramis diomedeae in having a special swimming bell with similarly dissociated canals. Thirdly, N. thetis, unlike N. diomedeae, but like Archisoma natans (which, of the Nectopyramidinae, has the most simple somatocyst (canal system), consisting of one longitudinal canal), has an open hydroecial cavity extending the whole length of the polygastric stage of the animal. There is also a sharp bend in the middle region of Nectopyramis thetis and Archisoma natans near the base of the stem, as in Rosacea plicata and R. cymbiformis. In these three ways it would seem that Nectopyramis thetis and Archisoma natans are more closely related than the two so-called species of Nectopyramis are to each other. Bigelow & Sears (1937) accepted my suggestion that Nectopyramis is a Prayid genus, and gave some more excellent figures of the type species N. thetis. It would be best perhaps to include Archisoma Bigelow (1919) mentions a name N. albatrossi. I have failed to find it elsewhere, and it appears to be a nomen nudum, possibly the manuscript name he used for N. diomedeae before publication of his 'Albatross' Report. 78 DISCOVERY REPORTS natans in the genus Nectopyramis, but we cannot exclude N. diomedeae, because the branches of the somatocyst of its polygastric stage can be homologized with those of N. thetis (see p. 84). A new species N. spinosa was published by Sears (1952) after this report was written. I have dealt with it on p. 86. Nectopyramis thetis Bigelow, 1911a. Attention has been drawn on p. 77 to the apparent close relationship between this species and Archisoma (= Nectopyramis) natans, and to its differences from Nectopyramis diomedeae. The eudoxids of N. thetis and N. natans each have a flattened dorsal (posterior sensu Bigelow, anterior sensu Leloup) surface (Leloup, 19326, calls it 'une crete mediane'), up to which reaches a small branch of the ascending canal (not shown by Leloup). But whereas there are in the eudoxid of N. thetis two lateral projections of the margins of the facet, each served by a lateral (sensu Leloup, transverse sensu Bigelow) canal, in N. natans both projections and lateral canals are absent. But in N. pyranm also the branch of the ascending canal reaches the dorsal surface. C long. Text-fig. 35. Nectopyramis thetis. Growth stages of eudoxids and homologies of bracteal canals with those of Abylids: A after Leloup, B, C, D after Bigelow. E, Abylopsis tetragona. The eudoxids, or stem groups, of the following Prayids are now known : Praia dubia, P. reticulata, Rosacea cymbiformis, R. plicata, Maresearsia praeclara sp.n. Amphicaryon spp. indet., Nectopyramis thetis, N. diomedeae and N. natans. They are all homologous, but Amphicaryon eudoxids, of which Eudoxia tottoni Leloup, 1934a, is one, appear to me much more primitive in that the bract is simple, helmet-shaped, and carries only a pair of longitudinal (sensu Leloup) bracteal canals running on each side of the bracteal cavity. But with this difference is correlated the fact that the bracts of Amphicaryon only extend on one side of the stem, whilst in all the rest of the genera they are attached athwart the stem, so that the half of each bract which corresponds with the whole Amphicaryon bract bears a similar pair of bracteal canals in the walls of the bracteal cavity. The ascending and descending bracteal canals, absent from the Amphicaryon bract, appear to have arisen as the bract extended on to both sides of the stem and increased in thickness above it. The ascending canal in each case arises from one of the longitudinal canals, though not on the same side in all Prayids. Bigelow (1911a) had SYSTEMATIC AND BIOLOGICAL ACCOUNT 79 at once noticed the homologies in the bracteal canals of Nectopyramis thetis and species of Rosacea. I know of no descriptions or figures showing exactly how the Prayid bract is attached by its muscular lamellae, and my own efforts to elucidate this point in the living specimens of Rosacea cymbiformis failed though I have lately found a canal linking the stem-cavity with the bracteal canal system in young, preserved eudoxids of Rosacea cymbiformis. There is a superficial resemblance between the bracteal canals in Prayids and in Abylids, a quadri- partite arrangement in each case. But whereas the paired longitudinal canals of the Prayid bract run athwart the stem, in the Abylid bract they clasp it. The resemblance can hardly be fortuitous. It may be well to record a note on the homologies of the early growth stages of bracts of the eudoxid of Nectopyramis thetis, originally figured by Bigelow, and on figures of mature bracts by Leloup and by Bigelow (Text-fig. 35). Only the canals C. lot. (Text-fig. 36B, C) are wanting in the Abylids. Bigelow (191 1 a) in his first figure (3) showed three crests or ridges R 1, R2, and R3. I have studied similar bracts still attached to the stem lying in the hydroecial canal. R3, which runs down to the tip of canal C4 (labelled C3 in Bigelow's fig. 4) is what Bigelow later called the anterior side. The other two ridges are the margins of the dorsal facet (Bigelow's posterior side). They start from the promi- nence over the tip of the ascending bracteal canal (C. lat.), run over the tips of the two horizontal canals and down to Bigelow's postero-dorsal margin of the hydroecium. In well-preserved bracts these ridges are preserved. In less well-preserved material they become rounded and obscure. In Bigelow's fig. 4 of his original paper (1911a) only one transverse canal can be seen, labelled C7. The other canal, shown nearby under a prominence, is the ascending canal C3. The descending canal C4 is labelled C3. N. thetis was not amongst the specimens sent to me taken by Beebe in his many deep hauls off Bermuda in the years 1929, 1930, 1931. There are, so far, thirteen records in 'Discovery' hauls in the Tropical and South Atlantic and in the South Indian Ocean. Altogether, four polygastric and seven- teen eudoxid specimens have been taken, in seven cases in closing hauls, as well as an eudoxid taken in 1951 and another in 1952 by 'Scotia' to the north-west of Ireland. Hitherto only three polygastric specimens had been recorded (see Bigelow & Sears, 1937) and eight eudoxids, often in an unsatisfactory condition. They were taken in the Bay of Biscay, Bay of Cadiz and between the Azores and Canaries ; always in open nets towed at considerable depths. Now from Indian Ocean Stations I have to record the capture by R.R.S. 'Discovery II' of five polygastric specimens and eleven eudoxids. ' Discovery ' Stations Depth (m.) Closing-net Locality Material !5°7 135° ° South-east of Cape of Good Hope, South Indian Ocean IE. 1568 1400-0 South-east of Cape of Good Hope, South Indian Ocean 4E. I57i 500-0 Indian Ocean, between Durban and South Madagascar IE. !573 800-0 Indian Ocean, between East African coast and South Mada- gascar 2E. 1581 600-0 Indian Ocean, East African coast 2P. 158S 500-0 Indian Ocean, East African coast IP. 1587 1250-800 X Indian Ocean, East African coast IE. J749 1050-600 X South Indian Ocean IP. J754 900-400 X South Indian Ocean IP, IE. J755 1700-0 South Indian Ocean IE. P. = polygastric. E. = eudoxid stage. x = closing net. 80 DISCOVERY REPORTS Seven other captures, all made by ' Discovery ' in the Tropical and South Atlantic are as follows : ' Discovery ' Stations Depth (m.) Closing-net Locality Material 86 iooo(-o) South Atlantic, west of Cape of Good Hope IP., IE. 89 iooo(-o) South Atlantic, west of Cape of Good Hope IE. IOOB 900-1000 X South Atlantic, west of Cape of Good Hope IE. 677 2000-0 South Atlantic IE. 2042 2800-2300 X South Atlantic IE. 2063 1150-600 X South Atlantic, west of Cape of Good Hope IP. 2639 1200-600 X Tropical Atlantic, off west coast of Africa IP., 2E. P. = polygastric. e. = eudoxid stage. Recent 'Scotia' captures are: 'Scotia' Stations Depth (m.) Closing-net Locality Material J951 Haul 363 !952 Haul 1305 Haul 1 1 97 1000-250 500-0 700-0 X North-east Atlantic, north-west of Ireland 580 38' N., 8° 21' W. 55° 28' N., 90 42' W. IE. IE. 2P. P. = polygastric. E. = eudoxid stage. The characteristic relationship of Leloup's 'formation speciaW (Bigelow's 'bilobed vesicle') with the bracteal and pedicular canals of the eudoxid is complex. I give two new figures which show that in the particular specimen figured, the pedicular canal of the special nectophore of the eudoxid arises from the ventral side of the central organ, and that all the ' canals ' at this point do not meet at a point (Text- fig. 36 B). The 'central organ', which lies in the hydroecium, is substantially the same as figured in Leloup's diagram (1933, pi. 1, fig. 13), and in Bigelow & Sears (1937, fig. 5). I believe that it serves as a food reservoir. Its upper wall forms a pouch, which spreads over on to the base of the gastrozooid. In a recently captured specimen ('Scotia' 1951, haul 363) the bilobed vesicle figured by Bigelow can be seen clearly. There is a ring of deeply staining tissue (? Bigelow's pedicular plate) between the central organ and the basal bulb (basigaster) of the gastrozooid which unites everything together, and from which the gonophores spring. The special nectophore is attached by a long semicircular adhesion to the junction of what I will call the semicircular bracteal canal (Text-fig. 36, C.br.s.) and the central organ. But the cavity of the bracteal canal system is isolated from that of the stem in the fully formed eudoxids of Prayids. In young stages of Rosacea cymbiformis I have seen a canal connecting the bracteal canal system to the stem, but later that disappears. The pedicular canal of the special nectophore of Nectopyramis thetis does not arise, as it can appear to do, from the semicircular bracteal canal, but from the central organ itself. This is what might be expected since I have discovered that in Rosacea plicata the central organ of the eudoxid represents a development of the remains of the stem. The central organ of Nectopyramis thetis varies in its state of development. Where it appears triangular (Text- fig. 3 5 A) I think it is ageing, and what looks like a descending canal is only a constricted part of the SYSTEMATIC AND BIOLOGICAL ACCOUNT 81 sac-like structure. Some of the better preserved eudoxids of Rosacea plicata from 'Discovery II' Station 1617 show a homologous central organ and central organ complex. The central organ in R. plicata is globular and is filled with a substance that may be of importance in controlling the density Text-fig. 36. Eudoxid of Nectopyramis thetis, from 'Scotia' 1951 St. 363. A, foreshortened dorsal view; B, lateral view; C, ventral view, x 5-5. .Angop ■C/at. -C.rad v ~Eud.Br. Angd Cd.-I. -CA/r ■CM s mm Text-fig. 37. Nectopyramis thetis. A, dorso-lateral (antero-lateral sensu Bigelow, postero-lateral se?isu Leloup) view of eudoxid from 'Discovery' St. 89, to show central organ and canals. B, dorso-lateral view of polygastric stage, x 6-6. of the eudoxid. The post-larva of R. plicata is provided with a similar organ. In that species, as the nectophore grows, this part of the pallial canal decreases in size, and becomes the pedicular canal of the muscular lamella attaching the nectophore to the stem. A similar somatocyst-organ can be seen in my figure (Text-fig. 38) of the nectophore of Nectopyramis natans, and in Moser's (1924a) figure of the larva of Rosacea cymbiformi;. 82 DISCOVERY REPORTS In some Diphyid eudoxids there is, between the bracteal somatocyst and the basigaster of the gastro- zooid, a cylindrical connecting piece (see Vogt, 1854; pi. 19, figs. 1, 12). In living specimens of Chelophyes appendiculata which I anaesthetized at Villefranche, it measures o-6 mm. in length and o-i mm. in diameter. To its middle and distal half are attached the successively developed gonophores. This peduncle is the homologue of the central organ described above, and represents the remains of the part of the stem to which the gastrozooid was originally attached, or perhaps the peduncle of the gastrozooid. Nectopyramis natans (Bigelow), 191 lb, = Archisoma natans Bigelow, 191 16. An eudoxid of this interesting Prayid was taken by 'Discovery II 'between 1650 to 950 m. at St. 1586, off Italian Somaliland. The polygastric stage has never been recorded. I take this opportunity, therefore, to mention two polygastric specimens, one taken by 'Discovery' at Station ioi, off South- west Africa, in a closing net at a depth between 2480 and 2580 m. ; and the other by S.Y. 'Rosaura' in 1937, at a depth between 750 m. and the surface, between St Paul and the Brazilian coast. Rlong C rod lot _--/ — C Pc -Riot Rlong Crodd CpoCsom) Cz' Text-fig. 38. Nectopyramis natans, polygastric stage. A, foreshortened view from lower end to show longitudinal ridges; B, dorso-lateral view, x 7. The eudoxid was described and figured by Bigelow (191 lb, 1913), and has been taken several times since. Curiously enough the polygastric stage has much the same long-bow shape as the eudoxid. The 'Discovery' polygastric specimen is 2 cm. in length, shaped like a long-bow, pointed above at one end, and truncate at the other. The hydroecium runs practically the whole length. At mid-length it contains the large gastrozooid and tentacle, and what are probably the buds of a second and third. On the dorsal surface are three longitudinal ridges extending from the apex to the nectosac, which lies about a quarter of the total length above the posterior truncate end. There is also a pair of lateral ridges running from the apex, past the nectosac to the truncated end. There is a single longitudinal SYSTEMATIC AND BIOLOGICAL ACCOUNT 83 canal, the somatocyst, running from end to end. It lies just underneath the dorsal wall of the hydroecium, except for the upper 5 mm. where it leaves the hydroecial wall and runs deeper in the mesogloea. In the middle region of the specimen the hydroecium is deepest, and the longitudinal canal has a corresponding bend, as in the better known Rosacea. Just at this point, and at the base of the primary gastrozooid, is an ovate structure found also in post-larvae of R. plicata, R. cymbiformis and Hippopodius hippopus. It is similar to the ' central organ ' found in the aboral end of some Prayid eudoxids. It may act as a food reservoir or perhaps control density. The dorsal and ventral canals of the nectosac arise independently, one at each end of the nectosac, and the lateral ones form a pair in the middle, as in the special swimming bell of the eudoxid. The lower edge of the nectosac is distinctly undercut. The structure of the tentacle of the polygastric specimen has not been well preserved, and the nature of the tentilla has not yet been elucidated. Study of a well-preserved eudoxid from 'Discovery' Station 287 (800-1000 m.) off the west coast of Africa enables me generally to confirm Bigelow's (1931) account, and to add a little to it. The two longitudinal hydroecial canals which arise as in Nectopyramis thetis, one on each side of the central organ, join one another distally as in the type, though in another specimen from ' Discovery ' Station 100 (2500-2000 m.) west from Cape Town they do not. There is a ' central organ ' (Leloup's ' formation speciale ') very similar to that in N. thetis, with a loop of the canal system — the semicircular bracteal canal — running round the ventral side. The central-organ sac has probably burst, so that its junction with the pedicular canal of the special nectophore1 cannot be seen, but this pedicular canal arises in approximately the same place as in N. thetis. The chief difference between the eudoxids of Nectopyramis nutans and Nectopyramis thetis is that in N. thetis there have grown out a ventral and two dorso-lateral processes of the bract, each served by a branch of the canal system. The tentilla of the eudoxid of Nectopyramis nutans; viewed in situ, resembles those of Nectopyramis thetis and differ very markedly from those of N. diomedeae. The relationship of this species have been dealt with above. Nectopyramis diomedeae Bigelow, 191 ib. (Plate V, figs. 2-3.) Only three polygastric and eight eudoxid specimens have ever been recorded, all from the Eastern Tropical Pacific (Bigelow, 191 1 b, 193 1). I can now report that the species occurs also in the Atlantic, where Beebe took seven polygastric and twelve eudoxid specimens off Bermuda. The specimens noted in my list of Siphonophora taken by Beebe in 193 1 {Nectopyramis sp.nov.?) belong to the species N. diomedeae. 'Discovery II' took two more polygastric specimens in the South Indian Ocean, in 320 lat., at Stations 1749 and 1755 ; two more off South-east Africa at Stations 1567 and 1568 ; and two further polygastric specimens off East Africa at Stations 1585 and 1587. 'Discovery II ' took eudoxids in the South Atlantic at seven stations between Station 679 at 260 S. lat. and Station 1555 at 390 S. lat. ; and again at Station 1999 (640 S.) between South Georgia and the Ice Edge. From the records so far available I estimate that the species lives at a depth of from 650 to 1600 m. or a little more. The polygastric stage has been taken, so far as we know at present, in only one closing net haul (1060-600 m.) ; the eudoxid in eight, varying in depth between 625-675 m. and 1500-2000 m. Beebe's seven hauls were open hauls to the surface after long horizontal hauls. His deepest net (1829 m-) contained a polygastric specimen when it came to the surface, as did his 1646-m. net, but there must be some doubt about the depth at which the captured specimens were living. Polygastric stage. There is the usual Prayid plan of a median longitudinal canal (Text-fig. 39 A, C.med.long.) that passes round the dorsal side of the hydroecium. Below, it gives off the pedicular 1 The special nectophore bears a manubrium and eggs, and is therefore a gonophore. The second gonophore probably remains in a reduced condition as far as its general shape and subumbrella are concerned. 84 DISCOVERY REPORTS canal (C.ped.) of the nectosac and then continues distad along the face of the hydroecium. Above, it leaves the hydroecial wall and continues (C.pa2) towards the apex of the bract, giving off four or more long, alternate lateral branches. At the point where C.pa2 leaves the hydroecial wall the longitudinal median canal forms on either side, as in N. thetis, a curved lateral canal, which runs on the wall of the hydroecium. On the left side it forks, one complicated branch (Bigelow's C.h.l.) running on the left wall of the hydroecium, and the other giving rise to C.pa3, which very shortly after the fork gives off a downward branch (Bigelow's C.h.r.) on the right wall of the hydroecium. Returning again to the base of C.pa2, the longitudinal canal, on the right side of C.pa2, gives off C.pa1, which almost at once gives off a descending branch, C.h.l.2, that runs on the left wall of the hydroecium. The problem is how to homologize the canal systems of N. thetis and N. diomedeae. My tentative solution is that the small C.h.l.2 in N. diomedeae is the true left hydroecial canal, and that Bigelow's C.h.l. is a secondarily developed canal. We should then see in both species a similar arrangement — an ascending canal C.pa2, a right and left ' dorsal ' canal C.pa3, C.pa2, and a right and left hydroecial canal C.h.r., C.h.l.2. In N. diomedeae the canal system is more complicated still, but the basic arrangement is similar to that in N. thetis. The whole nectophore is twisted, so that the nectosac opens on one side instead of dorso-basally, and the branches of the somatocyst have become secondarily displaced. Bigelow's (191 1 b) description of the system, as far as the lower part and the free upper part of the longitudinal median canal is concerned, agrees with mine. After this Bigelow (191 \b, p. 192) says, 'at the point at which it [the ascending canal] bends apically it gives off a branch which runs over the left-hand face of the hydroecium [his C.h.l. of fig. 4] and a short distance further on gives off a second trunk which passes over its right-hand face [his C.h.r.]. These two canals of the hydroecium. . .'. In my view only one of these, Bigelow's C.h.r., is the homologue of an hydroecial canal (the 'right' hand one) in N. thetis. Bigelow did not figure or describe what I take to be the true, short, simple, left-hand hydroecial C.h.l.2. The 'explanation' of Bigelow's (191 lb) plate 1 (which does not mention fig. 6, part of the eudoxid) corrects certain lettering in figs. 3 and 4. It should be noted that the lettering C.pa1 and C.pa2 also in these figures should be transposed. The 'descending branch', to which Bigelow refers on p. 192, can be seen at the top on the right in his figs. 3 and 4, descending from C.ped. (marked on the plate C.pa3). With these trifling corrections made it will be more easy to enjoy Bigelow's descrip- tion and figures, and to understand the homologies of this complex canal system, which he figures so beautifully. To clarify my remarks, I give a new figure (39 A) of part of the canal system, and reproduce a comparable one (39 B) by Bigelow. Since the adult nectophore of N. diomedeae retains in its nectosac the type of radial canals found in larvae of other Prayinae (its own larva is unknown), and its hydroecium has the pocket-shape so characteristic of larvae, it is reasonable to regard it as a neotenous form. Secondarily its somatocyst has become very much complicated (see Text-fig. 39). The tentilla of N. diomedeae (Text-fig. 39C), though partly described by Bigelow (191 ib), have never been figured. They are relatively large and of peculiar construction when mature, borne on long, loosely coiled pedicels. The apical part consists not of a terminal filament but, as in the larval tentillum of Physophora hydrostatica, of a small hollow cap of ovate nematocysts that appear to be stenoteles. This cap is from 0-2 to 0-25 mm. in diameter, and is perched on the apex of a large hemi- spherical cap of longer scimitar-shaped nematocysts. The larger cap measures from 0-35 to 0-4 mm. in diameter. Into it penetrates a tapered endodermal core, round which, externally to the cap, is twisted irregularly a long, double, structureless, elastic cord. Proximal to this cord is a thickened opaque length of ectoderm. It is not possible at the moment to make a further examination of the SYSTEMATIC AND BIOLOGICAL ACCOUNT 85 structure of these tentilla, but in the young tentillum the larger cap of the elongate nematocysts is narrower and one-sided, and it only secondarily acquires its final hemispherical shape. The smaller cap in its final form is provided with prominent cnidocils. In younger stages it is generally turned to one side. The tentilla of Nectopyramis thetis are totally different. There the main battery of nemato- cysts is confined to one side, and there is a long spirally coiled terminal filament. In N. diomedeae there are a number of irregular surface ridges which are shown on PL V, figs. 2 and 3. C.po? = Cpa' Text-fig. 39. Nectopyramis diomedeae. A, the branching somatocyst-complex in the region of the hydroecium of the polygastric stage. Specimen taken by Wm Beebe at Bermuda; B, the same, after Bigelow (191 lb, pi. 1, fig. 4) taken in Eastern Tropical Pacific; C, a tentillum. Eudoxid. The 'central organ' is well figured by Bigelow (19116; pi. 1, figs. 5 and 6). In fig. 6 it can be seen quite clearly that the pedicular canal of the large nectophore arises from the central organ and not from the nematocyst canal system, which is a closed system in the full-grown eudoxids of Prayids. What cannot be seen so clearly in Bigelow's figures is that there is a semicircular bracteal canal as in Rosacea ; that the hydroecial canal diverges from it at about the point marked in Bigelow's fig. 6 with a 'pointer' labelled C.br. ; and that in some specimens at least the ascending canal lies on the wall of the hydroecium for a short distance before passing to the apex. Its point of origin from the semicircular canal is not directly opposite that of the ascending canal, which does not arise from the central organ. As usual in Prayids, the central organ is closely attached to the semicircular canal, and the fact that their lumina are quite separate is not easy to determine. This description will show at once that the bracteal canal system of N. diomedeae can be homologized with that of N. thetis and of Rosacea spp. In Nectopyramis diomedeae there are no lateral canals at the junction of the semi- circular and hydroecial canals. 86 DISCOVERY REPORTS Nectopyramis spinosa Sears, 1952. (Plate V, fig. 1.) ? = Hippopodius cuspidatus Moser, 1925. During a period of many years I have picked out from 'Discovery' catches three very curious Prayid nectophores, which I think may be what Moser (1925) described and rather inadequately figured as ? Hippopodius cuspidatus. A fourth was taken in a closing net fished from 1000 to 250 m. by ' Scotia' in 1951 to the west of Ireland in W. long. 170 03', and a fifth also by ' Scotia' in 540 15' N., 1 40 32' W. at the same depth. The nectophores are of two kinds, and there are three of the first and two of the second. The first has a very minute obsolescent nectosac (Text-figs. 40 A, B) and the second a much larger functional nectosac (Text-fig. 40 C). In both kinds the radial canals arise separately as in Nectopyramis thetis. v^r-s ~ "iy^Cradd. Rbuc Nsobs Text-fig. 40. Nectopyramis spinosa Sears. A, dorsal view ; B, lateral view of nectophore of first type, with obsolescent nectosac, 'Scotia' 195 1 Cruise, St. 363 ; C, lateral view of nectophore of second type, with functional nectosac, 'Discovery II' St. 1639. *9-5- Although there is some variation in the arrangement of spinose ridges in the first group of three nectophores with obsolescent nectosac, there are elements in their arrangement that can be compared with that in the second group of two nectophores with the larger nectosac. For instance we find the same dorsal-median and what I will call ' buccal ' ridge in both groups ; but more material is needed to complete the morphological account. Nectopyramis (including Archisoma) is the only Prayid genus known in which the nectophores bear ridges. We may have here, in the existence of two types of nectophore, a phenomenon similar to that found in Amphicaryon, where one nectophore has an obsolescent nectosac. This would be of special interest seeing that other species of Nectopyramis have only one nectophore as far as we know, though presumably their ancestor had more. It is possible that one type of nectophore represents a caducous larval, and the other the definitive nectophore. In favour of this interpretation is the fact that the somatocyst or pallial canal is extended well beyond the obsolescent nectosac in the first group of necto- phores. This extension is characteristic of larval Prayid nectophores. In the second group the pallial canal stops short at the ventral radial canal of the nectosac, just as it does at the pedicular canal in definitive nectophores of Prayids. Since I wrote this note Dr Mary Sears has published (1952) a note on this siphonophore, of which ' Dana ' took six specimens — one with the large and five with the obsolescent nectosac. Three were taken in the Indian Ocean, two in the South Pacific and one in the Tropical Atlantic. In none of them were the stem or appendages preserved. Dr Sears regards it as a new species, and in the light of the SYSTEMATIC AND BIOLOGICAL ACCOUNT 87 information at present available this is convenient, since there must remain doubt about the identity of Moser's specimen. At the same time the number of siphonophore species is not large; and the spinose condition is so unusual that there is some probability that Moser was really reporting (if not accurately describing) the same species. Material. Depth (m.) Material 'Discovery II' Station 1179 'Scotia' 1951 Station 500-250 1 nectophore of first type with gastrozooid and tentacle 363 37i 1000-250 1000-250 1 nectophore of first type 1 nectophore of first type 'Discovery II' Station 68 1 1639 1000-750 2400-1 150 1 nectophore of second type with gastrozooid, tentacle and bud 1 nectophore of second type with gastrozooid and tentacle Dr Sears and I have been exchanging information for some time — a fortunate circumstance in days when publication is so long delayed. As soon as Dr Sears informed me that she was publishing a note on a new species N. spinosa I guessed that she was referring to the spinose Prayid I had for so long been studying. I sent her a figure of a nectophore with obsolescent nectosac and she confirmed it. But at the same time she asked if I had the nectophore as well as the ' bract ' ; and in her account, which at the time was in the press, she again refers to the ' bract ', by which she means the bract of the eudoxid. Describing its hydroecium she says that closely associated with it is a long tubular somatocyst, and adds that on two of her specimens two thread-like branches are given off dorsally and appear distally to have a 'globular connexion '. It is quite clear to me that the 'globular connexion ' is the obsolescent nectosac, and that ' the two thread-like branches of the somatocyst ' are the dorsal and ventral radial canals of the nectosac, so characteristically dissociated in N. thetis, instead of arising with the lateral radial canals from a common pedicular canal as in most Prayids. In the very well-preserved specimen (PL V, fig. 1) from 'Discovery II ' Station 1 179, the opening of the obsolescent nectosac is visible as well as the ring canal. In the three specimens to. which the gastrozooids are still attached the 'central organ' to which I have referred (p. 81) in my notes on other Prayids, is visible. It is comparatively large in the specimen from ' Discovery 1 1 ' Station 68 1 , and spreads down on each side of the base of the gastrozooid . The tentilla of the same specimen are well preserved and show a peculiar spheroidal swelling at the junction of pedicel and saccus. They have been examined only with a binocular microscope under very bright illumination, because it is not advisable to dissect the only three complete specimens. Rosacea [? Q. & G.] Bigelow & Sears, 1937. After many years study of both living and preserved siphonophores and of the literature I have come to the conclusion that we only know two species of the genus which Bigelow (1911^) and Bigelow & Sears (1937) rightly or wrongly called Rosacea. The distinguishing features of these two species, R. cymbiformis and R. plicata, is that the eudoxids have large gonophores and no special nectophore, and a somatocyst canal system of the type figured by Bigelow (191 lb, pi. 2, fig. 4). A third alleged 88 DISCOVERY REPORTS species of Rosacea, referred to by Bigelow as R. medusa Metchnikoff (1870), belongs to a separate genus. I have re-examined the Russian1 account and figures of Metchnikoff, and repeat some points of interest referred to in his Russian description. Praya medusa is described as being the most transparent of siphonophores, and is very difficult to catch. It differs from P. diphyes Vogt, says Metchnikoff, in its smaller size, shape, as figured by him, and equality of development of the nectophores, which do not overlap as in P. maxima. There is less gelatinous substance and there is a wide velum. Most important is the mention of little tentacles and ' eye-spots ' on the edge of the nectophores as well as on the edge of the special nectophores of the eudoxids, in the latter of which they number from twenty-five to thirty. Metchnikoff does not mention or figure the four radial subumbral canals of the main nectophores but says that the vascular system is 'the same as in other species '. There is a single (pallial) vessel running from top to base and ending in a small vesicle (somatocyst). The shape of the comparatively small gonophores can be seen in Metchnikoff's figures. Since this species was recorded from the Mediterranean, with whose fauna most of the siphonophore pioneers dealt, it may be useful to give my conclusions about the identity of the various Prayids found in the Mediterranean. There appear to be at least six species belonging to four genera in the Mediterranean : Genus A. Rosacea (Q. & G.) Bigelow, 191 \b Species 1. R. cymbiformis Delle Chiaje, 1822. Species 2. R. plicata ([?] Q. & G.) Bigelow, 191 ib. Genus B (new name required). ?= Desmophyes Haeckel, 1888 a (A pyriform branch of the eudoxid's somatocyst present.) Species 3. Praya diphyes Kolliker, 1853 ft. Diphyes Brajae [sic] Vogt, 1851. Praya diphyes Vogt, 1854 not pi. 16, fig. 3 =R. cymbiformis). Praya diphyes Graeffe, i860 (Taf. I, fig. 2). Praya diphyes Fewkes, 1880 (pi. ill, fig. 1). Praya fill 'for mis Keferstein & Ehlers, 1861 (pi. v, figs. 8-1 1). Genus C. Lilyopsis Chun, 1885 Species 4. Lilyopsis rosea Chun, 1885 (with meandering lateral canals in (main) nectophores). Praya diphyes Graeffe, i860 (Taf. I, figs. 1 a, b, c; eudoxids). Lilyopsis diphyes Moser, 1917, 19246. ? Praya medusa Metchnikoff, 1870. Species 5. Lilyopsis sp. (Fewkes), 1880 (pi. ill, fig. 2) (with straight lateral canals in (main) necto- phores). Genus D. Stephanophyes Chun, 1891 Species 6. Stephanophyes superba Chun, 1891. Forms closely resembling species 3 (as ' Rosacea plicata ') and 6 are recorded also from Misaki by Kawamura2 (191 5); species 4 by Bedot (1896) from Amboina; and species 6 from the Canaries by Chun (1891). In the course of many years during which I have been searching plankton samples for Siphonophora, I have found outside the Mediterranean — in the Atlantic, Indian and Antarctic Seas — many Rosacean 1 I have to thank Dr S. B. Markowski for translating parts of Metchnikoff's paper. 2 I have to thank Dr Mary Sears for having been instrumental in producing a valuable translation by Rodney Notomi of most of Kawamura's inaccessible Japanese work. SYSTEMATIC AND BIOLOGICAL ACCOUNT 89 nectophores all of which, except a few of Rosacea cymbiformis, appear to belong to one species, R. plicata. The only allied species that I know is the common Mediterranean one, R. cymbiformis (Delle Chiaje). In both species there are only four simple, subumbral radial canals, the lateral pair of which take a meandering course along the edge of the nectosac. A (first definitive) nectophore of R. plicata from 'Discovery II ' Station 1617, shows an interesting variation in the path of the lateral radial canals. Similar variations are not uncommon. This variation consists of irregular cross-branches connecting the major loop of the lateral radial canal with other parts of the subumbral canal-system. Sometimes only blind diverticula are seen. This variation is of particular interest in connexion with the description by Leloup (1934a) of R. (Prayoides) inter- media. Leloup's specimen (fig. 4A), which I select as the holotype of R. (Prayoides) intermedia Leloup, appears to show a larger larval nectophore — 'larval' because there is no descending branch of the somatocyst — from which projects a first definitive nectophore with deep hydroecium. In the absence of any indication of the looped Rosacea type of lateral radial canal I am inclined to regard this form as too much unlike R. cymbiformis and R. plicata to be included in the same genus. Moser (1925) described and figured a few nectophores from the Tropical Atlantic which she called R. tuberculoid.. Several poorly preserved specimens of R. plicata that I have seen look rather like Moser's figures, but I have never found a well-preserved one looking like these figures. I do not think 1 tuberculata' represents a distinct species. The synonymies of the two species of Rosacea, are, in my opinion, in a state of confusion, a fact that is not surprising in view of the unsatisfactory nature of so many of the earlier figures and descriptions. The first species, R. cymbiformis, was figured unmistakably by Delle Chiaje (1822, Tav. 50, fig. 3; 1841-4, Tav. 33, fig. 1); Leuckart (1854, pi. 11, figs. 18-24); Vogt (^54, pi- 16, fig. 3); and Haeckel (18886, pis. 31, 32), though Haeckel's figures of nectophores are poor. I know of no recognizable old figures of the second species, R. plicata. Rosacea plicata (? Q. et G.) Bigelow, 191 1. The nectophores of R. plicata fall into three categories: (1) larval caducous nectophores, in the hydroecia of some of which are to be found buds of (2) secondary (definitive) nectophores. Larger detached nectophores of this category are the ones most commonly found, and are those figured by Bigelow & Sears (1937, figs, n, 12). In the hydroecia of a few have been found buds of (3) tertiary nectophores. Nectophores of category 3 have also been found detached. The species to which all these specimens belong appears to be the one that Bigelow & Sears figured and described so well in their 'Thor' Report (1937) on the Mediterranean siphonophores, and for which species, rightly or wrongly, they used the name R. plicata Quoy & Gaimard. The eudoxids are large, up to 2 cm. in length, at first sight globular, and three-lobed, much as in R. cymbiformis.1 The right antero-lateral lobe in R. plicata bends round distally to the left, and in front of the distal end of the left lobe — a feature not shown by Bigelow (191 1 b, pi. 2, fig. 4) in R. cymbiformis, although it often occurs in that species. In the early growth-stage it can be seen that the proximal side of the bract of R. plicata is convex, whilst the distal side is flattened, as in R. cymbiformis. One noticeable and constant difference between eudoxids of the two species is in the position, on the shorter left, lateral hydroecial canal, from which the unpaired dorsal canal springs. In R. cymbiformis this position is proximal to the small lateral 'spur' canal, as clearly figured by Bigelow (191 1 b, pi. 2, fig. 4), whilst in R. plicata the position is distal to — in Bigelow's figure, to the left of- this small lateral spur. 1 See PL IV, figs. 6-7. 9° DISCOVERY REPORTS Cbrspvr Cbrv Ch.tr C.M.I: C-h.Lr Text-fig. 41. Rosacea plicata. Two views of an eudoxid from 'Discovery II' St. 1617, 395-155 m., *.$■$. Chlr. Cbr.d- .. Cdrspur Text-fig. 42. Rosacea plicata. Eudoxid from 'Discovery II' St. 1617, 395-155 m., x8-3. SYSTEMATIC AND BIOLOGICAL ACCOUNT 91 In R. plicata the right-hand longitudinal hydroecial canal of the bract is relatively much longer than the left. The caducous, larval nectophores grow to a diameter of at least 12 mm. Some of the first known larvae of this species, together with at least one eudoxid, were taken by ' Research ' in the Bay of Biscay in 1900; but all were recorded as larvae of Hippopodius hippopus. In the larval nectophores the lateral radial canals are somewhat sigmoid in shape and do not meander as in the definitive nectophores. Also there is no descending branch of the somatocyst, so that the attachment of the muscular lamella lies entirely above the junction with the pedicular canal that leads to the nectosac (see Text-fig. 32 A). C pa Cpo Cpa ~-Cped 2 mm Text-fig. 43. Rosacea plicata. A, B, two views of a larva from 'Discovery II' St. 1617, 395-155 m., X7; C, a second definitive nectophore 32 mm. in length, from the same station, x 4-2. The slit-like orifice of the hydroecium is less deep and narrower than its main cavity. The whole hydroecium is reminiscent of the hydroecium in the neotenous nectophore of Nectopyramis diomedeae. A constant character that distinguishes larval nectophores of this species [Rosacea plicata) from those of Hippopodiids is the kink caused by the ' central organ ' in the dorsal side of the hydroecium. The nature of this central organ, which has very little affinity for haematoxylin, borax carmine or Sudan III, is uncertain, but seems unlikely to be yolky. It is very like the ' central organ ' of the eudoxid, which appears to be a development of the small piece of stem that remains after separation of the individual 92 DISCOVERY REPORTS eudoxids. The irregularly triangular shape of the gonophore in transverse section is due to the fact that it fits into a space between the gastrozooid and two lobes of the bract. There is no special swimming nectophore in the eudoxid, and no sign of vestigial tentacles on the edge of the nectosac of either the polygastric or eudoxid stage. Bigelow & Sears (1937) say that the location of the point of attachment of the pedicular ('stalk') canal of the 'inferior (older)' nectophores is characteristically much closer to the apex of the bell in R. plicata than in R. cymbiformis. I have not been able to confirm or refute this statement. But their fig. 9 shows a connected pair of first and second definitive nectophores that have been twisted back into their proper relative positions after fixation, and some tearing of tissue may possibly have taken place. It is not always easy to distinguish the ' stalk ' canal from folds of the muscular lamella. What Bigelow & Sears call the 'inferior' ('older') nectophore appears to me to be the second definitive and therefore younger nectophore. The location of the point of attachment of the stalk-canal of the first definitive nectophore is just above the angular bend in the dorsal wall of the hydroecium, and not opposite the pedicular canal of the nectosac as figs. 9 and 12 of Bigelow & Sears seem to indicate. The eggs, which may number up to six, measure about 1-4 x 1-75 mm. Hitherto unrecorded closing-net and upper-layer catches are listed on pp. 93 and 94. Only about 16% of the nets of these stations have been examined. The depth at which the greatest number (23) of the seventy-seven 'Discovery' closing-nets took the polygastric stage of Rosacea plicata was about 300 m. The upper water-layers down to about 300 m. were fished at these stations three times as frequently (358) as the deeper water-masses (115), but only 16% of this total number of catches have been examined so far. At 'Discovery II ' Station 1581, off the East Coast of Africa, the species was living in water at a temperature of 12-200 C. and salinity 34-88-35-26 %0 at a depth of 50-0 m. At the Antarctic Ice Edge, Station 1639, where it also occurs, the temperature was about o-i8-i-36° C. at 2400-1 150 m., and the salinity 34-68-34-74 %0— a fairly wide range. On 28 June 1951 'Scotia' took a first definitive nectophore and two larvae in a closing-net between 1000 and 250 m. to the west of Ireland. This is the most northerly record for the species, which was also taken by H.M.S. 'Research' in 1900 in the Bay of Biscay, 300-0 m. and recorded as Hippopodius hippopus. Amphicaryon Chun, 1888 a Nomenclature. It seems to be quite clear from a study of the original wrappers of the journals in which the generic names were published that Mitrophyes Haeckel dates from 29 May 1888, and Amphicaryon Chun from 15 November 1888. But since Amphicaryon has been in common use since 191 1, it will be better to have the name entered in the list of nomina conservanda than to replace it by the little known Mitrophyes. Bigelow (1911ft) defines the genus as ' Prayidae with two nectophores of very unequal size, the older one degenerating. . .'. He adds that though we should expect the smaller enclosed nectophore of Amphicaryon acaule to be the younger, its relatively larger size in younger stages, when it overlaps the nectophore, which later encloses it, and its subsequent relative degeneration, shows that it is in reality the older, and that the younger overtakes it by its much more rapid growth. I believe that the vestigial nectophore is the younger, because in Prayids the buds of successive nectophores first appear inside the hydroecium of the older. No one has hitherto recorded the larval stages of A. acaule, but a series of three larvae of A. acaule mentioned (p. 95) shows that the larval nectophore is retained as the functional, adult one, and that the reduced, shield-shaped nectophore arises as a bud inside its hydroecium. I recognize three species of Amphicaryon, and am taking this opportunity to publish a new generic name Maresearsia1 for a giant new Amphicaryonine from the Atlantic Ocean (see p. 97). It has recently been identified by me in the Indian Ocean too. 1 Named after Dr Mary Sears. SYSTEMATIC AND BIOLOGICAL ACCOUNT 93 Material 'Discovery and Discovery II' Stations Depth (m.) Polygastric Eudoxid Larval 8 N 200 600-700 X — — IOOA TYF 625-675 X — — IOOB TYF 900-1000 X X X 1500-1550 X — — IOOC TYF 260-310 X X X 45°-55° X X — 2500-2000 X X — 102 N70V 500-250 X X — 750-500 X — — 103 N70V 500-200 X — — IO7 N70V 750-500 — X — 500-250 — X — 850-900 X — — II7 N70V 500-300 X X — 129 N70V 500-250 — X — 276 TYF 150-0 X — — 324 N70V 500-250 — X X 335 N70V 500-250 — X — 33° N70V 500-256 X X — 343 N70V 500-250 — X — 344 N70V 500-250 — X — 391 N450H i2oo-i3oo(-o) X — — 407 450 H 800-950 X — X 5i9 N 100 B 39C»-I37 X X — 588 N 100 B 460-150 X — — 592 N 100 B 350-124 X X — 594 N 100 B 435"l65 X — — 595 N 100 B 380-133 X X — 668 TYFV 500-250 X X — 750-500 X X — 1 500-1000 X — — 671 TYFV 500-200 X X — 750-500 — X X 1000-750 — X — 679 TYFV 750-500 X — — 701 TYFB 242-0 X — — 848 — 270-196 X — — 892 N 100 B 220-100 X X — 946 — 270-120 — X — 970 — 380-110 X X — !°55 — °-5 X — — i57i — 1 400- 1 000 X — — 1581 N70V 500-250 — ? — 50-0 X X — 1583 N70V 50-0 — — — 100-50 ? — — 250-100 — X — 500-250 — ? — 1584 N70V 50-0 — — — 100-50 — ? — 250-100 — X — 500-250 — — — 1585 TYFB 500-0 X — — 1400-700 — ? — 1586 TYFB 530-0 X — — 1630-959 — — — N70V 50-0 — — — 100-50 — — — 250-100 — — — 500-250 X — — 94 DISCOVERY REPORTS = present, absent. Material 'Discovery II' Stations Depth (m.) Polygastric Eudoxid Larval 1587 TYFB 450-0 — — — 1250-800 — — — 1588 N70 V 50-0 — — — 100-50 — — — 250-100 — — — 500-250 — — — 1589 N70V 50-0 — — — 100-50 — — — 250-100 — — — 500-250 X — — 1616 N70V 250-500 X X — 500-250 X X X 1617 N 100 B 395-^5 X X X 1627 TYFB 580-400 X — X 1629 TYFB 350-220 X X X 1632 N100B 33°^l65 X X X 1633 N70B i 100-875 X — — i635 TYFB 460-320 X X — 1636 N70B 380-150 — X — 1639 TYFB 2400- 1150 X — — I723 TYFB 800-500 X — — 1728 TYFB 800-350 X X — 1729 N70B 430-310 X X — 1730 N 100 B 420-230 — X X 1747 N450 1 200-400 X — — 1754 TYFB 900-400 X — — 1766 N450B 1200-600 X — — 2023 N 100 B 750-400 X X X 45°-I5° X X X 2024 — 650-400 X — — 2025 N100B 750-400 X — — 2038 TYFB 1 200-850 X — — 2059 N4SoB 1 900- 1 400 X — — 2066 N450B i9oo-ISS° X — — 2636 TYFB 950-550 X X — 2639 — 1200-600 X — — 2646 TYFB 1 500-800 X — 'William Scoresby' Stations WS61 N70V 500-250 — X — WS63 — 500-250 — X — WS69 — 750-500 X X ~ 'Scotia' Stations 1951 363 — 1000-250 X — X 540N.>i5°W. (approx.) Amphicaryon ernesti sp.n. It has long been clear to me that two distinct species may have been confused under the name A. acaule. At the same time, except for two specimens from Tristan da Cunha.and west of the Cape Verde Islands attributed to Mitrophyes peltifera by Moser (1925) (Haeckel's M. peltifera), a third species of Amphicaryon, has not, since his day, been recognized again, and has been regarded as a problematic species. Over a period of many years I have sorted out specimens of two species of SYSTEMATIC AND BIOLOGICAL ACCOUNT 95 Amphicaryon. The first A. acaule, has been well figured by Bigelow (191 ib, pi. 4). The reduced nectophore is embraced by the larger one, and its atrophied nectosac has the characteristic shape illustrated by Bigelow. Its closed entrance is in contact with the surface, and its four radial canals are distinct. The upper parts of the lateral-radial canals of the larger nectophore are simple and not branched as in ernesti and peltifera. In the second, a new species, to which I give the name ernesti,1 the smaller nectophore is not partially enclosed by the larger one, whose lateral radial canals are branched in the basal region, and the entrance to its atrophied nectosac lies at some distance from the surface, but is connected by a fine strand of tissue. Its ventral radial canal becomes a network covering the ventral wall, and its lateral canals disappear and leave only the dorsal canal. These differences in the radial canals appear to be constant. The whole arrangement is often difficult to see, and sometimes looks rather like Haeckel's figure (1888 a, p. xxviii, fig. 2) of Mitrophyes peltifera. For years I thought it must be M. peltifera, but in March 1952 I found (in a haul from 'William Scoresby' Station 459) an Amphicaryon specimen that is different from both acaule and ernesti, but closely resembles Haeckel's fig. 2 of Mitrophyes peltifera. All the radial canals of its larger nectophores are simple. In its shield-shaped second nectophore, whose outer wall is concave, can be clearly seen, on the ventral side, a three-branched canal, but there is no sign of a nectosac. Some days later I found two other exactly similar specimens in plankton from 'Discovery II' Station 1580, TYF B, 450-0 m. I feel hopeful, therefore, that one day more material of Haeckel's species will come to hand. Bigelow (1918) reported 30 specimens of ' A. acaule ' from the W. Atlantic. But they differ from that species as here recognized and agree with this new species in two of the main characters: (1) the smaller nectophore is not enclosed by the larger; (2) the lateral radial canals are branched. I suspect that he was dealing with A. ernesti. A fourth, new, giant Amphicaryonine species from the West Coast of Africa is described below. Both its nectosacs are functional. The eudoxid described by Leloup (1934a) as Eudoxia tottoni is the eudoxid of a species of Amphi- caryon, probably A. acaule. Material. 'Discovery II' Station 2648, 500-0 m. (Text-fig. 44, type specimen); Station 698, 470-0 m., 2 'larger' nects.; Station 702, 230-0 m., 4 ex., 1 larger nect. ; Station 1374, 230-0 m., 1 ex., 1 'larger', 1 'smaller' nectophore; Station 1584, 100-50 m., 1 larva, 175 mm. in length (Text- fig. 45) Station 2679, 2 ex. The larva is the first, as far as I know, of any species of Amphicaryon to be recorded. The lateral radial canals are similar to those of ernesti and are not simple as in acaule. The larval nectosac is similar in shape to that of the 'larger' of the adult nectophores. The pedicular canal is so short that it has almost disappeared. I think that the 'larger' nectophore in species of Amphicaryon must be the persistent larval one, and that it is not caducous as it is in Rosacea. A slightly older larva of Amphi- caryon acaule from 'Discovery II' Station 1583 (550-250 m.) also shows the bud of the smaller (obsolescent) definitive nectophore growing out from the spot marked N2B in Text-fig. 45 A. On the side of the central organ opposite to this bud, both larvae show the budding zone of gastrozooids. Both also show the median growing apex of the pallial canal or somatocyst, XB. Two other larvae of A. acaule from ' Discovery II ' Station 1581 (600-0 m.) show quite clearly that the first nectophore bud becomes the obsolescent nectophore of the adult, and that the ' larger ' nectophore of the adult is the persistent larval nectophore. 1 In honour of my esteemed colleague Ernest White, who has given me invaluable assistance in all my work since 1918, and who has undertaken so much of the routine work involved in examining thousands of plankton samples and in sorting and curating enormous siphonophore collections. I have to thank him for very great assistance in preparing and checking the text of this report. 96 DISCOVERY REPORTS The figured holotype A. ernesti (Text-fig. 44) bears the Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 5. A/sobs C rod/at C.rodx Cradd -Nsobs Text-fig. 44. Amphicaryon ernesti sp.n. Two views of the holotype specimen, x 16. Ns.lorv. c.rod.d. C.rod.v. Gz.m. Text-fig. 45. Amphicaryon ernesti sp.n. Three views of a larva from 'Discovery II' St. 1584, 100-50 m., x 29. Amphicaryon peltifera (Haeckel, 1888). Mitrophyes peltifera Haeckel, 1888 a. A single specimen was taken at 'Discovery II' Station 1584 in an open haul from 100-0 m. All the radial canals of the larger nectophore are simple. There is no nectosac in the shield-shaped second nectophore, but the characteristic three-branched canal which distinguishes it from A.acaule and A. ernesti can be seen. As mentioned above, it has seemed best not to use Haeckel's slightly prior name Mitrophyes for the genus since it has so long been out of use, but to attempt to have the well-known name Amphicaryon placed on the list of nomina conservanda. Three other specimens of this little-known species which I have had for study are from : 'Discovery II' Station 1580, 450-0 m., 2 ex. 'William Scoresby' Station 459, 135-0 m., 1 ex. SYSTEMATIC AND BIOLOGICAL ACCOUNT 97 Maresearsia praeclara gen.n., sp.n. For many years past I have been familiar with a remarkable Prayid, obviously related to Amphi- caryon acaule Chun and to the new species ernesti, as well as to Haeckel's Mitrophyes peltifera. Polygastric stage. The two nectophores, both of which are functional, though one is somewhat reduced or less developed than the other, fit in to one another to form a globular mass about 2 cm. in diameter. The basal part of the lateral and radial canals of the nectosac are hypertrophied as in Nlorv Nvest Som srnm Text-fig. 46. Maresearsia praeclara gen.n., sp.n., lateral view, x 6-6. Beebe, Bermuda Oceanographic Expedition, 1929, i828(-o) m. The smaller nectophore has been restored to its original position. the other two species. The gastrozooids are large, and measure 3-5 mm. in length. A study of the figures (Pis. VI, VII fig. 1; Text-figs. 46, 47) will make an extended description of the nectophore unnecessary. Eudoxid stage. Many eudoxids are still attached to the holotype by long stalks. The bract is similar to that found in Amphicaryon acaule, but there is a lateral notch, and the two hydroecial canals curve round on themselves in a nearly circular course. The gonophores are of a type usually found in Prayids and are too sac-like to warrant detailed description. A minute description of the eudoxid is unnecessary at this stage, and figures (PI. VII, figs. 2-4; Text-fig. 48A) will be sufficient for its identification. „ , 13 9« DISCOVERY REPORTS Nlarv C rod /at Cradv — ~C.rad.ti. ^Nsobs Text-fig. 47. Maresearsia praeclara gen.n., sp.n., view from above, x 6-8. _ im.m. Ns.larv Text-fig. 48. Maresearsia praeclara gen.n., sp.n. A, eudoxid from holotype specimen taken at 'Discovery II' St. 2636, 950-550 m., x 13-5 ; B, pallial canal, somatocyst and attachment of muscle of polygastric stage, ' Discovery II ' St. 2636, x 57 ; C, damaged polygastric specimen from 'Discovery II' St. 695, to show pallial canal and somatocyst. SYSTEMATIC AND BIOLOGICAL ACCOUNT 99 Material. Bermuda, Coll. Wm Beebe; 1645-1828111., two pairs of nectophores (separated), two large nectophores, one eudoxid. 'Discovery II' Station 695, 370-0 m., one connected pair, 1 cm. in diameter. Station 2636, 950- 550 m., one pair of nectophores (separated) with stem groups. Station 2639, 1200-600 m., one pair of nectophores (separated). Distribution. Hitherto the species had been found only in the Atlantic Ocean. 'Discovery II' Station 1580 (450-0 m.) in the Indian Ocean. Holotype. One larger nectophore — a smaller, now detached, probably belongs — with attached stem groups, 'Discovery II' Station 2636 (Pis. VI, VII; Text-fig. 48 A). Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 6. HIPPOPODIIDAE Hippopodius hippopus (Forskal), 1776. The four peculiar teeth of the nectophore of H. hippopus, and probably the two of Vogtia glabra are homologous with, and vestiges of, more numerous teeth found in other more spiny species of Vogtia. Bigelow ( 1 9 1 1 b, p. 2 1 1 ) states that the nectosome of Hippopodius hippopus is not effective as a swimming organ. Certainly, if retracted preserved nectophores alone are studied, this is an understandable deduction to make. But in life the upper, oral side of the nectosac has a very different appearance, and takes part in a scooping action (as it fills with water) reminiscent of a movement of the appendages of a cirripede. There is a reduplication of the dorsal part of the nectophore (see Chun, 1897a, fig. 11) which gives the thin upper wall of the nectosac freedom of movement. Observation with a lens showed me that it is chiefly the upper, dorsal half of the subumbrellar musculature that contracts. There is a distinct line of demarcation visible between the velum and subumbrella. The whole swimming action of the nectophore is reminiscent of Rosacea cymbiformis. The contraction of the nectosac of Hippopodius hippopus appears to be an advance upon that of Rosacea. The extent of the living nectosac of Hippopodius hippopus would surprise an observer who is familiar only with preserved and contracted material. I refer elsewhere (p. 138) to the 'smoky' appearance assumed by nectophores of this species when they are stimulated, a phenomenon mentioned also by Korotneff (1884) and by Chun (1888 a). When undisturbed the nectophores are quite trans- parent, a condition to which they return in about ten minutes after stimulation has ceased. The four teeth may show this ' smoky ' appearance even when unstimulated. When the nectophores, some time after detachment, become moribund they finally become ' smoky '. The change appears to take place in the ectoderm, as Chun also observed. This species is one of the siphonophores most easy to keep in confinement. A specimen with six or more nectophores, taken at Villefranche at 9 a.m. 19 May 1949, was still alive and pulsating at 7.45 p.m. on 23 May, the room temperature being about 200 C. By 28 May it had given up swimming and had lost three nectophores but was still transparent and able to turn 'smoky' when removed (in one piece) to a watch-glass. At 7.25 p.m. on that day I dropped the specimen into Bouin fixative, whereupon one more nectophore became detached. Detached nectophores can swim for days. The female gonophores were particularly active swimmers when detached, but as stated elsewhere (p. 77) I was unable to fertilize and rear the eggs. A specimen of H. hippopus taken at Villefranche on 18 March bore fifty trichocercous cercariae (with eyes), probably the larval stage (Cercaria setifera Monticelli) of the trematode Lepocreadium album Stossich.1 Some had made tubular tunnels into the jelly and lost their tails (metacercariae). 1 See Dollfus, R.-Ph. (1925) and Palombi, A. (1937). 13-2 ioo DISCOVERY REPORTS Chun (18880) said1 that the postembryonal stages of Hippopodins which he had previously observed from considerable depths in the Mediterranean, he had succeeded in capturing in an unbroken series at the surface near Orotava, Tenerife. He gave a good description of a stage that he had been able to study on three occasions in detail, and figured it well on p. 1 150. He did not at that time figure the other stages of development because he intended shortly afterwards to publish a full description of the postembryonic development of Hippopodius. It was not, however, for twenty-five years that he did so (1913, p. 35), though he had meanwhile published two figures of young stages. The sequence of developmental stages can be seen in the following published figures: Metchnikoff (1874), P^ ri> figs. 5-8; Chun (18886), pi. 2, figs. 1, 3; Moser (19240), pi. 3, fig. 2; Chun (1913), figs. 5, 6; Chun (18970), figs- °^> 6c; Chun (18880), p. 1150; Chun (1913), figs. 7, 8. The relationship of the two budding zones in the adult was best illustrated by Chun (18970, ng- IJ)5 and Richter (1907), pi. 28, fig. 27. Vogtia Kolliker, 18536 The four specific names spinosa, pentacantha, serrata, and glabra are used in the sense defined by Bigelow & Sears (1937), except that in pentacantha there are sometimes spines around the base of the median dorsal prominence, so that the facets are not entirely smooth. There is a good deal of variation in the amount and kind of spinosity from specimen to specimen, and probably from the first to the last of the series of nectophores on one specimen, but I agree that there are four species. None of the species were found in the Red Sea area, but Vogtia glabra was taken by Mr Fraser-Brunner in the Gulf of Aden. Some notes on the larvae of V. serrata will be found on pp. 73-77 in the special section on post-larvae of Prayidae and Hippopodiidae. Vogtia glabra Bigelow, 1918. (Plate IV, fig. 2.) The following interesting observation on phosphorescence in this species was made by Mr P. M. David: 'When bringing the 70 cm. net inboard after the 1000-750 m. haul at ' Discovery II ' Station 2890, an animal was observed phosphorescing on the \ in. netting above the closing band. It was seen to be part of a siphonophore, and was transferred to a small jar of sea-water. When taken into a dark cabin, it was seen to luminesce when prodded, or when the water in the jar was agitated; the light which was greenish-blue lasted several seconds before fading. The animal broke up after a while, but the component parts were still able to light up, the light seemed to be evenly spread throughout each part.' The preserved specimen consisted of ten nectophores of definitive shape and four buds, three of them still attached to the contracted stem and appendages. All the nectophores, the largest of which measures 17 mm. in diameter, are now quite transparent. The only other observation on luminescence in Hippopodiidae of which I am aware was made by Korotneff (1884, pp. 281, 282) on Hippopodius hippopus. He correlated luminescence by night with a smoky colour by day. Sulculeolariinae subfam. nom.nov. Galettinae Stechow, 1921 Sulculeolaria Blainville, 1830 (no published specific names or figures cited). Type species S. quadrivalvis Blainville, 1834 (figured). Galeolaria Blainville, 1830 (preoccupied). Galetta Stechow, 192 1. 1 English translation in Ann. Mag. Nat. Hist. London (6), III, p. 214, i88q. SYSTEMATIC AND BIOLOGICAL ACCOUNT 101 Nomenclature. Blainville (1830) named one species of Galeolaria (preoccupied, renamed Galetta) australis without any indication as to its identity, except references to three names that were only manuscript names in 1830. Blainville's generic description of Galeolaria refers to fairly firm gela- tinous, regularly and bilaterally symmetrical, subpolygonal or oval animals that have a large posterior ostium; a velum and a two-winged subostial lamella; a muscular subumbrella occupying much space inside; and a stem, borne on the anterior-superior face that issues from a bilabiate hydroecium. Various authors have since used the name australis for at least three separate species, Diphyes biloba Sars, 1846, G. turgida (Gegenbaur), 1853, G. chuni Lens & van Riemsdijk, 1908, and a new one to which I give (p. 108) the name Sulculeolaria angusta sp.n. We are, however, faced with an insuperable difficulty in the use of the name Galetta Stechow, 1921 (nom.nov. for Galeolaria Blainville, 1830, preoccupied Lamarck, 1818), for any of these species. One legalistic solution of the whole problem is as follows : The ' specific ' names cited by Blainville under Galeolaria in 1830 were all manuscript names. Opinion 46, rendered by the International Commission on Zoological nomenclature, seems to be applicable here, and we have to ask ourselves what is 'the first species published in connexion with the genus' after 1830; for it will be this species that becomes ipso facto the type species of Galetta Stechow, 1921. The answer is: (1) Galeolaria australis Quoy & Gaimard, 1833. But this is specifically indeterminable, as I will show on pp. 101-4, and therefore to be rejected. (2) G. qiiadridentata Quoy & Gaimard, 1833. So that if we reject G. australis on the grounds that it is specifically indeterminable, G. qiiadridentata Quoy & Gaimard, 1833 becomes the type species of Galeolaria Blainville, 1830 (preocc, =Galetta Stechow, 1921). But since I shall show (p. 109) that zoologically G. qiiadridentata Q. & G. 1833 is the same species as Sulculeolaria quadrivalvis Blainville 1834, 1 and since this species S. quadrivalvis Blainville, 1834, is, as I showed in 1932, the type species of Sulculeolaria Blainville, 1830, by virtue of opinion 46, Inter- national Rules of Nomenclature, it follows that Galetta Stechow, 1921, becomes a synonym of Sulculeolaria Blainville, 1830. Zoology. From a zoological point of view this may be a good thing, for Sulculeolaria has been used lately for those species which have teeth round the ostium of the anterior nectophore, and Galetta for those that have none. But I shall show (p. no) that sometimes Sulculeolaria quadrivalvis has no teeth; and my Text-fig. 50 A, B, D shows that Sulculeolaria [Galetta] biloba has a peak on the dorsal side of the ostium, which really represents undivided dorsal teeth. What generic name is the working systematist to use? I dislike upsetting the involved synonymy of siphonophores again, but regret that I must follow the logical course of using the name Sulculeolaria Blainville, 1830, for all the species, whether toothed or not. The subfamily will have to be renamed Sulculeolariinae. I hope that systematists in future will find that the identification and naming of Sulculeolariine Diphyids is fairly easy. It certainly has not been so in the past. The next reviewer of the genus will have a sounder basis from which to start. 'Galetta australis' G. australis Bigelow, 191 ib ( = 'G.' biloba Sars, 1846). G. australis Bigelow & Sears, 1937 ( = 'C turgida Gegenbaur). 'G. australis' auct. refers to several distinct species. More work is needed to confirm my views on how many are represented, and what is their morphology. Quoy & Gaimard's original ' Astrolabe ' 1 The S. quadrivalvis cited by Blainville in 1830 is only a manuscript name. Figures of it were published by Blainville in 1834, at which date it becomes a valid specific name. 102 . DISCOVERY REPORTS specimen came from the Southern Indian Ocean (360 32' S. lat., long, unknown), so that the name of their species, if applicable at all, should be used for one that occurs there. From the official narrative of the ' Astrolabe ' expedition it is difficult to follow the track from Trinidad Island1 in the South Atlantic, round the Cape to Albany, Western Australia. On 3 September 1826 'Astrolabe's' position was 370 17' S. lat., 270 W. long. On 13 September it was decided to make for Albany, and on 20 September she was following the parallel of St Paul Island in the South Indian Ocean, six or eight miles to the northward of which she probably passed on 21 September. Albany was reached on 7 October. As she met with particularly violent weather in the eastward passage of the Indian Ocean it seems unlikely that any collecting was done there although sea-temperature records were made. On the return journey the course taken was through Sunda Strait to Mauritius and Reunion and down to the Cape. It seems therefore more probable that the latitude, 360 32' S., given for the capture of the original ' G.' australis was where the ship's track cut this parallel, somewhere off Algoa Bay, particularly as Quoy & Gaimard say that it was here and off Western Australia that plankton was abundant in the Indian Ocean. An examination of hauls made by R.R.S. 'Discovery II' in this area, namely at Station 1567, 370 50' S., 350 47' E., and Station 1568, 340 48' S., 340 28' E., reveals the presence of at least three species of 'Galetta' that systematists in general would report as 'G.' australis. The differences are seen at once in the lamellae of the posterior nectophores, but it was only after completing this report that I was able to discriminate between anterior nectophores. One type of posterior nectophore can be distinguished by having a rounded, undivided lamella, as in the Mediterranean species ' G.' turgida. The second has a thickened, notched lamella, as described below for ['G.'] Sulculeolaria angusta sp.n. The third kind of posterior nectophore is similar in profile to that figured by Bigelow (191 ib, pi. 5, fig. 9), and the lamella has the 'duplex curve', figured by Bigelow (191 lb, pi. 5, fig. 9 bis). This is the well known and adequately figured ' G.' biloba (Sars), 1846 (see Text-figs. 49-51). Quoy & Gaimard's (1834) figure of 'G.' australis might represent any one of the three. The absence of any indication of a somatocyst from this old figure may or may not be significant. If given only the basal (articulating) end of a well-preserved posterior nectophore of 'G.' biloba, ' G.' turgida or ' G.' angusta ' it would now be possible for a specialist to identify it. If the lamella was in good shape too, any competent systematist could do so. Bigelow & Sears (1937) gave a key for the identification of the anterior nectophores of the 'australis' group of ' Galetta '. My criticism of this is that Leloup's species meteori is more clearly related to Lensia subtilis and should be excluded. ' Galetta ' chuni is now so well known that it gives no trouble. 'G.' turgida, common in the Adriatic, has a minute somatocyst and a divided 'basal wing' in the anterior nectophore and a rounded one in the posterior. The posterior nectophores are very character- istic, and each end exhibits characters that are very useful to the systematist. Bigelow & Sears (1937) found the task of identification of posterior nectophores more difficult than that of the anterior ones. The morphological features of specimens attributed by various authors to this ' species ' australis, and over which there has been some confusion, are as follows : The somatocyst of the anterior .or superior nectophore may be minute or nearly as long as the 'mouth-plate' or lamella. The posterior nectophore may have a comparatively ' broad baso-ventral sector ' when viewed laterally, or may be long and slender, and may or may not have lateral, longitudinal ' wings '. The ' mouth-plate ' or lamella of the posterior nectophore may be long or short ; notched, or have a ' duplex curve ', or be convexly rounded at the end. Bigelow (191 1 b) is the authority for the use of the name australis in recent years, having provided (pis. 5 and 6) both good photographs and not such good drawings of a species of ' Galetta ' bearing 1 Position: 20° 30' S, 290 20' W. SYSTEMATIC AND BIOLOGICAL ACCOUNT 103 that name in a work which is one of the standard references for the group. The legends to the figures do not indicate the locality of the particular specimens, but presumably they came from the Eastern Tropical Pacific. Much as I dislike having to criticize any part of Bigelow's delightful work which I admire so much, there is one drawing, beautiful in its execution, which I mistrust, fig. 1 of his plate 6. The only part of it which I query is the stiff base of the lamellar plates. The rest of the drawing appears to represent 'G.' biloba (Sars), but in that species, as reference to my Text-fig. 50 of the base of the anterior nectophore shows, the dorsal end of the lateral surfaces for articulation with the posterior nectophore reaches well out below the ostium of the nectosac, in fact to a point not far short of its centre. This means that the basal portion of the lamella or mouth-plate of the anterior nectophore in 'G.' biloba is much more rigid than the thin distal part, which in preserved specimens is usually bent back. Bigelow's fig. 1 of plate 6 does not show this feature, and is, I suspect, not correct. It is not at all easy to draw this region even with a camera lucida, because it is only rarely that a specimen is preserved in a satisfactory condition, and living material is not readily available. Then again this fig. 1 of Bigelow does not show the little pit in the roof of the base of the anterior nectophore that houses the tip of the stalk, a minor but a very characteristic feature of 'G.' biloba. It is of course possible that Bigelow has accurately drawn a representative of a Pacific species that I have not seen, but we are finding more and more that most siphonophores have very wide distributions and my experience is based on hauls from a very extensive oceanic area. Besides, Bigelow's photograph 8 on plate 5, though not strictly a side-view, shows all the characters of 'G.' biloba that I have mentioned. The somatocyst and the basal pit are diagnostic, though the somatocyst does not show very clearly. Kawamura (19 15, fig. 10) tried, unsuccessfully, to fit together drawings 1 and 3 of Bigelow's plate 6, but had to distort fig. 1, and left no space for the appendages. Bigelow & Sears (1937) attributed 514 anterior and 673 posterior nectophores from the Mediter- ranean to this ' species ', australis. All of the posterior ones showed (in lateral view) ' the broad baso- ventral sector' characteristic of both biloba and of turgida (see Bigelow, 191 ib, pi. 5, fig. 9, lateral view — there are two figs. 9), and all had an undivided basal wing, which is a character of turgida. Bigelow & Sears 's (1937) fig. 26 shows typical bases of turgida nectophores. The great breadth of this baso-ventral sector in 'G.' biloba and 'G.' turgida, which carries, at the proximal end of each ventral wing, a long (dorso-ventrally) and rounded articulating surface, necessarily implies that the base of the corresponding anterior nectophore is equally broad dorso-ventrally and that the margins of the hydroecium are hollowed out to take these articulating surfaces which transmit the thrust of the posterior nectophore. This condition is particularly well marked in the Mediterranean species that I refer to below as 'G.' turgida. Bigelow & Sears (1937) also stated that these posterior nectophores from the Mediterranean have undivided basal wings, but not, as I have already remarked, whether they have the duplex curve mentioned on p. 35 of their 'Thor' report and figured by Bigelow (1911ft, pi. 5, fig. 9 bis). This is important because in recent years Dr Gamulin has sent me from the Adriatic some well-preserved specimens of anterior and posterior nectophores of a ' Galetta ' of which the posterior ones possess a convexly rounded lamella, neither notched as in chuni nor with the duplex curve of Bigelow's figure of ' australis', but like Gegenbaur's (1854) pi. XXIII, fig. 3. In fact I believe Gamulin's specimens are the ' G.' turgida of Gegenbaur. The anterior nectophores have a much smaller somatocyst than Bigelow's figures (1911ft, pi. 5, fig. 8; pi. 6, fig. 1) of australis, and resemble those in fig. 26 of Bigelow & Sears (1937), which are in fact specimens of 'G.' turgida Gegenbaur. • At the end of September 1952, when the draft of this report had been completed, I was able to re-examine, through the courtesy of Dr Kramp of the Copenhagen Museum, a number of the anterior and posterior nectophores, separated into vials, from 'Thor' Stations 206 (1910) and 297 (191 1). io4 DISCOVERY REPORTS As I suspected, they were all, with the exception of one rogue {Sulculeolaria quadrivalvis), repre- sentatives of S. turgida (Gegenbaur). I carefully examined, after staining, eighty-five posterior (inferior) nectophores from Station 297 and found that they all possessed the simply rounded, undivided lamella of S. turgida (Gbr). The fifty-five anterior (superior) nectophores that I examined from the same station all had the minute somatocyst and conformation of the base of the nectophore characteristic of S. turgida. The specimens were not in very good condition, and I did not check the presence, or run of the lateral commissural canals. The problem has been to link up the anterior and posterior nectophores of the various ' Galetta ' forms so as to define species, and then to see if we can apply the specific name australis to any of them. But the natural desire to retain Quoy & Gaimard's name should not blind us to the danger of a perpetuation of the present state of confusion. This review of ' Galetta ' australis, then, shows that whilst it is possible that Quoy & Gaimard's specimen came from off Algoa Bay, this is by no means certain. Also that in this area are to be found three species of ' Galetta ' which can be separated by their posterior nectophores as well as by their anterior ones. Quoy & Gaimard only figured the anterior nectophore of 'G.' australis, and that very inadequately. It seems doubtful on the whole whether it will ever be possible now to use the name australis correctly. I have now found the North Atlantic 'G.' biloba (Text-figs. 49-51), with which I am well acquainted, in the Indian Ocean. It is worth while mentioning this to confirm Bigelow's statement that the identity of the Atlantic biloba with an Indo-Pacific species has been sufficiently established by comparison of series from the different oceans. It is a dioecious species which has been taken abundantly by the Marine Biological Association, Plymouth, in the Celtic Sea; by the Irish Fisheries Board; and at Valencia Island by Miss Delap, who noted its ability to jump out of the water, made interesting sketches of it and its developmental stages and presented specimens to the British Museum. It is of interest to record that had it not been for a chance invitation to Robert Prescott of Latymer School, a member of the Museum's Junior Naturalist Club, to do some plankton sorting, the presence of 'G.' biloba in the Indian Ocean might not have been demonstrated, and the whole question of ' Galetta ' species might have remained in confusion. [Galetta] Sulculeolaria biloba (Sars), 1846. Galeolaria australis Bigelow, 191 lb, pi. 5, fig. 8; Bigelow, 1918. Galeolaria australis Candeias, 1929 (part C, fig. 3). Non Galetta australis Bigelow & Sears, 1937 (? = 'C turgida). Sars's (1846) figures of this species are very accurate. His fig. 16 shows not only the characteristic somatocyst, but also the minute indentation in the base of the nectophore or roof of the hydroecium that houses the terminal end of the stem. Sars's fig. 17 of the young posterior nectophore shows, under a lens, the characteristic ' duplex curve ' of the lamella. I feel obliged to discard the name Galetta australis which Bigelow has used for biloba as being impossible to apply to any zoological concept, and to use for this widely distributed and adequately figured species its original name Sulculeolaria biloba (Sars), 1846. Anyone who is familiar with the North Atlantic species would say at once of Sars's figure ' that is the species ', whereas we shall probably always be in doubt about Quoy & Gaimard's Galetta australis,1 which was figured in such a way that it might represent any one of several Sulculeolariine species known to-day. Distribution. Under the name G. australis Q. & G. Bigelow (1918) reported on thirty specimens of this species from surface and intermediate hauls in the Western Atlantic ; and could find no distinc- tion between them and series from the North and South Pacific. Lens & van Riemsdijk (1908) 1 Bigelow held the contrary opinion. SYSTEMATIC AND BIOLOGICAL ACCOUNT 105 reported on a species under the name G. biloba Sars from the Malaysian area, but their figure is not good. Through the courtesy of Prof. Dr H. Engel, Director of the Zoologisch Museum, Amsterdam, I was able to re-examine a little tube full of 'Siboga' specimens labelled ' 119 E 1 Sulculeolaria biloba Sars '. These must be Lens & van Riemsdijk's specimens, now in alcohol, from ' Siboga ' Station 1 17 A, lat. i° 15' N., 1230 37' E. I could not recognize the figured specimen, but picked out seven anterior and seven posterior nectophores that were the best preserved. Most of the specimens were in very poor shape. The seven posterior nectophores had bilobed lamellae with a 'sigmoid curve'. One or two were preserved well enough for me to be able to decide that though small specimens they corre- sponded sufficiently well with those of G. biloba from the North Atlantic. I am satisfied that the C rod lat C.com iQmm Cradd Som Text-fig. 49. Sulculeolaria biloba. Three anterior nectophores taken by M.B.A. Plymouth in Celtic Sea, July 1937. x 4-4- seven anterior nectophores that I selected represented G. biloba Sars, though they were in poor condition. The somatocysts were not easily visible, but I think that I could just detect in each the characteristic arrangement of the somatocyst and the ' basal pit ' for the proximal end of the stem, such as I have described. These seven anterior and seven posterior selected nectophores could hardly represent any other species known to me. 'Discovery II ' took specimens at Stations 1567 and 1568 as reported above (p. 102) and very recently thirteen anterior and nineteen posterior nectophores were sorted out from a surface-haul from 'Discovery II' Station 2688 on the '90° E.' line of stations in lat. 200 19' S. The average length of the nectosac of these anterior nectophores is 8-8 mm. I am quite satisfied therefore that the Atlantic species Sulculeolaria biloba occurs in the Indian as well as the Pacific Ocean. It is possible that the North Atlantic specimens reach a greater size than the tropical Indian, but we have not sufficient data yet to be able to judge. 14 io6 DISCOVERY REPORTS Morphology. Lateral commissural canals are, I think, always present. In S. turgida they are sometimes absent. There are two inconstant characters about the commissural canals: (i) the position on the dorsal radial canal from which they spring — sometimes opposite each other, but usually the right-hand one given off before the left; (2) their course before junction with the lateral radial canals. Crodd Ccom C rod lot Crodd -Ccom Crodd Text-fig. 50. Sulculeolaria biloba. Basal parts of anterior nectophores taken by M.B.A. Plymouth in Celtic Sea, July 1937. A, dorsal view; B, D, lateral views; C, ventral view, x 8. The ventral side of these lateral radial canal-loops always springs from the pedicular canal-junction (Text-fig. 50 C) of all four radial canals, and never from the ring canal as described by Bigelow & Sears for a Mediterranean species (? S. turgida) which they called Galetta australis (1937, fig. 26 A). It is not easy to see this junction in a lateral view. One side (left) of the base of the posterior nectophore is higher than the other (Text-fig. 51 B). In consequence the cavity round the central boss of the base of the anterior nectophore is asymmetrical, as can be seen in my Text-fig. 50 C. The width of the base SYSTEMATIC AND BIOLOGICAL ACCOUNT 107 of the paired lamellae or mouth plates is greater than that of the base, and there is a groove between outer edge of lamella and base, shown in my lateral views. The somatocyst is probably ovate in life, as it is in the best preserved specimens. In those in which it is shrunken there is left an ovate outline in the jelly, seen perhaps only in certain lights. The upper end of the stem complex projects upwards into a small pocket in the base of the nectophore. This minor feature appears to be constant and characteristic, and was shown in Sars's original figure, as well as by Bigelow (191 ib, pi. 5, fig. 8). The base of the mouth-plate or lamella of the anterior nectophore is thicker than the terminal part, and projects out nearly to the centre of the ostium. iQmm C ped Text-fig. 51. Sulculeolaria biloba. Posterior nectophore from Celtic Sea, M.B.A. Plymouth 1937 Cruise, St. 6. A, C, dorsal and lateral view of the mouth-plate, x 2-5 ; B, ventrolateral view of proximal end to show triangular, articulating facets, x 3-5 ; D, posterior nectophore, x 3-5. Bigelow's ' baso-ventral' sector indicated between k and / in fig. B. [Galetta] Sulculeolaria turgida (Gegenbaur), 1854. Gegenbaur (1854) described and figured (under the name Diphyes turgida) a species of 'Galetta with no obvious somatocyst in the anterior and an undivided, rounded lamellato the posterior necto- phore from the Straits of Messina. Candeias (1929) found the anterior nectophore of a similar species, and in 1932 I reported on similar specimens from the Great Barrier Reef. Since then I have often seen these forms. In recent years Dr T. Gamulin has been sending me fine specimens of both nectophores of this species from the Adriatic. Since a species of this sort was long ago described from the Mediterranean and can still be recog- nized there, I think it well to retain the name turgida for it. The two points about the alleged absence of a somatocyst and the undivided lamella of the anterior nectophore can both be explained, I still think, by the difficulties encountered in trying to see the details of these transparent animals. Dr Sears herself says ' both characters which make the species . . . unique. . .are such that they could have been overlooked easily'. The most characteristic thing about the Mediterranean species is the rounded and undivided lamella of the posterior nectophore (Text-fig. 52), which Gegenbaur clearly figured. The larva was figured by Gegenbaur (1853, pi. 16, fig. 20) and labelled in error, Diphyes sieboldii. On 15 April 1950 Dr Mary Sears published a paper on siphonophores from the Marshall Islands. In it appears a review of this subject and a description of the deformed anterior nectophore of a similar species under the new name Galetta bigelowi in honour of Dr H. B. Bigelow that great contributor to our knowledge of the group, and himself the foremost authority on it. I have found similar specimens and think it is a distinct species. io8 DISCOVERY REPORTS S. turgida occurred in the Indian Ocean as follows (depth in metres follows the number of the Station) : A. Off south-east and east coasts of Africa. 'Discovery II' Stations: 1371, 146-0; 1373, 135-0; 1374, 230-0; 1375, 210-0; 1568, 1400-0 and 5-0; 1573, 800-0; 1581, 600-0; 1585, 500-0; 1587, 450-0; 1589, 600-0. B. On 900 E. line of ' Discovery II' Stations: 2688, 121-0; 2894, 142-0; 2895, 54-0. C. In the hauls made on the 320 S. line of ' Discovery II ' Stations 1736-66, S. turgida was not taken. The species has been taken by 'Discovery' and 'Discovery II' in the Atlantic at the following Stations: 89, ioo(-o); ioob, 0-5; 292, 110-0; 673, 340-0; 676, 290-0; 677, 420-0; 678, 360-0; 680, 260-0; 681, 500-250; 682, 375-0; 699, 370-0; 701, 242-0; 702, 236-0; 703, 358-0; 708, 208-0; 712, 224-0; 713, 200-0; 2635, 280-0; 2648, 500-0. Z mm .P/.M. i mm Text-fig. 52. Sulculeolaria turgida, posterior nectophore, 'Discovery II' St. 1581, x 11. A, ventral view; B, lateral view. [Galetta] Sulculeolaria angusta sp.n. The outline of the anterior nectophore (Text-fig. 53 C) has the general Sulculeolaria form. The halves of the long ' mouth-plate ' or lamella of the anterior nectophore have a roundly pointed posterior edge as can be seen in a well-preserved specimen after staining.1 The somatocyst is minute, but varies in size. The profile of the posterior nectophore is long and narrow, without the ' broad baso-ventral sector ' of Bigelow. The lamella of the posterior nectophore is roundly notched as in Galetta chuni, but is thicker in its basal half. About the centre of the mouth-plate appears a variable little prominence. It lies at the terminal end of the thick basal part of the lamella. Between this and the notch is a thinner area. No lateral ridges or wings have been seen, even on well-preserved specimens of the posterior nectophore. 1 It is generally advisable to stain Sulculeolariinae. With specimens preserved in formalin which has been buffered with hexamine, this may be done fairly quickly, after one or two washes in distilled water, by adding two or three drops of Delafield's haematoxylin to a watch-glass of distilled water. SYSTEMATIC AND BIOLOGICAL ACCOUNT 109 The type material was taken by R.R.S. 'Discovery II' at Station 1586 in an oblique haul from 550 m. to the surface, and consisted of twenty-seven anterior nectophores and twenty posterior, of which I select one posterior nectophore Brit. Mus. (Nat. Hist.) Register No. 1950. 6. 20. 1 as the holotype. It is not possible to select an anterior nectophore as well, because we do not know for certain whether any were connected in life with the holotype. Seven posterior nectophores were taken also at 'Discovery II' Station 1374 in an oblique haul from 230 m. (off the coast of Natal). The posterior nectophore figured by Candeias (1929, fig. 6) probably belongs to this new species. 2 mm.\ NCh Text-fig. 53. Sulculeolaria angusta sp.n. A, B, lateral and dorso-lateral views of holotype posterior nectophore, Brit. Mus. Reg. No. 1950.6. 20. 1., 'Discovery II' St. 1586,550-001., x 8-5 ; C, anterior nectophore probably of this species, 'Mabahiss' St. 61, night, o m., x 8-5. k-e = baso-ventral sector. Sulculeolaria quadrivalvis Blainville, 1830. The question whether Sulculeolaria quadrivalvis and S. quadridentata are synonyms of one and the same species was last reviewed by Bigelow & Sears (1937). They concluded (1) that there was an unbridged gap between the bidentate and the extreme quadridentate forms of the anterior nectophore; (2) that the bidentate form occurs in both sexes, and that the only quadridentate colonies of which the sex has been definitely determined chanced to be female; (3) that it is possible that the quadridentate form is a sport that appears more often in female than in male colonies ; (4) that it seems wisest to recognize two species. no DISCOVERY REPORTS ' Discovery II ' took 144 well-preserved anterior nectophores of various sizes, together with a similar number of posterior ones on 1 1 November 1 95 1 at Station 2905 in the Red Sea. Fifty-five are bidentate, and twenty-eight are of the extreme quadridentate form. Fifty-eight others have a pair of lateral teeth of various lengths (which I am unable to record at the moment). In my opinion they bridge the gap between the two forms satisfactorily. Three have no teeth at all but resemble this species more than they do S. chuni. My conclusions are that (1) it is not a question of the number of teeth but of the total development of the ' denticulate character '. For in bidentate forms the two teeth may be quite small, whilst in the four-dentate form they never are; and the inner distal corners of the mouth-plate of the bidentate form are generally rounded and not denticulate, whilst in the four-dentate form they are always denticulate, so that there is a gradual change from the condition where only two dorsal teeth are present and the mouth-plates are rounded, to the extreme four-dentate form with long dorsal and lateral teeth and 'denticulate' mouth-plates; (2) only one species is to be recognized. The name quadrivalvis has priority. It is possible that a genetical explanation of the variability in number and size of these periostial teeth might be found in a chromosomal deletion or in a preponderance of a suppressor gene. Fortunately this species is fairly abundant at times in the Mediterranean, and also in Valencia Harbour, Ireland, at either of which places genetical experiments might be tried in order to confirm the view that only one species is represented. S. quadrivalvis would then be established as one species of siphonophore whose characters were not so clear cut as most others. Lensia Totton, 1941 Since writing on the species of Lensia in 1941, I have repeatedly found specimens of them all, including Leloup's interesting little L. meteori, which I had not at that time seen. My grouping into species appears to be sound in the light of twelve years' further work, and only three further new species have appeared. Nineteen of the twenty-one known species are found in the Indian Ocean, two having been previously recorded there. I have figured fourteen of these before (1932, 1941). Bigelow & Sears (1937) figured L. conoidea, L. fozvleri and L. multicristata; and Leloup (1934a) figured L. meteori. My present identifications are in accordance with these figures and related descriptions. Lensia hotspur Totton, 1941. The syn types from ' Discovery ' Station 100 exhibited a very small hydroecial cavity which included the base of the somatocyst. In the Red Sea specimens the base of the somatocyst generally projects down slightly below the baso-lateral margins (Text-fig. 54C-E). At 'Manihine' Aqaba Station 13 (135-0 m.) four well preserved anterior nectophores, between 4-0 and 4-5 mm. in length, were taken. Two of them are of this same type, and two of the type in which the base of the somatocyst does not extend below the baso-lateral margin. From 'Manihine' Red Sea Station 8 came 19 anterior necto- phores containing both types and intermediates in which the base of the somatocyst is on a level with the baso-ventral margin. Associated with anterior nectophores from the Gulf of Aqaba were posterior nectophores and eudoxids, which unfortunately cannot be described and figured at the present time, except to say that the mouth-plate of the posterior nectophore has a rounded notch in its distal edge, and that the upper end of the left ventral ridge is not square but angular. Specimens of this species though abundant are nearly always in a poor condition. But systematists in future would do well to watch for variation in this respect, and to make certain that we are not dealing with more than one species. It is certainly clear now that the 'Carnegie' specimen drawn by me (1941, fig- 13) belongs to a different species, which has since been taken abundantly by H.M.S. ' Challenger ' in the E. Tropical Pacific in October 1950. The new species I have named in honour of the ship. L. hotspur occurs side SYSTEMATIC AND BIOLOGICAL ACCOUNT m by side in the Red Sea with L. subtiloides, but can at once be distinguished by the fact that there is practically no stalk to its relatively shorter somatocyst. Lensia challenged sp.n. Thirty anterior nectophores of what appears to be a distinct new species of Lensia were taken by Surgeon Lt. D. O. Haines, R.N. in H.M.S. 'Challenger' (Cdr. G. S. Ritchie, D.S.C., R.N.) off the coast of Southern California on 21 October 1950 at Station 104, 23° 51' N, 1120 59' W. The species occurred also at Stations 103, 105, 106, 107, 108, 109 and 114, the first being in 2i°4i' N., iii°43' W., and the last in 310 12' N., 1170 31' W. At all these stations the specimens were in a layer of presumed plankton located by echo-sounding. Temperatures in the layer varied from 570 to 61 ° F., and the depths from 37 to 220 m. At night the presumed plankton layer was observed to remain at about 37-55 m. as seen on the echo-sounding recorder and by day the layer was occasionally seen on the trace at about 274 m. A number of bathythermograph temperature/depth records taken at the time all show a similar pattern — 80-670 F. at the surface, a degree or so less at 27-5 m., 56-5-62° (once 72°) at 56 m. and 49-5-56° at 131 m. Captures of Lensia challenged sp.n., by H.M.S. 'Challenger' in the E/S layer of plankton off S. California in October 195 1 are given in the following table. It shows the descent before sunrise. Sunset Station Depth of layer (m.) T°F. Surface E/S Layer + 12 min. + 16 + 22 + 25 + 30 + 36 -3° Sunrise 114 io5 108 103 109 106 107* 104 73 64-91 9i 73 IOI 37-64 91* 220 67 75-5 69-5 80 69'5 76 70 78-5 53 55"5 52 S7'5 52 61-5 57-5 * The haul at this depth was probably made a few metres above the E/S layer. Commander Ritchie (1952) stated that migrations appear to depend on light, and not on various depth/ temperature changes, though Moore (1943) makes the generalization that among Siphonophores there is a tendency in the Florida Current for a change in day level in relation to temperature, the animals there moving upward when cold water comes closer to the surface. Herdman (1953) too shows that there appears to be a certain measure of agreement between the depths at which temperature-dis- continuities occur and those of the various echoes recorded. But he adds that it is a matter of con- jecture whether this connexion depends entirely on the purely physical boundaries, or on possible concentrations of falling detritus or animal life checked in its fall by a sudden change in density. According to Cdr. Ritchie the rate of change of depth for descent averaged about 4-26 m. per minute, the movement being very regular over five days. The ascent in the evening was more ragged, small groups sometimes appearing to come up ahead of others, and speeds of ascent varied from 2-74 to 0-91 m. per minute. Open plankton nets were lowered into the mid-level of the E/S layer at night. No fish were caught, and the fine ' texture ' of the echo-sounding record obtained led Cdr. Ritchie to believe that the echoes came from a vast layer of plankton rather than from fish feeding upon that plankton. Lensia challengeri bears some resemblances to both L.fowleri Bigelow and L. hardy Totton. From L.fowleri it can be distinguished by a projection, shaped like the peak of a cap, that is found on the baso-dorsal margin of the anterior nectophore, which has the appearance in lateral view of a dorsal ii2 DISCOVERY REPORTS tooth (Text-fig. 54A). The hydroecium is less deep than that in both L. fowleri and L. hardy. The somatocyst, which alters shape, presumably in accordance with the amount of fat in store, may be spheroidal or egg-shaped, often much flattened, but its base remains close to the basal facet of the nectophore, and it usually reaches to the ventral facet. The canal leading immediately to the somato- cyst is well on the baso-dorsal side. In dorsal view it will be noticed that the mouth-plate on the left (morphologically the right-hand one) is much broader than the other (Text-fig. 54 B). Both halves of the mouth-plate are much less deep than in L. hardy and L. fowleri. It may be found advisable later on to include the three species fozvleri, hardy and challengeri in a separate genus. The eudoxids of all are known but have not all been described and figured. The holotype specimen (Text-fig. 54A, B) bears the B.M. (Nat. Hist.) Register number: 1953.8.11.1. Cpa cped Text-fig. 54. Lensia challengeri, sp.n., anterior nectophore from 'Challenger' St. 104, Oct. 1950; A, lateral view, x 12; B, dorsal view of mouth-plate, x 35. Lensia hotspur, anterior nectophore, Gulf of Aqaba, Manihine St. 1 ; C, antero-lateral view of base, X40; C1, dorsal view of mouth-plate; D, lateral view, x 12; E, lateral view of base, x 25. Lensia cossack Totton, 1941. This species was reported by Browne from Chago, Mauritius, Farquhar and Amirante under the name of Diphyes subtiloides. I have re-examined eight of the eleven anterior and the two posterior nectophores which are now in the British Museum Collection. Lensia subtiloides (Lens & van Riemsdijk), 1908. The record by Browne from Chagos, Mauritius, Farquhar and Amirante is not for this species but for Lensia cossack. I have re-examined the specimens in the British Museum Collection. L. subtiloides and one of its associates, Diphyes chamissonis are not widely spread over the oceans. They were the only two really abundant Siphonophores that the Great Barrier Reef Expedition of 1928-9 found in the Barrier Reef lagoon, where Lensia subtiloides became relatively very rare when the SYSTEMATIC AND BIOLOGICAL ACCOUNT „3 temperature dropped below 240 C. In the Red Sea off Port Sudan, ' Manihine ' found them associated in five out of ten hauls (Text-fig. 55 D). H.M.S. 'Weston' took an abundance of specimens off Kamaran, Red Sea on 9 April 1936 at a surface temperature of 280 C. Once more the species was associated with Diphyes chamissonis. This species evidently thrives at a high temperature. Lensia leloupi1 sp.n. (Text-fig. 55 C). Since there is a new species of Lensia with which L. subtiloides might be confused a description is now given of the anterior nectophore. The nectophore differs from that of L. subtiloides in having a deeper hydroecium or longer mouth- plates. In other respects it is very similar, though larger. After comparison with Muggiaea delsmani it might be thought that Lensia leloupi was only another species of Muggiaea with still shallower Text-fig. 55. A, Muggiaea kochii, off Plymouth ; B, M. delsmani sp.n., Java Sea ; C, Lensia leloupi sp.n., ' Discovery' St. 277 ; D, Lensia subtiloides, Great Barrier Reef. A, C x 16; B X20; D x 18. hydroecium. But Lensia subtiloides and L. leloupi agree in (1) showing a notch on the edge of the left mouth-plate ('right' if viewed from the dorsal side, apex uppermost); (2) the origin of the somatocyst from a point on the hydroecial roof to the left of the highest point, and not nearly so close to the nectosac as in Muggiaea spp. ; (3) the tip of the nectosac being so near to the apex of the nectophore. This third character is found in M. bargmannae but not in the other three Muggiaea species. The type specimens of this new species Lensia leloupi were taken at ' Discovery ' Station 277 near Cape Lopez in the Gulf of Guinea, and not far from the mouth of the Ogowe river. No salinities are recorded from this station, but at Station 279, closer still inshore, the salinity was 33-08 %0 at the surface, a comparatively low salinity for Siphonophores to tolerate. The figured holotype bears the Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 1. Lensia multicristata (Moser), 1925. This species was recorded by Browne from Chagos and Amirante under the name Diphyes bigelowi. New Indian Ocean records: specimens of Lensia multicristata have been identified from the following 'Discovery II' Stations in the Indian Ocean: 847 (270-196 m.); 1743 (2100-1150 m.); 1566 (1350- 1 Named for Dr E. Leloup. D XXVII 15 ii4 DISCOVERY REPORTS o m.); 1568 (1400-0 m.); 1571 (500-0, 1400-1000 m.); 1581 (500-250, 1750-600 m.); 1585 (500-0, 1400-700 m.); 1586 (250-100, 550-0, 1650-950 m.); 1587 (450-0 m.); 1588 (500-200 m.); 1589 (600-0 m.). The other new 'Discovery' records for the species are: Stations 87 (1000-0 m.); 89 (1000-0 m.); ioob (900-1000 m.); 100c (260-310, 2500-2000 m.); 102 (250-0 m.); 282 (300-0 m.); 297 (200- 300-0 m.); ' Discovery II ' 670 (470-0 m.); 671 (2000-0 m.); 673 (1500-1000, 750-500 m.); 675 (500- 250 m.); 677 (1 500-1000, 470-0 m.); 679 (750-500 m.); 681 (500-250, 250-0 m.); 682 (375-0 m.); 684 (75°-500 m-); 69° (iS00-°» 500-25°. 46°-° m-); 693 (75°-5°°> 5°°-25°> 25°-° m-); 695 (37°-° m-); 696 (750-500, 500-250 m.); 697 (460-0, 450-0 m.); 698 (470-0 m.); 699 (500-250, 370-0, 250-0 m.); 702 (236-0 m.); 703 (358-0 m.); 845 (242-180 m.); 847 (270-196 m.); 848 (270-196 m.); 899 (350- o m.) ; 965 (3 10-106 m.) ; 968 (250-106 m.) ; 969 (250-100 m.) ; 1 178 (3 10-0 m.) ; 2068 (700-0 m.) ; 2635 (280-0 m.); 2636 (950-550 m.); 2639 (1200-600 m.); 2646 (1500-800 m.); 2647 (310-0 m.); 2648 (1450- 950 m.); 716 (212-0 m.); 2084 (1500-800 m.); 943 (128-0, 250-100, 356-130, 500-0 m.); 944 (500- 750 m.); 967(306-145 m.). The unfortunate nomen nudum L. profunda of Totton (1936, p. 235) refers to specimens of L. multi- cristata Moser from Bermuda (Beebe Collection). Surtort Surf art Lensia conoidea Keferstein & Ehlers, 1861. Bigelow & Sears (1937) published some fine figures of this species and valuable further notes. Their figure (fig. 35) shows the apical view of the posterior nectophore, but I find that in specimens from off the Irish Coast, from 'Discovery' Station 100c, and from ' Manihine ' South-west Ireland Station 4, the apico- ventral corner of the right hydroecial wing is not like Text-fig. 56. Lensia conoidea. Apical view of posterior the left one, which is figured correctly. The right one nectophore, x 34, to show the correct shape of the artic- is of a square section as shown here in Text-fig. 56. ulating surfaces. 'Manihine', South-west Ireland, St. 4. Lensia subtilis var. (Chun), 1886. Quite arbitrarily I am referring to fourteen specimens (Text-fig. 57 B, C, D) as a variety of this common species. They come from 'Discovery II' Stations 1585-7. The stalks of their somatocysts are relatively longer than in L. meteori (Text-fig. 60), and the wall of the hydroecium instead of being nearly vertical for most of its extent as in L. meteori, lies at 450 and more nearly approaches the condition found in L. subtilis. I report these specimens under the name subtilis var., pending the discovery of more specimens and hesitate to set up a new species. It may be significant that from 'Discovery II' Stations 1585 and 1586 at which eight of these intermediate type specimens were taken no other specimens of L. subtilis were found. Also at 'Discovery II ' Station 1587, where two specimens of L. meteori were found as well as several of this intermediate type, again no other specimens of L. subtilis were found, although L. subtilis was common at most stations off the East African coast. The fourteen specimens were found at a rather greater depth than those of L. meteori, but there may be no significance in this. I give sketches of six anterior nectophores of L. subtilis from Aqaba for comparison (Text-fig. 57 A). SYSTEMATIC AND BIOLOGICAL ACCOUNT 115 The stalk of the somatocyst of a Mediterranean specimen measured 0-025 mm- m diameter. The passage of an oil droplet up this canal may distend it. The expanded gastrozooids of the same specimen measured o-6-o-8 mm. in length and 0-15 mm. in diameter. The diameter of the distal half, the proboscis, was only 0-08 mm. In one complete specimen there was a bud of a third nectophore. B C Text-fig. 57. Lensia subtilis, anterior nectophores. A, a series of six specimens from the Gulf of Aqaba, Red Sea, 'Manihine' St. 17, x 21 ; B-D, specimens from 'Discovery II' St. 1587. Magnification: B x 21 ; C x 18; D x 11. Vei. Som. C pa Text-fig. 58. Lensia subtilis, anterior nectophores from Villefranche. A, B, C, three stages in contraction of the nectosac, x 12; D, basal view of a nectophore, x 19. In 1932 I reported that there was some doubt about the appearance of the eudoxids of this species. It is very abundant at Villefranche and I give a figure now (Text-fig. 59). A useful field mark at Villefranche for recognition of gonophores of this species is the yellow colour of the tip of the manubrium. 15-2 n6 --I4--W Phc Text-fig. 59. Lensia subtilis, eudoxids from Villefranche. A, B, two views of a female gonophore, x 24 ; C, latero-ventral view of a young eudoxid, x 24; D, E, two views of a bract, x 25. Text-fig. 60. Lensia meteori. A, B, D, somatocysts of Adriatic specimens (Dr Gamulin); C, ventral view of somatocyst of a twisted specimen from Port Sudan, ' Manihine ' ; E, F, specimens from South-west Ireland, ' Manihine ' St. 4 ; G, specimen from 'Discovery II' St. 1587. Magnification: A X47; B X38; C-D X20; E-G x 20. SYSTEMATIC AND BIOLOGICAL ACCOUNT 117 Lensia meteori (Leloup), 1934a. This charming but rather rare little species occurred at half of ' Discovery's ' East African Stations, and was represented by a single specimen only at four ' Manihine ' Stations off Port Sudan. I have found it once or twice in hauls from 100 m. or more at Villefranche, where Leloup also recorded it; and it occurs also off Split in the Adriatic, from which area Dr T. Gamulin has sent me specimens. It is a species that can easily be overlooked. As mentioned earlier in these notes, I believe that Leloup's meteori is closely related to L. subtilis, and not to species of 'Galetta', with which Bigelow & Sears (1937) linked it. Lensia subtilis has unlooped lateral canals in its posterior nectophores, as in all known species of Lensia, whereas in ' Galetta ' spp. the lateral canals are looped, or bend back on themselves. It seemed at one time that Lensia meteori was a very distinct species, but at ' Discovery II ' Stations 1585-7 fourteen specimens were found (Text-fig. 57 B, C, D) which are intermediate between L. meteori and L. subtilis. They are dealt with as a variety of L. subtilis in my notes on that species. Three anterior nectophores were taken by 'Manihine' in July 1950 off south-west Ireland. A single anterior nectophore found by ' Manihine ' off Port Sudan has a somatocyst which is pro- duced laterally as shown in Text-fig. 60 C. Lensia meteori (Leloup) New Records Station Depth (m.) Specimens 'Discovery II' 1567 1350-0 1568 1400-0 I58l 100-50 250-100 1584 250-100 1586 250-100 500-250 1587 450-0 2 1588 250-100 500-200 'Manihine', Gulf of 5 183-0 Aqaba 6 — 10 183-0 13 J35-° Port Sudan 11 366-0 S.W. Ireland, 4 183-0 3 50°48'N., ii° 4' W. 'Michael Sars' 8ia (? 1000) 1 48° 1' N., 39° 46' W. Lensia subtilis var. (-> L. meteori) 'Discovery II' Station 1585 1586 1587 Depth (m.) 500-0 1400-700 55°-° 1650-950 450-0 Specimens n8 DISCOVERY REPORTS Lensia reticulata sp.n. This is a small multistriate species with cross-connecting ridges, the whole pattern of ridges being somewhat irregular. The hydroecium is well formed and has a ventral slit; the somatocyst is short and ovate in outline, seen from the ventral side, but measures less across the dorso-ventral than across the latero-ventral axis. An aboral, dorso-lateral, horizontal ridge connects up the basal ends of half a dozen longitudinal ridges on either side of the nectophore as in Lensia exeter. Material. Three anterior nectophores, two from 'Discovery II' Station 1585, 1400-700 m. (Text-fig. 61), the smaller of which is chosen as the holotype of the species, and one from ' Discovery 1 1 ' Station 1586, 1400-0 m. So/n Text-fig. 61. Lensia reticulata sp.n. from 'Discovery II' St. 1585, 1400-70001. A, D, F, the whole, the apex and the base of a larger specimen ; B, E, H, C, G, the whole, two views of the base and two of the apex of a smaller specimen. Magnification : A X9; B x 18; C-H x 29. At first sight it might be supposed that the reticulation was due to shrinkage. I am quite satisfied that it is a specific character. Most of the facets are deeply concave, as can be seen from the sketches of the apical parts of two specimens (Text-figs. 61 C, D). The material, as is usually the case with small delicate specimens from the depths, is only moderately well preserved, and the reticulated ridges could only be examined after staining. This is one of the three new species of Lensia to come to light since I reviewed the species in 1941. The figured holotype bears the Brit. Mus. (Nat. Hist.) Register No. 1952. n. 19. 2. Eudoxia macra sp.n. This new eudoxid of unknown parentage is easily recognized by its peculiar bract and long gono- phore which has no prominent teeth but a deeply pigmented tip to its manubrium. Measurements of holotype. Length overall 3-2 mm. ; length of bract 1-25 mm. ; length of gonophore 2-5 mm. SYSTEMATIC AND BIOLOGICAL ACCOUNT 119 Using again the terminology that I employed in 1932 (p. 318) the headpiece of the bract would be described as small, the phyllocyst comparatively small and near the apex. In optical section the head and neck shield resemble those of Diphyes bojani. The sutural surface is in a groove. In fact the whole bract has not evolved far beyond the developmental stage through which Diphyid eudoxid bracts pass. The edges of the sutural surface tend to lap over one another as if lapping round the stem. These ventral edges are cut away like the sides of a man's tail-coat till they meet the dorsal edge of the bract. The tip of the bract is rounded and conical, and the sutural surface does not reach the apex. The eudoxid, then, may be regarded as more primitive than that of a species such as D. dispar. Man o.sm/n Text-fig. 62. Eudoxia macra sp.n., a new eudoxid from 'Discovery II' St. 1588, 250-100 m., whose complete polygastric stage is either unknown or has not been identified. C x 46. The eudoxid has one characteristic, namely the rounded, conical tip of the bract, that recalls the apex of the anterior nectophore in Clausophyes ovata, and of the eudoxid figured (1 861) by Keferstein & Ehlers. But Clausophyes is taken in much deeper water than this Eudoxia macra, and the eudoxid of Clausophyes appears to have longitudinal bracteal canals. Gonophore. There are two dorso-lateral and two ventro-lateral radial canals, and two slight rounded ridges overlie the dorso-lateral pair. There are two closely opposed hydroecial folds, one deeper than the other. Distally these folds become less and less deep till they disappear as slight ridges before reaching the level of the velum. Neither mouth-plate nor teeth are present. The proximal end of the gonophore does not appear to articulate with the bract itself. It is sometimes rounded and shaped like i2o DISCOVERY REPORTS the apex of the bract, and sometimes flattened on its ventral side. The female gonophore bears from fourteen to sixteen eggs. There is not enough material for a study of the details of the tentilla. Material. Two eudoxids, six bracts, fifty-four gonophores. 'Discovery' Station Net Depth (m.) Material IOOC TYF 260-310 2 gonophores 702 TYFB 236-0 1 gonophore IS67 N70B 135°-° 1 eudoxid, 6 gonophores 1568 N70B 600-0 4 gonophores 1581 N70V 50-0 100-50 1 gonophore 1 gonophore 1583 N50V 1 00-0 1 gonophore 1584 N70V 50-0 100-50 1 gonophore 2 gonophores 1586 N70V 100-50 250-100 500-250 2 bracts, 3 gonophores 1 bract, 3 gonophores i gonophore 1587 . N70 V 450-0 8 gonophores 1588 N70V 100-50 250-100 500-200 1 bract, 2 gonophores 1 eudoxid, 1 bract, 3 gonophores 1 bract 1589 N70B 600-0 4 gonophores 2647 TYFB 310-0 9 gonophores 2929 N70V 75°-500 2 gonophores The complete eudoxid from 'Discovery II' Station 1588 is chosen as holotype. It bears the Brit. Mus. (Nat. Hist.) Register No. 1952. 11. 19. 3. Muggiaea atlantica Cunningham, 1892. Russell (1934) published interesting information on the occurrence of the two hitherto known species of Muggiaea. He thought that both are essentially inhabitants of less contaminated coastal waters, a fact which he suggested would account for their absence from the collections of Oceanographical Expeditions. This species occurred abundantly near the surface off South Africa at 'William Scoresby' Stations 444-460. Bigelow & Sears (1937) most recently summarized our knowledge of the distribution of M. atlantica. The 'Trior' records made these authors think that it is definitely a neritic species, as are probably also M. kochii, Diphyes chamissonis, Lensia subtilis and L. subtiloides. Fraser-Brunner took Muggiaea atlantica in the Gulf of Aden (Chart, p. 10) off Aden and Alayu, localities which bear out the idea that it is definitely a neritic species. 'Manihine' took a single nectophore at each of three stations, 10 miles, 35 miles and 40 miles east of Sanganab Lighthouse, off Port Sudan. All the hauls were made from depths of 365 m. to the surface in January 195 1. SYSTEMATIC AND BIOLOGICAL ACCOUNT Muggiaea bargmannae1 sp.n. Many years ago I recorded in my manuscript notes on the distribution of Siphonophores along ' Discovery IPs ' 300 meridian line of stations the occurrence of a new species of Muggiaea at Station 663, in a haul from 500 to 250 m. Recently, whilst identifying Siphonophores from part of ' Discovery LPs' o° line of stations, I found this species again. I therefore re-examined much of the ' Discovery ' material of Dimophyes arctica, which the new Muggiaea superficially resembles, and found in hauls from the Antarctic more specimens which had been confused with those of Dimophyes arctica. Unless material is well preserved or stained, confusion of this sort is quite easy, since the new species lacks the ridges found in Muggiaea atlantica and M. kochii (the apex of the nectosac is closer to the Sorn Cradd Vel Som Text-fig. 63. Muggiaea bargmannae sp.n. A, B, holotype specimen, 'Discovery II' St. 2012, 100-50 m. ; C, baso-ventral parts of four paratype specimens from same station; D, E, apical and basal views of holotype. All figures x 12. apical end of the nectophore than in those species). But the mouth-plate is divided, unlike its counter- part in Dimophyes, and the ventral slit in the hydroecial wall is in the mid-ventral line and is shorter than in Dimophyes, so that in lateral view the ventral wall of the hydroecium extends further towards the base, as in the other two species of Muggiaea. This point will be readily appreciated if comparison be made between the ventral view of M. bargmannae given in my Text-fig. 63 B and that of Dimophyes arctica given by Chun (1 897 ft) in his fig. 4 of Taf. 1. With the knowledge of this new species I am inclined to think that Dimophyes whose affinities have always been obscure, is most closely related to Muggiaea, especially since its posterior nectophore is obsolescent. In M. atlantica and M. kochii the posterior nectophore has ceased to develop; and no posterior nectophore has so far been found associated with the new species, whose eudoxid has not yet been certainly identified. 1 This species is named after Dr Helene Bargmann. 16 122 DISCOVERY REPORTS M. bargmannae differs from both M. atlantica and M. kochii and the other new species M. delsmani described below in that the tip of the nectosac approaches much nearer to the apex of the whole nectophore and also in that the convex longitudinal folds into which the nectophore is thrown on contraction of the nectosac are not surmounted by crest-like ridges. Another point to note is that towards the ostium of the nectophore the two lateral folds almost meet, so that the dorsal fold is most apparent in the apical two-thirds of its course. The length of the somatocyst is about half the distance between the tip of the nectosac and the tip of the hydroecium. In well-fed specimens such as that illustrated in Text-fig. 63 A the somatocyst is sausage-shaped and its base projects ventrally over the top of the hydroecium. In this state it contains much oil. When not distended with oil, its diameter may measure only one-third or less, as shown in Text-fig. 63 C. The hydroecium is shallower than that of M. kochii, but not as shallow as that of M. delsmani. Records of capture of Muggiaea bargmannae sp.n. Stations Depth (m.) Nectophores 'Discovery' 117 208 250-100 800-0 1 5 ' Discovery II ' 519 39°"! 37 1 663 500-250 1000-750 3 2 671 1 000-0 1 1048 340-140 '5 1772 520-250 1 2010 100-50 5 250-100 5 2012 100-50 250-100 500-250 6 (including the figured holotype) 1 3 2014 1 00-0 100-50 3 1 250-100 9 2015 100-50 250-100 500-250 1 H 1 2017 100-50 250-100 500-250 750-500 1 1 1 1 2024 250-100 650-400 750-500 1 2 1 'William Scoresby' WS 550 WS 551 121-0 1 21-0 7 70 Examination of ' Discovery II ' hauls on both the o° and 300 lines of stations showed that M. barg- mannae occurred only to the southward of the Antarctic convergence. Of the other six records five are from the Antarctic, but Station 1772 was made in lat. 360 S., and the haul that contained one young specimen was made from 520 to 250 m. This record should be treated with reserve. The most southerly record I have at present is at 'William Scoresby' St. WS 551, in 68° 17^' S., where it was SYSTEMATIC AND BIOLOGICAL ACCOUNT 123 taken in a haul from 121 m. to the surface in temperatures of from 0-84° to 0-30° C. and salinities of from 3 3-99 %0 to 33-17 %0. The recognition of this new species of Muggiaea brings up the total of Siphonophores found in the Antarctic to fifteen. They are: Pyrostephos vanhoeffeni, Marrus antarcticus gen. et sp.n., Stephanomia convoluta, Diphyes antarctica, Muggiaea bargmannae sp.n., Dimophyes arctica, Lensia havock, L. achilles, Rosacea plicata, Vogtia serrata, Thalassophyes crystallina, Crystallophyes amygdalina, Heteropyramis maculata, Chuniphyes multidentata, and Ch. moseri sp.n. The first five and Vogtia serrata are confined to the Antarctic. The figured holotype from 'Discovery II' Station 2012 bears the Brit. Mus. (Nat. Hist.) Registei No. 1952. 3. 18. 1. Muggiaea delsmani sp.n. (Text-fig. 55 B.) In December 1937 I received for identification from Dr H. C. Delsman of Hilversum a small collection of Siphonophores from the Java Sea. From a station at 50 57' S., 1080 23' E. came specimens of Diphyes chamissonis, Lensia subtiloides and a new species of Muggiaea which I have named after Dr Delsman and describe below. An interesting point about the record is that the salinity was low for Siphonophores. DrDelsman's note read 'salinity surface 327%,,, bottom (45 m.) 3i-6%0. . .a few other stations where the salinity was slightly higher viz. about 33-3 %0. All these salinities, however, have been determined with an areometer which possibly gave slightly too low values. At any rate the salinity in the Java Sea hardly ever surpasses the 34 %0 and at an average is not higher than 33 %0 '. M. delsmani can be distinguished at once from M. kochii, which also has a short somatocyst, by its shallower hydroecium, whose upper wall is more nearly horizontal than in M. kochii. As in other species of Muggiaea the base of the somatocyst lies close to the ventral wall of the nectosac. The usual five ridges are present, and the tip of the nectosac, as in M. atlantica and M. kochii does not come as close to the apex of the nectophore as it does in the new species M. bargmannae. The distal edges of the two mouth-plates are rounded, and there is no sign of a notch in one of them as there is in Lensia subtiloides. Reference to Text-fig. 55 B makes further description of the nectophore unnecessary. The eudoxid has not been recognized as yet. The figured holotype bears the Brit. Mus. (Nat. Hist.) Register No. 1937. 12. 9. 1. Dimophyes arctica (Chun), 1897. The only previous records from the Indian Ocean are: (1) One complete polygastric specimen (not one eudoxid as stated by Browne, 1926) recorded by Moser (1925, p. 390). (2) One anterior nectophore from a locality 8° 16' S., 5 1° 26' E., between Providence and Alphonse ; open net, 900-0 fm., recorded by Browne (1926). The species was taken in seventeen of the forty-eight ' Discovery II ' hauls that I have examined from the west Indian Ocean ; it was probably breeding at eleven of them. Because of the possible value of this species as an 'indicator' of water masses the details of the catches are given on p. 124. The species was not found further north than 060 05' N. lat. It was breeding in a depth not greater than 250 m. as far north as 07° 42-1' S. lat. at a temperature somewhere between 1170 and 17-21° C. I have accumulated data on the temperature and salinity tolerances of D. arctica from 117 hauls of closing-nets made at sixty ' Discovery ' stations (outside the Indian Ocean) as set out on pp. 1 25 and 1 26. Using the higher temperature of the lowest range of any closing haul as the minimum and the lower temperature of the highest range as the maximum, I find that the known range of temperature is from —1-13° to 13-26° C. 16-2 124 DISCOVERY REPORTS Similarly the known range of salinity is from 34-08 %0 to 35-33 %0. From data given for 179 'Meteor' hauls where Dimophyes arctica was taken I have computed a known range of salinity 34- 14-36-57 %0, and of temperature 0-14-22-05° C. The data given on pp. 125-126 are representative, but not good enough perhaps, for drawing very definite conclusions. To do that it would be necessary to examine a much larger number of catches, and especially ones taken at all times of the year, such as those taken at full stations on the meridians of o° and 20° E. New records of catches o/D. arctica in the Indian Ocean {i.e. north of subtropical convergence) By interpolation from values 'Discovery IF Station Net Depth (in.) Catch given in Station List T. °C. S./bo 1587 TYFB 450-0 1 anterior nectophore — — 1586 N70 V 500-250 8 eudoxids 10-3-13-1 35"!3-35-2 1586 N70V 250-100 1 eudoxid 13-1-23-60 35-2-35"63 1586 TYFB 1650-950 2 anterior nectophores 3-8-7-2S 34-87-34-11 1586 TYFB 55°"° 5 anterior nectophores — — 1585 TYFB 1400-700 2 eudoxids — — 1585 TYFB 500-0 5 anterior nectophores, ? 2 gonophores — — 1584 N70V 500-250 1 ? posterior nectophore, 1 bract of eudoxid 9-25-11-92 34-98-35-06 1584 N70V 250-100 8 anterior nectophores (juv.), 1 eudoxid 11-92-21-42 35-06-35-29 1583 N70V 500-250 1 anterior nectophore, 6 eudoxids 8-64-11-64 34-85-35-°6 1583 N70V 250-100 1 anterior nectophore, 1 eudoxid 11-64-20-67 35-o6-35'28 1581 N70V 500-250 ? 4 gonophores 8-29-11-69 34'83-35-°7 1581 N70V 250-100 2 anterior nectophores, 1 eudoxid 11-69-17-21 35-°7-35-3° 158 1 TYFB 1750-600 1 anterior nectophore 3-09-7-68 3476-34'77 1581 TYFB 600-0 4 anterior nectophores — — 1568 N70B 1450-0 2 eudoxids — — 1567 N70B 135°-° 5 anterior nectophores — — The data indicate that D. arctica has a much greater range of tolerance than Diphyes antarctica, or at any rate that the habits of the two are quite different, for it would seem that the differences in their geographical distribution are mainly brought about by their resorting to different depths at various times, or at different stages in their life cycle, for this would bring them into currents which flow in different directions at various depths. But Mackintosh (1934) did not find any marked diurnal variation in the catches of the N 100 B nets, and there is probably little diurnal migration up or down of the polygastric stages upon which Mackintosh worked. On 17 November 1952, after this report was prepared, I received from Dr Fraser of Aberdeen some siphonophores taken in June of this year, and amongst them an undoubted posterior nectophore of Dimophyes arctica taken in a horizontal tow at 20 m. at ' Scotia ' Station 1 199. The temperature and salinity recorded at that depth were 12-05° C. and 35-39 %0. Bigelow & Sears (1937, fig. 81) plotted the frequency of capture of certain siphonophores, including D. arctica in ' Meteor ' hauls in Atlantic water of known minimum and maximum temperature. The main distribution for D. arctica was from 2° to 14° C. with a maximum around 30 C. At the sixty ' Discovery ' Stations mentioned on p. 125 the most frequent occurrence was between — i° and 6° C, with a very small marked maximum between i° and 2° C. (Text-fig. 64). SYSTEMATIC AND BIOLOGICAL ACCOUNT 125 Station Net Depth (m.) Date s.% Temp. ° C. WS22 70 V 1000-750 30. xi. 26 34-70-34-69 1-14/1-82 WS26 70 V 500-250 18. xii. 26 34-57-34-30 2-48/1-00 WS26 70 V 750-500 18. xii. 26 34-63-34-57 1-95/2-48 WS 29 70 V 500-250 19. xii. 26 34-56-34-40 2-02/1-33 WS29 70 V 250-100 19. xii. 26 34-40-33-96 1-33/0-60 WS30 70 V 250-100 19. xii. 26 34-43-34-95 i-55/-o-7o WS 30 70 V 500-250 19. xii. 26 34-60-34-43 2-04/1-55 WS36 70 V 750-500 22. xii. 26 34-63-34-54 1-99/2-15 WS38 70 V 750-500 22. xii. 26 34-69-34-58 1-90/2-00 WS38 70 V 500-250 22. xii. 26 34-58-34-40 2-00/1-73 WS44 70 V 500-250 8. i. 27 34-52-34-30 2-04/1-71 WS6I 70 V 250-100 18. i. 27 34-48-34-05 1-70/ -0-06 WS6I 70 V 500-250 18. i. 27 34-67-34-48 1-95/1-70 WS63 70 V 500-250 20. i. 27 34'54-34-33 2-09/1-86 WS67 70 V 250-100 20. ii. 27 34-J7-33-95 1-78/1-16 WS67 70 V 1000-300 20. ii. 27 34-63-34-23 2-13/2-09 WS69 70 V 500-250 22. ii. 27 34-18-34-12 2-42/3-11 WS70 70 V 1000-750 23. ii. 27 34-28-34-20 2-94/3-25 WS70 70 V 500-250 23. ii. 27 34-19-34-18 3-76/4-16 WS70 70 V 250-100 23. ii. 27 34-18-34-16 4-16/5-01 WS no 70 V 500-300 26. v. 27 34-6o-34-39 2-07/1-88 WS 388 70 V 400-250 16. ii. 29 34-5!-34-47 -i-i3/-i-24 n 70 V 500-250 16. ii. 26 34'54-34-22 2-02/1-34 102 70 V 250-100 28. x. 26 34-81-35-03 10-28/12-36 102 70 V 750-500 28. x. 26 34-34-34-42 3-85/5-77 102 70 V 500-250 28. x. 26 34-42-34-81 5-77/10-28 102 70 V 1000-750 28. x. 26 34-46-34-34 3-24/3-85 103 70 V 500-250 30. x. 26 35-33-3451 13-26/15-97 107 70 V 500-250 4. xi. 26 34-21-34-10 3-60/4-45 107 70 V 750-500 4 xi. 26 34-27-34-21 3-03/3-60 107 70 V 250-80 4. xi. 26 34-10-34-01 4-45/5-87 117 70 V 250-100 17. xi. 26 34-60-34-07 1-28/1-43 129 70 V 780-500 19. xii. 26 34-67-34-59 1-96/2-02 129 70 V 500-250 19. xii. 26 34-59-34-40 2-02/1-46 133 70 V 270-100 20. xii. 26 34-47-33-96 1-83/0-68 137 70 V 700-500 22. xii. 26 34-69-34-6o 1-86/1-94 J37 70 V 250-100 22. xii. 26 34-40-33-96 i-63/-o-20 138 70 V 500-250 22. xii. 26 34-66-34-47 1-98/1-83 138 70 V 750-500 22. xii. 26 34-71-34-66 1-82/1-98 138 70 V 1000-750 22. xii. 26 34-72-34-71 1-64/1-82 I5I 100 H 500-625 16. i. 27 34-60-34-64 1-84/1-85 !5l 70 V 750-500 16. i. 27 34-66-34-60 1-77/1-84 300 70 V 750-500 20. i. 30 34-72-34-65 1-84/1-85 300 70 V 500-250 20. i. 30 34-65-34-44 i-85/i-54 300 70 V 1000-750 20. i. 30 34-69-34-72 1-63/1-84 301 70 V 750-500 20/21. i. 30 34-71-34-63 1-85/1-92 301 70 V 500-250 20/21. i. 30 34-63-34-46 1-92/1-71 302 70 V 500-250 21. i. 30 34-59-34-34 1-90/1-19 302 70 V 250-100 21. i. 30 34-34-34-93 I-I9/-0-22 3°3 70 V 750-500 21. i. 30 34-66-34-56 I-92/I-88 303 70 V 500-250 21. i. 30 34-56-34-41 1-88/1-35 3°3 70 V 1000-750 21. i. 30 34-72-34-66 1-69/1-92 304 70 V 1000-750 21. i. 30 34-69-34-67 i-75/!-93 3°4 70 V 500-250 21. i. 30 34-59-34-35 1-89/1-38 305 70 V 500-250 21. i. 30 34-58-34-36 1-98/1-51 305 70 V 750-500 21. i. 30 34-68-34-58 1-97/1-99 311 70 V 500-250 24. i. 30 34-49-34-29 1-97/1-60 320 70 V 750-500 30. i. 30 34-69-34-63 1-76/1-69 126 DISCOVERY REPORTS Station Net Depth (m.) Date S-%o Temp. ° C. 320 70 V 500-250 3°- i- 3° 34-63"34"32 1-69/0-78 320 70 V 1000-750 3° i. 30 34-70-34-69 1-53/1-76 321 70 V 750-500 3° i. 30 34-69-34-62 1-74/1-84 321 70 V 500-250 30 i. 30 34-62-34-32 1-84/0-87 322 70 V 500-250 31 i. 30 34'6°-34-35 1-91/1-34 322 70 V 1000-750 31 i. 30 34-71-34-69 1-67/1-83 323 70 V 500-250 31 i. 30 34-59-34-41 1-96/1-58 324 70 V 500-250 1 ii. 30 34-57-34-38 1-81/1-46 325 70 V 500-250 1 ii. 30 34-47-34-29 2-10/1-60 334 70 V 500-250 4 ii. 30 34-54-34-31 1-97/1-42 334 70 V 1000-750 4 ii. 30 34-71-34-67 1-89/2-02 334 70 V 750-500 4 ii. 30 34-67-34-54 2-02/1-97 336 70 V 500-250 5 ii. 30 34-54-34-34 1-96/1-60 336 70 V 250-100 5 ii. 30 34-34-34-04 1-60/0-32 336 70 V 750-500 5 ii. 30 34-68-34-54 1-96/1-96 337 70 V 75°-5°° 5 ii. 30 34-68-34-57 1-97/1-96 337 70 V 500-250 5 ii. 30 34-57-34-36 1-96/1-60 339 70 V 250-100 5 ii. 30 34-27-33-96 1-20/1-30 342 70 V 500-250 7 ii. 30 34-65-34-48 1-86/1-67 343 70 V 500-250 7 ii. 30 34-65-34-49 1-98/1-60 344 70 V 500-250 7 ii. 30 34-59-34-4I 2-00/1-48 344 70 V 750-500 7 ii. 30 34-66-34-59 1-94/2-00 345 70 V 180-100 8 ii. 30 34-3°-34-i° 1-07/0-38 353 70 V 750-500 9 ii. 30 34-66-34-57 1-95/1-96 353 70 V 500-250 9 ii. 30 34-57-34-31 1-96/1-27 354 70 V 500-250 9 ii. 30 34-64-34-44 1-97/1-74 354 70 V 250-100 9 ii. 30 34-44-34-05 1-74/0-28 354 70 V 750-500 9 ii. 30 34-69-34-64 1 -88/i -97 355 70 V 500-250 9 ii. 30 34-64-34-48 i-9i/i-77 355 70 V 250-100 9 ii. 30 34-48-34-07 1-77/0-00 355 70 V 750-500 9 ii. 30 34-67-34-64 1-73/1-91 355 70 V 1000-750 9 ii. 30 34-69-34-67 1 -47/1 73 356 70 V 750-500 10 ii. 30 34-69-34-66 1-62/1-83 356 7oV 1000-750 10 ii. 30 34-70-34-69 1-37/1-62 357 70 V 75°-5°° 10 ii. 30 34-67-34-61 1-84/1-91 357 70 V 500-250 10 ii. 30 34-61-34-44 1-91/1-62 358 70V 1000-750 11 ii. 30 34-70-34-69 1-61/1-77 358 70 V 500-250 11 ii. 30 34-67-34-41 1-89/1-38 358 70 V 750-500 11 ii. 30 34-69-34-67 1-77/1-89 588 100 B 460-150 13 >• 31 34-7I-34-43 1-61/0-30 661 TYFV 750-500 2 iv. 31 34-68-34-68 0-62/0-86 661 TYFV 500-250 2 iv. 31 34-68-34-60 0-86/0-59 661 TYFV 3000-2000 2 iv. 31 34-66-34-67 — 0-15/0-00 661 TYFV 1000-750 2 iv. 31 34-67-34-68 0-46/0-62 663 TYFV 500-250 5 iv. 31 34-65-34-47 1-14/0-40 663 TYFV 1000-750 5 iv. 31 34-67-34-67 0-69/0-93 663 TYFV 750-500 5 iv. 31 34-67-34-65 0-93/1-14 663 TYFV 2000-1500 5 iv. 31 34-67-34-67 0-19/0-43 663 TYFV 1 500-1 000 5 iv. 31 34-67-34-67 0-43/0-69 666 TYFV 500-250 J7 iv. 31 34-56-34-32 1-88/0-94 671 TYFV 1000-750 22 iv. 31 34-4i-34-3i 2-53/2-63 671 TYFV 1 500-1 000 22 iv. 31 34-63-34-4I 2-51/2-53 671 TYFV 750-500 22 iv. 31 34-3I-34-I9 2-63/2-69 671 TYFV 500-250 22 iv. 31 34-I9-34-23 2-69/4-94 673 TYFV 750-500 24 iv. 31 34-28-34-41 4-08/6-31 673 TYFV 1000-750 24 iv. 31 34-28-34-28 3-20/4-08 847 100 B 270-196 11 iv. 32 34-78-34-86 10-05/11-22 942 100 B 350-110 31 viii. 32 34-56-34-58 8-40/8-97 1048 100 B 340-140 3° xi. 32 34-65-34-35 0-70/ -0-63 SYSTEMATIC AND BIOLOGICAL ACCOUNT 127 Mackintosh (1934) classed D. arctica as one of the species typical of the coldest Antarctic regions such as those of the Bellingshausen and Weddell Seas, where it was plentiful, and said that it rarely or never approaches the convergence. He gave the average number of anterior nectophores per haul between the mean summer isotherms — i-oo° to - 1-99° as 14-4, and between similar isotherms 6-oo'J to 6-99° as 1-40. Compared with an abundant Antarctic species like Calanus acutus (1900 per haul) Mackintosh found that Dimophyes arctica was only moderately abundant in the Antarctic (1-90 per haul). io1- I2V I4U I5°C. Number of times recorded temperature is minimum possible. : i Number of times recorded temperature 15 maximum possible Text-fig. 64. Frequency of capture of Dimophyes arctica in water of known upper and lower limits of temperature by 'Discovery' closing-net hauls. Black tally-marks show (upper series of columns) the number of times that the indicated temperature was the minimum in which the specimens could have been taken ; and (lower series) how often the temperature was the maximum possible for the same group of hauls (in). Superimposed, in the form of open circular tally-marks are some of Bigelow & Sears (1937, fig. 81) records (53) for 'Meteor' closing hauls that yielded this species (six of their records, covering temperatures ranging from 160 to 260 C, are omitted). The class boundaries used in the two sets of records are staggered by half a degree Centigrade. The reason for giving two series of columns is that each closing-net capture was made at an unknown depth in limited columns of water whose upper and lower levels were at different temperatures. In June 1952 'Discovery II ' took D. arctica in closing-nets in the North Atlantic at Stations 2927, 2929 and 2935 at depths between 1500-100 m., 1000-750 m. and 750-500 m. Dr Cooper informed me (in lit.) that the mean temperature at Station 2929 (440 N., i5°W. approx.) for depths between 746 and 977 m. was 10-19° C, and that tne water was 'Mediterranean' or 'Gulf of Gibraltar' water, formed in the Atlantic west of Gibraltar by mixing of warm, deep, saline water from the Mediterranean with North Atlantic central water. D. arctica is not known to occur in the Mediterranean. Chelophyes appendiculata (Eschscholtz), 1829. (Plate IV, figs. 1,3.) Although this is one of the most common Siphonophores it was, until lately, the only common diphyid that could not be associated in preserved plankton collections with an eudoxid. As Moser (1925) said, that there should still be doubt on this point was difficult to believe. But Bigelow, i28 DISCOVERY REPORTS as long ago as 191 1, said that the eudoxid was so well known as to require no description. The explanation of this paradox is that Bigelow mistook the eudoxid of Eudoxia mitra for that of this species. In the Mediterranean Chelophyes appendiculata is very common. Its seasonal and geographical distribution was worked out by Bigelow & Sears (1937), although 'Thor' did not provide them with any material of the eudoxid stage. On searching the literature for descriptions and figures of eudoxids that might be those of Ch. appendiculata we find first of all that Leuckart (1853) described and figured Eudoxia campanula, a common eudoxid which he found at Nice. I have found it near there too. The immature eudoxids he described and figured well in figs. 17 and 19 of his plate ill, but the details of the loose gonophore (his fig. 19) are not so well shown. He came to the conclusion, without being able to prove it, that it could be the eudoxid only of his Diphyes acuminata (Chelophyes appendiculata). In this I agree with him. Leuckart's paper was shortly afterwards followed by Gegenbaur's (1853) figure and description of Eudoxia messanensis from Messina. He does not mention E. campanula Leuckart because his work was already in the press when Leuckart's appeared, though he does refer to the paper in an appendix. The bract does not show the deep cavity, nor the gonophore the long apophysis that are characteristic of the eudoxid of Chelophyes appendiculata. I think that Eudoxia messanensis may be the eudoxid of Lensia conoidea; it is not that of Chelophyes appendiculata. Eudoxia lessonii Eschscholtz (1829) is neither convincingly figured nor does it form part of a well-defined fauna, so that we cannot be sure of its identity. The identity of McCrady's (1857) E. alata has been discussed by Moser (1925). I agree that there seems no reason to believe that it is the eudoxid of Chelophyes appendicidata. Schneider (1896, p. 581) copied Gegenbaur's figure (1853, pi. 16, fig. 20) of the larva of what is now known as Galetta turgida but was, in error, then called Diphyes sieboldii by Gegenbaur ; and labelled it D. appendiculata. Five pages later he gives it its proper name, D. turgida. At the same time, on plate 45, Schneider figured an eudoxid of unknown parentage, apparently from Naples, as the eudoxid of D. appendiculata. The gonophore, from its size and the minuteness of the gonadial part, would appear to be spent. It can scarcely be the eudoxid of any Mediterranean diphyid except Lensia conoidea. The figure certainly does not represent the eudoxid of Chelophyes appendiculata. Moser (1925), p. 245 says that the eudoxid of Ch. appendiculata so closely resembles, especially in its young stages, that of Muggiaea kochii that they are difficult to distinguish, but that it comes to exceed the latter in size. She notes in the eudoxid of Chelophyes appendiculata the presence of a rela- tively long, flattened apophysis to the gonophore, its freedom from the bract and the shallowness of its suture and of the cavity under the bract; her accompanying Text-fig. 39 is not characteristic of the eudoxid of Ch. appendiculata which has a deep bracteal cavity. She says (on p. 240) that since the eudoxid of Muggiaea kochii, Galetta subtilis and G. truncata (Lensia conoidea) were all found at Villefranche a determination of the differences was possible. She figured them on pis. I and IV. Her figure (pi. xii, fig. 8) is a good representation of the eudoxid of Chelophyes appendiculata. Moser did not succeed in getting Ch. appendiculata to separate off its cormidia till the month of May, when one specimen gave off thirty eudoxids in two days. These remained alive under observation for ten days. Moser (1925) figured these immature eudoxids on pi. xn in figs. 4-7. In the oldest of the figured specimens the bract had metamorphosed into the eudoxid condition. Her fig. 8 of an eudoxid of unknown parentage taken in a plankton haul gives a good representation of the common Diphyid eudoxids that I found at Villefranche in March, April, May and June 1949, and which I believe to be those of Ch. appendicidata. These eudoxids were shedding their ripe, active gonophores SYSTEMATIC AND BIOLOGICAL ACCOUNT 129 at the end of June 1949. I have watched the detached gonophores extruding eggs, and have also observed the spent gonophores. My attempts to raise the polygastric generation from wild male and female gonophores, kept in an improved Harvey rocker (see Rees & Russell, 1937), failed in my 1949 and 1950 experiments at Villefranche. I described and figured a very similar, if not the same, eudoxid as Eudoxia russelli in 1932, from material brought back from the Great Barrier Reef of Australia. It occurred there on 20 September 1928 at one station at the Barrier Reef lagoon, and on 20 October 1928 at another station outside Trinity Opening. At these two stations the polygastric stages of both Chelophyes appendiculata and Ch. contorta were present. In December 1948 Fraser-Brunner took some more eudoxids off Aden, very similar to Eudoxia russelli in structure and associated with the polygastric stages of Chelophyes contorta. Ch. appendiculata was very much less common in Fraser- Brunner's catches, ten specimens in one catch against 359 of Ch. contorta in five. In the one catch both specimens occurred. Using my 1932 terminology, the left sutural ridge of the eudoxid bract of Ch. appendiculata, taken wild at Villefranche, does not bear such a marked crest as the right, and the left ridge terminates before reaching the basal edge of the bract. Although a trivial character, this serves as a very useful field-mark for distinguishing between eudoxids of Ch. appendiculata and Lensia conoidea when breeding from wild eudoxids. Another such field-mark for use in the Mediterranean is the colour of the tip of the manubrium of the gonophore of an eudoxid. It is yellow in L. subtilis and pink in three species of Muggiaea. I have no written notes on the colour in Lensia conoidea or Chelophyes appendiculata. Vogt (1854), a pioneer who made very valuable observations at Nice, figured three eudoxids of unknown parentage whose nature he did not understand. He thought that they had developed from eggs and that they would grow into polygastric specimens of his Galeolaria aurantiaca ( = Sulculeolaria quadridentata). The largest one figured, without a mature gonophore, on his plate 21, probably presents the eudoxid of Chelophyes appendiculata. Up to the present time, therefore, I think it fair to say that we have no detailed description or figure of the full grown eudoxid of Ch. appendicidata except for Eudoxia campanula Leuckart (1853); Diphyes sieboldii Moser (1925, pi. xn, fig. 8 — an eudoxid caught at Villefranche); and Eudoxia russelli Totton (1932). A Mediterranean calycophore whose eudoxids might be mistaken for those of Chelophyes appendicu- lata is Lensia conoidea (referred to by Moser and also Leloup as Galetta truncata). But it is much larger and the bract lacks the distinct cavity found in Eudoxia campanula and E. russelli, and the gonophore has a truncate articulating end, as in most species of Lensia. There remains the task of differentiating between the eudoxids of Chelophyes appendiculata and Ch. contorta, which species were taken together, both by ' Mabahiss ' and ' Discovery II ' in the Indian Ocean. The length of a free swimming, mature, male gonophore of Ch. appendiculata taken at Villefranche (28 March 1949) was found to be 4-5 mm. A full-grown bract measured about 2-5 mm. in length and 1-25 mm. in (maximum) diameter. Observations on the behaviour of this very rapidly swimming species were made by me at Villefranche. When a complete, living, polygastric specimen sinks in a container of water it does so with its main axis vertical. The small posterior nectophore only is used to maintain vertical height, and contracts feebly at intervals to do so. The animal also cruises about slowly in a horizontal position in the same way without contracting its larger anterior nectophore until the high-speed 'escape action ' is called for, when the anterior nectophore comes into use. It was not possible to see what the posterior nectophore was doing when the animal was travelling at speed. 17 i3o DISCOVERY REPORTS Moser (1925, p. 243) said that there was no ventral radial canal in this species. Actually there is a very short one, measuring about o-i mm. in length, which runs to the circular canal, for the latter lies that distance within the inner boundary of the velum. The first part of the lateral radial canals are somewhat irregular, with a tendency to branch, and run parallel with the circular canal before rising up towards the apex. Som. Chelophyes contorta (Lens & van Riemsdijk), 1908. This species appears to be much more abundant in the West Tropical Indian Ocean than its ally Ch. appendiculata, though they occur side by side at most of the stations. During a recent examination of seven of the bottles of plankton brought back by M.Y. 'Manihine' from the Gulf of Aqaba I have identified young specimens of Ch. contorta and eudoxids of the Endoxia rnsselli type. Since Chelophyes appendiculata was not found in the Gulf of Aqaba it is probable that these eudoxids are of the species contorta, but I have not yet been able to distinguish them from those of Ch. appendiculata. They are both of the type described by me (1932) as Eudoxia russelli. This species is not often taken, but it was one of the common species taken by the Great Barrier Reef Expedition; and Russell & Colman (1935) showed that it could not apparently withstand the conditions of the environment in the Barrier Reef lagoon during the period of lower salinity. We find it in the Indian Ocean in some catches associated with Enneagonum hyalinum, one of the three Barrier Reef species that was definitely able to withstand lowered salinity, and so was possibly well suited for life in coastal waters. New Indian Ocean records. ' Discovery II ' Stations : iS67(i3S°-°m-); i-SI1 (500-0 m.); 1573 (800-0 m.); 1 58 1 (50-0, 100-50, 600-0, 1750-600 m.); 1582 (1900-1850 m.); 1583 (50-0, 100-0, 100-50, 250- 100 m.); 1584 (50-0, 100-50, 250-100, 500-250 m.); 1585 (500-0, 1400-700 m.); 1586 (50-0, 250-100,550-0, 1650-950111.); 1587(450-0, 1250-800 m.); 1588(50-0, 100-50, 250-100, 500-200 m.); 1589 (50-0, 100-50, 250-100, 600-0 m.). Diphyes chamissonis Huxley, 1859. 'Manihine' took D. chamissonis at six stations, fixed at 5, 10, 20, 35 and 40 miles east of Sanganab Lighthouse off Port Sudan, in open nets (366-0 and 183-0 m.), in January 195 1. At Station 6, ten miles east of the Lighthouse, twenty-one nectophores and seven eudoxids were taken; the largest catch. At five stations it was associated with Lensia subtiloides, an association that has been noted before. The species was previously recorded from the Red Sea by Leloup (19320). He based his record on one nectophore and an eudoxid in the Zoologisches Museum, Hamburg. He recorded another necto- phore in the same Museum from Ceylon. H.M.S. 'Weston ' found this species (both stages) abundant off Kamaran, Red Sea, in April 1936, at a surface temperature of 280 C. Text-fig. 65. Chelophyes contorta. Two views of the anterior nectophore, x 12. Specimen from off Alayu, Gulf of Aden. SYSTEMATIC AND BIOLOGICAL ACCOUNT 131 'Discovery II ' records from the west Indian Ocean are: Station 1581, TYF B, one bract of eudoxid (1750-500 m.); 1589, TYF B, one anterior nectophore (600-0 m.); 1589, N 70 B, three anterior nectophores (600-0 m.). The species was not present in any of the catches made on the 320 S. line of Stations. Chuniphyinae subfamily nov. Chuniphyes Lens & van Riemsdijk, 1908 For many years I suspected that I had been dealing with more than one species. A review of all the material after the chance finding of a well-preserved specimen from 'Discovery II' Station 1639 revealed that there are two quite distinct species, the well-known multidentata Moser, and a new species described below. The interesting eudoxid stage (Eudoxia problematica Moser) of one or other species was first related to the polygastric stage by Leloup (19340). It can now be said to be well known. Lens & van Riemsdijk had for study only one very poorly preserved anterior and posterior necto- phore. Very often specimens of these mid-water species arrive at the surface in a damaged condition, since the somatocyst frequently bursts into the subumbral cavity of the nectosac. Their description and figures leave very little doubt as to which species Lens & van Riemsdijk found and named multidentata. Moser (1925) appears to have had both species under examination since she says that the somatocyst may have either a cylindrical or a butterfly-shaped base. She figured (1925, pi. 23, fig. 1) the anterior nectophore of the new species under the old name. Chuniphyes multidentata Lens & van Riemsdijk, 1908. There are two characteristics of the anterior nectophore figured by Lens & van Riemsdijk which indicate that they had before them the species with the butterfly-shaped somatocyst-base, namely the location of the junction of the two lateral ridges at some distance from the apex, and the position of the rounded protuberance for the attachment of the posterior nectophore. This protuberance is nearer to the ostium of the nectosac than in the second species. I give a new figure of a well-preserved specimen (Text-fig. 66 B). Chuniphyes moserae sp.n. (Text-fig. 66 A.) The anterior nectophore of this species, which was figured quite characteristically by Moser (1925, pi. 23, fig. 1), shows the sub-apical junction of the lateral ridges and the location of the protuberance for the attachment of the posterior nectophore above the mid-level of the nectosac. Moser, however, regarded all specimens as belonging to one species, multidentata Lens & van Riemsdijk. I will not attempt to give an expanded description of the new species here nor show how it differs from the already known one. My figures of the anterior nectophores of the two species will make this clear. That there are differences between the posterior nectophores as well is probable, since there appear to be posterior nectophores of Chuniphyes of at least two kinds. One has markedly asymmetrical ventro-basal teeth and thick mesogloea separating the hydroecium from the nectosac, and is the kind generally associated with Ch. multidentata. The other has sub-equal ventro-basal teeth and a thinner sheet of mesogloea between the hydroecium and the nectosac. It is this kind that I think will prove to belong to Ch. moseri. No doubt the matter will be cleared up one day when we are lucky enough to find well-preserved, complete specimens. The holotype (anterior nectophore) was taken by R.R.S. 'Discovery II' at Station 1639 in lat. 580 35' S.,long. 920 06-2' E. between depths of 2400 and 1150 m. and bears the Brit. Mus. (Nat. Hist.) Register No. 195 1. 3. 24. 1. Twenty-seven anterior and twenty-three posterior nectophores were taken by Beebe in 1929 and 1930 off Bermuda in thirty-three open tow-nets that had been fished for i32 DISCOVERY REPORTS some hours horizontally at depths varying between 900 and 1800 m. Nineteen of the anterior and fifteen of the posterior nectophores were taken between 1450-1650 m., and as many (7) at 1800 m. as (8) above 1450 m. The figures for Ch. multidentata off Bermuda are comparable. Chuniphyes spp. have been recorded from at least fifty-eight other 'Discovery' Stations, varying in depth between 250-100 and 2480-2580 m., but specific identifications have not been re-checked so far. Text-fig. 66. A, Chuniphyes moserae sp.n., holotype, 'Discovery II' St. 1639, 2400-1 150 m., xg; B, Chuniphyes multidentata, 'Discovery II' St. 407, 800-900 m., x 5. The contracted gastrozooids of a specimen from ' Discovery ' Station 2006 are very small. A typical one measured 1-12 mm. in length and 0-37 mm. in diameter. The tentilla are minute and their structure has not been examined or described. Occurrence in the Indian Ocean. 'Discovery II' Station 1587 off Italian Somaliland, 1250-800 m., two anterior nectophores; 1581, off Zanzibar, 1750-600 m., three anterior, two posterior nectophores. SYSTEMATIC AND BIOLOGICAL ACCOUNT 133 Clausophyes Lens & van Riemsdijk, 1908 There appear to be two described species, ovata Keferstein & Ehlers, 1861 (figured), and galeata (not galatea) L. & van R., 1908, well figured by Bigelow 1913. I have also seen material of what appears to belong to a new species from 'Discovery II' Station 2084. Clausophyes has obvious affinities with Chuniphyes. Clausophyes ovata (Keferstein & Ehlers), 1861. (Text -fig. 67.) Eudoxid. 'Discovery II' (Station 1567) took two anterior and one posterior nectophore, all in very poor condition, in a haul from 1350 m. to the surface, together with an interesting though poorly preserved bract of an eudoxid. This eudoxid bract has a rounded apex, and from the phyllocyst arise a pair of bracteal canals. A free eudoxid of Clausophyes has never been described, and the stem groups have been figured only by Keferstein & Ehlers (1861), when they first described the polygastric stage. The presence of the pair of bracteal canals would seem to link this bract of a free eudoxid with those of Keferstein & Ehlers's attached stem-groups of C. ovata. Chi 2. mm 1 Text-fig. 67. Clausophyes ovata. Bract of eudoxid. 'Discovery II' St. 1567, 1350-0 m., x 25. Material. 'Discovery II' has taken this species in twenty-three closing-nets so far examined, at depths ranging from 3000-2000 to 310-260 m. It seems to be characteristic of the deeper water. My identification of ovata is based on the posterior nectophore which has a straight, not emarginate basal edge to the mouth-plate. The left hydroecial fold fits into a notch at the basal end of the right- hand fold. The notch is bounded on the inner side by a prominence which forms a good recognition mark. In a new species from 'Discovery II ' Station 2084, whose description I am unable to complete at the moment, the posterior nectophore is intermediate in size between the smaller ovata and the very large galeata. The right-hand hydroecial fold has a very large flap midway along its length, which I have never seen in ovata and galeata. 134 DISCOVERY REPORTS Here is a list of hitherto unpublished records of capture of C. ovata in closing-nets by R.R.S. 'Discovery' and R.R.S. 'Discovery II': Station Depth (m.) Station Depth (m.) 1 00 A 625-675 J753 2900-1400 IOOC f 2500-2000 \ 260-310 2057 I45°-7°° 2059 1 900- 1 400 IOI 850-95 2061 1900-1500 138 100-750 2063 1150-600 666 3000-2000 2065 1 600- 1 400 671 2000-1500 2084 1500-800 675 2000-1500 2636 95°-55° 696 750-500 2639 1 200-600 . 699 3000-2000 2646 1500-800 1639 2400-1 150 2648 I45°-95° 175° 2000-1900 Clausophyes galeata L. & van R., 1908. Clausophyes galatea (sic) Bigelow, 1913. 'Discovery II' took a very well-preserved posterior nectophore of C. galeata at Station 1639 off the Antarctic Ice Edge in the Australian basin, in a haul from 2400 to 11 50 m. The posterior end is thick and gelatinous. The emarginate mouth-plate being concave on the dorsal side and convex, with a pronounced median conical projection, on the ventral. The basal ends of the hydroecial folds extend further than in ovata. The basal end of the right hydroecial wing is thick and trihedral, and is over- lapped by the smaller left-hand one. Inside the hydroecium, just below the group of appendages, there is on either side a finger-shaped projection not shown in Bigelow's figure. The musculature of the nectosac is not complete, but the radial canals can be seen in places. The laterals, after ascending obliquely, run down towards the circular canal but turn up again before reaching it to form a second parallel loop. They finally run down again near the ventral radial canal. This nectophore measures 43 mm. in length. Crystallophyes amygdalina Moser, 1925. The only published records of this rare mid-water species are those of Moser's three original types from off the Antarctic continent, and of two specimens from South Georgia reported by Leloup ('Meteor' Expedition, 1 934*2). 'Discovery' and 'William Scoresby' took seventy-three anterior nectophores and twenty-three specimens of an associated posterior nectophore in thirty-seven closing- nets at twenty-nine 'Discovery' stations. The posterior nectophore (Text-figs. 68, 69 B, C) is un- described and appears to belong to this species. 'Discovery I IV 'Atlantic Line' series of stations (661-99) showed that the species was present only at Stations 661, 663 and 671. Like most small, deep-water siphonophores the specimens were usually in poor condition, but one or two better preserved specimens have enabled me to add to our knowledge of its morphology, especially of the previously unknown posterior nectophore. Moser's (1925) figure of the anterior nectophore does not show the lateral ridges, but she does indicate the characteristic curved ostial end of one. The very minute mouth-plate (Text-fig. 69, Pl.M.) does not curve in under the velum as suggested in Moser's figure. There is a short pedicular canal, as shown in my figure, at the junction of the radial canals. In a well-preserved specimen, what appears to be the scar of attachment of the muscular lamella can be seen continuing for nearly 1 mm. distad to the pedicular canal. There is some variation in the arrangement of the junction of the ridges SYSTEMATIC AND BIOLOGICAL ACCOUNT i35 Records of capture of Crystallophyes amygdalina in closing-nets by R.R.S. ' Discovery', 'Discovery IV and ' William Scoresby' [These records do not take into account some 30,000 plankton samples so far unexamined. Only one specimen was taken in each haul unless otherwise stated. Posterior nectophores have not been included.] Depth (m.) 3000- 2000 2500- 2000 2400- 1150 2000- 1500 1650- 950 1500- IOOO H5o- 95° 1400- 700 1000- 75° 800- 500 800- 350 75°- 500 500- 250 460- 320 39°- J37 395" J55 I380- J5° ' Discovery II ' Stations 'Atlantic Line' 661 663 671 'Discovery' and 'Discovery II' other Stations 100c 117 125 129 I51 335 337 355 35° 357 358 5J9 1616 1617 l635 1636 1639 1723 1728 2648 'William Scoresby' Stations WS30 WS36 WS38 WS67 'Discovery' Indian Ocean Stations 1585 1586 X x4 X X X x8 x X xi8 X X X *3 ><3 X X X X X 136 DISCOVERY REPORTS at the apex of the anterior nectophore. Sometimes all five meet in a point, but the junction of the dorsal and lateral ridges may be separated from that of the ventrals. A comparison of the figure of Chumphyes moserae (Text-fig. 66 A) and of the figure of Crystallophyes amygdalina will show an obvious relationship between the two. But in Chumphyes the dorsal and lateral ridges divide below to the apex. Text-fig. 68. Crystallophyes amygdalina. Lateral and ventral views of posterior nectophore, 'Discovery II' St. 357, 750-500 m., x 13. The posterior nectophore of Crystallophyes, hitherto undescribed, is flattened dorso-ventrally in the apical region, and the five ridges do not meet in a point; the ventrals join the laterals at a short distance below the apex. The laterals form prominent pointed teeth at the base. The lateral radial- canals leave the dorsal and ventral radial-canals at a point about a quarter of its length distal to the apex of the nectosac. Only after suitable staining is the nature of the ventral ridges apparent. At about the level of the mid-length of the nectosac, the ventral ridges form pointed flaps, between which and their distal continuations open grooves (Text-fig. 69 B) are found. We do not know at present how the anterior and posterior nectophores are articulated, but it would appear as if at some point there is a grasping of some part of the anterior by the posterior nectophore. SYSTEMATIC AND BIOLOGICAL ACCOUNT 137 Quite recently, since this report was drafted, it has been found that C. amygdalina was present at 'Discovery II' Ice-Edge-Cruise Stations between 59 and 980 E., and at another Ice-Edge Station 1723- A Cp°-~ Crad/at P/K4 -~P/M C.ptd Text-fig. 69. Crvstallophyes amygdalina. A, lateral view of anterior nectophore, 'Discovery II' St. 663, 1000-750 m. ; B, C, ventral views of posterior nectophores; B, 'Discovery' St. 129, 500-250 m. ; C, 'Discovery II' St. 663, 750-50001., x 14. Heteropyramis maculata Moser, 1925. (Text-fig. 71.) Moser described this species from specimens taken in seven open vertical hauls from deep water made between the Cape Verde Islands and Ascension Island. At one station on the equator 'Gauss' took a great number of nectophores in a vertical haul from 3000 m. Leloup (1934a) recorded it from the South Georgia region. Moser commented on the strong dorso-ventral flattening of the nectophore, on the open hydroecium running the whole length of the nectophore, and on the long somatocyst seated on the apex of the nectosac, the apex of which only reached the middle of the nectophore. Moser also drew attention to a number of opaque (pigmented) patches found, one at the apex, one at the basal end of each lateral ridge, and from two to five more, according to size in between, on each lateral ridge. The lower two are sometimes small, as if in process of formation. Such opaque spots are known only in the nectophore and the alleged eudoxid of this species. The phenomenon may be related to similar ones occurring in several other siphonophores. For example, Bassia bassensis has long been known to have opaque ridges.1 Also in Diphyes bojani parts of the ridges of the special nectophore of the eudoxid are sometimes opaque, and in Chuniphyes multidentata the same phenomenon is common. It is probably related also to the phenomenon occurring in Hippopodius hippopus and in the anterior nectophores of some species of Abylids, where preserved nectophores 1 Huxley (1859) said that they were dark blue when the animal was alive. 18 138 . DISCOVERY REPORTS are commonly opaque. At Villefranche I have observed that the nectophores of undisturbed specimens of H. hippopus are quite transparent, but that when disturbed they immediately become opaque. After a rest they once more become transparent. As far as is known no research has been carried out upon this phenomenon. But I have recently found a description of this same phenomenon by Chun (1888 a) which he observed whilst working in the Canaries. He said that the turbidity occurred in the ectoderm of Hippopodius, and added that a related phenomenon observed in the eudoxid of a Ceratocymba was due to the appearance of extremely fine granules in the neighbourhood of the horn-like canals of the somatocyst. These granules later spread throughout the mesogloea, but disappeared again in about half an hour if the eudoxid was left in repose. In 'Discovery' hauls so far examined I have seen eleven 'spotted' anterior nectophores. Neither the number, size, degree of opacity nor symmetry of their opaque spots is constant, if my identifica- tions of these small and ill-preserved nectophores are correct. In the better preserved of these nectophores of Heteropyramis maadata examined, nine opaque spots can usually be seen, one each at the apex and basal end of each dorso-lateral ridge, and three intermediate ones on each of these ridges. I have observed as few as eight spots and as many as sixteen in all. In the following diagram the kind of variation that exists in the opaque spotting is shown. ise? 1567 2636 264B IOOC 26 46 693 Text-fig. 70. Heteropyramis maadata. Diagram of distribution of pigment-spots in specimens from various 'Discovery' Stations. A serious doubt has arisen in my mind as to whether H. maadata and Thalassophyes crystallina1 are not really one and the same species. The difficulty is that the spotted form Heteropyramis maadata is rare, and the twelve nectophores that I have examined were none of them sufficiently well preserved to enable me to construct a very reliable drawing of the side-view. Moser, who described the necto- phores as flattened dorso-ventrally did not attempt to draw such a lateral view. Consequently, it is difficult to compare her published figures of the nectophores of the two forms. The ventral facets are similar, and both have two small incomplete ventro-lateral ridges, one on each side of the opening of the hydroecium. I can see no significant external differences in shape or size. As to the somatocyst, the absence of opaque spots — in Thalassophyes — on the lateral ridges does seem to be associated with a large, well-preserved and vacuolated somatocyst, the proximal part of which lies obliquely to the main axis of the nectophore; the distal part, which is a fine tube, lies roughly at right-angles to the proximal part. In one or two of the spotted forms the somatocyst can be partially seen. It appears to be thin-walled, never well preserved and straighter. Perhaps the terminal part of the somatocyst of spotted forms approaches more nearly to the apex of the nectophore. Two nectophores from 'Discovery' Station 2636 have somatocysts of this second type, but no opaque spots other than small, faintly discernible opaque areas at the apex of each. It may possibly be found in future that opaque spots can be present or absent in nectophores of Heteropyramis maadata. In this connexion my observations on the temporary opacity of the nectophores of Hippopodius hippopus should be borne in mind. It seems to be all the more probable that in the nectophores, opaque spots may be present 1 Text-fig. 72. SYSTEMATIC AND BIOLOGICAL ACCOUNT 139 or absent, because that is just what is found in the very peculiar and characteristic eudoxids of Heteropyramis maculata (see Text-fig. 71) described and figured by Moser (1925), and since found again in some numbers. It is possible, but scarcely likely, that there are two eudoxids of the same remarkable shape but belonging to two distinct species, one with opaque spots and one without. We find eudoxids of both kinds associated in hauls with both the spotted nectophores of H. maculata and the unspotted ones of Thalassophyes crystallina. The chief reason why I shall continue for the present to recognize nectophores of two species is that often quite well-preserved specimens of T. crystallina are found, but none of them ever have spots, whereas the spotted nectophores of Heteropyramis maculata are always poorly preserved and usually twisted and deformed. It is possible of course that shrinking and deformation are correlated with opacity. But recently a good specimen of H. maculata was taken by Scot. Fish. Res. V. ' Scotia'. Chuniphyes, too, is frequently much deformed, and is well known for its opaque ridges. The externally well-preserved spotted nectophore of Heteropyramis maculata taken at the end of June 1951 by ' Scotia' in a closing-net — 1000-250 m. — in lat. 540 15' N., long. 140 32' W., enables me to confirm the resemblance of H. maculata to Thalassophyes crystallina. The hydroecium is not open to the apex, as stated by Moser, and the nectophore is not flattened dorso- ventrally, but has the usual Diphyid pyramidal form. The ventral side of the nectosac of the ' Scotia ' specimen is broken open, and the shape of the base of the somatocyst, which may have burst into the nectosac, is difficult to determine. The pedicular canal cannot be distinguished nor can the lateral radial canals. In profile the apex of the nectosac is very much like that of Th. crystallina. It is shown as pointed in Moser's figures. It may be noted that Moser kept these closely similar, if not identical, forms wide apart in two separate families of her artificial classification, calling the former a Monophyid and the latter a Diphyid. It is true that I have found no posterior nectophore that I can closely associate with the spotted nectophores called Heteropyramis maculata, but I think that the relationship between the two species must be a very close one, if indeed they are separate species. From the Indian Ocean I have one nectophore, one bract and two gonophores of eudoxids, all bearing the characteristic opaque spots of H. maculata. The stations are given below : Station Depth (m.) 500-250 1400-700 450-0 Material 'Discovery II' 1S84 1585 1587 1 bract, 1 gonophore 1 gonophore 1 nectophore Further unpublished records (other than from the Indian Ocean) for this spotted species taken in closing-nets are : Station Depth (m.) Material 'Discovery' and 100 c 45O-550 1 nectophore 'Discovery II' 693 500-350 1 nectophore 696 500-250 2 nectophores 699 500-250 1 bract of eudoxid 2636 950-550 2 nectophores, 1 bract of eudoxid 2646 1 500-800 2 nectophores 2927 750-500 2 nectophores, 2 eudoxids 'Scotia', 195 1, 54° 15' N., 1000-250 1 nectophore 14° 32' W. 18-2 i4o DISCOVERY REPORTS Similar records for this spotted species in open hauls are : Station Depth (m.) 350-0 Material 'Discovery II' 1567 1 nectophore, 1 gonophore of eudoxid 2648 500-0 1 nectophore British Antarctic 'Terra 285 600-0 1 bract of eudoxid Nova' Expedition 191 0-13 'Carnegie VII' 31° 54' S., 88° 17' W. 64 1 000-0 1 nectophore Moser (1925, p. 122) gave a complicated account of the gonophore, which in some respects does not fit the facts as I have checked them. It is important to separate descriptive matter from hypothesis about phylogeny. The simple fact seems to be that the dorsal and lateral radial canals of the nectosac of the gonophore do not underlie the respective ridges, but are disposed much as in abyline gonophores. It is quite unnecessary, therefore, to put forward such a fantastic hypothesis as that used by Moser. So/71 /Vi Text-fig. 71. Heteropyramis maculata. A, eudoxid, 'Discovery' St. 129, 950-780 m. ; A', diagrammatic transverse section of gonophore, B, C, bract of eudoxid, ' Discovery II ' St. 2927, 750-500 m. ; D, nectophore, ' Discovery II ' St. 2927, 750-500 m. All figures x 15. She supposes that a four-ridged gonophore of Diphyid type (Moser's text-fig. 28 1) is rotated through 450 on its longitudinal axis and that the right ventral ridge then comes into collision with the stem and divides into two to give a five-angled gonophore. She denied what seems to be the truth that there is present an extra dorsal ridge as in Abylinae. Both polygastric and eudoxid stages of Heteropyramis suggest affinities with Chimiphyes and Abylinae. In particular the gonophore has a median dorsal ridge, and the two interesting basal vessels of the phyllocyst not only can be homologized with the pair of basal vessels in the phyllocyst of the peculiarly small bract of Chuniphyes spp. but also with the 'horns' of the phyllocyst of Ceratocymba spp. and other Abylids. SYSTEMATIC AND BIOLOGICAL ACCOUNT 141 On 8 June 1952, after the completion of this report, 'Discovery II' took two even more perfect specimens of Heteropyramis metadata together with two 'spotted' eudoxids at Station 2927. The accompanying Text-fig. 71 of one of these nectophores shows the very close similarity between nectophores of this species and those of Thalassophyes crystallina. In fact the only significant difference seems to be in the presence or absence of the spots. The distribution of opaque spots on the two eudoxids is as follows : an apical spot on the bract, a spot at the base of each dorso-lateral ridge of the bract, and a spot at each end of the two dorso-lateral ridges of the gonophore. Thalassophyes crystallina Moser, 1925. (Text-fig. 72.) Moser (1925) described this species from two anterior nectophores only, one from 2000 m., at ' Gauss ' Station in the Antarctic, and the other from an unknown locality. It has not been recorded since except by Leloup (1934c), who identified one anterior nectophore taken at 'Meteor' Station 56, from a depth of between 800 and 600 m., to the north-west of South Georgia. It occurred at 'Discovery II ' Stations 1568, 1585, 1588 and 1589, in the Indian Ocean. At Station 1585 it occurred in a closing-net haul made between 1400 and 700 m. ; at Station 1588 in a closing-net haul made between 500 and 200 m.; and at Station 1589 in two open hauls from 600 m. to the surface. Since this is a little-known species, I give all the other unpublished records that I have accumulated : 'Discovery' Stations, closing-net hauls, iooa (625-675 m.); 100c (450-550 m.); 102 (750-500 m.); 107 (1000-750 m.). 'Discovery IV 300 (1000-750 m.); 311 (500-250 m.); 343 (500-250 m.); 344 (500-250 m.); 357 (750-500111.); 661 (750-500, 1000-750111.); 663(750-500, 1000-750, 1 500-1000 m.); 666 (1000-750 m.); 671 (1500-1000, 2000-1500 m.); 1617 (395-155 m.); 1624 (350-210 m.); 1627 (580-400 m.); 1633 (1100-875 m.); x^39 (2400-1 150 m.); 1723 (800-500 m.); 2024 (650-400 m.); 2636 (950-550 m.); 2639 (1200-600 m.); WS 26 (1000-750 m.); WS 36 (750-500 m.). Anterior nectophores have been taken in closing-net hauls: Metres 'Discovery' No. of Station specimens 2000-1500 671 3 1500-1000 663 2 671 6 1400-700 IS85 1 1000-750 300 1 661 8 663 1 107 1 WS26 2 750-500 102 1 357 2 661 5 663 2 WS36 1 675-625 IOOA 3 650-400 2024 1 55°-45° 100c 2 500-250 311 1 343 1 344 1 500-200 1588 1 'Scotia', 1951, Haul 363 1000-250 — 1 142 In open hauls: DISCOVERY REPORTS Stations Depth (m.) Material 'Discovery' 87 iooo(-o) 1 nectophore 100 475(-°) 3 nectophores 208 8oo(-o) 1 nectophore 'Discovery II' 669 20oo(-o) 1 nectophore 1589 600-0 2 nectophores British Antarctic ('Terra Nova' 285 600-0 2 nectophores Expedition), 1910-13 Irish Fishery Board SR 175 (40 miles N.W. of Eagle I., Co. Mayo, Ireland 1097-0 2 nectophores Som H C pa Text-fig. 72. Thalassophyes crystallina. A, lateral view of specimen from 'Discovery II' St. 1588, 500-200 m., x 16; B, C, lateral and ventral views of a specimen from 'Discovery' St. 102, 750-500 m., x 16. The nectosac in this species is only half the length of the nectophore, and the dorsal extension of the somatocyst is found just above its apex. Comments have been made above, in the notes on Heteropyramis maculata, upon the close relation- ship between the two species. ABYLIDAE I recognize the family Diphyidae to be distinct from the Abylidae, which contains two subfamilies : (1) Abylinae (for Abyla and Ceratocymba), and (2) Abylopsinae (for Abylopsis, Bassia and Enneagonum). On 15 June 1953, when I was awaiting the galley-proofs of this report, I received Dr Mary Sears's ' Revision of the Abylinae ' of which she had already been kind enough to allow me to see an advance copy. I had tried to persuade Dr Sears not to publish the descriptions of what I believe to be eight abnormal specimens as the type material of new species and genera. The new names in question are : (1) Pseudabyla irregularis; (2) Pseudabyla dubia; (3) Pseudocymba asymmetrica; (4) Pseudocymba SYSTEMATIC AND BIOLOGICAL ACCOUNT 143 anomala ; (5) Abylopsoides ventralis; (6) Abylopsoides dorsalis; (7) Abylopsoides basalis;(8) Pseudabylopsis anomala. Judging by figures all the specimens so named belong to species of Abyla, Ceratocymba or Abylopsis whose development has been abnormal. I have come across such specimens very occasionally myself. They will be of interest to those who study organogeny, but I feel that it was unnecessarily confusing to give them new specific and generic names. From what little attention I have been able to devote to their analysis I am inclined to believe that, for instance, in Pseudocymba asymmetrica (Sears's text-figs. 23 D, E) the sagittal area of the anterior nectophore has been suppressed, and that there has been an overgrowth of one of the sides bordering on it; and that in Pseudabyla dubia (Sears's text-fig. 17F) one of the margins of the dorsal facet has been suppressed. But we have no knowledge, at present, of what organizers may have been upset during the ontogeny of the specimens. In my opinion if the specimens represented new species, we should find in anterior nectophores of known species just that sort of constant tendency towards variation which has the functional potentiality that seems to appear in the hydroecial folds of the posterior necto- phores. Also specimens of Abyla spp., Ceratocymba spp. and Abylopsis spp. are very abundant, but these crazy specimens are very rare, and none of them have the appearance of normal healthy speci- mens : it is not their rarity alone that makes me think that they are freaks. I can only record my regret that Dr Sears finally decided to make this abnormal material into the types of new species and genera, since I fear it will cause much confusion amongst systematists. Dr Sears has brought together much old and a great deal of new data on Abyline species, but the animals appear to be exceptionally variable, and this is only the starting point of our inquiries. The figures are very clearly drawn, but were not made, I understand, with a camera lucida. Those of anterior nectophores do not perhaps bring out the fact that asymmetry is the rule. My own analysis (p. 144) of the homologies of ridges in posterior nectophores leads me to affirm that in Ceratocymba spp. the dorsal ridge is present, whilst it is absent in Abyla spp. Dr Sears evidently holds the opposite view. For the sake of any systematist who tries to compare Dr Sears's descriptions with my own I should mention that in her very useful figure 5 (p. 14) the following synonyms must be borne in mind: Dr Sears's terminology: Terminology of this report: Dorsal ridge Left lateral ridge Left lateral ridge Right lateral ridge Right ventral wing Left ventral wing Left ventral wing Right ventral wing Right lateral tooth Left lateral tooth Abundant material of Abylid eudoxids is available in the Brit. Mus. (Nat. Hist.) Collection. Systematists would be very grateful for a revision which would enable them to identify and relate them with their corresponding polygastric stages. Abylinae Abyla The species of Abyla, as that name has been generally used, fall naturally into two groups, which should be given generic status. The chief characteristics of the two groups are as follows : in the species of the Ceratocymba group, namely C. sagittata, C. dentata and C. leuckartii, the eudoxid is a 'cymba', and the posterior nectophore has left and right dorso-lateral ridges terminating on the lateral teeth. There is a short, median-dorsal ridge terminating in the dorsal tooth. In the species of the trigona- group namely Abyla trigona, A. carina, A. haeckeli, A. bicarinata and A. schmidti, the eudoxid, where it is known, is an ' amphiroa' '; and in the posterior nectophore, what appears to be the homologue of 144 DISCOVERY REPORTS the left1 ridge in the Ceratocymba group, runs down to the mid-dorsal tooth, and not to the left tooth. In species of Abyla (sensu stricto), but not in those of Ceratocymba, the ventro-lateral facet of the anterior nectophore is divided by a horizontal ridge. As far as I can follow Moser (1925, p. 314) she said that the condition of the teeth and ridges at the distal end of the posterior nectophores of C. sagittata and C. lenckartii is the most primitive and most like Diphyes, and I agree with her. In Ceratocymba dentata she said that the median dorsal (Moser's ventral) ridge has shortened and is displaced to the right, together with the middle (dorsal) tooth. In C. dentata I have not seen such a displacement to the right, but I have seen a slight displace- ment towards the left. In C. leuckartii, however, the middle tooth is much closer to the left lateral than to the right lateral ridge. Moser appeared to believe that in the trigona-group the dorsal (Moser's ventral) ridge and dorsal tooth had disappeared altogether, that the tooth at the oral end of the left lateral ridge is the left lateral tooth, and that the tooth furthest to the left is a new structure, the 'pseudo-tooth'. In reality I think that the middle tooth and the distal end of the left lateral ridge have come together so that the left lateral tooth has no corresponding ridge (PI. IX, fig. 4). The beginning of the process can be seen in C. leuckartii, where the left ridge and middle tooth are close together. In fact, in C. leuckartii Moser's joint left and right ('ventral ') dorsal facets are presentjustastheyarein C. dentata. But the middle (median dorsal) tooth is at the left margin of this joint dorsal facet in the trigona-group instead of in its middle line as in C. dentata. One might almost say that in the trigona-group the reduced median dorsal ridge and tooth have coalesced with the left lateral ridge. Moser thought that the dorsal (her ventral) tooth had completely disappeared, together with its ridge, and that a ' pseudo- tooth' had been developed on the left side to compensate for this. But there is no need to invent a 'pseudo-tooth', for of course in all abylines there are three dorsal teeth at the distal end of the posterior nectophore, and it seems to be so much more probable that they are homologous structures in all species. One of the most striking features of the nectophores of abylines is the peculiar angle between the dorso-ventral planes of the anterior and posterior nectophores. Viewed from aft (PI. IX, fig. 4), the distal end of the posterior nectophore in some species is twisted clockwise through an angle of from 550 to 900 so that the dorsal sides of the two nectophores are no longer opposite one another. The twisting is accompanied by a progressive broadening of the left wing of the hydroecium in species of the trigona-group. Here, too, the distal edge of the broad left wing is marked by a double row of spines, an inner and an outer row as can be seen from a distal and ventral view. These two rows of spines are much better marked in Ceratocymba dentata (PI. X, figs. 4, 5). The starting-point of this evolutionary trend of axis-twisting and wing-broadening can best be seen in the homologous structures of the posterior nectophore of Diphyes dispar, where the distal edges of both right and left wings of the hydroecium are formed into a flattened area without any trace of spines. In that species the ventral edges of both hydroecial wings reach the level of the velum and then bifurcate, turning distad to form the flattened area just described. No specimens of those species of the trigona-carina type occurring in the Atlantic have been found in the 'Discovery' material from the Indian Ocean, though Dr Sears (1953) has recorded the capture of A. trigona from that ocean at 'Dana' Stations 3920, 3921 and 3955. The specimens consisted of one 1 My terminology differs from that of Bigelow & Sears (1937, p. 4) because I use the same orientation (oral and aboral) for an adult as for a larva, where it is self-explanatory. Unfortunately these animals progress with the aboral end forwards. By a very old convention the ventral side of a larva is the one on which budding first appears. 'Ventral' has thus become to be synonymous with adaxial, and 'dorsal' with abaxial. As far, then, as orientation of posterior nectophores of diphyids is concerned my 'left' and Moser's 'left' corresponds with Bigelow's 'right' and Sears's 'right'. SYSTEMATIC AND BIOLOGICAL ACCOUNT H5 colony and seven anterior nectophores. In my opinion, anterior nectophores of A. schmidti cannot be distinguished with certainty from those of A. trigona. Abyla schmidti (Text-figs. 73, 76, 77), though common in the Tropical Indian Ocean has not been noted hitherto in the North Atlantic except at Nant Protch Text-fig. 73. Abyla schmidti Sears. A, specimen from 'Discovery II' St. 1581 (base of right-hand side slightly restored), x6-4; B, C, D, posterior nectophores ; B, 'Discovery II' St. 1587, x6-4; C, 'Mabahiss' St. 131, x6"4; D, 'Discovery II' St. 1585, x 6-4. In fig. D, the break in the apophysis indicates shortening of total length. ' Dana ' Station 4762. This record perhaps needs checking. There is however an allied form in the western South Atlantic (Text-fig. 74). So far as I can tell at present, there appears to be more than one species of Abyla with anterior nectophores similar to those figured by Quoy & Gaimard (1827, pi. 2, B, figs. 1-8), Gegenbaur (1859), D XXVII r9 146 Npost Text-fig. 74. Abyla sp. aff. trigona from Tropical Atlantic. A, complete specimen from 'Discovery II' St. 711; B, C, D, different views of a specimen from 'Discovery II' St. 709. A-D x 4-2. Whydj W.hydl. iQmm Protch>* ^^v=^-=^r^^>< To.vr To.v.1. Text-fig. 75. A, Abyla trigona, posterior nectophore, 'Discovery II' St. 2067, Canaries, X42; B, Abyla carina, part of 'Challenger' material, posterior nectophore; C, Abyla tottoni, basal part of posterior nectophore, x 5-2, 'Discovery II' St. 1 178, eastern South Atlantic; D, Abyla trigona Q. & G., comb of right ventral hydroecial wing of posterior nectophore, part of type material (Mus. d'Hist. Nat. Paris); E, same, dorsal view of base of posterior nectophore to show ventral teeth, x 5-5. SYSTEMATIC AND BIOLOGICAL ACCOUNT 147 and Bigelow (191 1 b) for A. trigona. Fortunately Quoy & Gaimard's types have been preserved at Paris (Text-fig. 75 D, E) and thanks to the good offices of the authorities there,1 have been re-examined by both Dr Sears of Woods Hole and myself. At the moment, I have found no reliable criteria for separating the anterior nectophores of species of the trigona-group, though they are associated with several readily distinguishable types of posterior nectophores. As stated before, the distal (oral) end of the left (larger) hydroecial fold of the posterior nectophore of species in this trigona-carina group of species carries two marginal rows of spines. It appears that there is a tendency for the outer margin to overgrow the inner margin, so that a protuberance is formed as figured by Moser (1925) for A. bicarinata. The relative size and shape of this protuberance, together with the number of spines on the margins and the number of large spines or teeth on the infolded flap, ' Leiste ' or ' comb ' of the right hydroecial wing, may well be found to have specific significance. As a result of long and tedious work it appears that, in the Indian Ocean, there are six abyline species, Ceratocymba dentata, C. leuckartii, Abyla haeckeli, A. trigona and the two new ones A. schmidti and A. ingeborgae. Dr Mary Sears kindly kept me informed in letters of her intention to publish a revision of the Abylinae and gave me some of her views on the subject. I sent Dr Sears copies of some of my figures, and also one or two specimens. Dr Sears also sent me copies of her figures and finally a copy of her text. Her paper (1953) reached me on 15 June when I was awaiting galley-proofs. In it, one of my supposed new species will be found described and named after Dr Johannes Schmidt. Abyla schmidti Sears 1953. (Plate VIII). Posterior nectophores resembling one taken by ' Gauss ' to the south-east of Madagascar and figured by Moser (1925, pi. xix, figs. 7-9) were taken at several East African ' Discovery II ' Stations between lats. 70 S. and n°32'N. With two detached anterior nectophores Moser named her posterior nectophore A. bicarinata. In one haul only ('Discovery II' Station 1581, 600-0 m.) an anterior nectophore was found still attached to a posterior one of this kind, and this complete specimen (Text-fig. 77 A) I had designated the holotype of a new species, now schmidti Sears. Dr Sears has shown that Moser's figured anterior and posterior nectophores belong to separate species, the pos- terior one is A. schmidti. I now choose the larger of Moser's anterior nectophores as the holotype of A. bicarinata Moser. Moser, did not, I think, figure the velar end of the posterior nectophore very critically, unless the specimen was a dead, disintegrating one such as one finds sometimes in the plankton; but her fig. 7 shows a characteristic, distal, chin-shaped protuberance of the left hydroecial wing. It extends considerably more distad than in any of the ' Discovery ' specimens, even than in those which have begun to disintegrate and were presumably mature. I have not so far found posterior nectophores like those described and figured above (Text-figs. 73, 76, 77) as A. schmidti, except in material from the Indian Ocean, and certainly not from the region of the Canaries, where the types of both A. trigona and A. carina were found. Specimens of those two species (probably one and the same) grow to a size much greater than any of the Indian Ocean specimens that I have seen. But specimens of an intermediate type (Text-fig. 74) were found in the Tropical Atlantic. Much Abyla material is now available for some student of the future. Eudoxids. No entirely satisfactory figures have been published of the eudoxid (Amphiroa Blainville) of any species of the trigona-group of Abyla to which A. schmidti belongs. The earliest, Blainville's (1834) coPy °f Lesueur's unpublished drawing of a Bahaman specimen named alata by Blainville (1830), omits details of the hydroecial wings and folds or Leisten. Huxley's (1859) figures are more 1 I have to thank Dr A. Franc for much help in this matter. 19-2 I4g DISCOVERY REPORTS characteristic, whilst Gegenbaur's (1859) are not so satisfactory as Huxley's. Haeckel (18886) gave more elaborate drawings of A. carina, but the gonophores were not shown with sufficient detail. Bedot (1896) figured the only bract he saw at Amboina, but not the gonophores. Lens & van Riemsdijk (1908) again failed to give details of gonophores. Moser (1925) was the first to give a detailed sketch of the gonophores, but did not show an optical section of the mouth-plate from the side, a detail that is necessary for specific identification. In Part v of Kawamura's (191 5) paper on Caliconectid Siphono- phorae (translated in typescript by Rodney Notomi, and a copy generously presented to me by Woods Hole Oceanographic Institute), there are some useful original figures of Abylines. Figs. 27 and 28 of plate xv, labelled A. trigona from Misaki are good, but these again omit essential details of the gonophores. Whydr W.hydl. Xf- Tov.1 Tovr Text-fig. 76. Abyla schmidti. A, B, C, D, different views of a detached anterior nectophore from 'Discovery II' St. 1588 250-100 m., x 5-5. Note the slight asymmetry characteristic of all such nectophores; E, F, G, three views of distal end of a posterior nectophore from 'Discovery II' St. 1587. For comparison with the new Indian Ocean species of Abyla schmidti I give figures of the following: (1) type specimens of A. trigona Q. & G., kindly lent by the Museum d'Histoire Naturelle, Paris (PL XII) ; (2) type specimens of A. carina Haeckel (Text-fig. 75 B) ; (3) a specimen of A. trigona from ' Discovery II ' Station 2067 (Text-fig. 75 A), all these three from the Canaries region of the Atlantic; (4) a specimen of a related species A. tottoni Sears from 'Discovery II' Station 1 178 in the eastern South Atlantic (Text-fig. 75 C); (5) another related but unnamed new species from 'Discovery II' Stations 709 and 711 in the western South Atlantic (Text-fig. 74). Dr Sears states that both Quoy and Gaimard and Haeckel were confusing two (the same two) species in their accounts of A. trigona and A. carina. Her argument is that (1) Quoy and Gaimard's type anterior nectophores were too small to have been linked to the type posterior nectophores, which probably do not belong to them, (2) Quoy and Gaimard's type anterior nectophores resemble (a) some SYSTEMATIC AND BIOLOGICAL ACCOUNT 149 of Haeckel's Challenger [not type] specimens, and (b) some ' Dana ' specimens, some of both (a and b) of which have small inferior nectophores attached within the hydroecium. (3) These 'Challenger' and ' Dana ' young inferior nectophores differ from those inferior nectophores amongst Quoy and Gaimard's types. (4) One of Haeckel's 'Challenger' colonies (complete) has an inferior nectophore resembling Quoy and Gaimard's types. (5) It thus seems certain that these are the same two species in both Quoy and Gaimard's and Haeckel's samples. But Dr Sears does specify the resemblances and differences alluded to. She goes on to say that (6) Lens and van Riemsdijk (1908) can be considered Nam Apoph Whydr Whydl Tad Prol ch Tad Tovr To.vl Text-fig. 77. Abyla schmidti Sears. A, specimen from 'Discovery II' St. 1581, x 5-5; B, lateral view of anterior nectophore, 'Mabahiss' St. 131, x 5-4; C, view from above of some specimen as A, x 5-5 ; D, baso-lateral view of posterior nectophore from 'Discovery II' St. 1585, x 5-6. to have designated the superior nectophores of Quoy and Gaimard as the types of Abyla trigona, because they compared their specimens of anterior nectophores of ' Abyla trigona ' [in my opinion probably A. schmidti, A.K.T.] with those of Quoy and Gaimard and found them to be identical ' in all respects' [Dr Sears refers Lens & van Riemsdijk's posterior nectophores of 'Abyla trigona' to A. schmidti]. (7) Since the inferior nectophores of Quoy and Gaimard's types apparently belong to the same species as the complete colony and three superior nectophores in Haeckel's ' Challenger ' sample [in B.M. Coll.], she considers it appropriate to designate these as carina. From this I gather that Dr Sears regards Haeckel's ' Challenger ' anterior nectophores (but not the posterior ones) and Quoy and Gaimard's type posterior nectophores (but not anterior ones) as Abyla i5o DISCOVERY REPORTS carina, nevertheless she regards Haeckel's carina, if I understand her correctly, as trigona, yet she uses Haeckel's figure of his type carina as one of her criteria for distinguishing the posterior nectophore of carina. As other criteria for separating the posterior nectophores of the species, Dr Sears refers to (i) the ventral teeth, elongate, heavily serrated, straight and sharp in trigona but 'stubby' in carina; (2) comb with 6-8 teeth in trigona, 9-10 teeth in carina; (3) margin of ventral wing with heavy and prominent teeth in trigona, but not so robust and jagged in carina, and scarcely more than strong serrations ; (4) width of the nectophore half its length in trigona but nothing in its general shape to New records of Abyla schmidti Depth (m.) Polygastric Eudoxid Anterior Posterior Complete nectophore nectophore Complete Bract Gonophore 'Discovery II' Station *374 230-0 — — 1 — — — 1581 100-50 — 1 — — — — 600-0 1 4 — 30 — J5 1582 i900-i85o(-o) — 7 4 — — — 1583 250-100 — — — 1 — — 1585 500-0 — 4 1 10 — 10 1400-700 — — — — — 1 1586 1650-900 — 2 1 — — — 55o-o — 1 3 1 2 8 100-50 — — — 1 — 1 50-0 — — 1 — — 1 1587 1250-800 — 1 1 — — — 450-0 — 4 2 1 1 1 50-0 — — 1 — — — 1588 250-100 — 1 1 — — — 1589 600-0 — — 1 — 1 — 175° 1900-0 — — 1 — — — 'Mabahiss' Station 131D 2500 — 1 1 — — — 1500 — 2 2 — — — 500-0 — 1 — 2 — 1 131A 600-0 — i — 2 — — distinguish that of carina; (5) an overgrowth of the outer ventro-basal margin of the larger of the ventral wings in carina. In her own words ' the inner row of nine or ten small spines is parallel to the basal margin close to the two ventral teeth but ventrad it gradually swerves inwards a short distance from the margin crossing the inner surface to merge with the outer row at the ventral margin, much as is shown in one of Haeckel's drawings (1888&, pi. 35, fig. 1)'. We know nothing of Haeckel's actual type specimens, though the ' Challenger ' ones still exist at the Brit. Mus. (Nat. Hist.). Therefore we have to depend entirely for data on the ' Challenger' figures and description of the Lanzerote type specimens. I am not satisfied as yet that A. carina Haeckel is not a synonym of A. trigona Quoy and Gaimard. Chun (1888 a) said that A. trigona occurred in great swarms in the Canaries at the end of February and in March. Even the eight anterior nectophores of Haeckel's type material show variation in the shape of the dorsal tooth and in other respects. In the adults of both species the general shape of the nectophores is the same, there are nine or ten teeth on SYSTEMATIC AND BIOLOGICAL ACCOUNT 151 the comb of the right ventral wing, and the shape of the ventral teeth is the same. It is true that in a young posterior nectophore of one of the 'Challenger' specimens of A. carina which is only just emerging from the hydroecium of the anterior nectophore, the ventral teeth are relatively longer, narrower and more spiny, but this difference is probably a matter of age and growth. In order satisfactorily to delimit species of the A. trigona group, much statistical analysis will have to be made on large samples of swarms. Dr Sears has already published a description of another new species of this group. Abyla tottoni Sears, 1953. (Plate IX.) Dr Sears recorded the capture in 1930 of 27 complete colonies, 30 anterior and 19 posterior necto- phores of Abyla tottoni at three ' Dana ' stations in the subtropical South Atlantic, as well as one anterior nectophore from between ' Timor and New Guinea '. 'Discovery II' took 12 complete colonies, five anterior and six posterior nectophores and three eudoxids of this species in 1933 at Station 11 78 close to one of these 'Dana' stations; and a young specimen (post. nect. 15 mm. 1.) some 1500 miles further west in the same region at Station 686. The whole margin of the larger ventral wing may be serrated. There may be from 12 to 13 teeth on its inner basal margin, and from 1 1 to 13 teeth on the basal margin of the smaller wing, the ventral corner of which, between comb and base, is more rounded than shown by Dr Sears (fig. 15 B). The number of comb teeth may be as large as 12. But it is difficult to count comb teeth at the apical end, and to agree on the number when there are smaller intermediate teeth, or the final apical one is not well marked. This appears to be a distinguishable species, and I had already marked it as a new one. Abyla haeckeli Lens and van Riemsdijk, 1908. Until Dr Sears published her revision of the Abylinae I had assumed that all anterior nectophores of Abyla, in which the apico-ventral facet was subdivided by a transverse ridge, belonged to the species haeckeli. Whilst this report was in the press I re-examined my figured material. All the anterior nectophores from ' Discovery II ' E. African stations are relatively small, and have the type of ventral facet described and figured by Dr Sears for haeckeli. All the specimens from ' Discovery II ' Atlantic stations are relatively large, and have a ventral facet of the type described and figured by Dr Sears for ingeborgae. I am hoping that these two forms will prove to be quite distinct species, though Browne's specimens seem to be intermediate. Dr Sears (1953) has published a figure of a young posterior nectophore for the first time. She appears to have examined 10 colonies, 553 anterior nectophores and 51 posterior nectophores from 38 'Dana' hauls. This should be a good sample, and it is hoped that some one will have time to establish the range of specific variation. I have re-examined the three anterior nectophores described by Browne (1926). Two of them measure 6-i mm. in length, the third measures 6-9 mm. One of these, specimen 'A' is 4-6 mm. wide, but two are rather squashed, and their width cannot be measured. Specimen 'A' is in good condition and the horizontal ridge in side view is only slightly above the middle of the somatocyst (i/i-i). Dr Sears's critical proportions of the pentagonal ventral facet, namely width-between-middle-corners/ length-from-middle-corners-to-tip are as follows: 1/0-65, ll°'75> 1/0-82. The proportions of the facet figured by Dr Sears (fig. 13 D, E) for A. haeckeli were 1/0-57 and f°r A. ingeborgae i/i-i. On the basis of these last two characters, as well as that of the ventral view, Browne's specimens are intermediate between Dr Sears's figs. 13 D and E. The total size, the relative size of the apico-ventral facet, and the size and position of the lateral projections lead me to conclude that they belong to A. haeckeli. iS2 DISCOVERY REPORTS It is certain however, that specimens of Abyla spp. will not all fit neatly into Dr Sears's proposed categories. I have also re-examined the anterior nectophore recorded by me (1932) from the Great Barrier Reef. The proportions of the pentagonal facet are 1/0-63, an<^ I confirm that it belongs to A. haeckcli. It happens to be another of those partly injured or disorganized freaks that Dr Sears has brought to notice. Abyla ingeborgae Sears, 1953. Complete specimens of this species at 'Discovery' Stations 295 and 297 and at 'Discovery II' Station 699, which enabled me to identify loose posterior nectophores (Text-fig. 78) from Stations 706 and 2639. Until the publication of Dr Sears's paper, I had identified all these Atlantic specimens as A. haeckeli, but anterior nectophores of the two types, haeckeli and ingeborgae, can sometimes be separated at once. A good figure of the full grown posterior nectophore of A. haeckeli is needed for comparison with that of A. ingeborgae. ro./r_ -Rl. -R.n ---Tad --Whydl. To. v.l. Text-fig. 78. A, B, posterior nectophores of Abyla ingeborgae Sears, x 3-8, from ' Discovery II' Sts. 706 and 2639 (1200-600 m.) respectively, each associated with anterior nectophores of that type; C, basal view of posterior nectophore of holotype specimen of Abyla schmidti Sears from 'Discovery II' St. 1581, x 4. The solution of this question of a distinction between A. haeckeli and A. ingeborgae may lie in the fact that a species, whose distribution covers such a wide area as that including the N. Atlantic and Indo-Malaysian oceans, may appear as a larger form in the N. Atlantic, and that problems in allometric growth may be involved. Sulculeolaria biloba appears to be another such species. Ceratocymba dentata (Bigelow), 1918. (Plates X, XI.) This Abylid has been taken in the Indian Ocean only at 'Discovery II' Stations 1585 and 2696, though ' Discovery II ' took it many times in the Atlantic. It has been recorded only by Bigelow (1918), who made the original record of the species from two North-west Atlantic Stations, about 3 20 N. lat. and 70-720 W. long., and by Moser (1925) west of the Cape Verde Islands under the synonym Abyla quadrat a. SYSTEMATIC AND BIOLOGICAL ACCOUNT 153 The only other known and still unrecorded, catches, are of eudoxids taken by Beebe at Bermuda; and of both stages taken by ' Discovery ' at Stations in the South Atlantic. These ' Discovery ' Stations lay between lat. 320 S. and the Cape Verde Islands. They include some Stations (280-298) on a line starting from the West Coast of Africa on the Equator and running westwards and northwards to the Cape Verde Islands; and Stations on another line running up the 300 mer. from 280 30' S. lat. to the Cape Verde Islands; as well as Station 81 at the same south latitude, but nearer the Cape. Also there has recently come to my notice a specimen of Ceratocymba dentata taken by Mr J. S. Colman in 1937 during a cruise in Lord Moyne's yacht 'Rosaura' in the Guinea current. The locality was 70 27' N. lat., 230 08' W. long. It was taken in a 2-m. stramin net fished from 1000 m. to the surface. T 2. mm I — ' ■ I ■ Text-fig. 79. A, C, Ceratocymba sagittata, a posterior nectophore from 'Discovery II* St. 676; A x 3-4; C x 9-8; B, D, Ceratocymba leuckartti, two views of a posterior nectophore from 'Discovery II' St. 694; B x6-5; D x 12-5. Polygastric stage. The lengths of the nectophores of a complete polygastric specimen from 'Discovery II ' Station 2635, 280-0 m., measured from either end to the notch on the dorsal side of the posterior nectophore, where the ventro-basal tooth of the anterior nectophore locks the two in position, are : anterior nectophore, 1 1 mm. ; posterior, 55 mm. The length of the nectosac of the posterior nectophore is 45 mm. ; its diameter at the forward end is 3 mm., increasing to 4-5 mm. just before the terminal bulge, which measures 5 mm. in diameter. There are nine or ten functional stem groups, the terminal three of which show the growth of the eudoxid bract and two gonophores. On one of the bracts, still attached to the stem, in the long hydroecial canal of the posterior nectophore, can be seen the two ridges as found in the full-grown bract and shown in Text-fig. 80. The identity of loose bracts is therefore established. The temperature in which Moser's type of quadrata was taken at the surface on 7 November west of the Cape Verde Islands, must have been about 260 C. i54 DISCOVERY REPORTS The surface temperature at 'Discovery II' Station 2635 was 27-40° C. but it is uncertain at what depth and temperature the species was living, since even at 60 m. at a nearby Station (2640) five days later, there was a drop to 20° C. « Op Q R lot. I Text-fig. 80. Ceratocymba dentata. Lateral and dorsal views of the bract of the eudoxid, x 4-4, from 'Discovery' St. 286 (see Text-fig. 81). C p?tf Cpo C.pa „ Cpecf --To.l ~--Jov./. Text-fig. 81. Ceratocymba dentata, gonophores of eudoxid from ' Discovery' St. 286, x 4; A, B, dorsal views of female and male gonophores; A'B', the same gonophores locked together, viewed from side of attachment to bract (see Text-fig. 80). Eudoxid: A large bract, still bearing an immature female gonophore ('Discovery' Station 286) measures : length, 1 1 mm. ; breadth, 19 mm. ; depth, 14 mm. A pair of detached gonophores from the same haul (mirror images of each other) were possibly carried by the same bract. The larger is female SYSTEMATIC AND BIOLOGICAL ACCOUNT 155 and measures 23 mm. in length, the smaller, male, 20 mm. They fit accurately together, as shown in Text-fig. 81, the hydroecial wings of the larger clasping those of the smaller at the upper and lower sides of the whole eudoxid. The pedicular canals of the two gonophores arise near the forward end on the borders of their upper hydroecial wings, so that they must arise just below the forward end of the swan-shaped somatocyst1. In fact, scars of corresponding shape can be seen in this position on the bract. The hydroecial wings of the gonophores surround the gastrozooid and its tentacle. This combined hydroecial cavity narrows towards its after end, and at its periphery there are two large curved teeth, one above the other, and one belonging to each gonophore. These large teeth are in positions about two-thirds of the way down the obliquely running hydroecial folds, which bear a few more much smaller distal teeth. Such teeth are commonly found on the hydroecial fold or 'Leiste' of gonophores and of posterior nectophores of many genera and species. In the posterior nectophore of this species dentata there are about two dozen teeth on the left hydroecial fold, the ter- minal one as usual being the largest. Their function perhaps is to prevent entanglement of the numerous side branches of the tentacles, or to remove prey. In structure the gonophores of Ceratocymba dentata are perhaps more complicated than those of any other species of Siphonophore, a fact that strengthens the belief that it is the most highly evolved Calycophore. Abylopsinae Abylopsis tetragona (Otto), 1823. In 1 95 1, nine open hauls containing Siphonophores were made by ' Manihine ', six of them during the night of 16-17 January between five miles and forty miles east of Sanganab Lighthouse, off Port Sudan, from either 183 or 366 m. to the surface. An examination of the catches showed that A. tetragona was abundant and actively breeding. At three other 'Manihine' Stations 1 (P), 2 and 3, further north, fifty-six polygastric specimens, and 276 eudoxids were taken. _ - Som. Nsobs Tov.r Whydr Whydl Cmb Text-fig. 82. Abylopsis tetragona. The two nectophores locked rigidly together as in life. Xf2. Since the ' Manihine ' catches appear to be some of the first records of Siphonophores from the Red Sea it was thought worth while to compare the absolute abundance per unit volume of water of this species in the Red Sea ('Manihine' records), in the Mediterranean ('Thor' records) and in the warm tropical Atlantic ('Meteor' records). Bigelow & Sears (1937) went into the question of the last two in some detail. Here I need only give the comparable figures and repeat the caution about accepting figures based on a small number of open hauls, where the straining efficiency of the net, the 1 Clearly shown in Text -fig. 80. 156 DISCOVERY REPORTS speed of hauling and the occurrence of swarms are not known precisely. As far as ' Manihine ' records are concerned it is assumed that an open 1 m. stramin net was hauled obliquely from either 183 or 366 m. to the surface whilst the ship went slowly ahead for about 100 metres. Abundance of Abylopsis tetragona per unit volume of zvater 1. 'Thor', Mediterranean, western basin (Bigelow & Sears, 1937). Mean frequency of specimens, at the level of maximum abundance, at Stations where the species occurred : Polygastric Eudoxid Winter Late June Late August Maximum (late August) 1 per 180 cu.m. 1 per 21 cu.m. 1 per 25 cu.m. 1 per 2 cu.m. 1 per 1 100 cu.m. 1 per 20 cu.m. 1 per 152 cu.m. 1 per 0-2 cu.m. 2. ' Meteor ', warm tropical Atlantic. Frequency at several levels, calculated from Bigelow & Sears's tables (based on those of Leloup & Hentschel, 1935): Polygastric Eudoxid 0-50 m. 50-400 m. 1 per 75 cu.m. 1 Per 375 cu-m- 1 per 60 cu.m. 1 per 375 cu.m. 3. 'Manihine', Red Sea, off Port Sudan, 16/17 January 1951. (T. ° C. estimated at 21-5-22; equivalent to a late June surface temperature at Villefranche, West Mediterranean) : Polygastric Eudoxid Mean frequency at six stations (183 or 366-0 m.) Maximum frequency (St. 4, 183-0 m.) 1 per 4 cu.m. 1 per 2 cu.m. 1 per 1-4 cu.m. 1 per 0-4 cu.m. It appears that Port Sudan would be a favourable spot at which to attempt to observe the species throughout its life cycle, an objective that has not so far been achieved. The importance of this is to ascertain whether the small anterior nectophore is the primary nectophore or a secondary hetero- morphous one. Haeckel's attempt at Lanzerote in 1867 failed. So also did my own at Villefranche in 1950, where, at that time, there were never more than two or three ripe gonophores available on odd days, and then generally one sex or the other and not both at once. Text-fig. 83 illustrates the arrangement of the two hydroecial folds, for comparison with the arrangement in Abyla spp. A distinctive character is that the broad, left-hand fold, the free edge of which Moser called the left, pseudo-dorsal ridge, is held in place at one point between two parts of the right-hand fold, i.e. between a mamma-like structure and the comb or' Leiste ', at a position just proximal to the terminal parting of the two folds. Moser's figure (1925, pi. xx, fig. 4) and her description (p. 330) does not make this clear. It should be remembered that owing to her fantastic ' Urmedusentheorie ' she called the morphological dorsal side of a posterior nectophore ' ventral '. The ' Leiste ' or comb, which Moser described as arising from the 'dorsal' wall of the nectosac should, I believe, properly be considered as part of the free, ventral margin of the right hydroecial fold, as can be seen at any rate proximally in the apophysis or portion of the posterior nectophore that is enclosed in the hydroecium SYSTEMATIC AND BIOLOGICAL ACCOUNT 157 of the anterior one. This comb or ' Leiste ' is the homologue of the comb or ' Leiste ' in Abyla spp. The hydroecium of the posterior nectophore has been displaced to the left side in the proximal half. I observed at Villefranche that the tip of the tentacle consists of a sinker in the form of a disc, around the edge of which are from 13 to 16 nematocysts. A similar sinker occurs at the tip of the tentacle of Rosacea cymbiformis. Cmb- - To.L- Tovl. A C.radd radlat ~i~~C ' radv - - Whydr Whydr Text-fig. 83. Abylopsis tetragona, posterior nectophore, to show how the left ventral hydroecial fold is held rigidly in place by the right. A, latero-ventral view, x 5 ; B, a section just above the locking tooth of the right wing, x 8-8 ; C, apical view, x 8-8. REFERENCES Agassiz, A., 1863. On Nanomia cara. Proc. Boston Soc. nat. Hist, ix, pp. 180-1. 1865. North American Acalephae. 111. Cat. Mus. comp. Zool. Harv. No. 11, pp. 201-13, text-figs. 331-50. Ankel, W. E., 1951. Phyllirrhoe, Mnestra und die Probleme der Gemeinschaften des ozeanischen Pleustals. Verh. dtsch. zool. Ges. 1 95 1, pp. 396-404, 6 text-figs. Bedot, M., 1893. Revision de la Famille des Forskalidae. Rev. Suisse zool. I, pp. 231-54. 1896. Les Siphonophores de la Bale d'Amboine. Rev. Suisse zool. Ill, pp. 367-414, pi. 12. 1904. Siphonophores provenant des campagnes du yacht Princess- Alice (1892-1902). Result. Camp. sci. Monaco, xxvn, pp. 1-27, 4 pis. Berrill, N. J., 1930. On the Occurrence and Habits of the Siphonophore, Stephanomia bijuga (Delle Chiaje). J. Mar. biol. Ass. U.K. xvi, pp. 753-5, 1 text-fig. Bigelow, H. B., 1904. Medusae from the Maldive Islands. Bull. Mus. comp. Zool. Harv. xxxix, pp. 245-69, 9 pis. 1911a. Biscayan Plankton Collected during a Cruise of H.M.S. 'Research', 1900, Part xin. The Siphonophora. Trans. Linn. Soc. Lond. (Zool.), x, pp. 337-58, pi. 28. 191 1 b. The Siphonophorae. Reports Sci. Res. Expedition to the Eastern Tropical Pacific. . .Albatross. . .XXIII. Mem. Harv. Mus. comp. Zool. xxxvm, No. 2, pp. 173-402, 32 pis. 1913. Medusae and Siphonophorae Collectedby the U.S. Fisheries Steamer ' Albatross' in the North-Western Pacific, 1906. Proc. U.S. nat. Mus. xliv, pp. 1-119, 6 pis. 191 8. Some Medusae and Siphonophorae from the Western Atlantic. Bull. Mus. comp. Zool. Harv. lxii, pp. 363-442, 8 pis. 19 19. Contributions to the Biology of the Philippine Archipelago and adjacent regions: Hydromedusae, Siphonophores and Ctenophores of the 'Albatross' Philippine Expedition. Bull. U.S. nat. Mus. c, i, 5, pp. 279-362, 5 pis. 193 1. Siphonophorae from the ' Arcturus' Oceanographic Expedition. Zoologica, N.Y. vm, No. 11, pp. 525-92, text-figs. 185-220. Bigelow, H. B. & Sears, M., 1937. Siphonophorae. Rep. Danish oceanogr. Exped. Medit. 11, Biology, H. 2, pp. 1-144, 83 text-figs. Blainville, H. M. de, 1830. Zoophytes. Diet. sci. nat. Paris. LX, pp. 1-548, 68 pis. 1834. Manuel d'actinologie ou de zoophytologie, pp. 1-694, I0° P's- 20-3 i58 DISCOVERY REPORTS Bohm, R., 1878. Helgolander Leptomedusen. Jena. Z. Naturw. XII, pp. 68-203, ° pis. Brandt, J. F., 1835. Polypus, Acalephas Discophoras et Siphonophoras, nee non Echinodermata continens. Prodromus descrip- tionis animalium ab H. Mertensio in orbis terrarum circumnavigatione observation, Fasc. I, pp. 1-76. St Petersburg : Graeff. Brooks, W. K., 1886. The Life History of the Hydromedusae : A discussion of the Origin of the Medusae, and of the Significance of Metagenesis. Mem. Boston Soc. nat. Hist, m, No. 12, pp. 359-430, 4 pis. Browne, E. T., 1926. Siphonophorae from the Indian Ocean. Trans. Linn. Soc. Lond. (Zool.) (2), xix, pp. 55-86. Candeias, A., 1929. Note sur quelques siphonophores Calycophorae de Madere. Bull. Soc. portug. Sci. nat. x, No. 23, pp. 269-84, 12 text-figs. Carter, G. S., 1949. The Turbellarian Theory of the Cnidaria. Being a Review of Turbelarijska Teorija Knidarijev', by J. Hadzi. Sci. Progr. Lond. xxxvn, No. 145, pp. 136-9. Chamisso, A. de & Eysenhardt, C. G., 1821 . De animalibus quibusdam e classe vermium Linneana . . . ,fasc. 2. Nova Acta Leop. Carol, x, pp. 343-74, 10 pis. Chun, C, 1882. Vber die cyklische Entzvickelung und die Verwandtschaftsverhaltnisse der Siphonophoren. S.B. preuss. Akad. Wiss. lii, pp. 1155-72, 1 pi. 1885. Vber die cyklische Entwickelung der Siphonophoren. S.B. preuss. Akad. Wiss. xxvi, pp. 511-29, 1 pi. - 1886. Vber Bau und Entzvickelung der Siphonophoren. S.B. preuss. Akad. Wiss. xxxvm, pp. 681-8. 1888a. Bericht iiber eine nach den Canarischen Inseln im Winter 1887-1888 ausgefiihrte Reise. S.B. preuss. Akad. Wiss. xliv, pp. 1 141-73. 18886. Die pelagische Thierwelt in grdsseren Meerestiefen und ihre Beziehungen zu der Oberfldchenfauna. Bibl. Zool. Stuttgart. 1, Heft 1, pp. 12-17, 1 pi. 1891. Die Canarischen Siphonophoren in monographischen Darstellungen. 1. Stephanophyes superba und die Familie der Stephanophyiden. Abh. senckenb. naturf. Ges. xvi, pp. 553-627, 7 pis. 1897a. Vber den Bau und die morphologische Auffassung der Siphonophoren. Verh. dtsch. zool. Ges. vii, pp. 48-1 11, 29 text-figs. 18976. Die Siphonophoren der Plankton-Expedition. Ergebn. Plankton Exp. 2, K.b., pp. 1-126, 8 pis. Kiel and Leipzig. 1913. Vber den Wechsel der Glocken bei Siphonophoren. Ber. sachs. Ges. (Akad.) Wiss. LXV, pp. 27-41, 8 text-figs. Claus, C, i860. Vber Physophora hydrostatica nebst Bemerkungen iiber andere Siphonophoren. Z. wiss. Zool. x, pp. 295-332, 3 Pls- 1878. Vber Halistemma tergestinum n.sp. nebst Bemerkungen iiber den feinern Bau der Physophoriden. Arb. zool. Inst. Univ. Wien. I, pp. 1-56, 5 pis. Cunningham, J. T., 1892. On a Species of Siphonophore observed at Plymouth. J. Mar. biol. Ass. U.K. (N.S.) 11, No. 3, pp. 212-15, 2 text-figs. Dana, J. W., 1858. On a New Species of Medusa related to Stephanomia, Crystallomia polygonata. Mem. Amer. Acad. Arts Sci. (N.S.) vi, pt. 2, pp. 459-60, 1 pi. Dawydoff, C, 1928. Traite d'embryologie comparee des invertebres, xiv + 930 pp., 509 text-figs. Paris. Deacon, G. E. R., 1952. The 'Manihine' Expedition to the Gulf of Aqaba. II. Preliminary Hydrological Report. Bull. Brit. Mus. (Nat. Hist.) Zool. 1, No. 8, pp. 159-62, 4 text-figs. Delle Chiaje, S., 1822. Memorie sulla storia e notomia degli Animali senza vertebre del Regno di Napoli. Atlas, 109 pis. Napoli. 1829. Memorie sulla storia e notomia degli Animali senza vertebre del Regno di Napoli. iv, i-viii+ 1-14 pp. Napoli. 1841. Descrizione e notomia degli Animali Invertebrati delta Sicilia citeriore osservati vivi negli anni 1822-30. vi-vn, Atlas, 181 pis. Napoli. Delsman, H. C, 1923. Beitrdge zur Entwickelungsgeschichte von Porpita. Treubia, m, pp. 243-66, 28 text-figs. Dollfus, R.-Ph., 1925. Liste critique des cercaires marines a queue setigere signalees jusqu a present. Trav. Sta. zool. Wimereux, ix, pp. 43-65, 6 text-figs. Eschscholtz, F., 1825. Bericht iiber die zoologische Ausbeute wahrend der Reise von Kronstadt bis St Peter und Paul. Oken's Isis, xvi, pp. 733-47, pi. v. 1829. System der Acalphen, 1-190 pp. 16 pis. Berlin. Fewkes, J. W., 1880. Contributions to a Knozvledge of the Tubular Jelly-fishes. Bull. Mus. comp. Zool. Han', vi, pp. 127-46, 3 Pis- 1882. Notes on Acalephs from the Tortugas, with a description of New Genera and Species. Bull. Mus. comp. Zool. Harv. ix, pp. 251-89, 7 pis. 1888a Studies from the Newport Marine Zoological Laboratory. On certain medusae from New England. Bull. Mus. comp. Zool. Harv. xm, pp. 209-40, 6 pis. - 18886. On a new Physophore, Ploeophysa, and its Relationships to other Siphonophores. Ann. Mag. nat. Hist. (6), I, pp. 317-22, 1 pi. REFERENCES 159 Forskal, P., 1775. Descriptiones animalium . . . quae in itinere orientali observavit, post mortem editit Carsten Niebuhr. Hauniae, 20 + xxxiv + pp. 15, 1 map. 1776. Icones rerum naturalium. . .post mortem auctoris . . .editit Carsten Niebuhr. Hauniae, pp. 15, 43 pis. Garstang, W., 1946. The Morphology and Relations of the Siphonophora. Quart. J. micr. Sci. Lxxxvil, pt. II, pp. 103-93, 57 text-figs. Gegenbaur, C, 1853. Beitrage zur naheren Kenntniss der Schwimmpolypen (Siphonophoren). Z. wiss. Zool. v, pp. 285-344, 3 pis. 1854. Ueber Diphyes turgida n.sp., nebst Bemerkungen iiber Schwimmpolypen. Z. wiss. Zool. v, pp. 442-54, 1 pi. 1857. Versuch eines Systemes der Medusen, mil Beschreibung neuer oder wenig gekannter Formen; sugleich ein Beitrag sur Kenntniss der Fauna des Mittelmeeres. Z. wiss. Zool. vm, pp. 202-73, 4 pis. 1859. Neue Beitrage sur naheren Kenntniss der Siphonophoren. Nova Acta Leop. Carol, xxvn, pp. 331-424, 7 pis. Graeffe, E., i860. Beobachtungen iiber Radiaten und Wiirmer. Denkschr. schweiz. naturf. Ges. VII, pp. 1-59, 10 pis. Gronberg, G., 1898. Beitrage zur Kenntniss der Gattung Tubularia. Zool. Jb. Anat. XI, pp. 61-76, 2 pis. Hadzi, J., 1944. Turbelarijska Teorija Knidarijev. Dela mat. prir. Akad. Ljubljan. Ill, pp. 1-239. Haeckel, E., 1869a. Zur Entwickelungsgeschichte der Siphonophoren. Natuurk. Verh. Prov. Utrechtsch Genoots. i, VI, pp. 1-120, 14 pis. 18696. tJber Arbeit stheilung in Natur und Menschenleben, pp. 1-40, 1 pi., 18 text-figs. Berlin. 1888a. System der Siphonophoren . . .. Jena. Z. naturw. xxn, pp. 1-46. - 18886. Report on the Siphonophorae. Rep. sci. results H.M.S. 'Challenger' Zool. xxvm, pp. 1-380, 50 pis. Hartlaub, C, 1907. Craspedote Medusen, 1 Teil, 1 Lief., Codoniden und Cladonemiden. Nordisches Plankton, Lief. 6, xn, pp. 1-135, 126 text-figs. Herdman, H. F. P., 1953. The deep scattering layer in the sea: Association with density layering. Nature, Lond., clxxii, p. 276. Huxley, T. H., 1859. The Oceanic Hydrozoa. . . Voyage of H.M.S. 'Rattlesnake', pp. 1-143, 12 pis. London: Ray Society. Hyman, L. H., 1940. The Invertebrates: Protozoa through Ctenophora, 1st ed., pp. 1-726, 221 text-figs. New York and London : McGraw-Hill Book Company, Inc. Jesperson, P., 1935. Quantitative Investigations on the distribution of Macroplankton in different Oceanic Regions. Dana-Report, No. 7, pp. 1-44, 28 text-figs. Kawamura, T., 1910. ' Boxunira' and ' Katsuwo no Eboshi'. Rhizophysa and Physalia. Zool. Mag., Tokyo, XXII, pp. 445-54, 1 pi. and 5 text-figs. 1911a. Genus Agalma and Crystallomia. Zool. Mag., Tokyo, xxm, pt. 1, No. 267, pp. 1-10, 1 pi. 19116. ' Shidarezakura Kurage' and ' Nagayoraku Kurage' Cupulita picta Metschnikoff and Agalmopsis elegans Sars. Zool. Mag., Tokyo, xxin, pp. 359-63, 1 pi. 1915. Calycophorae III. Zool. Mag., Tokyo, xxvn, pp. 317-24, 1 pi. Keferstein, W. & Ehlers, E., 1861. Beobachtungen iiber die Siphonophoren von Neapel und Messina. Zoologische Beitrage gesammelt im Winter 1 859-1 860 in Neapel und Messina, pp. 1-34, 5 pis. Leipzig: Wilhelm Engelmann. Kolliker, A., 1853a. Bericht iiber einige im Herbste 1852 in Messina angestellte vergleichendanatomische Untersuchungtn, Ouallen. Z. wiss. Zool. iv, pp. 306-15. 18536. Die Schwimmpolypen oder Siphonophoren von Messina, pp. 1-96, 12 pis. Leipzig: Engelmann. Korotneff, A., 1884. Zur Histologie der Siphonophoren. Mitt. zool. Sta. Neapel. v, pp. 229-88, 6 pis. Kramp, P. L., 1942. The Godthaab Expedition 1928. Siphonophora. Medd. Gronland, lxxx, 8, pp. 3-24, 5 text-figs. - 1949. Origin of the Hydroid Family Corymorphidae. Vidensk. Medd. dansk. naturh. Foren. Kbh. cxi, pp. 183-215, 9 text-figs. Lamarck, J. B. P. A. de M. de, 1801. Systeme des Animaux sans vertebres. Paris, 432 pp. Leloup, E., 1929. Recherches sur Tanatomie et le de'veloppement de Velella spirans Forsk. Arch. Biol., Paris, xxxix, pp. 397-478, 3 pis., 6 text-figs. 1932a. Contribution a la Repartition des Siphonophores Calycophorides. Bull. Mus. Hist. nat. Belg. vm, No. 11, pp. 1-30, 3 text-figs. 19326. L'eudoxie d'un Siphonophore Calycophoride rare, le Nectopyramis thetis Bigelow. Bull. Mus. Hist. nat. Belg. vm, No. 3, pp. 1-8, 5 text-figs. 1933- Siphonophores calycophorides provenant des campagnes du Prince Albert 1" de Monaco. Result. Camp. sci. Monaco, Fasc. 87, pp. 1-67, 1 pi. i934tf- Siphonophores Calycophorides de I'Ocean Atlantique Tropical et Austral. Bull. Mus. Hist. nat. Belg. x, No. 6, pp. 1-87, 15 text-figs. 19346. Siphonophores de Madras (hides Anglaises). Bull. Mus. Hist. nat. Belg. x, No. 9, pp. 1-5. ■ 1941. A propos des Siphonophores du genre Anthophysa Brandt, 1835. Bull. Mus. Hist. nat. Belg. xvn, No. 47, pp. 1-7, 1 pi., 2 text-figs. Leloup, E. & Hentschel, E., 1935. Die Verbreitung der calycophoren Siphonophoren im Siidatlantischen Ozean. Wiss. Ergebn. dtsch. atlant. Exped. 'Meteor', XII, 2, pp. 1-3 1. !6o DISCOVERY REPORTS Lens, A. D. & Riemsdijk, T. van, 1908. The Siphonophora of the Siboga Expedition. Siboga Exped. ix, pp. 1-130, 24 pis., 52 text-figs. Lesson, R. P., 1826. Voyage autour du monde. . .sur la Corvette de sa Mcqesti, 'La Coquille', pendant les annees 1822, 1823, 1824 et 1825 Atlas, Zoophytes, 16 pis. 1838. Voyage Autour du Monde, etc. Zoologie, 2, part 2, 2nd division, pp. 155. 1843. Acalephes. Histoire naturelle des zoophytes, pp. 1-596, 12 pis. Paris: Librairie Encyclopedique de Roret. Lesueur, C. A. & Petit, N., 1807-11. Voyage de decouvertes aux Terres Australes. Atlas, 40 pis. Paris. Leuckart, R., 1853. Zoologische Untersuchungen. 1. Die Siphonophoren, pp. 1-95, 3 pis. 1854. Zur naheren Kenntnis der Siphonophoren von Nizza. Arch. Naturgesch. Jahrg. xx, 1, pp. 249-377, 3 pis. Linne, C, 1758. Systema naturae.. . .Ed. x, 1, 823 pp. McCrady, J., 1857. Gymnopthahnata of Charleston Harbor. Proc. Elliott Soc. Sci. I, pp. 103-221, 5 pis. Mackintosh, N. A., 1934. Distribution of the Macroplankton in the Atlantic sector of the Antarctic. Discovery Reports, ix, pp. 65-160, 48 text-figs. Metchnikoff, E., 1870. Contributions to the Knowledge of Siphonophores and Medusae (Russian). Mem. Soc. Amis Sci. nat. Moscou, vin, pp. 295-370, 6 pis. 1874. Studien iiber die Entwickelung der Medusen und Siphonophoren. Z. wiss. Zool. xxiv, pp. 15-83, 11 pis. 1886. Medusologische Mittheilungen. Arb. zool. Inst. Univ. Wien. VI, pp. 237-66, 2 pis. Moore, H. B., 19/3. Plankton of the Florida Current. II. Siphonophora. Bull. mar. Sci. Gulf & Caribbean, 11, No. 4, pp. 559- 73, 9 text-figs. Moser, F., 1917. Die Siphonophoren der Adria und Hire Beziehungen zu denen des Weltmeeres. S.B. Akad. Wiss. Wien, Math.- nat. Klasse, Abt. 1, cxxvi, Heft 9, pp. 703-63, 4 pis., 1 text-fig. and 1 chart. 1924a. Die larvalen Verhaltnisse der Siphonophoren in neuer Beleuchtung. Zoologica, xxvin, 1 Lief., Heft 73, pp. 1-52, 5 pis., 35 text-figs. 19246. Siphonophora in: Kukenthal & Krumbach. Handbuch der Zoologie, 1, pp. 485-521, text-figs. 462-511. I925. Die Siphonophoren der Deutschen Sudpolar Expedition, 1901-3. Dtsch. Sudpol. Exped. xvii, Zool. 9, pp. 1-541, 36 pis., 61 text-figs. Muller, O. F., 1776. Beschreibung zweier Medusen. Beschaft. Ges. naturf. Fr. Berlin, 11, pp. 290-8, 1 pi. Okada, Yok, 1932. De'veloppement post-embryonnaire de la Physalie pacifique. Mem. Coll. Sci. Kyoto, Ser. B, 8, No. 1, Art. 1, pp. 1-25, 11 text-figs., 1 pi. Otto, A. W., 1823. Beschreibung einiger neuen Mollusken und Zoophyten. Nova Acta Leop. Carol, xi, pp. 273-314, 5 pis. Palombi, A., 1937. II ciclo biologico di Lepocreadium album Stossich sperimentalmente realizzato. Osservazioni etologiche e considerazioni sistematiche sulla Cercaria setifera (nonjoh. Muller) Monticelli. Riv. Parassit. 1, 1, pp. 1-12, 1 pi., 2 text-figs. Peron, F., 1807-16. Voyage de decouvertes aux terres Australes, execute. . .pendant 1800-4. . .et redige par M. F. Peron. . ., 2 torn, [et Atlas]. Paris. Quoy, J. R. C. & Gaimard, J. P., 1827. Observations zoologiques faites a bord de ' V Astrolabe', en mai 1826, dans le detroit de Gibraltar. Ann. Sci. nat. x, pp. 5-21, 172-93, 6 pis. 833. Voyage de decouvertes de ' V Astrolabe \..deM. J. Dumont U Urville. Zool. IV, pp. 1-390, Atlas, Zoophytes, 1," Pis. 1-26. Paris. Rees, W. J., 1937. The development of the Hydroid Corymorpha nutans M. Sars from the egg. J. Mar. biol. Ass. U.K. xxi, No. 2, pp. 743-6, 2 text-figs. Rees, W. J. & Russell, F. S., 1937. On rearing the hydroids of certain medusae, with an account of the methods used. J. Mar. biol. Ass. U.K. xxii, pp. 61-82, 12 text-figs. Richter, W., 1907. Die Entwicklung der Gonophoren einiger Siphonophoren. Z. wiss. Zool. lxxxvi, pp. 557-618, 3 pis. 13 text-figs. Ritchie, G. S., 1952. H.M.S. Challenger's Investigations in the Pacific Ocean. J. Inst. Navig. v, No. 3, 251-61. Russell, F. S., 1934. On the occurrence of the Siphonophores Muggiaea atlantica Cunningham and Muggiaea kochi (Will) in the English Channel. J. Mar. biol. Ass. U.K. xix, pp. 555-8. Russell, F. S. & Colman, J. S., 1935. The Zooplankton. IV. The occurrence and seasonal distribution of the Tunicata, Mollusca and Coelenterata (Siphonophora). Sci. Rep. Gr. Barrier Reef Exped. 11, 7, pp. 203-76, 30 text-figs. Sars, M., 1846. Fauna littoralis Norwegiae, 1, pp. 1-94, 10 pis. Christiania. Schneider, K. C, 1896. Mittheilungen iiber Siphonophoren. II. Grundriss der Organisation der Siphonophoren. Zool. Jb. Abt. anat. ix, pp. 571-664, 3 pis., 32 text-figs. 1898. Mittheilungen iiber Siphonophoren. III. Systematische und andere Bemerhungen. Zool. Anz. XXI, pp. 51-7, 73~95> 114-33, I53"73' 185-200. Sears, M., 1950. Notes on Siphonophores. 1. Siphonophores from the Marshall Islands. J. Mar. Res. ix, 1, pp. 1-16, 2 text-figs. 1952. Notes on Siphonophores. 3. Nectopyramis spinosa n.sp. Breviora, No. 3, pp. 1-4, 3 text-figs. REFERENCES 161 Sears, M., 1953. Notes on Siphonophores. 2. A revision of the Abylinae. Bull. Mus. comp. Zool. Harv. 109, No. i, pp. 1-119, 29 text-figs. Sewell, R. B. Seymour, 1935. Introduction and list of stations. Sci. Rep. Murray Exped. No. 1, pp. 1-41, 1 chart. 1948. The free-swimming Planktonic Copepoda: Geographical distribution. Sci. Rep. Murray Exped. vm, No. 3, pp. 317-592, 24 text-figs. 2 charts. Stechow, E., 1921. Neue Genera und Species von Hydrozoen und anderen Evertebraten. Arch. Naturgesch. A. 87, 3, 248-65. Sverdrup, H. U., Johnson, M. W. & Fleming, R. H., 1946. The Oceans, their physics, chemistry, and general biology, pp. 1-1087, 265 text-figs., 7 charts. New York: Prentice-Hall, Inc. Totton, A. K., 1932. Siphonophora. Sci. Rep. Gr. Barrier Reef Exped. iv, No. 10, pp. 317-74, 36 text-figs. 1936. Plankton of the Bermuda Oceanographic Expeditions. VII. Siphonophora taken during the year 193 1. Zoologica, N.Y., xxi, 4, pp. 231-40. 1941. New species of the Siphonophoran Genus Lensia, 1941. Ann. Mag. nat. Hist. (11), vm, pp. 145-68, 29 text-figs. Vogt, C, 1 85 1. Zoologische Brief e, etc. I, pp. 138-41, 4 text-figs. Frankfurt. 1852. tjber die Siphonophoren. Z. wiss. Zool. ill, pp. 522-5, 1 pi. 1854. Recherches sur les animaux inferieures de la Mediterranee. 1. Sur les Siphonophores de la Mer de Nice. Mem. Inst. nat. genev. 1, pp. 1-164, 21 pis. Weismann, A., 1883. Die Entstehung der Sexualzellen bei den Hydromedusen. Zugleich ein Beitrag zur Kenntnis des Baues und der Lebenserscheinungen dieser Gruppe, pp. iii-xiii, 1-295, 24 P^s-> 2I text-figs. Williams, C. B., 1951. A note on the relative sizes of genera in the classification of animals and plants. Proc. Linn. Soc. Lond. clxii, 2, pp. 171-5, 2 text-figs. Woltereck, R., 1904. Vber die Entwicklung der Velella aus einer in der Tiefe vorkommenden Larve. Zool. Jh. Suppl.-Band VII, PP- 347-72. 3 Pls-> 6 text-figs. 1905. Beitrdge zur Ontogenie und Ableitung des Siphonophorenstocks, mit einem Anhang zur Entwicklungsphysiologie der Agalmiden. Z. wiss. Zool. lxxxii, pp. 611-37, 2I text-figs. INDEX TO GENERA AND SPECIES Abyla, Abylopsis, Agalma, Amphicaryon, amygdalina, Crystallophyes, angusta, sp.n. Galetta, Sulculeolaria, antarcticus, Marrus, gen.n., sp.n., appendiculata, Chelophyes, Archisoma = Nectopyramis arctica, Dimophyes, Athorybia, atlantica, Muggiaea, australis, Galetta, bargmannae, sp.n., Muggiaea, Bargmannia, gen.n., bijuga, Nanomia, biloba, Galetta, Sulculeolaria, cara, Nanomia carina, Abyla, Ceratocymba, challengeri, sp.n., Lensia, chamissonis, Diphyes, Chelophyes, Chuniphyes, Clausophyes, conoidea, Lensia, contorta, Chelophyes, Cordagalma, cordiformis, Cordagalma, cossack, Lensia, 143-147 crystallina, Thalassophyes, 157 Crystallophyes, 61-69 92 delsmani, sp.n., Muggiaea, 1.34 dentata, Ceratocymba, 108 Dimophyes, 55 diomedeae, Nectopyramis, 127 Diphyes, 123 edwardsii, Forskalia, 37 elegans, Agalma, 120 elongata, Bargmannia, gen.n., sp.n., 101 ernesti, sp.n., Amphicaryon, Eudoxia, 121 69 Forskalia, 52 104 galeata, Clausophyes, Galetta, 53-54 glabra, Vogtia, 144-150 152 haeckeli, Abyla, 1 1 1 Heteropyramis, 130 Hippopodius, 127-130 hippopus, Hippopodius, 131 hotspur, Lensia, 133 hydrostatica, Physophora, 114 130 ingeborgae, Abyla, 69 69 leloupi, sp.n., Lensia, 112 Lensia, 141 134 123 152 123 127 83 130 7i 61 69 94 118 70.73 i34 100- -109 100 151 137- -141 99' -100 99 no 42 152 113 no l62 macra, sp.n., Eudoxia, maculata, Heteropyramis, Maresearsia, gen.n., Marrus, gen.n., melo, Melophysa, Melophysa, meteori, Lensia, moserae, sp.n., Chuniphyes, Muggiaea, multicristata, Lensia, multidentata, Chuniphyes, Nanomia, natans, Nectopyramis, Nectopyramis, okenii, Agalma, orthocanna, Marrus orthocannoides, Marrus, gen.n., sp.n., ovata, Clausophyes, peltifera, Amphicaryon, Physophora, plicata, Rosacea, Porpita, porpita, Porpita, praeclara, Maresearsia, gen.n., sp.n., DISCOVERY REPORTS 118 137 97 55 41 40 117 120-123 113 I31 52-55 82 77 64 59-61 59 i33 69 42-46 89 33 33 97 quadrivalvis, Sulculeolaria, reticulata, sp.n., Lensia, Rosacea, rosacea, Athorybia, rubra, Stephanomia, schmidti, Abyla, serrata, Vogtia, spinosa, Nectopyramis, Stephanomia, subtilis, Lensia, subtiloides, Lensia, Sulculeolaria, tetragona, Abylopsis, Thalassophyes, thetis, Nectopyramis, tottoni, Abyla, trigona, Abyla, turgida, Galetta, Sulculeolaria, Velella, velella, Velella, Vogtia, 109 118 87 38 47 147 73-77, 100 86 46 114 112 100 J55 141 78 151 144-150 107 34 34 100 PLATE I Athorybia rosacea Forskal Villefranche, o metres, 9 May 1950 Unfinished sketches of the living animal by Mile Bauer, by permission of Professor Portmann. A. Batteries of nematocysts on the tips of the palpons. B. A single palpon with its pink tip. C. Bracts. Artist's notes: C, bract, fattish, delicate; D., (apex of pneumato- cyst) brownish stippling with green spots; E.t view from above (binocular); F.G., naked eye sketches; F., (pneumatocysts) metallic; G., (centre) opaque. All the rest very transparent. DISCOVERY REPORTS, VOL XXVII PLATE I • * ^ -< u~~ ■ " !"■'•"' , *,, i/'iCt~AT>~^i+ 9. f~us so 2 ■ \ . ' PLATE II Athorybia rosacea Forskal Water-colour sketch of a living animal, made by Sydney Parkinson in 1768, during Cook's first voyage. The specimen was taken two or three hundred miles south of the Cape Verde Is. The general colour used is cerise : outside the pneumatophore, reddish orange. The dark pigment outlining the topmost bracts and the lines of bracteal nematocysts was originally white. The white area on the outer edge of the pneumatophore represents the giant cells of the gas-gland. Nat. size right. Animal x 9. Tentacle x 20. Plates II and III reproduced by kind permission of the Trustees of the British Museum (Natural History). DISCOVERY REPORTS, VOL XXVII PLATE II \ *&£ I * ./ J ffl/e/iucL - rut/am . VWw luJw&vfiJiltcL. AJau\^aS OnH&)*%, Pfe& PLATE IV Figs, i, 3, Chelophyes appendiculata. From the surface plankton at Villefranche, x 25. The left hydroecial wing of the separated gonophore lies to the left under the right wing (out of focus). Fig. 2, Vogtia glabra. Larva taken at Villefranche by Dr Tregouboff, x 20. For explanation see legend to Text-fig. 33. Figs. 4, 5, 8, Forskalia edwardsii. Species with the yellow spot on the velum, showing the helically coiled stem; photographed alive in the laboratory of the Station Zoologique at Villefranche. x |, x \, x\. Figs. 6, 7, Rosacea cymbiformis. Photographed alive in the laboratory of the Station Zoologique at Villefranche to show the extended stem and the eudoxids. x $. DISCOVERY REPORTS, VOL XXVII PLATE i r L i-i r z, / n 3 u / PLATE IV Figs, i, 3, Chelophyes appendiculata. From the surface plankton at Villefranche, X25. The left hydroecial wing of the separated gonophore lies to the left under the right wing (out of focus). Fig. 2, Vogtia glabra. Larva taken at Villefranche by Dr Tregouboff, x 20. For explanation see legend to Text-fig. 33. Figs. 4, 5, 8, Forskalia edwardsii. Species with the yellow spot on the ^relum, showing the helically coiled stem; photographed alive in the laboratory of the Station Zoologique at Villefranche. x %, x \, x\. Tigs. 6, 7, Rosacea cymbiformis. Photographed alive in the laboratory of the Station Zoologique at Villefranche to show the extended stem id the eudoxids. x J. L J DISCOVERY REPORTS, VOL XXVII PLATE IV PLATE V Fig. i. Nectopyramis spinosa Sears, the nectophore with obsolescent nectosac, x 12. 'Discovery II' St. 11 79, 500-250 m. Figs. 2-3. Nectopyramis diomedeae, X4-6, Bermuda, 1646 m. Beebe. See also Text-fig. 39. DISCOVERY REPORTS, VOL XXVII PLATE V %. PLATE VI Maresearsia praeclara, gen.n., sp.n., larger nectophore, gastrozooids and eudoxids (still attached) of holotype specimen from 'Discovery II' Station 2636, 950-550 m. x 9-5. DISCOVERY REPORTS, VOL XXVII PLATE VI PLATE VII Maresearsia praeclara gen.n., sp.n. Fig. i. Gastrozooids and cormidia of holotype specimen, x 23. Fig. 2. An eudoxid of holotype specimen showing first gonophore partly severed from bract, together with a bud of the second gonophore. x 17. Fig. 3. A younger eudoxid, bract on the left, gonophore and bud on the right, borne on a peculiar stalk, x 17. Fig. 4. The bract of a young eudoxid. x 17. DISCOVERY REPORTS, VOL XXVII PLATE VII >^^^C- ■* \ £ I "\ PLATE VIII Abyla schmidti Sears, whole animal. 'Discovery II' St. 1581 Fig. 1 . View showing ventral side of the anterior nectophore and right side of the posterior nectophore. x 7. Fig. 2. End-on view, showing the ostium of the nectosac of the posterior nectophore, and its surrounding teeth, x 5-2. Fig. 3. View showing the dorsal side of the anterior nectophore and the left side of the posterior nectophore. x 7. Fig. 4. View showing the right side of the anterior nectophore and the postero-lateral aspect of the posterior nectophore. x 7. Fig. 5. Apical view of the anterior nectophore, with the large left-hand, hydroecial wing of the posterior nectophore in the background, x 7. Fig. 6. View showing the left side of the anterior nectophore and the right side of the posterior nectophore. x 7. DISCOVERY REPORTS, VOL XXVII PLATE VIII PLATE IX Abyla tottoni Sears, whole animal, x 4. 'Discovery II' St. 1178 Fig. 1. Apical view of the anterior nectophore, with the large left hydroecial wing in the right background. Fig. 2. View showing the ventral side of the anterior nectophore and the right side of the posterior nectophore. The left dorsal ridge is on the right running down to the median dorsal tooth. Fig. 3. View showing the dorsal side of the anterior nectophore and the left side of the posterior nectophore. The strong teeth on the 'comb' (or inturned edge of the right hydroecial wing) can be seen lying under the larger left hydroecial wing. Fig. 4. Basal view of the animal, showing the ostium of the nectosac of the posterior nectophore and its surrounding teeth. The large left hydroecial wing lies on the left, and on the right the left dorsal ridge can be seen joining the median dorsal tooth. The right dorsal ridge lies above, and parts of the anterior nectophore are seen in the background. Inset: the ostial region, showing (white pecked lines) the actual junction of the left ridge, R.l, with the median tooth, To.d., and its hypothetical former junction with the left tooth, To.l. DISCOVERY REPORTS, VOL XXVII PLATE r R. r. R. 1. Ms. W. hyd. 1. 1 Abor. D. 1_ N. ant. W. hyd. 1. Cmb. \ R. r. To. r. To. d. W. hyd. r. W. hyd. 1. r ** .a •bvrf .W .8n PLATE IX ^iy/a (oHom Sears, whole animal, x 4. 'Discovery II' St. 11 78 Fig. 1. Apical view ^f the anterior nectophore, with the large left ► hydroecial wing in the right background. .iodA Fig. 2. View showing the ventral side of the anterior nectophore and tjhe right side of the posterior nectophore. The left dorsal ridge is^fcn the right running down to the median dorsal tooth. .1 .SI pig. 3. View showing the dorsal side of the anterior "netta^hore and the left side of the posterior nectophore. The strong teeth on the 'comb' (or inturned edge of the right hydroecial wing) can be seen lying under the larger left hydroecial wing. Fig. 4. Basal view of the animal, showing the ostium of the nectosac of the posterior nectophore and its surrounding teeth. The large left ,b .oT hydroecial wing lies on the left, and on the right the left dorsal ridge can be seen joining the median dorsal tooth. The right dorsal ridge lies above, and parts of the anterior nectophore are seen in the background. Inset: the p^} nggion, showing (white pecked lines) the actual junction of the left ridge, R.I., with the median tooth, To.d., and its hypothetical former junction with the left tooth, To.l. .1 .n .i.Si .dmO .Si .b .oT ' .1 .oT .rn.M .1.0T .b.oT^--^ J J ■***•* .i.b.rf.W I .bV 1 .1 .Si _J DISCOVERY REPORTS, VOL XXVII PLATE IX PLATE X Ceratocymba dentata, whole animal. 'Discovery II' St. 2635 Fig. i . View showing the right side of the anterior nectophore and the ventro-left lateral aspect of the posterior nectophore, showing the remarkably long cylindrical nectosac of the posterior nectophore, and the obsolescent one of the anterior, x 2. Fig. 2. View showing the dorsal side of the anterior nectophore and the right side of the posterior nectophore. x 2. Fig. 3. View showing the left side of the anterior nectophore and the ventro-right lateral aspect of the posterior nectophore, showing the elongated left ventral tooth. Eudoxid groups can be seen in the hydroecium of figs. 1-3. x 2. Fig. 4. Basal view of the animal showing the ostium of the nectosac of the posterior nectophore and its surrounding teeth. In the back- ground parts of the anterior nectophore are seen. In the foreground above can be seen the left dorsal ridge of the posterior nectophore running down to join the left dorsal tooth, and on the left the right dorsal ridge running down to the right dorsal tooth, leaving the median dorsal tooth free. Below are the bases of the two ventral hydroecial wings, x 4. Fig. 5. Basal part of the posterior nectophore to show the oblique flattened facets, the left-hand one bearing a double fringe of spines, at the base of the ventral hydroecial wings, x 3-5. DISCOVERY REPORTS, VOL XXVII PLATE X PLATE XI Ceratocymba dentata, an eudoxid. 'Discovery II' St. 286 Figs. 1-2. Detached 3 and $ gonophores from the upper (attached) side. Their respective median dorsal teeth are facing each other. x4. Fig. 3 . Bract and gastrozooid of the eudoxid from the outer side, showing on the left the incomplete 'left' ridge. The complete ridge is the misplaced 'right' one. A third gonophore (?) is still attached, 1 1 mm. in length, x 2-8. Figs. 4-5. Detached $ and ? gonophores viewed from their ventral (opposed) sides, showing the large single teeth at the basal end of the left (and vice versa right) ventral hydroecial wings, also the unequal right and left ventral teeth, x 4. Fig. 6. Stephanomia rubra. Part of a denuded cormidium to show how the nectophores (one young bud visible) are attached to one side (nectosome, on the right) and the appendages to the opposite side (part of the siphosome, on the left), x 17. DISCOVERY REPORTS, VOL XXVII PLATE XI m PLATE XII Abyla trigona Quoy and Gaimard, syntypes (Musee d'Histoire Naturelle, Paris) Figs. 1-4. One of the anterior nectophores. Figs. 5-7. A second anterior nectophore. Figs. 8-10. A posterior nectophore. Fig. 8 by transmitted light, figs. 9-10 by reflected light. DISCOVERY REPORTS, VOL XXVIT PLATE XII w '* r [Discovery Reports. Vol. XXVII, pp. 163-200, August, 1954.] THE PELAGIC MOLLUSCA OF THE BENGUELA CURRENT PART I. FIRST SURVEY, R.R.S. 'WILLIAM SCORESBY', MARCH 1950 WITH AN ACCOUNT OF THE REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF LIMACINA BULIMOIDES By J. E. MORTON, Ph.D. CONTENTS Introduction page 165 Acknowledgements 165 I. Systematic ... 166 Ianthina ianthina .166 Ianthina globosa . . . ■ ■ ■ ■ ■ ■ ■ .167 Atlanta peroni . 169 Limacina inflata ........... 170 Limacina bulimoides . . . . . . ■ ■ ■ .171 Ecology — Depth distribution 172 Diacria trispinosa 175 Euclio pyramidata . . . 176 Cavolinia inflexa 176 Cymbulia peroni . . . . . ■ ■ • ■ • .176 Thliptodon diaphanus 177 Pneumodermopsis paucidens 178 Lamellibranch larvae •• . • .181 Cephalopod larvae 182 II. Reproductive system and sexual succession of Limacina bulimoides 183 (i) The Gonad and Genital Tract 184 (ii) The Genital Ducts 185 (iii) Sexual Succession 194 (iv) Sex in Gastropoda 196 Summary 197 References 198 T THE PELAGIC MOLLUSCA OF THE BENGUELA CURRENT PART I. FIRST SURVEY, R.R.S. 'WILLIAM SCORESBY', MARCH 1950 WITH AN ACCOUNT OF THE REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF LIMACINA BULIMOIDES By J. E. Morton, Ph.D. (Text-figs. 1-17) INTRODUCTION he Mollusca collected by the R.R.S. 'William Scoresby' during the first of her two surveys of the Benguela Current off the coast of South West Africa (March 1950— Stations WS 964-1002) form a small collection, representing in all eleven species of pelagic gastropods, one larval gastropod, two lamellibranch post-larvae (one species in considerable numbers), and some cephalopod larvae. Of the gastropods the ' pteropods ' are by far the largest group and they yielded two species of numeri- cal importance, the gymnosome Pneumodermopsis paucidens, and— in special abundance at certain points— the thecosome Limacina bulimoides. The latter was taken at five offshore stations lying to the west of the continental shelf, and at two stations was the dominant zooplankton species. The material secured gives valuable information on the depth distribution, reproductive activity and sexual succes- sion of this pteropod. Further, the fixation of the Limacina in 5 % formalin was in many cases good enough to enable detailed morphological work to be carried out on this species, and the second section of this report has been devoted to an account of reproduction in L. bulimoides, in which the genital system has been described in detail and the sexual phenomena of this mollusc discussed. Lamellibranch larvae occurred close inshore at Walfisch Bay, at the inshore station WS 989, and at the mouth of the Orange River. The cephalopod larvae, which were four in number, have kindly been examined by Dr W. J. Rees of the British Museum (Natural History) and I am grateful to him for allowing me to publish his comments on them. In addition to material from the South West African coast, one haul of Ianthina taken with a surface net in north equatorial waters off the north coast of the Gulf of Guinea is conveniently referred to here; while one of the cephalopods (from WS 1091) belonging to second survey material is included for comparative purposes in this report. For full hydrological data, reference should be made to the detailed Station List already issued for the 1950 cruise of the 'William Scoresby' {Discovery Reports, xxvi, 211-58). Preliminary accounts of the hydrological results of the expedition and of the distribution of phytoplankton in the South West African area have already been published (Currie, 1953 ; Hart, 1953). A table of positions and dates together with surface temperatures and salinities from the stations which yielded mollusca during the first survey is given on p. 166. ACKNOWLEDGEMENTS I am indebted to Dr N. A. Mackintosh, C.B.E., of the National Institute of Oceanography far per- mission to report upon this material. I must also record my appreciation of the kindness of Dr T. J. Hart, at whose suggestion the work was placed in my hands, and who was personally responsible for the heavy task of sorting and classifying the molluscan material from the general bulk samples. l66 DISCOVERY REPORTS I am also grateful to Dr V. Fretter for a great deal of helpful discussion on reproduction in the Gastropoda, and to Dr H. E. Bargmann and Mr T. J. Evans for their careful and kindly criticism of the manuscript. The greater part of this paper was completed while I was the holder of the University of London Table at the Plymouth Laboratory of the Marine Biological Association, and I have to acknowledge the kindness throughout of Mr F. S. Russell, F.R.S., and others of the laboratory scientific staff. Table i . Positions and hydrological data of stations yielding Mollusca* Surface Surface Station Position Date temp. °C. salinity " /bo WS976 From 220 50' S, n°38'E to 220 48' S, ii° 35' E 6. iii. 50 19-31 35'32 WS978 22°28'S, I2°42'E 7. iii. 50 18-16 35-23 WS980 22° 44' S, I4° 08' E 8. iii. 50 16-97 35-15 WS981 22° 44' S, I4° 20' E 8. iii. 50 16-69 35-J5 WS986 From 220 15' S, 130 06' E to 25°i3'S, i3°o8'E 10. iii. 50 17-10 35-oo WS987 25°i3'S, i3°43'E 10. iii. 50 I571 34-97 WS988 250 12' S, 140 22' E 11. iii. 50 16-26 35-o° WS989 25° 11' S, i4° 39' E 11. iii. 50 14-82 35-03 WS996 From28°4i'S, 130 25' E to 280 39' S, 13° 23' E 12. iii. 50 20-38 35-iS WS997 From 280 40' S, 140 06' E to 280 37' S, 140 08' E 13. iii. 50 20-18 35-20 WS998 280 40' S, 140 43' E 13. iii. 50 18-36 35-°! WS999 From 280 38' S, 140 59' E to 28°3i'S, i5°oo'E 13. iii. 50 lS-22 35-0° WS 1002 28°4o'S, 1 6° 14' E 14. iii. 50 14-23 33-°4 WS 1091 200 38' S, 12° 18' E 8. x. So 15-6 ~ * See Discovery Reports, supr. cit. I. SYSTEMATIC Class GASTROPODA Subclass Prosobranchia Order MESOGASTROPODA Superfamily Ptenoglossa Family Ianthinidae Ianthina ianthina (Linnaeus) 1767 Helix ianthina Linn., Syst. Nat. ed. 12, 1246. 1798 Ianthina violacea Bolten, Mus. Bolt. 75. 1815 Ianthina fragilis Lmk., Anim. sans Vertebr. (earlier ed.). 1835 Ianthina communis Lmk., Anim. sans Vertebr. (ed. Desch.), 9, 4. 1850 Ianthina planispirata A. Ad. & Rve., Zool. Voy. H.M.S. 'Samarang', p. 54, pi. xi, fig. 10. 1853 Ianthina britannica Leach in MS., Forbes and Hanley, iv, 260, pi. 133, fig. 1. 1858 Ianthina balteata Rve., Conch. Icon, xi, pi. 3, fig. 11. 1858 Ianthina depressa Rve., Conch. Icon, xi, pi. 3, fig. 14a, b. 1858 Ianthina trochoidea Rve., Conch. Icon, xi, pi. 2, fig. 10a, b. 1858 Ianthina roseola Rve., Conch. Icon, xi, pi. 1, fig. ia, b. 1858 Ianthina africana Rve., Conch. Icon. XI, pi. 2, fig. 8, 8a. SYSTEMATIC 167 Ianthina globosa Swainson 1822 Ianthina globosa Swainson, Zool. Illustr. II, pi. 85. 1828 Ianthina nitens Menke, Synopsis, 84. 1840 Ianthina rotundata Leach., MS. Dillwyn, Contrib. towards a History of Swansea, 59. 1840 Ianthina pallida Thompson, Ann. Nat. Hist, v, 96. 1844 Ianthina patula Phillipi., Moll. Sicil. II, 224, pi. 28, fig. 14. 1858 Ianthina striolata Rve., Conch. Icon. XI, pi. 5, fig. 24^, b. 1858 Ianthina iricolor Rve., Conch. Icon, xi, pi. 5, fig. 23 a, b. Two species of Ianthina are represented in this collection from the first survey, both occurring together at each of two places. On the second survey, specimens occurred at WS 1057 and WS 1058 in surface hand-nets. The samples consist of seven /. ianthina (of planispirata-balteata shape and markings) and four /. globosa (mauve-white to pale pinkish, and probably closest in form to Thompson's pallida). From north equatorial waters (o-6° 10' N, 1 1° 25' W), on the northern edge of the Guinea Current, came another sample, four /. ianthina — of the typical ianthina form — and two large typical /. globosa. The Biological Log-book records the presence of the siphonphore Porpita, upon which the Ianthina were feeding. One /. ianthina carried an egg-raft. These two species are very distinct, wide-ranging over most seas, and would seem very frequently to occur side by side. The systematics of the genus Ianthina are at present in course of full revision by Dr Dan Laursen, who has worked on the extensive Dana material. I have not seen these results, and have therefore in the present account merely assigned the two Benguela species according to the currently accepted view of Iredale (1910), who regarded Ianthina as containing four widely distributed species. The first is ianthina (Linn.) in which Iredale groups planispirata, trochoidea, britannica and depressa. Oliver (19 1 5) considered violacea Bolten to include both communis and balteata, and in addition to these species I have examined roseola, africana and fragilis, as represented in the Winckworth Collection at the British Museum, and would tentatively place all these forms in synonymy with ianthina. In shell features, the distinctive characters of this species are evidently its angled periphery and strongly retracted, sharply triangular apertural sinus. Typical specimens of the form ianthina are paler bluish white, above the periphery, and deeper blue below; the chief colour variation in the species is the development of a white band at the central part of the base of the shell, giving the planispirata- balteata pattern, of pale blue above, deep blue around and below the periphery and white at the base. In shell height, depressa and planispirata are the most flattened, africana and trochoidea the most elevated and conical, distinctly higher than broad. The second species, globosa Swainson, has a much more delicate shell; the whorls are rounded, the body whorl especially large, and bubble-shaped, with an extremely wide trumpet-shaped aperture in the largest adults. There are rather wavy growth lines and these are gently recurved at the periphery, to give a shallow apertural sinus, rounded or obtusely angled. In some forms, especially pallida, there may be low, rounded, longitudinal striae, best developed on the body whorl. The colour does not vary much, and is usually paler blue or mauve, and almost white in the form pallida. The two remaining species recognized by Iredale are exigua and umbilicata. Neither is represented in the Benguela material. I have provided outline sketches (Fig. 1) illustrating the leading shell characters of ianthina and globosa. All consideration of radula features, anatomy or life-history has been omitted. These aspects will be dealt with both by Dr Laursen, and also in a forthcoming paper by Professor Graham, on the 168 DISCOVERY REPORTS results of a study of the animals of British Ianthina and the effects of growth on shell contour. These results will together give a clearer view of the interrelations of the various forms included in the species of Ianthina, and in particular of the influence of life-history on shell form. Fig. I. Sketches of the shells of Ianthina ianthina (L.) (a, b, c) and of /. globosa Swainson (d) to show characteristic range of form, (a) I. ianthina. The typical communis-iorm (Winckworth Collection, Madras), (b) I. ianthina. The planispirata- form, view from above and in side view, (c) I. ianthina. View of the balteata-form, showing the periphery and strong labial sinus characteristic of the species (shell from WS 1057). (d) I. globosa. Three views of shells from the Winckworth Collection (Brit. Mus.). Order MESOGASTROPODA Unidentified prosobranch larval shells The three sets of inshore stations, WS 980 and WS 981, off Walfisch Bay, WS 988 and WS 989 off Sylvia Hill, and WS 1002 at the mouth of the Orange River, each yielded shells of a gastropod veliger larva and of lamellibranch larvae. The gastropods at all three stations were of the same species, which — owing to the retraction or loss of the velar lobes and damage to the shell after preservation— were SYSTEMATIC 169 impossible to identify accurately. The largest shells were of 3-3! whorls, turbinate, with globose, smoothly rounded whorls, the body whorl very large, and the aperture two-thirds the height of the shell, ovoid and entire. There were no special sculptural features, except for the presence in the body whorl of about six well-defined, close-spaced spiral cinguli, encircling the base of the whorl, well below the point of greatest diameter. Numerous earlier stages of various sizes were present. Dimen- sions of the largest individual: height c. 1-5 mm., diameter 1 mm. Superfamily Heteropoda Family Atlantidae Atlanta peroni Lesueur 1817 Atlanti peroni Lesueur, Journ. Ac. Nat. Sci. Philadelphia, 1. 1852 Atlanta rosea Souleyet, Voy. de la Bonite, Zoologie, 11, 373, pi. 19, figs. 1-8. 1949 Atlanta peroni Tesch, Dana Reports, xxxiv, p. 16, fig. 9. Table 2. Occurrence of Atlanta peroni Station Time Depth (m.) No. of animals WS 996 1858 to 1610 50-0 100-50 500-250 1000-750 1 adult, 22 juvenile 6 juvenile 1 juvenile 1 adult WS997 0105 to 0401 50-0 250-100 500-250 250-0 (extra sample) 3 adults, 1 juvenile 1 adult 1 adult 5 adults WS998 0850 to 0810 175-100 1 adult This is the only heteropod represented in the present collection. It is an abundant and widespread species, the commonest in Atlantic waters. Tesch (1949) discusses its distribution (p. 21). In the North Atlantic it dwindles south of 200 N and appears to avoid equatorial waters. In the Benguela material it was present on the southern line only, at three stations, WS 996, WS 997, WS 998. The largest specimens measured 4-5-5 mm., and there were in several hauls an abundance of juvenile shells, measuring up to 0-7 mm. Subclass Opisthobranchia Order PTEROPODA1 Suborder Thecosomata Superfamily Euthecosomata Family Limacinidae (=Spiratellidae) The continued use of the generic name Limacina in this paper may be held to require justification. Sherborn (1930) accepted Spiratella de Blainville (1817), as displacing Limacina as proposed by 1 The single order Pteropoda is employed here to include both the thecosomatous and gymnosomatous pteropods, in keeping with Thiele's system of classification (1931) which has been adopted throughout this section. It is not possible, however, to regard such an order as a natural grouping. Pelseneer (1887, 1888) conclusively demonstrated the separate origins of the two groups of pteropods, placing the Thecosomata at the end of the bullomorph series and regarding the Gymnosomata as being derived from an aplysiomorph stock. He based his classification of 1906 on this view. The present writer (1954a) has elsewhere pointed out— from a consideration of feeding and digestive mechanisms in the pteropods— that the reversion by Thiele and by Hoffmann (1938) to a single order Pteropoda cannot be defended on strict phylogenetic grounds. i7o DISCOVERY REPORTS Cuvier in Le Regne Animal (1817), apparently on the ground that the latter name was employed in the form ' les Limacines ' and was nowhere in that work explicitly proposed in latinized form. A second pteropod generic name, 'les Pneumodermones', established by Cuvier was saved by the addition of the latinized parenthesis ' (Pneumodermon) ', which followed the usual practice upon introduction of new generic names in Le Regne Animal, but was omitted, evidently inadvertently, in the case of Limacina. Lamarck in 1819 validly proposed Limacina in conformity with Linnaean practice in the Latin form. Unfortunately de Blainville (18 17) had already followed upon Cuvier with the introduc- tion of the term Spiratella, which thus by two years antedates Limacina Lmk. as the first validly pro- posed name for these pteropods. Such a loss of the Cuvierian generic name would be particularly unfortunate, in view of its long-established currency and of the obviousness of the editorial slip in the proposition of the name. The general practice in the nineteenth and early twentieth century was to employ Limacina without question: thus — for example — Pelseneer (1888), Meisenheimer (1905), Bonnevie (1913) and Vayssiere (1915). Following Sherborn (1930), however, two Continental works of first importance have adopted Spiratella — Hoffmann's opisthobranch treatise in Bronn's Tierreiches (1938) and Thiele (193 1) in his authoritative systematic account. In British oceanographic literature the use of Limacina has continued unchanged, and the name was most recently employed by Tesch in his two extensive works (1946, 1948). It is greatly to be doubted whether the progress of systematics is really assisted by excavations from the forgotten past. Here, one might have thought, was a name which could have rested secure and for which a suspension of International Rules might justly have been claimed. It is, however, one of the disadvantages of a rule of law that obedience is necessary even when the consequences appear most irksome. If it does not prove possible to add Limacina to the list of nomina conservanda, Spiratella must unquestionably take precedence of it and come into general use. In the meantime the familiar name Limacina is retained in use without apology in both the systematic and morphological sections of this paper. Limacina inflata (d'Orbigny) 1836 Atlanta inflata d'Orbigny, Voyage dans VAmeriqne meridionale, V, p. 174, pi. xii, 16-19. 1852 Limacina inflata Souleyet, Voyage de la 'Bonite', Zoologie, 11, 216, pi. xiii, figs. 1-10. 1888 Limacina inflata Pelseneer, Challenger Reports, XXIII, ii, p. 17. 1905 Limacina inflata Meisenheimer, Pteropoda, Wiss. Ergebn. D. Tiefsee-Exped. 'Valdivia', ix, 4. 1946 Limacina inflata Tesch, Dana Reports, v (28), p. 8, pi. i, fig. 1. Table 3 . Occurrence of Limacina inflata Station Time Depth (m.) No. of animals WS986 1118 50-0 100-50 250-100 500-250 750-500 4 17 6 1 This is the Limacina of the line to the north of the southernmost line; it occurred here at one off- shore station, WS 986, and locally outnumbered bulimoides, which accompanied it in the 0-50 m. haul only. The species inflata was represented by a relatively small number of animals, the shells having been for the most part dissolved by the formalin preservative. But from the size and structural characteristics of the animal there can be little doubt that inflata is the species to which it belongs. Moreover, this Limacina has been reported several times as accompanying bulimoides in great numbers in the Sargasso Sea, and the warm temperate Atlantic. The depressed spire of the animal is of i\ turns SYSTEMATIC 171 only, and the edge of the mantle is strongly prolonged forward, as in the overarching margin of the shell aperture, most characteristic of this species. The umbilicus was open, and the aperture wide and oval. An operculum was obtained which did not agree very well with the only available figure (Souleyet, loc. cit.), being rather more circular in shape. The depth distribution of inflata at this day-time station corresponds closely to what has also been established (see below) for bulimoides. It shows its greatest day-time concentration at 50-100 m., thinning out markedly below this level and also at the surface. Limacina bulimoides (d'Orbigny) 1836 Atlanta bulimoides d'Orbigny, Voyage dans VAmerique meridionale, v, p. 179, pi. xii, figs. 36-8. 1852 Limacina bulimoides Souleyet, Voyage de la 'Bonite", Zoologie, II, 224, pi. xiii, figs. 35-42. 1888 Limacina bulimoides Pelseneer, Challenger Reports, XXIII, ii, p. 30. 1904 Limacina bulimoides Tesch, Siboga-Exped. Repts. 27, iv, p. 13. 1905 Limacina bulimoides Meisenheimer, Pteropoda, Wiss. Ergebn. D. Tiefsee-Exped. 'Valdivia', ix. 1946 Limacina bulimoides Tesch, Dana Reports, v (28), p. 9, fig. 5, pi. i, fig. 4. Table 4. Occurrence of Limacina bulimoides Station Time Depth (m.) No. of animals WS986 1118 50-0 8 WS996 1858 50-0 93 100-50 !542 250-100 178 500-250 25* 75°-500 90* 1610 1000-750 6 WS997 0410 50-0 1116 100-50 468 250-100 3H2 500-250 J3 750-500 82 0105 1000-750 76 WS998 0850 175-100 80 WS 999 1125 50-0 5 100-50 1 1200 150-100 29 * Many empty. This species is one of the more abundant of tropical and subtropical Atlantic Limacinas, and its general distribution has been discussed by Tesch (1946). Of the high-spired species of the genus, bulimoides is distinguished from retroversa and australis by the oval aperture as distinct from the quadrangular. The umbilicus is constricted as in retroversa, and almost imperceptible as Pelseneer points out. From the subspecies balea, the taller form of retroversa, bulimoides is best distinguished by the smaller number of whorls, which never reaches ten as in the mature balea. The shape of the aperture is not always an easy character to interpret precisely: it may vary within a population with the age of the individual. In typical adult females (1-9 mm.) in the Benguela material the columella is always slightly concave towards the aperture; in earlier stages it may be straight or even convexly curved to the left. A good distinction, however, from all the taller forms of retroversa is the absence of a tentacular lobe on the anterior margin of the fin. From trochiformis , bulimoides is differentiated by its larger size, the presence of an apertural beak and the relatively longer spire. As Tesch says, the 172 DISCOVERY REPORTS shell is tall and turreted and quite transparent, sometimes exhibiting faint longitudinal striae inter- rupted by growth lines on the last whorl. The aperture is drawn out in front into a beak or rostrum, rounded or more or less pointed. In the Benguela material this is especially well marked in the younger stages, while tending to be obscured in the mature females. A distinctive character is the frequent marking of the suture line with light chestnut brown, which may also be present at the columellar edge of the aperture and around the vestigial umbilicus. A figure of a shell (mature female — 1-8 mm.) and of an operculum is given here (Fig. 2). Recorded distribution. According to Tesch this species is very generally distributed in the tropical Atlantic, being nearly always present in large numbers. Its centre of abundance is probably in the Sargasso Sea area ; it reaches the eastern United States seaboard and extends, according to Meisenheimer's chart, into the Mediter- ranean where, however, it is less abundant. Sparser records are from the Indian Ocean and the tropical Pacific. The Atlantic area of Limacina bulimoides appears to lie roughly between 400 N and 400 S. Comparing its distribution with that of L. inflata, Tesch found its greatest abundance per fishing unit in the upper 100 m. to be 243 against 280 of inflata in the Sargasso Sea. In the Benguela hauls of the first survey, this predominance is reversed and inflata is never present in large numbers. Off South West Africa buli- moides was never taken inshore and was confined in March to the five stations WS 986, WS 996, WS 997, WS 998 and WS 999, near the edge of the continental shelf. It is clearly an oceanic species avoiding the colder up-welled waters towards the coast. Fig. 2. Limacina bulimoides. (a) Shell of 'mature female' at stage 6. Shell height approx. 1-9 mm. (b) Operculum, on slightly enlarged scale. ECOLOGY— DEPTH DISTRIBUTION The samples from WS 996 and WS 997 were from an area poor in phytoplankton according to Dr Hart (1953).1 There was some evidence of recent heavy grazing, though the abundant green faeces present were probably attributable to euphausiids. Whatever may have been the economics of this population of Limacina, its presence in such numbers must have made greedy inroads on the phyto- plankton of the vicinity. Although presumably they had fed rather recently, the stomachs of almost all the numerous specimens sectioned were empty of recognizable contents. In one specimen, coccolithophores were present, which may represent an important part of the diet of this population, and which would, at c. 40 /x, be likely to be missed by the N 70 V. plankton net. Of predators probably feeding on Limacina, we may mention the abundance of Pneumodermopsis paucidens accom- panying Limacina at a depth of 50-100 m. at WS 997, and a few large Atlanta at the same station and at WS 996. Partial exclusion of Limacina at 50-100 m. at WS 997 (night station) is evidently not to be explained by the action of predators or by existing hydrological features. An explanation has been ventured below after considering the pattern of depth migration of Limacina bulimoides. Depth distribution. Stubbings (1937) regards the known facts of the diurnal depth migrations of thecosomatous pteropods as consistent with the idea of an optimum light intensity for each particular species. 'At dusk the animal follows the optimum towards the surface and after dark the stimulus to 1 Also in personal conversation. ECOLOGY— DEPTH DISTRIBUTION 173 ascend is removed and the animals probably sink back to the water layers offering the best combina- tion of other factors such as temperature and perhaps salinity. With the coming of dawn they rise to meet the optimum and follow it downwards.' Thus for many species of cavoliniids a double diurnal rhythm is established ; those liking a low light intensity will rise very near the surface at night and those preferring a higher intensity will lose the stimulus earlier and sink. At dawn the reverse takes place, those liking a low intensity receiving the stimulus earlier and preceding the others to the surface. Considering first the two stations WS 996 and WS 997, we find great abundance of Limacina buli- moides by day (4 to 6 p.m.) and by night (1 to 4 a.m.) (Fig. 3). At WS 996 the day-time distribution shows a maximum at 50-100 m. (1542) and relatively few (93) at 0-50 m. From 100-250 m. the numbers fall away markedly to 178, and the fall continues, decreasing fairly evenly down to 750-1000 m. where only six specimens were taken. The species is virtually excluded below 250 m., some of the shells counted here being dead. Of the three smaller day-time samples, no station yielded sufficient material to compare adequately with WS 996. At WS 986 (11 a.m.) (on the line immediately north of the southernmost one — the other four stations being on the southernmost line) there were only eight specimens taken and these were all at 0-50 m. At WS 999 at noon there was a small maximum (twenty-nine) at 100-150 m., and at WS 998 (8 a.m.) Limacina was found only at 100-175 m. where there were eighty. For the night distribution at WS 997 there were 11 16 just below the surface (0-50 m.), many fewer (468) at 50-100 m. and the greatest number anywhere recorded (3142) between 100 and 250 m. On the basis of much smaller samples than the above, Stubbings finds the vertical distribution of cavoliniids (with the exception of the abnormal Creseis virgula which he regards as not controlled by light intensity) to vary between 250 m. for Creseis acicula down to 650 m. for Clio citspidata and 700 m. for Diacria quadridentata, with Clio pyramidata, Cavolinia globulosa and Hyalocylia striata at inter- mediate optima from 400 to 600 m. All show a well-marked surface migration by night, and well- developed diurnal depth behaviour (see Stubbings, 1937, fig. 1, p. 27). When considering the depth distribution and light response of Limacina, it is important to keep its much smaller size in mind. A journey to the surface from 250 m. in the six hours after dusk would entail continuous travel at the rate of 2 ft. per minute, a performance greatly in excess of what I have found L. retroversa capable of maintaining for even a short time (Morton, 1954c). On grounds of size alone we should thus expect species of Limacina to be much less sensitive to changes in illumination than the larger cavoliniids, if only because they are not large or swift enough to swim away from regions of diurnal changes in light intensity. We might expect to find them remaining rather near the surface as grazers on phytoplankton and tolerant of strong light. If they abandon the surface too far by avoiding light during the day, they will be unable to reach it again by night. A similar reasoning may explain the preference for the upper layers of the two smaller and more slender cavoliniids Creseis virgula and C. acicula. These have a relatively tiny wing expanse and no doubt limited diurnal swimming powers. The cavoliniids occurring deepest are the larger and more robust ones. The greatest density of Limacina bulimoides by day at WS 996 is between 50-100 m. This is much nearer the surface than any optimum recorded for cavoliniids (with the exception of C. virgula), and this layer probably represents the maximum light concentration above which Limacina is not happy by day. With the doubtful exception of WS 886, L. bulimoides was never found predominantly in the upper 50 m. by day. The ninety-three individuals present at 0-50 m. at WS 996 were but a small portion of a population in very much greater concentration in the layer immediately below. The most numerous day-time samples came from 50 to 100 m., 100 to 150 m. and 100 to 175 m. Light response by large diurnal depth changes seems to be a feature that arose with the cavoliniids; the more primitive limacinids show it 174 DISCOVERY REPORTS 0 50- 100- 200- INDIVIDUAL5 i — i — i — i — i — i 300- LLI h400- i i i i i i 0 500 ?500- h600- Q_ Ld Q 700- 800- 900- 1000- 1000 J Fig. 3. Limacina bulimoides. Depth distribution diagrams for populations at the day-time station WS 996, and the night station, WS 997. (Insets) Histograms showing size composition of the population at WS 996 (50-100 m.), WS 997 (0-50 m.) and WS 997 (100-250 m.). ECOLOGY— DEPTH DISTRIBUTION i75 much less. It is perhaps worthy of note that L. helicoides, by far the largest species of Limacina (12-5 mm.) is also the only one found regularly at great depths (see Tesch, 1948). It follows that at the night station WS 997, worked between 1 and 4 a.m., we should not expect to find a great concentration of the population at the surface. Those species accustomed to a high con- centration of light should lose the impulse to come to the surface soon after dark, and would have begun to sink downwards by the early hours of the morning. Such a species will probably rise again early with the dawn, but hauls between 1 and 4 a.m. may be expected to coincide with the period of night when the stimulus to rise is weakest. From the greatest concentration by night at the 100-250 m. level, L. bulimoides would seem to show little upward movement— perhaps even a descent, from its day-time optimum. The greatest depth at night, however, is still less than that of any cavolinnd optimum (except Creseis virgula) by day, and the population of Limacina as a whole is probably not significantly responsive to light at all. There is an apparent tendency for a minority to rise to the surface at night. It is probable that only the larger individuals and strongest swimmers will be able to make this migration, and that these individuals will come not from the 250 m. level but from water layers of less than 100 m. depth. Further, the individuals at the surface at night may be expected to be mainly large ones. How does the size composition of the population at various depths at WS 997 agree with this con- clusion? By reference to the histograms (Fig. 3) for the levels 0-50 m. and 100-250 m., a great falling off is found in the percentage of small individuals near the surface— twenty-one as compared with forty-one for the added percentages of the two lowest size groups. Conversely there is an increase in the percentage of larger individuals at the top, thirty-five as compared with twenty-five for the added percentages of the two largest size groups. In the more homogeneous day-time population at WS 996, the comparative percentages are twenty-four for the two large groups and twenty-six for the two smallest groups. It may thus be possible to explain the constriction in the distribution diagram for WS 997; there is a tendency to depth division of the population at night, larger individuals tending to rise nearer to the surface. The bulk of the population including smaller individuals and those present in the deeper layers of the day-time distribution, are not able to reach the surface at night, and show little tendency to migrate at all. Perhaps with no light stimulus to hold them at day-time level, they tend to sink a short distance, which they do faster and more easily than they are able to compensate by swimming up. It would be interesting to look for similarly constricted distribution figures for the night behaviour of other planktonic organisms relatively feebly responsive to light. Superfamily Euthecosomata Family Cavoliniidae Superfamily Pseudothecosomata Family Cymbuliidae These two families, constituting the larger size thecosomatous pteropods, were rather poorly repre- sented among the mollusca of the first survey, the first by small numbers of juvenile individuals of three species, and the second by a single juvenile individual. Diacria trispinosa (Lesueur) 1821 Hyalaea trispinosa Lesueur, in de Blainville, Diet, de Sci. Nat. xxu, 65-83. 1905 Diacria trispinosa Meisenheimer, Pteropoda, Wiss. Ergebn. D. Tiefsee-Exped. 'Valdivia', ix, 27. 1946 Diacria trispinosa Tesch, Dana Reports, xxvm, p. 24, pi. iii, fig. 14. This species occurred only at WS 986, represented by three individuals at 50-100 m., and two at 100-250 m. The three first specimens are the young form, the Cleodora compressa of Souleyet, I76 DISCOVERY REPORTS terminating in a rounded embryonic shell (one of 2-5 mm. the other two of c. 1 mm.). The others are slightly older, representing 'stage B' of Tesch, 3 mm. in length. The species is, according to Tesch, eurythermic, ranging widely from the tropical into the temperate Atlantic. Table 5. Occurrences of Cavoliniidae and Cymbulidae Station WS986 WS986 WS996 WS997 WS 997 WS997 WS986 Time Depth (m.) Diacria trispinosa 1118 100-50 250-100 Euclio pyramidata 1118 1858 to 1610 0105 to 0410 1000-750 500-250 50-0 750-500 Cavolinia inflexa 0105 to 0410 250-100 Cymbulia peroni 1118 100-50 No. of animals 3 juveniles 2 juveniles 1 juvenile 2 juveniles 2 juveniles 1 juvenile 1 juvenile 1 juvenile Euclio pyramidata (Linn.) 1767 Clio pyramidata Linnaeus, Syst. Nat. ed. 12, 1, pt. ii. 1905 Clio pyramidata Meisenheimer, Pteropoda, Wiss. Ergebn. D. Tiefsee-Exped. 'Valdivia', ix, 21. 1946 Euclio pyramidata Tesch, Dana Reports, xxvm, p. 14, pi. ii, fig. 11. This is much the most numerous species of the genus, and is widely distributed throughout the Atlantic. Four records were obtained, all from the deeper layers and represented by a few very young specimens. They are not easy to identify after the erosion of the shell by formalin, and two measure only 1 mm., the larger ones about 3-5 mm. They are placed in pyramidata, since they show no trace of the enlarged foot characteristic of the rest of the species of the genus. Cavolinia inflexa (Lesueur) 1813 Hyalaea inflexa Lesueur, Mem. sur quelques Anim. Moll, etc., Nouv. Bull. Soc. Philom. Ill, 281-5, figs. i-v. 1905 Cavolinia inflexa Meisenheimer, Pteropoda, Wiss. Ergebn. D. Tiefsee-Exped. 'Valdivia,' ix, 35. 1946 Cavolinia inflexa Tesch, Dana Reports, xxvm, p. 30, pi. iii, fig. 21. This species occurred at WS 997 only, represented by very young specimens — 1 mm. long. It is to be regarded as a very common species from 400 N to 400 S — and it is worthy of remark that no more numerous or adult material of any of the above pteropods was available. Cymbulia peroni de Blainville 1 8 18 Cymbulia peroni de Blainville, Diet, de Sci. Nat. xn. 1946 Cymbulia peroni Tesch, Dana Reports, xxvm, p. 37, pi. v, figs. 27, 28. One juvenile specimen of the eastern Atlantic species Cymbulia peroni was taken, at 50-100 m. at WS 986. ECOLOGY— DEPTH DISTRIBUTION 177 Suborder Gymnosomata (=Pterota) Families Thliptodontidae and Pneumodermatidae Table 6. Occurrences of Gymnosomata Station Time Depth (m.) No. of animals Thliptodon diaphanus WS996 1610101858 500-250 1 Pneumodermopsis paucidens WS996 1610 to 1858 100-50 82 250-100 34 500-250 2 750-500 2 1000-750 3 WS997 50-0 7 250-0 6 Thliptodon diaphanus Meisenheimer 1905 Thliptodon diaphanus Meisenheimer, Pteropoda, Wiss. Ergebn. D. Tiefsee-Exped. 'Valdivia', ix, 59. 1926 Thliptodon diaphanus Pruvot-Fol, Moll. Pterop. gymnos. Res. Camp. Sci. Monaco, lxx, p. 28, pi. 3, fig. 79. 1942 Thliptodon diaphanus Pruvot-Fol, Gymnosomata I, Dana Reports, xx, 23, figs. 20-5. 1950 Thliptodon diaphanus Tesch, Gymnosomata II, Dana Reports, xxxvi, 22, figs. 10, 11. Thliptodon diaphanus was the only gymnosomatous pteropod beside Pneumodermopsis paucidens to be taken by the first survey, and it occurred as one specimen only at 250-500 m. at WS 996. Its recorded distribution shows it to be an essentially oceanic species with a wide range over all the warmer regions. In the Atlantic Ocean (Massy, 1917) it reaches the west coast of Ireland (520 N) while in the Pacific it seems to attain its southernmost extreme at 45 ° S on the New Zealand coast. It is a species of the upper layers as well as of the deep sea. The leading character of the genus Thliptodon is the enormously swollen head, which contains a pair of large bladders (' Schlundblasen ') and two very roomy hook-sacs, in which lie a number of golden-coloured hooks. The number of these is of taxonomic importance. Otherwise the only reliable specific character is the radula. The fins form a pair of narrow outgrowths at the constriction between the enlarged head and the rest of the body, and frequently — as in this specimen — both they and the lobes of the foot are difficult to see, being withdrawn into folds of the baggy skin. In the specimen dealt with here, the number of hooks in either sac is about fifteen, decreasing in length inwards as is the case in diaphanus. My figure of the radula does not perfectly agree with either of those of Pruvot-Fol (1926, 1942) or with that of Tesch (1950), but the differences are not hard to reconcile among these variant drawings. Pruvot-Fol's (1926) figure and my accompanying one (Fig. 4a) both show a narrow arcuate trans- verse bar, all that can be seen of the median tooth. Tesch (1950), however, is probably correct in figuring this tooth as a broader plate which is for the most part extremely thin, fragile and transparent, and impossible to see in my probably damaged specimen. The anterior edge of the tooth, however, is considerably more thickened than would appear from Tesch's figure, and I was able to detect a row of the finest denticulations, like those figured by Pruvot-Fol (1926). At the inner edge of the inter- mediate plate is a short but strong spine. In the specimen figured, this did not appear so strongly I?8 DISCOVERY REPORTS curved down as in Tesch's figure or in either of Pruvot-Fol's, and— as Tesch also claimed— the denticles are lacking from the spine of the intermediate plate. The rest of the antero-median margin is exactly as in Pruvot-Fol's first figure, and the three falciform laterals agree well with the previous illustrations. Differences "between this radula and that of Thliptodon rotundatus lie in the absence of the long curlew's bill spine at the inner end of the intermediate tooth, and the smaller width of the median. From both schmidti and gegenbauri, on the other hand, diaphanus is well distinguished by the greater width of the median tooth compared with that of the intermediate tooth. MED. LAT. Fig. 4. Thliptodon diaphanus. (a) The radula, showing the median plate and the intermediate and lateral plates of one side. INT., intermediate plate; LAT., lateral plate; MED., median plate, (b) Enlarged detail of the mesial edge of the intermediate plate, (c) A bundle of hooks from the hook-sac of one side. Pneumodermopsis paucidens Boas 1886 Pneumodermopsis paucidens Boas, Spolia Atlantica. Dansk. Vid. Selsk. Skrift. iv, 160, pi. vii, figs. 105-6. 1887 Dexiobranchaea paucidens Pelseneer, 'Challenger' Reports, xix. Pteropoda, 1, 17. 1926 Pneumodermopsis paucidens Pruvot-Fol, Moll. Pterop. gymnos. Res. Camp. Sri. Monaco, lxx, 12, pi. i, figs. 36-7. This species was very local in its occurrence in first survey material. It was taken in considerable numbers at 50-100 m. and at 100-250 m. at WS 996, and makes a second appearance at WS 997, where it was rather sparsely represented. It turned up nowhere else. Its greatest abundance at WS 996 may be related to the heavy concentration of Limacina bulimoides, since it is chiefly upon small thecosomes that these Gymnosomata are thought to feed. A close agreement is apparent between the depth distribution of P. paucidens at station WS 996 (Fig. 5) and that of Limacina at the same station. The night station, WS 997, did not yield sufficient numbers of individuals for an adequate comparison, but the day-time distribution of this gymnosome is in agreement with Meisenheimer's conclusion (1905, p. 98) that the Gymnosomata (' Tagtiere ')— unlike most of the thecosomes— approach the surface by day. ECOLOGY— DEPTH DISTRIBUTION 179 0- I0D- 200- 300- ^400 a: 500- F600- 700- 800- 900- 1000 -I Fig. 5. Depth distribution of Pneumodermopsis paucidens at the day-time station WS 996. T.AC. ■ME. DID. IOm. Fig. 6. Pneumodermopsis paucidens. The animal viewed from the right side, with the buccal apparatus fully extended. BA., row of basal suckers; BR., gill; CIL., posterior ciliated circle; DIG., area of digestive gland; GON., gonad; HS., hook- sac; LA., lateral lobe of foot; L.i, lateral sucker of first pair; L.z, lateral sucker of second pair; ME., median lobe of foot; SAL., salivary gland; T.AC., enlarged terminal sucker; V., position of genital opening (female); W., right wing. l8o DISCOVERY REPORTS The anatomy of the genus Pneumodermopsis has been best described by Pruvot-Fol (1924). The distinctive feature of the group is the presence of a membranous triangular outgrowth forming a lateral gill on the right side, somewhat beyond the point of greatest diameter. There is no posterior gill, and in some species— such as the present one— the posterior ciliated circle may persist long into Fig. 7. Pneumodermopsis paucidens. Outline drawings of a representative group of twelve specimens from station WS 996, showing some of the variations of form encountered in preserved material. adult life. The hook-sacs are extremely small, forming a pair of short, blunt outgrowths at the sides of the buccal mass, each containing a cluster of short chitinized hooks. The posterior (or median) lobe of the foot is long, tapering and slender. The species represented here is clearly P. paucidens Boas. Its chief distinguishing character is the number and arrangement of the suckers borne on the aceta- buliferous appendages inserted on the dorsal wall of the buccal cavity. There is a long median arm with a single large terminal sucker, more than twice the size of the lateral suckers, and much larger indeed than might be deduced from the diagrammatic illustration of Tesch's key (1950). Two pairs ECOLOGY— DEPTH DISTRIBUTION 181 of pedunculate lateral suckers are inserted upon the stalk of the large median sucker, and at the base of the stalk, inserted directly upon the wall of the buccal cavity and almost encircling it, is a single transverse row, consisting on either side of the middle line of five small, sessile suckers. Species of Pneumodermopsis appear in general to be of rather restricted distribution. P. paucidens was first described from the tropical Atlantic (Boas, 1886) and Massy (19 17) has recorded it as reaching the south-west coast of Ireland. Pruvot-Fol gives some further subtropical Atlantic localities, and mentions (1942) its occurrence in the Indo-Pacific. The considerable amount of material obtained from the Benguela survey yielded several rather well-expanded specimens by which identification was greatly assisted, and it has been thought worth- while to provide a figure (Fig. 6) for this species, since detailed illustrations appear to be altogether lacking. Like all the Gymnosomata, P. paucidens suffers great distortion upon preservation. The foot and wings are frequently lost from sight, disappearing into a roll of skin surrounding the middle of the body like a tyre. The buccal armature is but rarely well-extended. At times, however, even a mis- shapen specimen yields outlines rather characteristic of its species when preserved, and I have pro- vided here a series of outline figures representing the size range and the more typical modes of distortion found in a series of specimens from the larger sample at WS 996 (Fig. 7). Class LAMELLIBRANCHIATA Order EULAMELLIBRANCHIATA Two species of lamellibranch larvae occurred at the three sets of inshore stations. The first, lamelli- branch larva 'A', was from WS 981, off Walfisch Bay, where 540 specimens were taken in 0-50 m. The second species, lamellibranch larva 'B', was present off Sylvia Hill, at WS 988 (two larvae at 0-50 m., one at 50-100 m.) and at WS 989 (six larvae in 0-50 m.), and, much more numerously, at the mouth of the Orange River where a sample of about 600 was secured at WS 1002 (0-50 m.). The Station WS 1002, off Walfisch Bay, is in the area of an 'azoic zone' where the bottom deposit consists of a mud populated with sulphate-reducing bacteria (Currie, 1953 ; Marchand, 1928). It would appear that settlement of the larval lamellibranch taken in large numbers in the plankton hauls is scarcely possible on this ground, and Currie observes that during March oxygen depletion of the water was most marked near Walfisch Bay, water with less than i-o c.c. oxygen per litre extending in some places to the sea surface. Lamellibranch larva 'A' (Fig. 8) The largest individuals of this species consisted of a ' veliconcha ' made up of the two ' prodisso- conch ' shell stages, and measuring 290-300 fi in length. The smallest members of the population were late stages of the ' D-shaped larva', still possessing the straight-line hinge along about half the visible dorsal margin, and measuring 150-180 /x in length. In neither of the species of lamellibranch larvae were details of the hinge-teeth easily available for examination ; shells preserved in formalin were softened so that their shape became distorted on applying the methods for opening recommended by Rees (1950). Outline figures of the intact shells are, however, provided, in the hope that a further knowledge of lamellibranch larvae of this region or correlation with the occurrence of adults on shore may enable a more exact identification of this material. Details of hinge-teeth were available by examination of the transparent intact shell of the late D-shaped or straight-hinged stage. Here, although there is already some differentiation of hinge- teeth, the hinge structure is unfortunately not advanced enough to justify comparison with the stages figured for the main superfamilies by Rees (1950). The provinculum at this stage appeared to bear no teeth at all, but a simple lateral hinge system was present. In the left valve, lateral teeth are 3-2 i82 DISCOVERY REPORTS developed in the form of solid triangular wedges, accommodated in shallow depressions in the right valve. These teeth lie at either end of the featureless provinculum, two smaller teeth posteriorly, one large tooth anteriorly. Lamellibranch larva ' B ' (Fig. 9) This species — at the more advanced stages present — is of larger size than larva 'A'. Its shape is more orbicular with the beaks rather less prominent. The shell-length of the largest individuals is approximately 500 /x, and the smallest, which just passed the straight-hinged D-shaped stage, measure 150/x in length. Fig. 8. Shell of lamellibranch larva 'A' from inshore station WS 981, off Walfisch Bay. {Above) The hinge-line of the late ' D-shaped stage ', the right valve towards the observer. Fig. 9. Shell of lamellibranch larva ' B ' from inshore station WS 1002, off Orange River. Class CEPHALOPODA Subclass DlBRANCHIA Order DECAPODA Superfamily ARCHITEUTHACEA Four specimens of oigopsid larvae were present in the Benguela material, and — chiefly because of the incompletely known life-histories of oceanic decapods of the South Atlantic area — presented some difficulty in assigning to genera. Specific identification has not been ventured on, but by the kindness of Dr W. J. Rees, who has examined the material, I am able to provide the following notes. 1 . A cranchiid larva, approximately 5 mm. in length, with the characteristic features of a bursiform body, very short series of eight sessile arms, and protuberant or pedunculated eyes. The body is unpigmented, and the eyes reddish brown. The specimen 'compares very favourably with the juvenile Taonidium figured by Chun (1910), PI. 61, fig. 9'. Occurrence. WS 976, 250-500 m. ECOLOGY— DEPTH DISTRIBUTION 183 2 and 3. Two very juvenile larvae, not long out of the egg. The body is colourless, the eyes sessile and the arms very short and subequal. Beyond the position of these larvae in the Architeuthacea, it is possible to establish very little about them. Occurrence. WS 978, 350-500 m., one specimen, c. 3-5 mm.; WS 978, 250-500 m. (second haul), one specimen, c. 2-5 mm. 4. A larval histioteuthid, agreeing largely with the figure of a young stage of Calliteuthis reversa, given by Naef (1923), p. 355, fig. 174: 'Naef was working in the Mediterranean on two histioteuthids which he could separate because there are only two adult species known in that area. In the Atlantic there are at least six species which may have larvae of this kind.' Occurrence. WS 1091, surface to 100 m., one specimen, c. 75 mm. II. THE REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF LIMACINA BULIMOIDES It is to the work of Hsiao (19390, b) that we owe almost our whole knowledge of reproduction in Limacina. He gave an account of the reproductive system in L. retroversa and established that this animal is a protandrous hermaphrodite passing after a sexually undifferentiated period through a male stage, succeeded by a phase that is predominantly female. He worked with sufficiently large numbers to establish the incidence of sex phases among the size groups in a population at different times of the year. A knowledge of breeding in an ecological dominant like Limacina has two special advantages: we are able to trace the elements in a population made up of mixed components not always with the same relation between sex and size (see Redfield, 1939), and, with information about the mode of life and conditions of reproduction, we are able to predict the probable seasonal movements and activity of a given population. The Benguela material of L. bulimoides is especially good for the study of a large population of a single origin with an apparently uniform sexual sequence. We still lack comparative data for any other month, but— taken in the breeding month of March— the autumn Benguela material contains a good proportion of mature females and a complete section of other age groups. It is hoped in a second report to examine the late winter or early spring material from the second survey. Hsiao gave a rather brief account of the structure of the reproductive system with little histological or structural detail. Unfortunately, from the literature then available he was not able to take a com- parative view of Limacina in its relation to the other primitive opisthobranchs. The present material of bulimoides, though preserved in routine formalin, was found to be good enough in many cases for a study of histology and enabled a fuller report of the genital system to be prepared. The lack of living material was unfortunate, yet with such a large number of stages and with living British retroversa for comparison it was felt justifiable to redescribe the whole system. Protandrous hermaphroditism may well have been the original condition in the Gastropoda. In prosobranchs at the most primitive level it has been found by Orton (1909) in Patella and by Bacci (1947 a, b) in Fissurella nubecula and in Patella coerulea. In mesogastropods, hermaphroditism survives only in rather isolated cases, such as Valvata and Crepidnla. We may regard the Mesogastropoda, especially in their frequent atypical spermatogenesis, as sexually specialized gastropods. In the pulmonates and opisthobranchs previously known, hermaphroditism is of the simultaneous type and it appears, as for example in Helix and in higher nudibranchs, that the stages have been telescoped together. Morton (19546) presents evidence that the opisthobranchs and pulmonates came off close together from a prosobranch stock below the level of the present day Mesogastropoda, and in a study of the ellobiids he gives the first account of reproductive succession in a primitive pulmonate. In a 184 DISCOVERY REPORTS single season there is a well-spaced successive hermaphroditism, male to female. Limacina represents the first primitive opisthobranch so investigated, and its life-history may be tentatively regarded in all respects as typical of the early gastropods. The specialization of atypical sperm-formation has not been developed. Hsiao looked for the representatives of 'nurse sperms' in Limacina and not sur- prisingly did not find them. The masses of residual cytoplasm observed by him are too indefinite to suggest any such anomaly and have not been found in Limacina bulimoides at all. (i) The Gonad and Genital tract We have now a description of a primitive genital tract in a pulmonate (Morton, 19540) and in a primitive opisthobranch, from the earlier work of Guiart (1901), while a forthcoming account by Lloyd will make clear the condition in several bullomorphs. In Limacina we must look for features characteristic of primitive tectibranchs, with resemblances on the one hand to early pulmonates and on the other to prosobranchs. In primitive opisthobranchs we are probably near a point where the three major groups of gastropods draw very close together. The gonad consists of a long, tubular sac occupying in the mature female 3^ visceral whorls and the whole of the apex of the animal. It is not subdivided and bears in all its stages a general resemblance to that of L. retroversa. It leads forward alongside the liver into a narrow hermaphrodite duct. This opens into the glandular genital tract, which is the pallial genital duct of Fretter (1946), and lies on the floor of the mantle cavity, at later stages bulging strongly into it. It consists of an albumen gland and a mucous gland. From the albumen gland, in the mature female there runs back alongside the stomach a sausage-shaped receptaculum seminis, lying dorsally on the first visceral whorl and attaining almost the size of the stomach. The common genital aperture opens on the right side rather far forward in the pallial cavity, and from it a ciliated fold leads forward to the secondary male aperture on the head at the base of the right wing. The male aperture leads into a highly complicated invaginated penis in front of and quite separate from the pallial tract; and with the penis is closely associated a stout, tubular prostate gland. Condition of the gonad and sexual stages Hsiao (19390) divided his population into those with sexually undifferentiated gonads, those with 'pure male' gonads (designated 'functional males'), 'functional hermaphrodites', and finally 'herma- phrodite females'. It was recognized that no clear line of demarcation separated the males and females, but that they merged into a continuous series. No pure females were found among L. retro- versa, and functional females were held to be those where more than 75 % of the gonadial contents were oocytes or eggs. In bulimoides, however, it is possible to recognize a pure female phase. The gonad in animals of more than 1 -9 mm. shell height is usually cleared of all but a few lingering sperms. The female duct is now highly developed, the penis is lost and the sperm sac is full after copulation. The sexual succession is thus better spaced and the later stages further separated in bulimoides than in retroversa. We may list in Table 7 the sexual phases and their characteristics as recognized in L. bulimoides. For the three largest samples of L. bulimoides comprising 5800 individuals from stations WS 996 and 997, the following data were obtained (Table 8) for the percentage distribution of sexual stages with size. Seven size groups separated by 0-2 mm. intervals were obtained by breaking up the three samples by the use of graded and measured pipettes, and the results afterwards carefully checked by the measurement of representative samples from each group. From each size group a sample of approximately fifty individuals was then examined after bulk staining with Ehrlich's haematoxylin and REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 185 clearing in xylol. The condition of the gonad could be ascertained by rapid inspection of the cleared animal, and each specimen assigned to its stage in Table 7. The results were checked by correlating the appearance of the stained gonad with the condition of the genital tract and gonad in a large series of sectioned specimens. Table 7. Stages in the sexual succession of Limacina bulimoides 1. No sexual development at all. Gonad represented only by a small cap of undivided pregerminative cells. 2. Less than one gonadial whorl. An apical mass of lighter staining primary spermatogonia, which in lower part of gonad have begun to divide into darker-staining secondary spermatogonia. 3. Less than two gonadial whorls. Primary spermatocytes arranged in spherical clusters lying freely in the lumen of the gonad. Penis and prostate absent or very tiny. 4. More than two gonadial whorls. Gonad contains primary spermatocytes and spermatids. Stained appearance never shows lighter areas of bundles of sperm tails. Penis and prostate fairly large. Female ducts present, but usually non-glandular. 5. Mature male. Most of gonadial lumen occupied by platelets of mature sperms. Penis and prostate large. Oocytes which have now appeared, are increasing in size (c. 70-100 /x) but still parietal and with no dark granular yolk inclusions. Epithelium of female duct filled with undischarged secretion. Receptaculum seminis usually not visible. 6. Pure Female. Gonad containing entirely (or almost entirely) oocytes, with at times a few clusters of sperms left behind. Spermatogenesis entirely ceased. Many ova discharged into lumen of gonad, and of large size (c. 130/i) with dark-staining yolk inclusions. Penis and prostate lost. Mucous gland large and distended with secretion. Receptaculum seminis large and tightly filled. Table 8. Percentage distribution of sexual stages among size groups in Limacina bulimoides Sexual stage (%) Shell length in millimetres over i-8 1-8-1-6 1 -6-i -4 I'4-I-2 I -2-1 i-o-8 Less than o-8 1 — — — — 1 1 85 2 — — — — 6 89 !5 3 — — — 11 69 — — 4 5 17 8 61 41 5° 89 25 - 6 83 3i 9 — — — — The data of Table 8 have been expressed diagrammatically in Fig. 10 in a form easily comparable with Hsiao's diagram for L. retroversa (Hsiao, 1939&, p. 286). In Fig. 11 is further shown the distribution of the four sexual types, here designated females, mature males, developing males and sexually undifferentiated individuals, among the three largest samples from the March cruise. The form of the diagram should enable easy comparison with data obtained in the second survey, and with subsequent data obtained for this or other species of Limacina over the twelve months of the year. (ii) The Genital Ducts (Fig. 12) The animals used for description are chiefly those of stage 5 shortly before female maturity. The glandular genital duct already bulges into the pallial cavity, where later, at stage 6, it becomes pressed closely against the pallial mucous gland (PALL.G.) on the roof of the mantle, and filled with masses of colourless secretion. As it appears in Fig. 12, looking down upon it lying on the pallial cavity floor, the mucous gland, which is the most obvious part of the genital tract, consists of two limbs. The largest is the descending tubular limb (MUC.D.), very thick-walled, tapering forward to the common genital aperture (C.GEN.). It acts as an hermaphrodite genital duct, as in primitive hermaphrodite gastropods, and sperms traverse it before the passage of eggs. It is translucent white in colour. The 1 86 DISCOVERY REPORTS 100 90 PERCENTAGE 80 70 60 50 40 30 Z0 ID 0 Fig. io. Limacina bulimoides. Diagram showing the percentages of each sexual stage in the different size groups of the population. Combined data from the three largest samples at WS 996 (50-100 m.) and WS 997 (0-50 and 100-250 m.). Fig. n. Limacina bulimoides. March population at WS 996 and WS 997. Diagram showing the distribution of pure females, mature males, developing males and sexually undifferentiated individuals in the whole March population. REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 187 other limb of the mucous gland is a plump, horseshoe shaped crescent of glandular tissue (MUC.AC.) closest on the right side to the edge of the mantle cavity and the columellar muscle (COL.). Sperm would not appear to enter this part of the mucous gland and it forms an appendage to a deeper lying part of the genital duct. This deeper portion constitutes the albumen gland (ALB.), which consists of a wide atrium, thinner walled than the mucous gland. The non-glandular little hermaphrodite duct (L.HERM.) opens into it posteriorly. In front it leads forward between the two limbs of the mucous gland, into a thin walled pouch with very long cilia (FERT.). This pouch corresponds in position and in its histology to the fertilization pouch of primitive pulmonates, and forms an annexe to the albumen gland where the sperms come into contact with the egg. The albumen gland is lined with a rather opaque white epithelium, contrasting in appearance with the mucous gland. It differs from the same gland of most primitive pulmonates in still forming an integral part of the duct traversed by the genital products, and bulging freely into the pallial cavity. In even the most primitive pulmonates, as also in the higher opisthobranchs, the albumen gland forms a diverticulum which becomes quite removed from the main channel of the genital tract, and is haemocoelic in position. After copulation, the albumen gland in Limacina contains isolated sperms suspended in mucoid secretion. Leading back from it, close alongside the little hermaphrodite duct, runs the short tubular duct of the receptaculum seminis (REC). The receptaculum is well developed only at stage 6 when it becomes tightly distended with sperm and easily recognizable in cleared whole-mounted animals after staining. It forms a very spacious caecum, reaching a maximum dia- meter of more than 100 /x, and it apparently increases in size suddenly after one or more acts of copula- tion. It is thin-walled and lined with flattened squamous epithelium, entirely non-glandular and, so far as can be made out, non-ciliated. The large masses of sperm within it are arranged in rather wavy, parallel bundles, but they are never packed together in such distinct platelets as are seen in the lumen of the male gonad. The duct of the receptaculum is 15-20x1 across, consisting at first of flattened ciliated cells. Before its opening into the albumen gland, the duct itself becomes non-ciliated and albuminiparous. The whole sac is of the nature of an outgrowth from the albumen gland. It is homologous with the receptaculum of prosobranchs, and not with the bursa copulatrix, which is a separate structure opening from just within the common genital aperture, and forming the sperm storage sac in Aplysia (Eales, 1921) and in the majority of pulmonates. From the sketch given by Hsiao, the receptaculum in Limacina bidimoides seems to correspond to what he calls the 'vesicula seminalis' in retroversa, though this is a wrong name for it, since it does not store 'home' sperm derived from previous male gametogenesis. This is evident by its position as a diverticulum of the albumen gland, not as an outgrowth or distension of the little hermaphrodite duct as in the 'vesicula ' of pulmonates or prosobranchs. In Limacina the whole of the stored sperm produced by the animal in its male phase evidently remains in bundles in the cavity of the ovotestis at stage 5 ; further, the receptaculum does not appear until after the penis of the same individual has disappeared. Of the sperm sac described by him in this position, Hsiao states that 'masses of mature spermatozoa can be seen inside this organ in the case of the mature Limacina '. His account of the origin of these sperma- tozoa whether from the gonad of the same animal or from a partner at copulation is non-committal. Of the ' receptaculum seminis ' described as such by Hsiao in retroversa, I can find no representative in bidimoides. Hsiao states that this sac is connected with the ' basal portion of the outgoing duct ' and that it consists of two portions, the first thick-walled and much folded, containing sperms normal in structure and staining deeply with iron haematoxylin. The second is described as thinner walled and baggy in appearance; the sperms stain only very lightly and undergo degeneration. From the position of this diverticulum in Hsiao's figure and from the occurrence of sperm disintegration, it would appear to correspond to a true bursa copulatrix. i88 DISCOVERY REPORTS a * -o o E &£ &2 c 5 c „ to ^£2 a to T3 60 o c 2 n rt M M ™ O 60 to 6 „ co « « J3 M <0 N _D 60 ~" ' .2 o — — O. 3 ^ : ^ Si c +3 S 3 u .2 J= to rt s >> .3 4-» •S JJ CO T3 a a 4-» 1> - j= .3 ■" j= s'f O l> <£i 60 T3 .3 co >-. •5 ~ > t! o ^ 60_ « « +J -t-» M "3 o 5 S 60 S3 3 3 « « . C to J3 ° ~ E 3 u *c3 s* | g — V CO <0 CO " ^ J3 O . £ U to T3 3 tJ u £ -3 "° Ec S3S .2 c " GO U to s ^ C3 m u 2 "o ^ *3 2 o o o -a u cs N N 60 . X) J3 H -c -J to -g 2 - g rt o ■Co -a o -c to -jj en a to * to J3 kT a, c iv O « •£ "S -3 ° *-" 4-1 s §»§ ^.3 60 W o rt -a *-• en O O CO ^ J3 60 •3 rJ co 33 o S 3 ^ 2 o a o C a 60^ o J c "So to 3 ■a b ° =3 J3 co a. E 2 3 E 4" | O 3 .- o E c 3 O 60fe| o 3 ^ 2 ^ -3 a. "3 E rt a; S 9 0| ■a 13 C cs ,2 a 60 - o CO £^ 60^ C M 5 o o to a o 3 — O 60 Li ^ Ui C3 U Kr '-*— F £ G c _i: "eu — r <-l-t o 3 L5 c 0 CO u CO C- ~ < 7. 0 S I •o ^J pt s| s E S S REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 189 Of the histology of the glandular genital tract, which was not within the scope of Hsiao's paper, only a brief description is needed here. The epithelium of the albumen gland (Fig. 13 a) reaches c. 50 fi in height at stage 6. Its free surface becomes rather ragged by the discharge of secretion and only a few cells remain typical and intact, with the rounded tips which represent the usual condition at stages 4 and 5. Ciliated cells are small and interposed at rather wide intervals between the gland cells. Their nuclei are smaller and darker than those of the glands, and the tufts of fine cilia become dislodged and carried away at the secreting stage. The nuclei of the gland cells are 6-7 /x in length, densely chromatic and rounded to sausage-shaped. The contents of the cytoplasm bear a considerable resemblance to those in the albumen cells of Ellobiidae (Morton, 1954a). The secretory inclusions form masses of tiny granules of varying but minute size. They stain golden brown with eosin- Orange G, and especially towards the base of the cell they become enclosed in clear, non-staining vacuoles. A group of vacuolated spherules is sometimes found enclosed by a single larger vacuole. The secreting cells of the mucous gland (Fig. 136) increase in size from 20 to 30 ^ at stage 5 to as much as 150^ in the greatly distended gland at stage 6. Their nuclei are basal, in a single row, large and densely chromatic. Here again there is a close resemblance to the Ellobiidae. The contents of the cell stain deep purple with haematoxylin and are granular, and with the change to a secreting state, large ovoid, lightly staining droplets are formed, bulging from the cells into the lumen. The ciliated cells are extremely narrow; they fan out as triangular wedges at the free surface, and their small and dark staining nuclei are either compressed between the gland cells, or flattened closely against the top of a neighbouring cell. No eggs of bidimoides were taken in the Benguela hauls, nor does Hsiao give a description of the egg mass. Lebour (1932) has, however, observed the oviposition of British retroversa, and has described and figured the eggs as planktonic, floating in small gelatinous strips with the eggs scattered and widely separated. Each strip measures about 2 mm. in length and 0-64 mm. in breadth and both the matrix (which is obviously the secretion of the mucous gland), and the eggs are perfectly transparent. Some- times the strips are interconnected by a thread-like secretion which is presumably also a product of the glandular female duct. The penis, as examined at stage 5, is perhaps the most complicated part of the reproductive system, and Limacina, with others of the thecosomatous pteropods, possesses probably the most elaborately constructed penis in the opisthobranchs. As pointed out by Hsiao this organ is practically independent of the rest of the reproductive system, both in position and in structure. The seminal groove does not continue as a closed tube towards the male aperture at the base of the right wing, but passes forward as the ciliated groove and becomes continuous with the interior of a large ovoid pouch the penial sac, lying in front of the rest of the reproductive organs and bulging into the pallial cavity just behind the head (Fig. 14). It is first visible at stage 4, especially large at stage 5 and lost at stage 6. The penial sac is involuted with a lining wall elaborately ridged and furrowed, and this lining, or a portion of it, constitutes the penis itself when evaginated. There is no muscular papilla traversed by a closed vas deferens as in even the earliest pulmonates. Associated with the penial sac is a tubular prostate gland. This is intricately involved in the folds and furrows of the sac, and the whole structure in the fixed and contracted condition is exceedingly difficult to make out clearly. A good deal can be seen from cleared whole-mounted individuals, but in the absence of living material serial reconstruction is necessary to determine its details. Hsiao does not attempt a description, and in the other Thecosomata, such as the cavoliniids, Meisenheimer gives few details. Bonnevie (1916) provides an account of the penis and its associated structures in Cuvierina, a cavoliniid, and there is a briefly labelled drawing of the penis in longitudinal section in his figure of Cymbulia (1905, Atlas, pi. 13). Vayssiere (1915), figuring the opened penial sac, gives a picture of the folds of the interior without any interpretation. Tesch's 4-2 190 DISCOVERY REPORTS ALB. NU.CL IOO T NU.CL. NU.ALB. MUC ML/CD, d C.GEN: Fig- 13- Fig. 13. Limacina bulhnoides. Histology of the genital tract, (a) The albumen gland approaching the secretory condition at stage 6. (b) The mucous gland at stage 5, before the onset of secretory activity, (c) The short glandular epithelium lining the general cavity of the penial sac. (d) The prostatic epithelium, at stage 5, the cells opening into the prostate above, and the smaller cells forming part of the external epithelial sheath below. ALB., albumen spherules beginning to pass from the free surface of the cell; M.F., muscle fibre in wall of penial sac; MUC, mucous contents of a secretory cell of the mucous gland; NU.ALB., nucleus of albumen cell; NU.CL., nucleus of ciliated cell; NU.EP., nucleus of glandular epithelial cell; NU.PR., nucleus of a prostatic cell; PR., Prostatic epithelium; PR.EXT., external sheath of prostatic cells. Fig. 14. Limacina bulhnoides. Diagram of the relations of the gonad, genital ducts and penial sac, for comparison with Fig. 12. The penial sac is shown in broken outline as it appears at stage 5; the receptaculum (REC.) in fact never appears until the penis is lost, at stage 6. ALB., albumen gland; C.GEN., common genital aperture; FERT., fertilization pouch; GON., gonad; L.H.D., little hermaphrodite duct; MUC. AC, accessory limb of the mucous gland; MUCD., descending limb of the mucous gland; M.AP., male aperture; PE.S., penial sac; REC, receptaculum seminis; SEM.GR., ciliated seminal groove. REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 191 figures of a dissection of the penis in Limacina helicoides are also wanting in detail. A prostate is nowhere mentioned by these authors, though Hsiao correctly points out its presence in retroversa. Fig. 15 must serve to illustrate schematically the interior of the penis in L. bulimoides. It represents a stereogram of the whole sac opened longitudinally and slightly obliquely. The ciliated seminal groove enters at the male aperture (M.AP.) and through the same aperture the intromittent portion of the penis is extruded. The chamber of the sac into which the sperm flows from the seminal groove is to a large extent occupied by the thick tubular prostate, which is usually coiled in an S or doubled upon itself in the retracted and fixed condition. It is slung to the wall of the sac by a long, winding septum and the lumen of the penial sac consists of a series of narrow slits and spaces un- occupied by the prostate. The lumen of the prostate itself is a narrow slit, rounded or crescentic in cross-section. It receives the secretion from the lining cells and communicates apparently only at its proximal end with the sperm canal. So far as can be made out from reconstruction, the distal end of the prostate appears to end roundly and bluntly. At the bottom of the penial sac there is a flask-like expansion, with which the inner end of the prostate communicates, and towards which converge also the interstices of the sac surrounding the prostate. This flask-like expansion (LU.PEN.) is prolonged into a conical or pointed tube, perforate throughout and surrounded by a layer of vacuolated tissue, clear-staining and cartilage-like. While no doubt flexible, this structure appears to be the most rigid part of the male apparatus. It terminates at one side of the prostate, opening directly into the lumen of the penial sac, some distance inside the male aperture. Though in the absence of observations of copulation, it is difficult to reach a final conclusion, it would appear, from a comparison with similar ' cartilage '-supported structures in Clione, that this is the intromittent portion of the male apparatus. It is the only part that is not tied down by the insertion of the prostate and that is likely to be eversible from the male opening. Into it, as has been said, converge the folds leading the sperm inwards and between the coils of the prostate ; and at its base opens the duct of the prostate itself. The histological structure of this part of the penis is illustrated diagrammatically in Fig. 15. The main thickness of its wall is formed by a row of large wedge-shaped or columnar vacuolated elements. These are actually the ' cells ' of molluscan cartilage, a modification of connective tissue, resembling a stiff parenchyma, developed — as in the buccal mass — in most regions where stiffness combined with flexibility is required. The lumen of this part of the penis is lined with a low epithelium of cubical cells of the same glandular type (see below) as cover the rest of the interior of the penial sac. At the base of the sac, and spreading along its wall from the intromittent portion of the penis runs an extension of the cartilage-like connective tissue. It seems to be one of the characteristic features of the genital system of primitive tectibranchs that the ' prostate ' develops as a long appendage to the penis quite separate from the rest of the glandular genital duct. Such a condition is in strong contrast to that of the primitive Pulmonata, where the prostate— as in the prosobranchs— retains a close relation with the anterior part of the pallial genital tract. Here at first it forms part of a common mucous-prostatic duct, possessing a single lumen con- veying both eggs and sperm ; later — in the higher pulmonates it becomes cut off as a separate male channel. In both pulmonates and opisthobranchs the nature of the prostatic glandular cells appears to be essentially similar. Their function is probably identical, and there would seem to be no good reason to restrict the descriptive term 'prostate ', following the suggestion of Hubendick (1948), to the prosobranch or pulmonate type, and to exclude from this category the male accessory gland of tectibranchs. In Philine, according to the description by Guiart (1901), there is a tubular prostatic gland opening as an appendage from the base of the penis, with which it makes a double connection. 'Chez la Philine aperta la prostate se compose d'un long tube glandulaire qui forme de nombreux replis dans la cavite cephalique de l'animal, mais le canal central de cette prostate se trouve en rapport 192 DISCOVERY REPORTS PR. J. CART. I. COMM. 'CART. 2. LU PEN. GL.EP. PR. 2. PR.C. PR. EXT. M. A P. SEM. GR. Fig. 15. Limacina bulimoides. The penial sac containing the coiled prostate gland illustrated as a stereogram by the opening of the sac in an obliquely longitudinal plane. The prostate is represented as having been cut transversely at three points, PR.i, which ends bluntly in the direction of the observer, and PR.z which becomes continuous above the plane of the section with PR-2, forming the terminal part of the prostate lying at the base of the intromittent portion of the penis. The arrows suggest the course followed by the sperm from the seminal groove, through the interstices between the prostate and the wall of the penial sac, to the base of the intromittent portion. The inset diagram shows the shape of the intact prostate. 'CART. 1 ', vacuolated connective tissue, forming cartilage-like support of the intromittent portion of the penis; 'CART.z', 'cartilaginous' connective tissue of the wall of the penial sac; COMM., communication between the grooves of the penial sac and the base of the intromittent portion; GL.EP., glandular epithelium forming the general lining of the penial sac; LU.PEN., lumen of the intromittent portion of the penis; LU.PROS., lumen of the prostate; M.AP., male genital aperture at anterior end of penial sac; PR.C, prostate as indicated by dotted lines through wall of penial sac; PR.i, PR.2, PR-2, prostate gland; PR. EXT., investment of prostatic epithelium external to the tubular prostate gland; SEM.GR., ciliated seminal groove, running from common genital aperture to male opening; SEP., septum. REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 193 par un double conduit — d'une parte avec la gaine du penis par un conduit relativement court et d'autre part avec le penis par un tube excreteur formant une anse assez considerable.' In Haminea, Doridium and Gastropteron there is also a prostate of the appendage type, and a similar structure was described briefly by Tesch (1950) and further investigated by the writer (unpubl. obs.) in the gymnosomatous pteropod Clione limacina. It is apparently lacking in Aplysia, the pleurobranchomorphs and in the bullomorph Acera, and also in the primitive Actaeon, which in many features of its genitalia would seem to stand apart from other tectibranchs. The condition in the Thecosomata, with the prostate so closely involved in the actual penial sac, would appear to be unique. Its relations are likely to be much better understood after the examination of more species of pteropods, and especially after Lloyd's work has been published, which will give a full and detailed review of the genitalia in a wide range of opisthobranchs. The histology of the prostate in Limacina (Fig. 13^) is relatively simple; there is one type of gland cell and at no stage could ciliated cells or cilia be identified at all. The glandular cells are columnar, at stage 5, 75 ju, in length. The nuclei are large rounded and deeply staining, forming a single basal row. The inclusions in the cytoplasm vary in size extremely; the largest are about 10 /x across and the smal- lest extremely fine, much less than 1 /x. Their size in a given cell is relatively constant, and sometimes in a transverse section there appears a sector of cells containing smaller granules, very distinct from the rest. The free surface of the cells is fiat or bulges slightly into the lumen — it is broken only by the passage of contents from the mature cell. The staining reaction is little if any with haematoxylin, pink in eosin, and orange after eosin-Orange G. The epithelium of the prostate lies on a very thin con- nective tissue sheath, sparsely muscular. Outside the prostatic tube, facing towards the lumen of the penial sac, there is another layer of glandular epithelium (PR. EXT.). The base of this layer lies against the base of the prostatic epithelium, forming a sleeve wrapped around about half the circum- ference of the prostate along the best part of its length. The secretion from these cells is similar to that produced by the prostate, and it is discharged directly into the lumen of the penial sac. The height of this layer reaches 10/x, and the cell nuclei are rather flattened, only about half the size of the prostate nuclei. Towards the edge of the epithelial zone the height of the cells diminishes and they merge gradually with the very thin glandular epithelium, which seems to line the whole of the rest of the penial sac, being reflected both over the coils of the prostate and the ridges and furrows of the sac, and running through the intromittent portion of the penis itself. Here the cell height is less than 5 /x, and the cell boundaries are very indistinct. The cytoplasm is packed with tiny granules of secretion, some of them highly refractile, and staining like those of the prostate. The nuclei form a flattened basal row, and the free border of the cells, when fixed in formalin, tends to break away with the shedding of its contents. Whether it is ever ciliated is not easy to determine, but even in the best fixed material no cilia were detected anywhere in the prostate or penial sac. So far as is known, Limacina has never been observed in copulation, and our knowledge of pairing in a pteropod is restricted to the one species Clione limacina. It would be of great interest to examine the relations of the male organs in the everted condition and to find how the supposed intromittent organ is brought into use. We may, how- ever, reasonably suggest from examination of the structure of the penial sac the course taken by the sperms, which are presumably liberated freely into the general cavity of the sac when brought to it by the seminal groove. Together with prostatic secretion they enter the intromittent organ at its base. The secretion of the prostate, and that of the general lining of the sac evidently contribute the nutritive portion of the semen, and the sperm finds its way through the lumen of the sac by several rather tortuous paths between the coils of the prostate, to the base of the penis. The arrows in Fig. 1 5 suggest the general course taken by the sperm after it is received into the sac and before ejaculation. IO, DISCOVERY REPORTS (iii) Sexual succession In the colder waters of the North Atlantic according to Hsiao, sexual differentiation in L. retroversa commences in the winter months and sexually differentiated individuals appear as in ' pure males ' by the beginning of March. The predominance of pure males decreases towards the end of March and disappears in May, being replaced by a predominance of female animals. In April the percentage of females was forty and early in May the two sexual types were in equal numbers. Spawning had commenced by the end of April, which may be compared with Lebour's statement that in British waters the eggs are ripe for the greater part of the year, with a maximum in summer when each individual may apparently have had several broods. In retroversa from the Gulf of Maine the period of egg-laying thus reached its height at the end of the spring. The Benguela bulimoides were from autumn material. There were few cases of individuals having been spent in spawning and a relatively small number of very young individuals. The immature males may well belong to a com- ponent of the population resulting from spawning earlier in the spring of the same year. All sexual stages of the population were represented in the autumn material of the first survey and there is a very uniform correlation of successive sexual stages with increasing shell size throughout all the material examined. There is no indication, in the form of variable sex-size relations that this population was heterogeneous, as was demonstrated by Redfield (1939) in North Atlantic retroversa. Several questions arise regarding L. bulimoides which probably can be answered after examining material from the second survey, with a spring population available for comparison. First, Is there a differentiated breeding season in L. bulimoides, or does reproduction, as in the English retroversa, (as distinct from cold water Gulf of Maine material) take place all the year round? In Hsiao's retroversa a spring breeding season commences by the end of April, with a fairly rapid maturation of oocytes and a change in sex predominance during the preceding month. In L. bulimoides, where an autumn breeding season has been shown to occur, it would seem likely that reproductive activity in March may have been preceded by breeding in spring, possibly continuing without interruption during the summer months. This leads to the query, Are single individuals of Limacina able to undertake a second breeding season during the same year, or does the breeding life of a single Limacina extend beyond the first breeding year? From the sex-size analysis of the autumn population of bulimoides, there seems to be a single peak within a definite size group for each successive sexual stage. As yet there is no evidence for this species that a restitution of the male phase is possible in individuals in which the oocytes are spent. Such a form of sexual alternation occurs, as shown by Orton (1926) in the oyster ; and the present writer obtained evidence from a colony of the primitive pulmonate Carychium, which had laid eggs in June, that a restitution of sperm-producing tissue took place during the same year in the gonad of spent females. Further, Do seasonal differences occur in Limacina bulimoides, in the proportions of the various sexual stages at different sizes of the animal? Hsiao (1939b) suggested that small individuals in the hermaphrodite female condition in his 'population B' did not pass through a protandric phase of noticeable length, though other specimens of the same population had a distinct 'young male phase' preceding the female. He concludes that ' during the warmer part of the year sexual differentiation takes place sooner than in colder times and that the sexual phases may be telescoped together when development is speeded up in the summer'. This is in agreement with observations made by Coe (193 1 ) and Orton (1909; and other papers) on Ostraea. Such a trend towards the 'telescoping' of sexual stages in the warmer parts of the year points the way to the simultaneous hermaphroditism of higher opisthobranchs and pulmonates, where the female phase is pushed back until it is contem- poraneous with the male and the two types of sexual product reach maturity at the same time in the common ovotestis. REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 195 In view of the probably uniform breeding season of the English Channel retroversa and the Ben- guela bulimoides, with but slight acceleration in summer, it is probable that variation in the rate of sexual succession is less pronounced in Limacina from warmer seas. Geographically, Limacina is divisible into two groups, the smaller-sized tropical species, trochiformis, inflata, lesueuri and bulimoides, HEIGHT OF SHELL IN MM. •5 1-4 1-3 1-2 I- 1 1 1 I I L 09 0-8 _L 0:7 06_ 2-0 -[ 1 1 1 r 1-5 1-4 1-3 l-E H HEIGHT OF SHELL IN MM. Fig. 16. Diagram to summarize the course of the sexual succession of Limacina bulimoides. The topmost panel represents the growth of the gonad, the scale showing the number of gonadial whorls. The successive stages in gametogenesis are indicated by the types of shading. At the centre of the figure is shown the development of the male organs, as indicated by the epithelial height of the prostate (see scale); and of the female tract, expressed by the cell height of the epithelium of the mucous gland. Below is represented on the left the increase in size of the receptaculum seminis (total diameter, by the same scale), and the increase in the number of whorls of the animal, gonadial whorls being represented in black. and the larger species of higher latitudes like retroversa (which extends also into the warmer parts of the North Atlantic), balea, helicina and antarctica. The genus is well known to reach its highest development, as indicated by the size attained by the species, in colder seas. It is probable that in high latitudes the ' telescoping ' of the sexual succession, occurring in the summer months, may be more advanced than the primitive condition of a fully extended sexual succession in warmer seas. For example, in bulimoides 'telescoping' of the succession appears hardly to occur at all; but in cold water i96 DISCOVERY REPORTS retroversa, Hsiao at no time found any pure females. The final stage in succession has evidently been pushed back to occur simultaneously with the later part of the male phase, and the largest animals are ' hermaphrodite females '. In the autumn bulimoides there were as many as 15 % °f pure females with sperm absent, or present in the gonad only in the smallest traces. In these females, the penis is always lost. Such a condition may perhaps occur also in the ' hermaphrodite females ' of retroversa, though this appears to be nowhere explicitly stated by Hsiao. Within the family Limacinidae, it might be reasonable to expect that evolution— as in the great group of opisthobranchs as a whole — has moved away from complete protandry, towards a telescoping of the sexual succession in the direction of simultaneous hermaphroditism. One remaining species of Limacina, which dwells in cold waters, the very large, abyssal-planktonic helicoides, which may reach a diameter of 12-5 mm., appears from Tesch's account (1946) to have developed reproductive specializations not elsewhere found in the family. This species, as first noticed by Bonnevie (1913), is viviparous and, at the female stage, developing eggs and small embryonic replicas of the adult are found in the distended coils of the mucous gland. We may attempt to sum up the sexual development and succession in L. bulimoides in a diagram (Fig. 16) expressing the stages in the maturation of the gonad and the condition of the accessory genital apparatus against the growth of the shell. The lower panel represents the number of gonadial whorls in relation to the total number of whorls of the animal. This is a feature easy to deter- mine by external inspection. The gonad shows a regular increase up to 3^ whorls in the mature female at stage 6. It should be noted that because of the increase in size from the apical to the body whorl this relationship does not directly express the proportion of gonadial to other tissues. With practice in the examination of stained sections, the viewing of the gonads of cleared whole animals lightly stained in Ehrlich's haematoxylin yields accurate information as to the stage of sexual develop- ment. Parallel with gonad growth, Fig. 16 shows the stages in gametogenesis corresponding to the sexual phases established for this species in Table 7. The development of the male accessory organs is expressed in terms of the maturity of the prostate gland as determined by the height of its epithelial cells. In the female system, development is regarded as a function of the cell height of the epithelium of the mucous gland, while the total diameter of the tubular receptaculum is shown on the same scale — this organ increases in size not primarily by growth but by rapid inflation when the sperm is deposited at copulation. (iv) Discussion — Sex in Gastropoda With the description of sexual succession in a primitive opisthobranch and in a primitive pulmonate and with a knowledge of the changes undergone by the gonad in some primitive prosobranchs, we may venture very tentatively to reconstruct the evolution of sexuality in gastropods (see diagram, Fig. 17). Let us suppose protandrous hermaphroditism to be the earliest condition in the gastropoda at the point where, from evidence given elsewhere (Morton, 19536) the three great subclasses of gastropods draw closest together. We may suppose this to be somewhere near the archaeogastropod level of the prosobranchs. From such a condition, perhaps like that of Bacci's Fissurella nubecula today, and without any development of genital ducts, we first find among living and archaic archaeo- gastropods a tendency to the separation of the sexes in different individuals, which has probably been best realized in living trochids. Then, with the emancipation of the renal genital duct from its excre- tory function and the addition of a glandular genital duct developed from the pallial wall, the further advances are possible. In almost all the rest of living prosobranchs there is a tendency to complete separation of the sexes and bisexuality obtains in the great majority of species. Such cases of herma- phrodite prosobranchs as do remain (such as the Calyptraeidae and the Valvatidae), assumed in the past to be specialized, may on this interpretation be regarded as survivals of a primitive hermaphro- REPRODUCTIVE SYSTEM AND SEXUAL SUCCESSION OF L. BULIMOIDES 197 ditic condition. It is to be noted that in the great majority of mesogastropods — whether bisexual or hermaphrodite — there is a trend towards sperm dimorphism, and in some species the oligopyrenic sperms reach a grotesque size. [Fretter (1953)1 has demonstrated huge giant sperms in Clathrus transporting vast numbers of attached normal sperms.] Parthenogenesis, too, appears to occur in at least one species, Paludestrina jenkinsi. The sexual condition in the mesogastropods as a whole is therefore too specialized to make them easily acceptable as ancestors of the pulmonates or the opistho- branchs. The two latter groups have probably originated close together at a higher archaeogastropod level and have retained their hermaphroditism. Each group began with a protandrous sexual succes- sion and the succession also involves the pallial genital duct which, like the higher prosobranchs, these OPISTHOBRANCHIATA HIGHER PROSOBRANCHS PULMONATA ELLOBIIDAE [MORTON_ I9S3a] -~~CALYP TRAEIDA £ [ORTON_ 1909] FlSSURELLlDAE [BACCI-1947] LOWER PR0S0BRANCH5 Fig. 17. Diagram to illustrate the theory of the evolution of sex in the Gastropoda discussed in this paper. Black indicates a condition of protandrous hermaphroditism, stipple simultaneous hermaphroditism and white separation of male and female individuals. groups have acquired. The higher members of both subclasses have evolved in the direction of simultaneous hermaphroditism, with the sexual stages telescoped together and eggs and sperm pro- duced side by side as in Helix, for example, or in separate acini of the ovotestis (higher nudibranchs). A similar but less well-marked trend seems to have been to cut out localized breeding seasons and to produce eggs and sperm simultaneously all the year round. A more detailed discussion of the evolution of hermaphroditism in gastropods is intended in a later account, after the presentation of further work. It is already clear, however, that a study of Limacina is of great value for an understanding of the early condition with which the sexual evolution of many opisthobranchs and pulmonates would seem to have begun. SUMMARY The pelagic Mollusca collected during the Benguela Current Survey by the ' William Scoresby ', in March 1950, comprise the following species: Ianthina ianthina, I. globosa, Atlanta peroni, Limacina inflata, L. bulimoides, Diacria trispinosa, Euclio pyramidata, Cavolinia inflexa, Cymbulia peroni, Thliptodon diaphanus, Pneumodermopsis paucidens, as well as two species of larval lamellibranchs, one prosobranch larva, and several larval cephalopods. The numerically important members of the plankton were the two lamellibranch larvae at stations close inshore, and the two pteropods, Pneumodermopsis 1 Also in personal conversation. 5-2 i98 DISCOVERY REPORTS paucidens and Limacina bulimoides offshore. The latter was by far the most abundant species and greatly outnumbered the related L. inflata. Its ecology and distribution are discussed. There is evidence that the diurnal depth migrations performed by the Limacinidae are very much shorter than in the Cavoliniidae already investigated, and there appears at night to be a size-depth division of the population, the larger individuals making a journey to the surface, and the smaller individuals remaining predominantly lower down as in their day-time distribution. The gymnosomatous pteropod Pneumodermopsis paucidens is a species that has not been found commonly in past collections, and a taxonomic description and fresh figures are given. In the second part of this paper the reproductive system and sexual succession of Limacina buli- moides are described in detail. The genital ducts are typical of the lower level of evolution in both the opisthobranchs and the pulmonates, and consist of a little hermaphrodite duct and a glandular herma- phrodite tract comprising albumen gland, fertilization pouch and mucous gland. A receptaculum seminis forms a large diverticulum in the mature female. The penis is a complicated structure, and with it is associated a tubular prostate gland. These structures are for the first time described fully, with details of their histology. Limacina is a protandrous hermaphrodite and six sexual stages are recognized, all of which are present in the March population, ranging from sexually undifferentiated individuals of small size, through various stages of male development, to the pure females which are the largest individuals. A comparison is made between L. bulimoides and L. retroversa as reported upon by Hsiao, and a short discussion is added, dealing with the sexual succession in Limacina and the problem of the evolution of sex in the Gastropoda as a whole. REFERENCES Bacci, G., 1947a. L'inversione del sesso ed il ciclo stagionale della gonade in Patella caerulea L. Publ. Staz. zool. Napoli, xxxi, pp. 183-217. 19476. SulV ermafroditisma nelT Archaeogastropoda. Arch. Zool. Ital. XXXII, p. 329. De Blainville, 18 17. Diet. Sci. Nat. ix, p. 407. Boas, J. E. V., 1886. Zur Systematik und Biologie der Pteropoden. Zool. Jahrb. Jena, I, p. 311. Bonnevie, Kristine, 1913. Pteropoda. Sci. Repts. 'Michael Sars' N. Atlantic Deep Sea Exped. 1910, 3, (1) Zoology. 1916. Mitteilung iiber Pteropoden. I. Geschlechts-apparat von Cuvierina columnella Rang. Jena. Z. Naturw. liv, p. 245. Chun, C, 1910. Ober Cephalopoden der D. Tiefsee-Expedition. Zool. Anz. Leipzig, xxxm. Coe, W. R., 1931. Spermatogenesis in the Californian Oyster (O. lurida). Biol. Bull. Woods Hole, lxi, p. 309. Currie, R., 1953. Upwelling in the Benguela Current. Nature, London, clxxi, p. 497. Cuvier, G., 1817. Le Regne Animal. Paris, n, p. 380. Eales, Nellie B., 1921. Aplysia. L.M.B.C. Memoir, no. xxiv. Liverpool. Fretter, Vera, 1946. The Genital Ducts 0/ Theodoxus, Trivia and Lamellaria and a discussion on their evolution in the proso- branchs. J. Mar. biol. Ass. U.K. xxvi, p. 312. 1953. The Transference of Sperm in Gastropods. Proc. Linn. Soc. London (in the Press). Guiart, J., 1901. Contribution a I 'etude des gasteropodes opisthobranches et en particulier des ciphalaspides. Mem. Soc. zool. Fr. xiv, (5). Hart, T. J., 1953. Plankton of the Benguela Current. Nature, London, clxxi, p. 631. Hoffmann, H., 1938. Opisthobranchia, in Bronn's Klassen und Ordnungen des Tierreichs. Mollusca, II, vol. m. Hsiao, S. C. T., 1939a. The reproductive system and spermatogenesis of Limacina (Spiratella) retroversa. Biol. Bull. Woods Hole, lxxvi, pp. 7-25. — 19396. The Reproduction of Limacina retroversa (Flem.). Biol. Bull. Woods Hole, lxxvi, pp. 280-303. Hubendick, B., 1948. The Anatomy of Bulinus with a discussion of the term Prostate and its sense in the Basommatophora. Proc. Malacol. Soc. London, xxvn, pp. 186-96. I red ale, T., 1910. On Marine Mollusca from the Kermadec Islands and on the ' Sinusigera' apex. Proc. Malacol. Soc. London, IX, p. 74. Lebour, Marie V., 1932. Limacina retroversa in Plymouth Waters. J. Mar. biol. Ass. U.K. xvm, pp. 123-9. Marchand, J. M., 1928. Special Report No. 5, Fish, and Mar. Biol. Surv., Union of S. Africa. Massy, Anne L., 1917. The gymnosomatous pteropods of the coasts of Ireland. Sci. Proc. R. Dublin Soc. xv, p. 223. REFERENCES '99 Meisenheimer, J., 1905. Pteropoda. Wiss. Ergebn. Deut.-Tiefsee Exped. 'Valdivia', ix. Morton, J. E., 1954a. Studies on the Digestive and Reproductive Systems of the British Ellobiidae (Leucophytia bidentata, Ovatella myosotis and Carychium tridentatum). (In the Press.) 19546. The Evolution of the Ellobiidae with a discussion on the origin of pulmonates. Proc. Zool. Soc. London (in the Press). 1954c The Biology of Limacina retroversa. J. Mar. biol. Ass. U.K. xxxiii, pp. 297-312. Naef, A., 1923. Die Cephalopoden. Teil II. Systematische Morphologie der inneren Organisation. (Vergleichende Anatomie und Organogenese.) Fauna e Flora del Golfo di Napoli, Monografia, xxxv, p. 2. National Institute of Oceanography, 1950. Station List R.R.S. 'William Scoresby' 1950. Discovery Repts. xxvi, pp. 211-58. Oliver, W. R. B., 1915. The Mollusca of the Kermadec Islands. Trans. N.Z. Inst, xlvii, pp. 509-68. Orton, J. H., 1909. On the occurrence of protandric hermaphroditism in the mollusc Crepidula fornicata. Proc. Roy. Soc. London, B, lxxxi, p. 468. 1926. Observations and Experiments on Sex-change in the European Oyster (O. edulis). J. Mar. biol. Ass. U.K. xiv, P- 967- Pelseneer,P.,i887. Report on the Pteropoda collected by H. M.S. 'Challenger'. I. The Gymnosomata. Sci. Repts. 'Challenger', Zoology, xix, (iv), pp. 1-69. 1888. Report on the Pteropoda. . . 'Challenger'. II. Thecosomata; III. Anatomy. Sci. Repts. 'Challenger', Zoology, xxm, pp. 1-132. 1906. Mollusca, in A Treatise on Zoology (ed. by E. Ray Lankester, London), pp. 355. Pruvot-Fol, Alice, 1924. Etude de quelques Gymnosomes mediterr. des peches de VOrvet en 1921-22. Arch. Zool. exp. gen. Paris, lxii, pp. 345-400. ■ 1926. Moll, ptirop. gymnos. provenant des campagnes du Prince Albert I de Monaco. Res. Camp. sci. Monaco, LXX, pp. 1-60. 1942. Les Gymnosomes. I. Dana Rept. XX. Redfield, A. C, 1939. The History of a population of Limacina retroversa during its drift across the Gulf of Maine. Biol. Bull. Woods Hole, lxxvi, pp. 26-47. Rees, C. B., 1950. The Identification and Classification of Lamellibranch Larvae. Hull. Bull. Mar. Ecol. Ill, (19), pp. 73"io4- Sherborn, C. D., 1930. Index Animalium, 1801-1850. London, xxm, p. 6093. Stubbings, H. G., 1937. Pteropoda, in Sci. Repts. John Murray Exped. 1933-34. v. PP- x5-33- Tesch, J. J., 1904. The Thecosomata and Gymnosomata of the Siboga-Expedition. Leyden, 27 (iv), p. 13. 1946. The Thecosomatous Pteropods. I. The Atlantic. Dana Repts. xxvili. 1948. The Thecosomatous Pteropods. II. The Indo-Pacific. Dana Repts. xxx. 1949. Heteropoda. Dana Rept. xxxiv. 1950. The Gymnosomata. II. Dana Repts. XXXVI. Thiele, J., 193 1. Handbuch der systematise/ten Weichtierkunde, (11), Jena. Vayssiere, A., 1915. Moll, eupteropodes. Res. Camp. sci. Monaco, xlvii. [Discovery Reports. Vol. XXVII, pp. 201-218, August, 1954.] THE CIRCUMPOLAR CONTINUITY OF ANTARCTIC PLANKTON SPECIES By A. de C. BAKER National Institute of Oceanography CONTENTS Introduction page 203 Material and methods 203 Phytoplankton 204 Zooplankton 208 Identification of species 209 Distribution 213 Conclusions 214 Summary 216 Bibliography 217 T THE CIRCUMPOLAR CONTINUITY OF ANTARCTIC PLANKTON SPECIES By A. de C. Baker National Institute of Oceanography (Text-figs. 1-5) INTRODUCTION he Southern Ocean is a continuous circumpolar belt, the greater part of which is drifting eastwards under the influence of the prevailing westerly winds. It is clear from the widespread soundings taken by many ships that there is an uninterrupted belt of deep water, and Deacon (1937), Clowes (1938) and Mackintosh (1946) have shown that there is a circumpolar continuity in the relative positions and movements of the main water masses,1 in the distribution of nutrient salts and in the surface isotherms. As a general rule the physical features are arranged in zones (with local modifica- tions and distortions) in which uniform conditions persist in east and west directions and changes or gradients occur from north to south. Since the distribution of pelagic species may be expected to extend as far as the environment is uniform, it would be surprising if the distribution of planktonic species were not also circumpolar. Those who have examined collections of plankton from different parts of the Southern Ocean have observed that the familiar species can indeed be looked for in any longitude at least in the Antarctic zone, and the circumpolar distribution of the plankton is generally taken for granted. This, however, is a far-reaching assumption, and if it is true it may sometimes allow general conclusions based on samples from one sector of the Antarctic to be applied to other sectors. Although much diffuse evidence of circumpolarity may be found in the reports of various expeditions which have collected material during the past sixty years or so, it appears that no evidence has yet been assembled to show that circumpolarity is a general character of the distribution of the plankton of high southern latitudes. The present paper sets out to show that at least the common species of the macroplankton animals and of the phytoplankton have a circumpolar distribution in the Antarctic zone, and that this circum- polar distribution is continuous in so far as, with sufficient sampling, they can be found at one time or another in all longitudes. It goes little further than this, but the point is considered to be of sufficient basic importance to form the subject of a separate paper. MATERIAL AND METHODS The material considered here includes the most numerous Antarctic phytoplankton species, and the commoner animals of the macroplankton in the Antarctic surface layer. Material could also be found in collections and publications for examining the circumpolar distribution of the smaller planktonic animals, those of the deeper layers, and the mammals, oceanic birds, fishes, and benthos; but such a comprehensive investigation would involve more complex problems, and the work entailed would at this stage be out of proportion to the particular question at issue. 1 The Antarctic Bottom Water might be regarded as an exception. Although it can be traced all the way round, its potential temperature shows a gradient from west to east (Deacon, 1937, pi. XLiv) so that it does not have uniform circumpolar properties. 204 DISCOVERY REPORTS During the Discovery Investigations some thousands of plankton samples have been taken from stations spread over the whole of the Southern Ocean at all seasons of the year, the majority south of the Antarctic Convergence. For the present investigation an arbitrary selection of samples has been made from hauls between the surface and a depth of 250 m. in the Antarctic water, which means that they have been taken from within the limits of the Antarctic surface water (Deacon, 1933). Further particulars of the selection of samples are given below. As explained on p. 209 the material used in this paper is largely derived from the work of other members of the staff of the National Institute of Oceanography and the former Discovery Committee, who for other purposes have identified the species in a large number of samples. I have had the advantage of advice from various specialists also in checking the identity of zooplankton species. Their names are mentioned below on p. 209. My own work has been to examine such additional samples as were necessary, to assemble the material, and to present the conclusions. For the phyto- plankton, however, most of the work had already been done some years ago by Dr T. J. Hart, who has very kindly allowed me to use, in the section which follows, his own account of the phytoplankton, which he had prepared in anticipation of some more general study of circumpolar distribution. It is reproduced here with only slight modifications to adapt it to the present paper. Dr Hart has had much experience of the Antarctic diatoms and is confident of the specific identities of the forms listed in Fig. 2. PHYTOPLANKTON Among organisms other than diatoms, only Phaeocystis and Distephanus can be considered important in the plant population of the seas of the Antarctic zone. The general homogeneity of the vegetation is not always obvious from the results of a single voyage owing to seasonal changes in its constituents. Prolonged cruises by the research ships of the Discovery Committee have enabled us to distinguish between seasonal variation and differences in regional distribution with some certainty. In the course of this work it became so evident that all the more important species had a completely circumpolar distribution, that in a paper on phytoplankton periodicity in those seas Dr Hart stated it as a fact (Hart, 1942, p. 270). Nevertheless, no detailed exposition of common circumpolar distribution for Antarctic phytoplankton organisms has yet been made, and these notes are designed to supply one. The observations available to the south of the Antarctic Convergence have been grouped into intervals of 200 longitude and found sufficiently numerous for the selection of ten representative summer stations in all but two of these eighteen sectors. Between 1400 and 1600 E. only eight observa- tions are available and between 1800 and 160° W. only nine. Much material obtained during the second and fourth commissions of the R.R.S. ' Discovery II ' has not been fully examined, but the results presented here seem ample to prove the point at issue. In Fig. 2 the frequency of occurrence of each species in any one sector is shown, to the nearest 5 %, as a percentage of the number of samples examined in that sector. The stations selected are listed in Table 3, p. 217. In tabulating the results Dr Hart has considered all the more important phytoplankton species of the first four ecological groupings adopted in his paper on ' Phytoplankton Periodicity in the Antarctic ' (1942). It will be realized that in most sectors all the observations were obtained under strictly oceanic conditions. This alone accounts for such gaps as occur in the observed distribution of diatoms of the neritic/ice-edge group. Other gaps are probably due mainly to the rather rigid time limit adopted to keep the observations in several sectors fairly comparable; for example, the inclusion of autumnal stations would doubtless show a more complete distribution of Chaetoceros radicidum, but would greatly diminish the percentage frequency of occurrence of other more important species. The chief source of the observations is the series of qualitative counts from hauls with the Harvey PHYTOPLANKTON 205 * * * . ■ i ■ * • ^ ■ 160 W 180°E 160 140 Fig. 1 . Positions of the stations from which samples have been used to demonstrate the circumpolar continuity of the common phytoplankton and zooplankton species of the Antarctic surface water. +, Phytoplankton stations; O, Zooplankton stations; ©, Phytoplankton and zooplankton stations. 206 DISCOVERY REPORTS apparatus (NHP) during the circumpolar cruise of the R.R.S. 'Discovery II ' during the summer of 1937-8. This has been supplemented by published data from hauls with the Gran international net (N 50 v) obtained in 1929-31 by both the R.R.S. 'Discovery II ' and the R.R.S. 'William Scoresby '; by centrifuge counts, usually from six depths at each station, obtained during the third commission of the R.R.S. ' Discovery II ' (1933-5) '■> an^ by some of Mr J. W. S. Marr's field notes obtained during the fourth commission (1935-7)- Taking the ecological groups in order: Group I. Small oceanic pennate diatoms with Distephanus. Here it is seen at once that all three species have a very complete circumpolar distribution. Group II. Large diatom species: the solenoids, large Chaetocerids and two elongated oceanic pennate forms, twelve categories considered. Here Chaetoceros criophilum, Rhizosolenia alata, Dactylio- solen antarcticus, both main phase groups of Corethron and Synedra pelagica, all show complete circumpolar distribution, and there is no doubt that a wider choice of stations within the relevant dates between o° and 200 W. would complete the circle for that ' patchy ' species, Thalassiothrix antarctica. The species of Rhizosolenia, other than Rh. alata have a gap here and there, but they would almost certainly show continuous distribution if further observations become available. Group III. Neritic/ice-edge forms, seven categories considered. This group necessarily shows less continuous distribution than the others, for the majority of the observations were obtained under strictly oceanic conditions. It has already been shown (Hart, 1942, p. 285) that the importance of this group in the open Southern Ocean is confined to a particularly narrow time-interval immediately following the break up of the pack-ice. It is to this fact and the unavoidable absence of neritic observations in many sectors that the discontinuity of distribution shown here is due. Asteromphalus parvulus shows complete continuity, but this is the most nearly oceanic member of the group. Thalassiosira spp. and Nitzschia closterium, however, show only one gap each, and it is to be supposed that further work would show continuous distribution for the other species of this group also. Group IV. Oceanic Chaetocerids of medium size, eight categories. Gaps in the circumpolar dis- tribution of this group as represented here are almost certainly due to their time distribution. Dr Hart has shown (Hart, 1942, p. 293) that they reach their greatest relative importance during the post maximal midsummer decrease of the phytoplankton as a whole — Chaetoceros radiculum even later as already mentioned — and this fact made it impossible to select stations representative of the full spatial distribution of the group. Even so, there is very little doubt of the complete circumpolar distribution of all the categories. Chaetoceros atlanticum and Ch. dichaeta (type phase) show complete continuity from the present data. Ch. castracanei shows one gap, Ch. chunii and Ch. schimperianum only two each. From the data presented here the following species show complete circumpolar distribution within the Antarctic zone: Fragilariopsis antarctica*, Nitzschia seriata, Distephanus speculum, Chaetoceros criophilum, Rhizosolenia alata, Dactyliosolen antarcticus, Corethron criophilum, Synedra pelagica*, Asteromphalus parvulus*, Chaetoceros atlanticum, Ch. dichaeta. The majority of these are also completely cosmopolitan species, occurring in polar and sub-polar waters of both hemispheres ; most of them in specialized portions of the intervening sub-tropical and tropical waters as well, though not, as a rule, in the same phases. Only the species marked with an asterisk are rigidly confined to the far south, and even they 'overlap' occasionally into the sub- Antarctic zone, especially Fragilariopsis. Turning to species of which the circumpolar distribution seems certain though not completely shown by the present data we can add : Rhizosolenia antarctica* , Thalassiothrix antarctica* , Thalassio- sira spp., Nitzschia closterium, Chaetoceros castracanei. PHYTOPLANKTON 207 DEGREES EAST DEGREES WEST O BO IOO 120 140 160 180 160 140 120 IOO SO 60 40 20 C o1 (E N9 OF SAMPLES EXAMINED | IO FRACILARIOPSIS ANTARCTICA NITZSCHIA SERIATA V DISTEPHANUS SPECULUM CHAETOCEROS CRIOPHILUM RHIZOSOLENIA ALATA RH ANTARCTICA RH. CHUNII RH HEBETATA RH STYLIFORMIS RH. TRUNCATA DACTYLIOSOLEN ANTARCTICUS CORETHRON CRIOPHILUM DO., SPINELESS CHAINS SYNEDRA PELAGICA V THALASSIOTHRIX ANTARCTICA f THALASSIOSIRA SPP. ASTEROMPHALUS PARVULUS BIDDULPHIA STRIATA EUCAMPIA BALAUSTIUM CHAETOCEROS NEGLECTUM CH SOCIALE ^NITZSCHIA CLOSTERIUM CHAETOCEROS ATLANTICUM CH CASTRACANEI CH. CHUNII CH. CURVATUM CH DICHAETA DO.TENUICORNIS PHASE CH. RADICULUM ^.CH. SCHIMPERIANUM Fig. 2. Occurrence of species of phytoplankton in all longitudes around the Antarctic zone of the Southern Ocean, showing the percentage frequency of occurrence in samples taken within every 200 of longitude. 2o8 DISCOVERY REPORTS It is instructive to note the proportion of the species in the phytoplankton within the Antarctic zone. For the present purpose all the categories previously examined may be divided according to the measure of completeness of circumpolar distribution shown by the data examined here (Table i). The figures tabulated with them are mean percentages by counts from 177 observations obtained in the three main oceanic regions of the Antarctic zone over the periods of the main increase during 1937-8 and 1938-9. Table 1 . Probability of circumpolar distribution and mean percentages of each species in the phytoplankton population Circumpolar distribution Proved % Almost certain % Probable °/ /o Fragilariopsis antarctica 24-0 Rhizosolenia antarctica o-i Rliizosolenia chunii 0-2 Nitzschia seriata 11-8 Thalassiothrix antarctica 17 Rh. hebetata !-3 Distephanus speculum i'4 Thalassiosira spp. 2-4 Rh. styliformis >o-i Chaetoceros criophilum 6-5 Nitzschia closterium i-3 R/i. truncata >o-i Rhizosolenia alata 19 Chaetoceros castracanei i-3 Biddulphia striata >o-i Dactyliosolen antarcticus i-8 Eucampia antarctica o-5 Corethron criophilum IO'I Chaetoceros neglectum 17 Synedra pelagica 0-9 Ch. sociale 3-8 Asteromphalus parvulus 0-4 Ch. chunii 2-3 Chaetoceros atlanticum 40 Ch. curvatum 0-4 Ch. dichaeta 4-5 Ch. dichaeta, tenuicornis phase Ch. radiculum Ch. schimperianum 6-8 0-3 '•3 68-3 6-8 187 It will be seen that the rarer microscopic plants, not considered here, form only 6-2 % of the total phytoplankton ; 68 % of the organisms captured by the NHP have been shown to have a completely circumpolar distribution, while it is very highly probable that a further 25 %, representing some eighteen species, are truly circumpolar also. It thus appears that the statement 'almost all the important species have a completely circumpolar distribution ' is not exaggerated. The only important plant not dealt with is Phaeocystis brucei. Dr Hart has already explained (Hart, 1942, p. 266) that this organism defies existing methods of estimating even its relative abundance; further, that it has a very limited time distribution as a first colonist when pack-ice disperses. Thus we cannot show the distribution of the species with the precision possible when dealing with diatoms. All our available records of Phaeocystis brucei plotted on a circumpolar chart are shown herewith (Fig. 3). From this it appears (at the least) highly probable that this species has a completely circum- polar distribution also. ZOOPLANKTON The material used for the investigation of the zooplankton consisted of samples collected by the R.R.S. 'Discovery II' between 1930 and 1951, and the R.R.S. 'William Scoresby' during the years 1928-31. These samples were taken with the i-m. net hauled obliquely (N100B) from a depth of approximately 100 m. to the surface. The construction and operation of this net have been described by Kemp, Hardy & Mackintosh (1929). Where necessary the N 100B material has been supplemented with samples from similar oblique hauls with the 70-cm. net (N70B) and from vertical hauls with the 70-cm. net (N70V) from 100-50 m. and from 50-0 m. As in Dr Hart's treatment of the phytoplankton the circumpolar area has been divided into eighteen sectors, each of 200 longitude. The method was to choose about ten of the richer samples spaced as ZOOPLANKTON 209 evenly as possible in each of these sectors, and to note which species were present. There were never less than ten samples available in each sector. If some species were absent, then a few more samples would be examined. Hence, as will be seen in Table 2, the number of samples examined from each sector varies between ten and sixteen. The total amounts to 230 and the positions of the stations at which they were taken is shown in Fig. 1. The station numbers are given in Table 3, p. 217. Most of the samples used were taken during the southern summer (October-April) but in order to cover each sector satisfactorily it has been necessary to include twenty-eight winter stations. The sectors in which these occur and the relative numbers of winter and summer stations are shown in Table 2. Table 2. Proportion of winter to summer stations Sectors 0- 20° E. 20- 40° E. 40- 6o°E. 60- 8o°E. 80- 100° E. 100- 120° E. 120- 140° E. 140- 160° E. 160° E.- 180° Winter 6 1 0 0 0 0 8 3 0 Summer 10 10 12 H J3 12 2 12 12 Total 16 1 1 12 H x3 12 10 15 12 Sectors 180- 160° W. 160- 140° w. 140- 120° W. 120- ioo° W. 100- 80° W. 80- 6o°W. 60- 40° W. 40- 20° W. 20° W.- 0° Winter 0 I 2 1 0 2 0 I 3 Summer IS !3 10 12 11 11 11 10 12 Total 15 H 12 13 11 13 11 II iS IDENTIFICATION OF SPECIES During the years prior to the last war a large number of plankton samples were analysed by various members of the staff. This has greatly facilitated the present work, since just over half the data used here have come from samples in which the species were already identified. The methods (of fractioning, etc.) used in these analyses have been described by Mackintosh (1934), but it is necessary to give here some idea of the accuracy of identifications as they affect this present work. Where a large number of plankton samples have to be examined it is essential that a technique should be developed that enables each to be dealt with comparatively rapidly. Descriptions and reference specimens of the common Antarctic zooplankton were provided by specialists, and it was found that with the aid of these in the first instance the worker very soon became familiar with the general appearance and one or two diagnostic features of the larger and more characteristic forms, which enabled them to be identified by the naked eye. Thus it will be seen that the analyses are based on a familiarity with the species that are commonly found in the samples, and this obviously allows a slight risk of inaccurate identification where some of the forms are concerned. I have therefore made a further check on the identifications of the species. For this purpose specimens were taken from samples from each of the following four sectors of the Antarctic circumpolar area, 0-200 E., 80-1000 E., 180-1600 W. and 100-800 W. These specimens were then submitted to specialists for examination. As a result it was found that in all but two instances, which are considered below, the selected specimens had been correctly identified. I would like to take this opportunity of thanking the specialists concerned for their assistance. Captain A. K. Totton of the British Museum very kindly examined the Siphonophores and Mr N. Tebble of the British Museum the Polychaets. I would particularly like to thank Dr J. P. Harding, DISCOVERY REPORTS Fig. 3. Observed distribution of Phaeocystis brticei within the Antarctic zone of the Southern Ocean. IDENTIFICATION OF SPECIES 211 also of the British Museum, for his help with the Copepoda, a group that involved a great deal of work. Mr P. M. David of the National Institute of Oceanography examined the Chaetognatha and has been of assistance on several occasions, and Dr J. E. Morton of Queen Mary College has kindly examined the Mollusca. Dr T. J. Hart, as well as contributing a large part of this paper, has given me assistance with the Amphipoda. Mr N. B. Marshall of the British Museum has kindly read a draft of the paper and made some helpful comments. Finally I would like to thank Dr N. A. Mackintosh for his general guidance during this work. The following is a list of genera and species identified : Siphonophora Diphyes antarctica Moser Dimophyes arctica (Chun) Pyrostephos vanhoffeni Moser Anthomedusae Sibogita borchgrevinki E. T. Browne POLYCHAETA Vanadis antarctica (Mcintosh) Tomopteris spp. Chaetognatha Sagitta gazellae Ritter-Zahony S. planctonis Steinhaus Eukrohnia hamata (Mobius) Echinodermata Auricularia antarctica MacBride Ostracoda Conchoecia hettacra Miiller Copepoda Calanus acutas Giesbrecht C. propinquns Brady C. simillimus Giesbrecht Rhincalanus gigas Brady Pleuromatnma robusta (F. Dahl) Metridia gerlachei Giesbrecht Copepoda (cont.) Haloptilus ocellatus Wolfenden H. oxycephalic Giesbrecht Pareuchaeta antarctica Giesbrecht Euchirella rostromagna Wolfenden Heterorhabdus austritius Giesbrecht Candacia spp. Amphipoda Parathemisto gaudichaudi (Guerin) Primno macropa Guerin Vibilia antarctica Stebbing Eusirus antarcticus (Thomson) Cyllopus spp. Euphausiaceae Euphausia superba Dana E. frigida Hansen E. triacantha Holt and Tattersall Thysanoessa spp. Mollusca Cleodora sulcata (Pfeffer) Limacina helicina (Phipps) L. balea Moller1 Spongiobranchea australis d'Orbigny Clione antarctica E. A. Smith Tunicata Salpa fusiformis f. aspera (Chamisso) The samples (or the notes made by those who had already analysed them) were examined for the presence or absence of the genera and species on the above list. It will be seen that four organisms are listed only by generic names. These have little real value compared with identified species but nevertheless it seems worthwhile to include them as they appear to have a circumpolar distribution when grouped as genera. In the samples examined these genera are probably represented by, at most, two species each. Two common species of Tomopteris occur in the Antarctic, T. carpenteri and T. septentrionalis. It is not easy to separate them except when the former is adult. Thysanoessa spp. includes Th. macrura and Th. vicina, and again no attempt has been made to separate these two owing to the difficulty of distinguishing the younger stages ; Cyllopus spp. includes C. lucasii and C. magel- laniais; these are only easily distinguishable when fresh. Candacia spp. possibly includes more than 1 Dr J. E. Morton suggests that this is better referred to as L. retroversa balea (Fleming). The two forms intergrade very closely. 212 DISCOVERY REPORTS one species. Of these genera only Thysanoessa occurs in large numbers ; the others, although they appear regularly enough to show a circumpolar distribution, are rarely represented by more than three or four specimens in one sample. Siphonophora. Of the three Siphonophores included in the list the nectophores of Diphyes antarctica and Pyrostephos vanhoffeni are easily identified in the adult stage. The third, Dimophyes arctica was thought, until recently, to be equally straightforward, but Captain Totton (1954, p. 121) has found a new species which has previously been confused with D. arctica. It should be pointed out that the proportion of the new species to D. arctica is only small, only one being present among the twelve specimens presented to Captain Totton for examination, and that this by no means invalidates the conclusion that D. arctica has a circumpolar distribution. Anthomedusae. The single Medusan considered, Sibogita borchgrevinki is, as far as can be ascer- tained, the only species of this genus to be taken in the Antarctic. Polychaeta. The Tomopteris species have been mentioned above. In addition, Vanadis antarctica is the only species found commonly in the Antarctic surface water. Chaetognatha. The three members of this group found in the Antarctic surface water are easily distinguishable by their general appearance. Sagitta gazellae is the only flaccid species; this prevents confusion with Eukrohnia hamata which is firm, more or less opaque,1 has continuous lateral fins and has almost always a distinct oil globule in the gut (Fraser, 1952, p. 10). Sagitta planctonis is altogether a more rigid and opaque species with a stout tail region. Echinodermata. The pelagic Echinoderm larva found consistently in the Antarctic surface water is identified here as Auricularia antarctica but as the specimens have not been examined by a specialist it is not impossible that other species may be included under this heading. Ostracoda. The Ostracod Conchoecia hettacra would not at first sight appear to be easily and rapidly identified by the methods applied to the larger forms, and in fact because of its small size it has inevitably to be examined at least under a low-power lens. The shell shape and the arrangement of the shell glands then enable a confident identification to be made. This is not to say that all the members of this species including the young stages in the samples have been examined, but that sufficient specimens have been identified for the purpose of this work. Copepoda. On the whole the larger common Antarctic copepods are easy to recognize. Calamus acutus has a characteristic triangular shaped head when seen from the dorsal surface which prevents it from being confused with C. propinquus or C. simillimus. The two latter species of Calanus present certain difficulties. The adult C. propinquus is straightforward as it can be distinguished from C. simil- limus by size alone, but to avoid confusion between C. simillimus and juvenile C. propinquus examination of the fifth thoracic limb is necessary. This causes little delay where only the presence or absence of a species is to be decided, as only a few specimens need be examined, but where a count is being made the delay may be considerable. Again Pareuchaeta antarctica is readily identified when adult but confusion may occur between juvenile P. antarctica and P. biloba ; similarly Euchirella rostromagna and E. latirostris may be confused in the younger stages. Thus these two species, Pareuchaeta biloba and Euchirella latirostris are not included in the list of species under consideration. Of the remaining copepods only Candacia sp. presents any difficulty and as mentioned above this is listed only as a genus. Amphipoda. The members of this group included in the above list of species are, with the exception of Vibilia antarctica, readily identified. There was some doubt concerning the correct identification of V. antarctica and it was intended to compare specimens with the type specimen, but unfortunately 1 Remarks concerning the identification of organisms refer to their appearance after preservation in 10% sea water formalin. IDENTIFICATION OF SPECIES 213 this was not possible. However I have compared my specimens with those identified by Dr K. H. Barnard and as there is agreement the specific name antarctica has been used. As previously mentioned no attempt has been made to separate the two common species of Cyllopus. Euphausiacea. The Antarctic species of Euphausia have formed the subject of a previous Discovery Report (John, 1936) and the descriptions and figures in this enable the adults of the various species to be identified with confidence and usually with little trouble. Mollusca. The five species from this group that have been taken into consideration are, provided they are not badly damaged, easily identified by their general appearance. The shells of the two species of Limacina are rather fragile and are often broken in the samples but they are usually present in such numbers, particularly L. balea, that sufficient complete specimens may be found. Tunicata. The only Tunicate considered here is Salpafusiformis f. aspera. This is readily identified by the spinose ridges on the test. This characteristic is used by Thompson (1948) in a key to the Australian pelagic Tunicata as the distinguishing feature between S. fusiformis fusiformis and S. fusi- formis aspera. Excluding any possible errors inherent in the actual method of obtaining samples of the population with the nets, it will be seen from what has been said above that there are two sources of possible error in dealing with the plankton samples themselves. The first, the overlooking of smaller organisms, is generally corrected by the examination of additional samples. The second, error in identification, has been shown to be comparatively slight and usually to occur only in dealing with juvenile forms. Again, where only presence or absence is being considered, this has little effect, as there must be few samples containing nothing but juveniles of a species, and only one adult specimen is required for evidence of presence. DISTRIBUTION As in the section on phytoplankton, Figs. 4 and 5 show the frequency of occurrence of a genus or species in each sector expressed as a percentage of the number of stations examined in that sector. E.g. in the sector 0-200 E. Diphyes antarctica was present in twelve out of the sixteen samples examined and is thus shown as having a frequency of occurrence of 75 %. When all the 230 samples had been examined it was found that eight species still showed gaps in their distribution of between 200 and 6o° of longitude in certain areas, but it was thought that these gaps might be caused by insufficient sampling of a rather sparse population and not by their complete absence from those regions. Therefore additional samples were examined for the discontinuous species only, and, with the exception of Eusirus antarcticus, all such gaps disappeared. Where samples have been examined for only a few species these samples are not included under ' No. of samples examined ' and the presence of a species is indicated in the histograms by a pecked line. The purpose of the histograms is to show diagrammatically the circumpolar continuity of all the species whose distribution has been examined in this connexion, and this continuity is seen to hold good for all, except in so far as Eusirus antarcticus shows a gap of more than 200 of longitude. The variations in the height of the histograms are of no great significance, for they merely show the frequency of occurrence of a species in those samples which happen to have been chosen from among those available in each 200 sector. The only reason why this percentage occurrence is indicated in the diagrams is that it shows some species to be more nearly ubiquitous than others. Thus it may be noted that the following species occur in over 80 % of the total samples examined : Sagitta gazellae, Calanus propinquus, Eukrohnia hamata, Thysanoessa spp., Rhincalanus gigas and Calanus acutus. This sequence represents the order of frequency in which they were found, Sagitta gazellae being present in 96-0 % and Calanus acutus in 83-5 % of the samples. Taking the lower range of frequencies, Eusirus antarcticus and Sibogita borchgrevinki occur in only 9-5 and 15-0% of the samples respectively. The 2i4 DISCOVERY REPORTS species which show the highest percentage occurrence are not necessarily the most abundant species but rather those which are more evenly distributed. Species with a ' patchy ' distribution are liable to be absent from several samples even though they may be present in the sea not far from where the samples were taken. Again some species are confined either to the warmer or colder parts of the Antarctic water, and if, as in some instances, the available samples tend to be grouped in one part of the temperature range, a false impression is given of their occurrence in the sector as a whole. DECREES EAST DECREES WEST 0 2.0 4p 60 8p IOCI20 140 160 180 160 140 120 100 SO 60 40 20 p N9 OF SAMPLES EXAMINED SIPHONOPHORA DIPHYES ANTARCTICA DIMOPHYES ARCTICA PYROSTEPHOS VANHOFFENI ANTHOMEDUSAE SIBOCITA BORCHCREVINKI POLYCHAETA VANADIS ANTARCTICA TOMOPTERIS SPP. CHAETOGNATHA SACITTA CAZELLAE S . PLANCTONIS EUKROHNIA HAMATA ECHINODERMATA AURICULARIA ANTARCTICA MOLLUSCA CLECDORA SULCATA LIMACINA HELICINA LIMACINA BALEA SPONCIOBRANCHEA AUSTRALIS CLIONE ANTARCTICA TUNICATA SALPA FUSIFORMIS f.ASPERA Fig. 4. Occurrence of species of zooplankton in all longitudes around the Antarctic zone of the Southern Ocean, showing the percentage frequency of occurrence in samples taken within every 200 of longitude. CONCLUSIONS The zooplankton can be taken first. To begin with it is not claimed that the species dealt with here include every oceanic species of the macroplankton existing in the Antarctic surface water, and it must be made clear that there has been no selection of those which happen to have a circumpolar distribution. For this it is probably enough to point out that these species include all except two of those definitely identified by Mackintosh (1937) and listed on his p. 372 in order of abundance and CONCLUSIONS 215 without reference to their circumpolar distribution. The exceptions are Sagitta maxima and Euphansia vallentini, and these are not usually found in the Antarctic surface water. We are thus dealing with all except the rare species and a few which happen to be difficult to identify. DECREES EAST DECREES WEST O 20 40 60 80 lOO I2Q I4Q 160 ISO 160 UP I2Q IPO 6Q 60 4Q 2Q O N9 OF SAMPLES EXAMINED OSTRACODA CONCHOECIA HETTACRA COPEPODA CALANUS ACUTUS C. PROPINQUUS C. SIMILUMUS RHINCALANUS CICAS PLEUROMAMMA ROBUSTA METRIDIA CERLACHEI HALOPTILUS OCELLATUS H. OXYCEPHALIC PAREUCHAETA ANTARCTICA EUCHIRELLA ROSTROMACNA HETERORHABDUS AUSTRINUS CANDACIA SPP. AMPHIPODA PARATHEMISTO CAUD1CHAUDI PRIMNO MACROPA VIBILIA ANTARCTICA EUSIRUS ANTARTICUS CYLLOPUS SPP. EUPHAUSIACEA EUPHAUSIA SUPERBA E. FRICIDA E . TRIACANTHA THYSANOESSA SPP. 16 II 12 14 13 12 10 IS 12 IS 14 12 13 II 13 II II IS % -100 - so -100 - 50 —100 - so - 0 -100 - so -100 - 50 — 0 —100 - so —100 - so - 0 —100 - so J- —100 - 50 - 0 —100 - so - 0 —100 - so - 0 — 100 — so — 0 — 100 — so . —100 — 0 —100 - so —100 - so — 0 — 100 — so —100 - so —100 - so — 100 - so — 100 - 50 r 5° ■— 0 Fig. 5. Continuation of Fig. 4. It has been shown that, with the exception of Eusirus antarcticas, each of these species has been found at one time or another at least once in every 200 sector round the whole circle of 3600 south of the Antarctic Convergence. All of them, including E. antarcticus can properly be described as having a circumpolar distribution; and there can scarcely be any doubt that such gaps as appear in this species would be eliminated with further sampling. It is thus reasonable to infer that this circumpolar 216 DISCOVERY REPORTS distribution is continuous in the sense that there are no longitudes from which the species are excluded. This is not quite the same as to say that each species is represented by a single completely continuous circumpolar population throughout the year, for it is not impossible that at a time of year when the numbers of a species are reduced in the surface layer the population becomes broken up into isolated local stocks. It seems safe, however, to conclude that there is no persistent discontinuity in any longitudes. The distribution of the samples in time and space hardly leaves room for the possibility of several discrete populations drifting round through all longitudes but remaining separate from each other, and in any case it would be difficult to conceive of any mechanism in the water movements by which they could be kept permanently isolated from one another. To demonstrate on taxonomic grounds alone that there can be no reproductively isolated population of a species in different parts of the Antarctic it would be necessary to prove that there are no sub- specific differences in different sectors ; and in this connexion it must be remembered that in a wide- spread species distance alone may constitute a sufficient barrier to bring about genetic isolation and the development of subspecific differences between remote parts of the range, even in a continuous system of populations. No proof is offered here that the representatives of each species are everywhere subspecifically identical, but it can be said that no such differences have been observed, and in view of the uniformity of the properties of the Antarctic surface layer in a circumpolar direction, and the absence of any known environmental barriers which might interrupt the continuity, there is good reason to believe that each species has a circumpolar population forming a single unit in which little if any geographical variation can be expected. This does not of course rule out the possibility of two or more concentric circumpolar populations occupying different zones, partially isolated from one another and perhaps exhibiting subspecific differences. Very much the same can be said of the species of phytoplankton. Here rather more gaps appear in the histograms (in Fig. 2) than in those for the zooplankton, as Dr Hart adopted a rather more rigorous method of selecting the samples. The distribution of all the species can be described as circumpolar, it is certainly or almost certainly continuous in most of them (at least in the period of relative abundance), and it is highly probable that apparent gaps in the remainder would disappear with further sampling. The species dealt with in this paper do not include the smaller zooplankton forms, but they are an arbitrary selection of the total plankton in the sense that they are the common species taken in certain types of nets, and they include representatives of many taxonomic groups with different habits and modes of life. Evidence has been given that all of them are circumpolar, and there is at least a strong expectation that circumpolar continuity is a general rule for oceanic plankton species of the Antarctic surface layer. SUMMARY It has been suggested that, as the Southern Ocean is an uninterrupted circumpolar belt with more or less uniform conditions prevailing in east and west directions, the range of planktonic species of the Antarctic surface water may be expected to extend as far as these uniform conditions persist, i.e. to be circumpolar. Dr Hart has shown this to be true of the important phytoplankton species and discusses the reasons for the few gaps that appear when the selected data are plotted. The same is shown for the larger zooplankton species and an indication of the degree of accuracy of the specific identification is given. Histograms are drawn which demonstrate the occurrence of each species in a circumpolar direction and show the percentage of samples in which the species occur. BIBLIOGRAPHY 217 Table 3. Serial numbers of the stations on which Figs. 2, 4 and 5 are based 0— 20- 40- 60- 80- 100- 120- 140- 160E.- 180- 160- 140- 120- 100- 80- 60- 40- 20 W. 20° E. 40 ° E. 6o° E. 8o°E. ioo°E. I20°E. 140 E. i6o°E. 1800 i6o°W. 140° W. 120° W. ioo°W. 8o° W. 60 W. 4o°W. 20°W. o° Phytoplankton 2496 2089 1543 1627 2128 2159 2174 2187 1663 1283 1263 1255 571 566 1222 WS470 479 WS541 2498 2090 2101 1629 2129 2160 2175 2188 2198 2220 1265 2241 572 568 1223 WS471 480 WS543 2500 2091 2103 2116 2131 2161 2177 2189 2199 2221 1267 2242 1247 577 1224 1213 485 WS547 2502 2092 2104 2117 2132 2162 2178 2190 2200 2222 2230 2244 1249 1 24 1 1235 1214 486 WS 550 2504 2093 2106 2119 2134 2163 2180 2192 2201 2223 2232 2245 1251 1243 1236 1215 518 WS 552 2508 2094 2107 2122 2137 2164 2181 2193 2202 2224 2233 2247 1253 1245 1237 1491 523 WS 554 25 1 1 2095 2109 2123 2138 2165 2.83 2195 2204 2225 2235 2248 2254 1472 1239 1492 525 1513 2513 2096 2110 2125 2141 2166 2184 2196 2205 2227 2236 225O 2255 1473 1447 1493 1 198 1515 2517 2098 2113 2126 2142 2167 2185 2206 2229 2238 2251 2256 1474 1449 1494 1212 1517 2519 2100 2114 2127 2143 2168 2186 2207 2239 2253 2258 1475 1450 1495 1510 1519 Zooplankton 4S2 1357 856 8S8 862 1720 883 904 1662 1270 1263 1254 WS 503 575 549 547 WS 310 WS543 1161 1532 857 859 863 1725 884 9°5 1663 1271 1265 1255* 973* 975* 1216 6i3 362 461 1356 1533 1543 860 865 1726 885 906 1665 1273 1267 1256 1246 1240 1220 1213 532 1351 1521 1534 1546 861 1634 1727 886 1695 1667 1274 1287 1257 1248 1241 1221 1214 1198 1352 1525 1538 1547 1626 1642 1729 887 1697 1669 1282 1289 1258 1250 1242 1223 1326 1208 1383 1527 1540 1621 1627 2129 1722* 889 1700 i675 1283 1291 1259 1252 1244 1238 1327 1344 1385 1528 1548 1623 1628 2130 2161 890 1701 1677 1284* 1292 1294 1253 1306 1239 1328 1390 1386 1784 1550 2103 1629 2131 2163 891 2187 1679 1285 1293 1295 1301 1307 1314 1332 1501 1514 1785 2091 2106 1630 2134 2165 2175 2189 2201* 1647 2230 1296 i3°5 1309 1315 '337 1510 1515 1786 2097 2108 1631 2140 2166 2179* 2190 2202 1648 2232 1297 1464 1310* 1412 1492 15" 1519 1796 2374 2109 '633 2142 2169 2181 2191 2206 1651 2234 1299 2256 1311 1414 2298 2302 1520 1812 2110 2121 2810 2815 2192 2764 2758 2237 2835 2258 2269 2286 2006 2014 2123 2813 2816 2193 2767 2759 2741 2836 2260 2271 2287 2008 2016 2125 2195 2760 2832 2837 2009 2017 2197 2761 2309 2477 2763 Note. Further particulars of these stations (excluding the more recent Stns. 2758-2837) can be found in the Station Lists published in the Discovery Reports, Vols, in, pp. 1-132, iv, pp. 1-232, xxi, pp. 1-226, xxil, pp. 1-196 and xxiv, pp. 1-422. Stations marked with an asterisk are those from which samples were examined for one or two species only (see p. 213). BIBLIOGRAPHY Clowes, A. J., 1938. Phosphate and Silicate in the Southern Ocean. Discovery Rep., xix, pp. 1-120. Deacon, G. E. R., 1933. A General Account of the Hydrology of the South Atlantic Ocean. Discovery Rep., vn, pp. 171-238. 1937. The Hydrology of the Southern Ocean. Discovery Rep., xv, pp. 1-124. Fraser, J. H., 1952. The Chaetognatha and other Zooplankton of the Scottish Area and their value as Biological Indicators of Hydrological Conditions. Scottish Home Dept., Marine Research, 1952, No. 2, pp. 1-52. Hart, T. J., 1942. Phytoplankton Periodicity in Antarctic Surface Waters. Discovery Rep., xxi, pp. 261-356. John, D. D., 1936. The Southern Species of the Getius Euphausia. Discovery Rep., xiv, pp. 193-324. Kemp, S., Hardy, A. C. & Mackintosh, N. A., 1929. Discovery Investigations: Objects, Equipment and Methods. Discovery Rep., 1, pp. 141-232. Mackintosh, N. A., 1934. Distribution of the Macroplankton in the Atlantic Sector of the Antarctic. Discovery Rep., ix, pp. 65-160. 1937. The Seasonal Circulation of the Antarctic Macroplankton. Discovery Rep., xvi, pp. 365-412. 1946. The Antarctic Convergence and the Distribution of Surface Temperatures in Antarctic Waters. Discovery Rep., xxiii, pp. 177-212. Thompson, H., 1948. Pelagic Tunicates of Australia. Commonwealth Council for Scientific and Industrial Research, Australia (Melbourne). Totton, A. K., 1954. Siphonophora of the Indian Ocean, together with systematic and biological notes on related specimens from other oceans. Discovery Rep., XXVII, pp. 1-162. [Discovery Reports. Vol. XXVII, pp. 219-234, October 1954.] THE PLANKTONIC DECAPOD CRUSTACEA AND STOMATOPODA OF THE BENGUELA CURRENT PART I. FIRST SURVEY, R.R.S. 'WILLIAM SCORESBY', MARCH 1950 By MARIE V. LEBOUR, D.Sc. CONTENTS Introduction P<*ge 221 Acknowledgements 221 Systematic 221 Penaeidea 221 Caridea 223 Thalassinidea 226 Scyllaridea 227 Anomura 228 Brachyura ........■■•• 230 Stomatopoda 231 Literature 232 THE PLANKTONIC DECAPOD CRUSTACEA AND STOMATOPODA OF THE BENGUELA CURRENT PART I. FIRST SURVEY, R.R.S. 'WILLIAM SCORESBY', MARCH 1950 By Marie V. Lebour, D.Sc. (Text-figs. 1-6) INTRODUCTION THE R.R.S. 'William Scoresby' made two surveys in 1950 of the Benguela Current off the coast of South-West Africa. The plankton hauls so obtained were sorted by Dr T. J. Hart, who handed the Decapoda and Stomatopoda over to me. The material from the first survey (March, Stations WS 977-1002) forms the basis of this report. The second survey will be dealt with subsequently. The adult specimens consisted of two species of Gennadas and one of Pasiphaea. The rest were all larvae. One of the commonest of these was a species of Callianassa which occurred in several hauls, and the most frequent crab larva was a species of Ebalia, closely related to E. tuberosa. A pagurid larva was of special interest as it represents a type hitherto undescribed. It is here compared with one from Bermuda, very closely related and so far unknown. The Stomatopod larvae all belong to one species and are almost certainly Squilla armata which is the commonest form in these regions. ACKNOWLEDGEMENTS My best thanks are due to Mr F. S. Russell, Director of the Marine Laboratory for allowing me to work at Plymouth. SYSTEMATIC Penaeidea Penaeidae _ L. ,~ , . Aristeinae Gennadas guchristi (Caiman) Caiman, 1925, p. 6. Balss, 1927, p. 261 (as Amalopenaeus). Occurrence. St. WS 997, 1 000-0 m. Distribution. Known from near the Cape of Good Hope and in the East Atlantic. Gennadas similis Stephensen Stephensen, 1923, p. 12. Balss, 1927, p. 256 (as Amalopenaeus). Occurrence. St. WS 996, 1000-750 m. Distribution. South-east and East Atlantic. The original locality is the Bay of Cadiz (Stephensen). Gennadas sp., 1 protozoea Heldt, 1938, p. 122. Occurrence. St. WS 997, 50-0 m. Table i . List of Stations at which Decapods and Stomatopods were taken Station Position Date Depth in in. Species Larvae WS977 22°39'S. 120 16'E. to 6/7. iii. 50 50-0 ? Sergestes cornutus 1 elaphocaris I— II 22° 33' S. 120 17' E. 1 Sergestes (cf. arcticus) 1 mastigopus WS978 22°28'S. I2°42'E. 7. iii. 50 1000-750 Acanthephyra acanthetelsonis 1 late or post-larval WS 979 220 47' S. 130 35' E. 7. iii. 50 50-0 100-50 Callianassa Solenocera (cf. membranacea) Brachyrhyncha indet. Hippolytid (cf. Eualus sp.) Processa 1 Pasiphaea semispina (juv.) Hippolytid indet. Brachyrhyncha indet. Callianassa Processa 15 larvae I-V 1 protozoea III 3 zoea I 2 larvae I and V 1 larva VIII or IX 1 larva 1 zoea II 1 larva I 1 larva IV or WS980 22° 44' S. I40 08' E. 8. iii. 50 50-0 Solenocera (cf. membranacea) Solenocera (cf. membranacea) Brachyrhyncha indet. 1 protozoea II 2 protozoeae III 3 zoeae II, III and V 100-50 Callianassa Brachyrhyncha indet. 2 larvae (early) 1 zoea I WS987 25° J3' S. i3°43'E. 10. iii. 50 50-0 250-100 3 Pasiphaea semispina 1 Pasiphaea semispina (ovigerous) WS988 25° 12' S. 140 22' E. 11. iii. 50 50-0 Brachyrhyncha indet. 1 zoea I WS996 28°4i'S. i3°2s'E. to 12. iii. 50 100-50 Periclimenes? sp. 1 late larva 280 39' S. 130 23' E. Ebalia 1 zoea III or IV 250-100 Ebalia 1 megalopa ~ 1000-750 Sergestes arcticus 1 Gennadas similis (adult) 1 elaphocaris III WS997 280 40' S. 140 06' E. to 13. iii. 50 50-0 1 Sergestes sp. (juv.) (cf. diapontius) 1 mastigopus (juv.) 280 37' S. 140 08' E. Gennadas sp. Ebalia 1 protozoea 4 megalopae 250-100 Ebalia Ebalia 1 zoea IV 1 megalopa 1000-750 Ebalia 1 zoea I 1 000-0 1 Gennadas gilchristi (adult) Glaucothoe 2 larvae WS998 280 40' S. 140 43' E. 13. iii. 50 50-0 Ebalia Ebalia Brachyrhyncha indet. Callianassa 18 zoeae 1 megalopa 2 zoeae 25 larvae I-V 100-50 Callianassa Crab indet. Acanthephyra 1 larva I 1 megalopa 1 larva 175-100 Callianassa Ebalia 1 larva I 3 megalopae WS999 280 38' S. 140 59' E. to 13. iii. 50 50-0 Ebalia 1 megalopa 28°3i'S. is°oo'E. Callianassa Brachyrhyncha indet. 9 larvae II— III 1 zoea IV 100-50 Hippolytid (cf. Eualus) Callianassa 1 larva c. IV 1 larva III 150-100 Callianassa Ebalia 1 larva V 1 megalopa WS 1000 28°4o'S. i5°29'E. 13. iii. 50 50-0 Jasus Stomatopod 4 phyllosoma I 2 larvae WS 1001 28°4o'S. i5°56'E. 14. iii. 50 50-0 Stomatopod Callianassa Pagurid indet. ? Hippolytid lEualas sp. (cf. lebourae) 19 larvae 28 larvae I-V 1 larva I 1 larva I 1 post larva 100-50 Callianassa Ebalia Stomatopod 4 larvae I— III 1 megalopa 3 larvae WS 1002 28°4o'S. 1 6° 14' E. 14. iii. 50 50-0 Portunid Callianassa Hippolytid (cf. Eualus) Stomatopod 1 megalopa (large) 7 larvae I-IV 3 larvae 4 larvae SYSTEMATIC 223 SOLENOCERINAE Solenocera membranacea (Milne-Edwards) Heldt, 1938, p. 125. Occurrence. St. WS 980, 50-0 m., 1 protozoea, stage II; 2 protozoeae, stage III. Sergestidae ,, .. Sergestinae Sergestes cornutus Kroyer Gurney & Lebour, 1940, pp. 13-19. Occurrence. St. WS 977, 50-0 m., 1 elaphocaris, stage I changing to stage II. Sergestes sp., ef. arcticas Kroyer Hansen, 1922, p. 62. Occurrence. St. WS 977, 50-0 m., 1 mastigopus, late larval or juvenile. Sergestes arcticus Kroyer Gurney & Lebour, 1940, pp. 19-21. Occurrence. St. WS 996, 250-100 m., 1 elaphocaris, stage III. Sergestes diapontius Bate Hansen, 1922, p. 172. Occurrence. St. WS 997, 50-0 m., 1 mastigopus, late larval or juvenile, c. 15 mm. long. The specimen agrees with diapontius in the shape of the third maxillipede. It is apparently inter- mediate between diapontius and vigilax. I am inclined to place it in diapontius as this species is known to have a more southerly distribution than vigilax. Caridea hoplophoridae Acanthephyra acanthetelsonis Bate Kemp, 1939, p. 574. Occurrence. St. WS 978, 100-50 m., 1 larva, stage V; 1000-750 m., 1 post-larva or juvenile. The specimen from 1000-750 m. is probably a post-larva as it closely resembles the post-larva of Acanthephyra purpurea Milne-Edwards from Bermuda (see Lebour, in Gurney & Lebour, 1941). It is, however, clearly distinguishable by the number of dorso-lateral spines (in this case fourteen each side), the number in A. acanthetelsonis, according to Kemp, ranging from thirteen to nineteen. The adult characters are not all present, but the dorsal spines on the third to the sixth abdominal somites and the large hump on the third, continuing in the dorsal carina, are conspicuous. There are five spines dorsally on the rostrum and one below as in the post-larval A. purpurea, and the rostrum hardly reaches to the end of the eyes. Holthuis (1951) has recorded this species from West Africa (Gold Coast and Liberia) and Kemp gives its distribution as 'Central and South Atlantic about 140 N. to 280 S.' 224 discovery reports Pasiphaeidae Pasiphaea semispina Holthuis Holthuis, 1951, p. 9; 1952, p. 26. Occurrence. St. WS 987, 50-0 m., 3 specimens, c. 35-40 mm. long; 250-100 m., 1 specimen, c. 45 mm. long, in berry. This species was described from a single male of 66 mm. by Holthuis (1951) from off Angola (7° 35' S-> I2° 38' E> 235~46° m- depth, bottom mud, 17. iii. 46). Later, Holthuis (1952) recorded several more specimens, including females, from the South Atlantic. These records range from 50 52' S., ii° 43' 30" E. to io° 45' S., 130 E., therefore all came from much farther north than the 'William Scoresby's', which were collected between 200 and 260 S., thus greatly extending the known range of the species. The ' William Scoresby ' specimens came from depths of 50-0 to 250-100 m. and were 35-45 mm. long; Holthuis's from 74-500 m. and measured up to 70 mm. in length. The species apparently does not inhabit great depths. A young pasiphaeid, 7 mm. in length, occurred at St. WS 979, 100-50 m. It is almost certain that it belongs to this species for the spines on the fourth and sixth abdominal somites are beginning to show, the rostrum and appendages are similar and the merus of the first and second legs have no spines. The sixth abdominal somite is very long, a good deal longer than the fourth and fifth combined, but it frequently happens in decapod larvae that the length of this somite is longer in the young than in the adult. It seems likely that there are only a few stages between this and the newly hatched larva. The telson bears eight setae at its end but is slightly concave instead of convex at the margin. HlPPOLYTIDAE Several larvae are present but it is difficult to place the young stages, most of which resemble Hippolyte or Eualus. One first stage differs in having a distinct hump on the third abdominal somite and a very much elongated telson, and this may not be an Hippolytid at all (St. WS 1001, 50-0 m.). Two Hippolytid larvae indet., cf. Eualus (St. WS 979, 50-0 m.) and one Hippolytid larva indet. (St. WS 979, 100-50 m.) are at stage I, with humps on the third abdominal somite and a long telson, as is the larva from St. WS 1001 (50-0 m.). One Hippolytid post-larva, possibly Eualus lebourae, was described by Holthuis (1951) from West Africa. This is not unlike the post-larval stage of E. occulta (Lebour, 1936a). It is possible that the larva from St. WS 1001 may be a first young stage of this species {occulta) as the exopods have disappeared from the legs. There are two dorsal teeth on the rostrum, the telson has three pairs of dorso-lateral spines and three pairs at the end which is pointed. The second leg is incompletely segmented but there are three distinct divisions and an indefinite number (possibly four) not properly formed. The antennule is like that of E. lebourae but more elongated. Three Hippolytid larvae, stages II-IV, possibly Eualus, occurred at St. WS 1002, 50-0 m. Processidae Occurrence. St. WS 979, 50-0 m. 1 Processa larva, stage VIII or IX. St. WS 979, 100-50 m. 1 Processa larva, stage IV or V. These larvae are of the Processa edulis type with no dorso-lateral spines on the fourth abdominal somite (Lebour, 19366). Holthuis (195 1) describes several new species of Processa from the East Atlantic, not far from the present region. His P. intermedins seems to be the closest to P. edulis. SYSTEMATIC 225 Palaemonidae pontoniinae Periclimenes, sub-gen. Periclimenes sp. Occurrence. St. WS 996, 100-50 m., 1 late larval stage, Fig. 1. This late larval stage of a pontoniid is very interesting and apparently belongs to a species of Periclimenes, sub-gen. Periclimenes. It bears a close resemblance to Gurney's ' larva of an unknown genus' (Gurney & Lebour, 1941, p. 117, fig. 9m, n) from 'Discovery' Station 278 off North Africa, and probably belongs to the same genus. These peculiar larvae have the very unusual feature of a long antennular flagellum which surpasses the length of the body. These two specimens, belonging to Fig. 1. Periclimenes sp., late larval stage, c. 10 mm. long, a, showing complete antennule; b, the same, enlarged, antennular flagellum cut off; r, telson. separate species and possibly to separate genera, are apparently the only known instance of such a phenomenon. Gurney included his figure in order to compare it with Rhinchocinetes rigens, which in the late larval stages possesses an extremely long antennal flagellum comparable with the present antennular flagellum. In Gurney's figure the flagella are figured in each case (fig. 9//, m) as loosely held at the side and behind, but in the live Rhinchocinetes larva observed by myself (Gurney & Lebour, 1941, p. 119) the antennal flagella, nearly four times the length of the body, were carried straight out in front, the two together closely approximated, and having a ram-like appearance. In the present larva here described, the antennular flagellum is carried straight in front in an exactly similar way (Fig. 1 a). The purpose of these analogous organs is probably to help suspend the animal in the water-layer in which it habitually swims. Gurney's figure m and the present larva agree in many features besides the long antennular flagellum, and almost certainly belong to the same genus although they differ enough to be placed in separate species. The present larva is bent at the third abdominal 226 DISCOVERY REPORTS somite, but the body is apparently straight in Gurney's larva. In both, the rostrum is slender and unarmed and reaches beyond the eyes for about the length of the eye, the antennal scale reaches nearly to the end of the inner antennular branch, there are as yet no chelae on the first and second legs, there are very long setose exopods on the third maxillipede and on legs i and 2, in all cases longer than the endopods, and there are no exopods on legs 3-5 which are as yet not fully developed but are all of about equal length. The pleopods are small buds in Gurney's larva, but are slightly setose in the present specimen which is at a later stage. The sixth abdominal somite is as long as the third to the fifth inclusive. The sides of somites 1-5 are slightly pointed. In the present larva the third abdominal somite has a small median tooth and just behind it another tooth. These are apparently not present in Gurney's larva. He gives no size for the latter. The present larva measures nearly 10 mm. in length from the tip of the rostrum to the end of the telson, the antennular flagellum measuring 12-45 mm- m length. The telson is nearly straight-sided, the end being slightly indented in the centre and rounded on each side. It is armed with two pairs of dorso-lateral teeth and six setae terminally on each side of a small knob, the outer tooth small, the inner teeth setose on the inside only. A comparison with known larvae of this sub-genus shows that it must belong here. Very little is known of the larval stages of the sub-genus Periclimenes, genus Periclimenes. We do, however, know one, the last larval stage of a form attributed with a query to Periclimenes {Periclimenes) longicaudatus (Stimpson) (Gurney & Lebour, 1941, p. 146) which was followed through four moults to the third stage and was shown to belong undoubtedly to this sub-genus. Lebour (1949a) described the newly hatched larvae of Periclimenes {Periclimenes) iridescens Lebour. These are the only larvae so far known which undoubtedly belong to this sub-genus, and the late larva differs much from Ancylocaris whose late larvae are also known. The features of these indicate that they belong to some species of the sub-genus Periclimenes of the genus Periclimenes and a com- parison with the last larva of P. longicaudatus {}), shows certain features in common. For example, the body is bent at the third somite (as in the present larva), the fourth and fifth legs are short and similar and there are no exopods on legs 3-5. This last feature is, however, perhaps not of generic value as in many larvae of various genera, the number of exopods on the legs varies. The telson is of a palaemonid type and very like that of the larva of P. longicaudatus {}), and also like some of the larvae attributed in 1924 by Gurney to the Palaemonidae. It is of course impossible to be sure of the place of this larva, but it seems almost certain that it is a palaemonid of the sub-family Pontoniinae and belongs to the genus Periclimenes sub-genus Peri- climenes. There are several species of this sub-genus recorded from West Africa, but it may quite likely belong to a new species. Thalassinidea Callianassidae Occurrence (Stations). WS 979, 50-0 m., 15 Callianassa larvae, stages I-V. WS 979, 100-50 m., 1 Callianassa larva, stage I. WS 980, 100-50 m., 2 Callianassa larvae, early. WS 998, 50-0 m., 25 Callianassa larvae, stages I-V. WS 998, 100-50 m., 1 Callianassa larva, stage I. WS 998, 175- 100 m., 1 Callianassa larva, stage I. WS 999, 50-0 m., 9 Callianassa larvae, stages II— I II . WS 999, 100-50 m., 1 Callianassa larva, stage III. WS 999, 150-100 m., 1 Callianassa larva, stage V. WS 1001, 50-100 m., 28 Callianassa larvae, stages I-V. WS 1001, 100-50 m., 4 Callianassa larvae, stages I— III. WS 1002, 50-0 m., 7 Callianassa larvae, stages I-IV. A large number of Callianassa larvae (Fig. 2) were obtained, apparently all of the same species, but varying from stage I to stage V. They are of the type I of Gurney (1942) to which C. subterranea belongs (sub-genus Cheramus). The sub-genus Trypaea has a similar larva, those of T. affinis from SYSTEMATIC 227 California (Lebour, 1938) and T. australiensis from New South Wales (Dakin & Colefax, 1940) being known from the parent. There are species from South Africa which apparently belong to Trypaea and which were recorded by Barnard (1950), but the only callianassid, Callianassa guineensis, which undoubtedly comes from the region of the Benguela Current, was recorded by De Man (1928), and this belongs to the sub-genus Calochirus, whose larva is of a different type (type II of Gurney). It would seem therefore that the present larvae must belong to a species of Trypaea or Cheramus, and one that is common in the region. Fig. 2. Callianassa sp. St. WS 998, 3. x. 50. a, b, stage I, length 4-2 mm.; c, d, e, stage II, length 5 mm.;/, g, stage III, length 5-5 mm. Fig. 3. Jasus lalandii, stage I, length 1-5 mm. SCYLLARIDEA , , ... ,T , s Palinuridae Jasus lalandu (Lamarck) Gilchrist, 1916, p. 101. Gurney, 1936, p. 416. Occurrence. St. WS iooo, 50-rO m., 1 phyllosoma, stage I. St. WS 992, 100-0 m., 1 phyllosoma, ? stage VIII. Jasus lalandii is the common crawfish of the district. Phyllosoma, stage I, measured 1-5 mm. from the front to the end of the telson. It has been well described by Gilchrist (1916). A peculiarity of the phyllosomas of Jasus, unlike the other Palinuridae, is the absence of an exopod on the first maxillipede, except for a minute rudiment in some of the later stages. Gurney (1936) specially notes this as a character of all known phyllosomas of the Scyllaridae. The first stage of Jasus (which has been shown by Gilchrist to emerge from a naupliosoma) is very like some of the known first stages of Scyllarus 22g DISCOVERY REPORTS and Scyllarides, so much so that one is inclined to regard them as very closely related. In 1949* I described and figured a phyllosoma from Bermuda which I attributed to Scyllarides sp., stage I. Although like Scyllaris, in many ways it was more developed, especially the antennae which were segmented and branched, the telson also being much more advanced. In many ways it resembled Jasus which has not been recorded from Bermuda, but the telson of Jasus is different and the antennae reach beyond the eyes, the two branches being much longer. The third leg is also much longer, the thorax broader than long (in the Bermuda species it is pear-shaped and longer than broad). At least two species of Scyllarides occur in Bermuda and it is probable that my S. phyllosoma B belongs to one of these. The similarity to the Jasus larva is, however, very striking. Gurney's later stage of Jasus was 11 mm. long and was about stage VIII (Gurney, 1936, p. 421). He suggested that his specimen was about six weeks old. It was captured at a depth of 200-0 m. One of Gurney's specimens of this size was taken 900 miles from the nearest coast. The present specimen came from 93 miles off-shore (Fig. 3). Anomura Paguridea Paguridae Three species of pagurid larvae occur in the collection. The first represented by one damaged specimen only is of the Eupagurus type and is in the second stage (St. WS 1001, 50-0 m.). The second is repre- sented by two larvae of the Glaucothoe peronii Milne-Edwards type (see Gurney, 1942, p. 259) (St. WS 997, 1000-0 m.). It has been suggested by Bouvier (1905) and Thompson (1943) that these large Glaucothoes belong to the genus Sympagurus, the larva agreeing in many ways with the adult of this genus. It is probable that it belongs to some deep-water genus and not to a shore form. The third larva represents a type of pagurid larva not hitherto described. Two specimens were obtained from St. WS 1001, 50-0 m., two specimens from St. WS 1047, 100-0 m., and three speci- mens from St. WS 1047, 100-0 m. The two last hauls were made during the second survey, the material from which has not yet been fully worked out. All these larvae are in the first stage. One specimen only of a similar larva, but of a different species obviously closely related, was obtained in one of Dr Moore's hauls in Bermuda (ioo-i5ofms.). These elongated larvae belong to some pagurid, but are not at present identifiable with any known adult. They are peculiarly interesting as they differ entirely from any known form and they are worth a detailed description. The drawings of the Bermuda specimen were shown to the late Dr Gurney who suggested that the larva might belong to one of the symmetrical genera. The characters which these first larva have in common are the great length, elongated body, long rostrum, carapace rounded posteriorly with conspicuous grooves, antenna with a large spine at the base of the exopod, the endopod with three terminal setae, abdominal somites 2-5 with long curved latero-ventral spines directed backwards, telson deeply incised with the usual hair-like second seta of the pagurids and their allies, rounded knobs on rostrum and carapace, and on part of the abdomen and telson. I shall call these larvae from the Benguela Current and from Bermuda, larva A and B respectively. Larva A. Five specimens from the Benguela Current (Fig. 4). The colour does not show. Length of body 6 mm., transverse groove on carapace definite but imperfect. Rostrum reaching well beyond eyes for about twice the eye length, knobs on each side of the margin. Knobs scattered on carapace, not showing much on abdomen but present near the margin of the telson. Abdominal somites 2-5 with SYSTEMATIC 229 small teeth on each side above the large recurved spines. Telson deeply indented but not so broad as in larva B. Mandible, maxillule and maxilla of the usual pagurid type except that the posterior lobe of the exite of the maxilla has four setae, the end one on the posterior margin, which is unusually bare. Fig. 4. Pagurid larva A, St. WS 1001-1, 50-0 m., 14. iii. 50, length 6 mm. a, side view; b, head; c, telson; d, antenna; e, mandible;/, maxillule; £, maxilla. Fig. 5. Pagurid larva B, Dr Moore's cruise, St. 101, Dec. 1938, Bermuda, length 6 mm. a, dorsal; b, side view; c, antennule; d, antenna. Larva B. From Bermuda. Length 6 mm. (Fig. 5). Colour brilliant red with the main background yellow (as this was preserved but examined soon after capture, the colour may be slightly different in life). Transverse grooves on carapace very conspicuous. Rostrum similar to larva A with knobs along the sides and along margin of carapace and also on the sides. Knobs less conspicuous than in larva A 23o DISCOVERY REPORTS but here they are also along the abdomen. Telson more slender with a longer first spine, eye much larger. Abdominal somites pointed posteriorly but without small teeth above the lateral spines. Main spine of upper lobe of maxillule very conspicuous. It seems certain that these larvae are pagurids but of a type hitherto unknown. Gurney (1924, p. 156, fig. 62) figured and described a larva attributed to the Laomediidae, but later (1938, p. 156) regarded by him as a pagurid, which has features in some way resembling the present larvae. The abdominal somites 2 and 3 have backwardly directed spines, the rostrum is not unlike, the telson deeply indented. This larva of Gurney's is at stage III and therefore in many ways different, but the exite of the maxilla is armed on the posterior lobe in a similar way to the Bermuda larva. These larvae all show the thalassinid characters well and seem to emphasize the resemblance, which Gurney pointed out (1938) between the Upogebia and the Laomedia type of thalassinid larvae and the pagurid larvae. There are other pagurid larvae (at least three species) to be found in Bermuda, all of which have a deeply hollowed telson in the first stage and two of these are much alike. These were attributed to Clibanarius, as that is the commonest genus in Bermuda, moreover the C. misanthropus larva described and figured by various authors (Williamson, 1915 ; Hesse, 1876; etc.) is of this type. Although differing in many ways, especially in size, from the present larvae, they have certain features in common — telson much indented, rounded carapace, long and broad rostrum, endopod of antenna with three setae. It is probable that they all belong to the same group, and as Clibanarius belongs to the Eupagurus group, it follows that Gurney's remarks on the larvae do not apply here. We know too little of the pagurid larvae to divide them into groups with any certainty — and study of the newly hatched larvae of any known species is much wanted. Brachyura Occurrence (Stations). WS 979, 50-0 m., 3 Brachyrhyncha zoeae indet, stage I. WS 979, 100- 50 m., 1 Brachyrhyncha zoea indet., stage II. WS 980, 50-0 m., 3 Brachyrhyncha zoeae indet. These are peculiar in having a very large telson spine externally on the fork. WS 980, 100-50 m., 1 Brachyrhyncha zoea indet., stage IV. WS 996, 100-50 m., 1 Ebalia zoea, 3rd and 4th stages. WS 996, 250-100 m., 1 Ebalia megalopa. WS 997, 50-0 m., 4 Ebalia megalopae. WS 997, 250- 100 m., 1 Ebalia zoea, 4th stage, 1 Ebalia megalopa. WS 997, 1000-750 m., 1 Ebalia zoea, 1st stage. WS 998, 50-0 m.; 18 Ebalia zoeae (1st to last), 1 Ebalia megalopa, 2 Brachyrhyncha zoeae indet., 1st stage. WS 998, 100-50 m., 1 Brachyrhyncha megalopa indet. WS 998, 175-100 m., 3 Ebalia megalopa. WS 999, 50-0 m., 2 Ebalia zoeae, 1 Ebalia megalopa, 1 Brachyrhyncha zoea indet., 4th stage. WS 999, 150-100 m., 1 Ebalia megalopa. WS 1001, 100-50111., 1 Ebalia megalopa. WS 1002, 50-0 m., 1 large Portunid megalopa indet., having the last legs with setose swimming paddles. A number of crab zoeae occur in the collection and some megalopae. The commonest of these, both zoeae and megalopae, belong to a species of Ebalia. The remainder are nearly all Brachyrhyncha, some of which can be identified as portunids. Adult Ebalia live on the bottom and these larvae closely resemble E. tuberosa (Pennant), but differ from it in having more setae on the last pleopods of the megalopa (Lebour, 1928). No colour was present in the preserved specimen of the zoea and consequently it could not be distinguished from that species. E. tuberosa is widely distributed and has been recorded from the Atlantic as far as the Azores and Canaries. It may well be that the present form is closely related to it. SYSTEMATIC 231 Stomatopoda Five adult Squilla, identified as S. armata Milne-Edwards and confirmed as such by Dr I. Gordon of the British Museum, were taken at St. WS 990, net. o.T.c, depth 128 m. Position: from 250 35' S., 140 27' E., to 250 36' S., 140 24' E., off west coast of Africa. Occurrence (Stations). WS 1000, 50-0 m., 2 larvae (intermediate stage). WS 1001, 50-0 m., 19 larvae (from first pelagic stage to latest seen, presumably stage IX). WS 1001 , 100-50 m., 3 larvae (1 stage IV, 1 stage VII, 1 stage VIII). WS 1002, 50-0 m., 4 larvae (intermediate stages). Fig. 6. Squilla armata}, St. WS 1001, 14. iii. 50, 50-0 m. a, alima, first pelagic stage, length 47 mm. ; b, late alima, probably last (rostrum and posterior carapace spine broken), length 20 mm.; c, the same, raptorial claw; d, uropod; e, end of telson. Several Stomatopod larvae were obtained in the plankton round about the area at which the adult Squilla armata occurred, and almost certainly belong to this species (Fig. 6). The oldest larva had the raptorial claw with seven teeth, some of which were still covered but were quite distinct. The uropods are also very like this species. The oldest larva, c. 20 mm. long, is either at the last or penultimate stage and no older specimens were seen. The other stages range from the first pelagic stage, through several intermediate stages to this late stage. They correspond very closely as a series to those of 232 DISCOVERY REPORTS Squilla desmarestii Risso, as described by Giesbrecht (1910), and are possibly identical with Alima bidens of Claus (1871) and of Brooks (1886) which, according to Gurney (1945), are probably synonymous with A. longicandata Jurich (1904). The larvae are certainly much alike except that A. bidens has no dorsal spine on the carapace. (Only two specimens were described by Brooks, and it is possible that the spine may have been broken which seems to be a common occurrence, or alternatively it may have disappeared in the last stage.) Jurich's A. longicaudata, about stage VII, which was obtained in the Benguela Current, is almost certainly the same as the present form and we may presume that it belongs to Squilla armata. Description of first pelagic stage (Fig. 6 a) Length 47 mm. Long spiny rostrum about equal in length to the carapace. Conspicuous anterior lateral spines, slightly spiny, about one-third the length of the rostrum. Posterior lateral carapace spines spiny, about half the length of the carapace. Posterior median spine on carapace short. Carapace covering two thoracic somites and a half. Five abdominal somites without lateral spines. Four pairs of setose pleopods. Raptorial claw with unarmed dactyl, except for a spine underneath the cuticle, the propod with fine spines internally. Telson with a series of teeth laterally, the first two separated by a space larger than the others, c. twenty-four spines internal to the terminal lateral spines. The following stages progress in a manner similar to those of Squilla desmarestii. The latest seen which corresponds to stage IX of that species still has a small medium posterior spine on the carapace, the anterior lateral spines have dwindled and the rostrum and posterior lateral spine are shorter. None of them has spines. The carapace is about three times as long as broad. The thoracic limbs are all well developed and there are gills on maxillipedes 1-4. There are six abdominal somites and five pairs of setose pleopods, somites 1-5 with long lateral spines. The raptorial claw has a long spine and a series of small spines on the propod internally, and the dactyl has four free spines and three still covered. The telson has a small tooth just behind the lateral tooth and there are nineteen spines each side, between the intermediate and sub-median teeth, and thirty-six between the sub-median teeth. Length of larva from tip of rostrum to end of telson 20 mm. Uropods very like those of the adult S. armata. As S. armata is the common Squilla of this district and has much the same distribution as jfasus lalandii and as only one Squilla larva occurred in the plankton hauls, and in one case together with a first larval stage of jfasus lalandii, it seems almost certain that these larvae belong to Squilla armata, which is known from a wide area including the region of the Benguela Current (Balss, 1938). It has been described by several authorities including Bigelow (1895) and Kemp (1913). Although it is nearly certain that these larvae belong to S. armata, occurring as they do in exactly the same locality as the adult, it is just possible that they may belong to Caiman's S. africana (Caiman, 1916). His specimen of the adult came from the West African coast, from the Congo northwards, and he shows that Jurich's (1904) S. empusa from the Congo belongs to this species, as do other specimens from West Africa in the British Museum. All, however, are obtained farther north than the present specimens and it seems more probable that these latter belong to S. armata. LITERATURE Balss, H., 1927. Macrura der Deutsche/! Tiefsee-Expedition, 3. Natantia. Teil B. VViss. Ergeb. deutsch. Tiefsee-Expedition 'Valdivia', xxm, pp. 247-75. Barnard, K. H., 1950. Descriptive Catalogue of S. African Decapod Crustacea. Ann. S. Afr. Mus. xxxvm. Bouvier, E. L., 1905. Nouvelles Observations sur les Glaucothoes. Bull. Mus. Oceanogr. Monaco, no. 51, pp. 1-14. Calman, W. T., 1925. On Macrurous Decapod Crustacea collected in South African Waters by S.S. 'Pickle'. Rep. Fish & Mar. Biol. Survey, iv. Special Report 3, pp. 1-26. SYSTEMATIC 233 Dakin, W. J. & Colefax, A. N., 1940. The Plankton of the Australian Coastal Waters of New South Wales, Part I. Pub. Univ. Sydney, Part I of Zool. Monograph 1, pp. 1-395. De Man, J. S., 1928. The Thalassinidae and Callianassidae collected by the ' Siboga' Exp. 'Siboga' Exp. Rep. xxxix, a, b, pp. 1-183. Gilchrist, J. D., 1916. Larval and Post-larval stages of ]asuste\andii (M.-Ed.). J. Linn. Soc. Lond. Zool. xxxm, pp. 101-24. Gurney, R., 1924. Decapod Larvae. Nat. Hist. Rep. 'Terra Nova' Exped. Zool. vm, No. 2, part ix, pp. 37-302. 1936. Larvae of Decapod Crustacea. Part m. Phyllosoma. 'Discovery' Reports, xn, pp. 400-40. 1938. Larvae of Decapod Crustacea, Part v. Nephropsidea and Thalassinidea. 'Discovery' Reports, XVII, pp. 291-344. 1942. Larvae of Decapod Crustacea. Ray Society Monograph, pp. 1-306. Gurney, R. & Lebour, M. V., 1940. Larvae of Decapod Crustacea, Part vi. The Genus Sergestes. 'Discovery' Reports, xx, pp. 1-68. 1941. On the Larvae of Certain Crustacea Macrura, mainly from Bermuda. J. Linn. Soc. Lond., Zool., xli (no. 277), pp. 89-181. Hansen, H. J., 1922. Crustaces Decapodes (Sergestides) provenant des campagnes des yachts ' Hirondelle' et ' Princesse Alice' (1885-1915). Res. Camp. Sci. Monaco, fasc. lxiv, pp. 1-232. Heldt, J. H., 1938. La Reproduction chez les Crustaces decapodes de la Famille des Peneides. Ann. Inst, oceanogr. Monaco, n.s. XVIII, fasc. 2, pp. 1-206. Hesse, M., 1876. Description des Crustaces rares ou nouveaux des cotes de France, Arb. 5. Ann. Sci. Nat. Zool. (6), pp. 1-42. Holthuis, L. B., 195 1. The Caridean Crustacea of Tropical West Africa. 'Atlantide' Report, no. 2, Sci. Res. of the Danish Exp. to the Coasts of Tropical West Africa, 1945-6, pp. 7-187. 1952. Crustaces decapodes macrures. Exp. Ocean. Beige dans les Eaux cotieres africaines de l'Atlantique sud (1948-1949). Res. Sci. in, fasc. 2, pp. 1-88. Kemp, S., 1939. On Acanthephyra purpurea and its Allies (Crustacea Decapoda: Hoplophoridae). Ann. & Mag. Nat. Hist. ser. ii, iv, no. 24, pp. 568-79. Lebour, M. V., 1928. The Larval stages of the Plymouth Brachyura. Proc. Zool. Soc. Lond. part 2, pp. 474-560. 1936 a. Notes on the Plymouth Species of Spirontocaris. Proc. Zool. Soc. Lond. pp. 89-104. 19366. Notes on the Plymouth Processa. Proc. Zool. Soc. Lond. pp. 609-17. 1938. The newly hatched larva of Callianassa affinis Holmes. Proc. Zool. Soc. Lond. CVIII, pp. 47-8. 1949a. Some New Decapod Crustacea from Bermuda. Proc. Zool. Soc. Lond. cxvm, pp. 1 107-17. 19496. Notes on some larval Decapods (Crustacea) from Bermuda. Proc. Zool. Soc. Lond. cxx, pp. 369-79. Stephensen, K., 1923. Decapoda Macrura. Rep. Danish Oceanog. Exp. 1908-10, to the Mediterranean and adjacent Seas, 11, D 3, pp. 1-252. Thompson, E. F., 1943. Paguridae and Coenobitidae. Sci. Rep. 'John Murray' Exp. 1933-34, VII> PP- 412"26- Williamson, H. C, 1915. Nordisches Plankton, Lief, xvm, Decapoden, I. Teil (Larven). Kiel and Leipzig, pp. 315-588. LITERATURE STOMATOPODA Balss, H., 1938. Stomatopoda. In Bronn's Klassen und Ordnungen des Tierreichs. 5. Abt. 1. Buch 6, Teil 11, pp. 1-178. Bigelow, H. P., 1895. Report upon the Crustacea of the order Stomatopoda collected by the Steamer 'Albatross' between 1885 and 1891, and on other specimens in the U.S. National Museum. Proc. U.S. Nat. Mus. XVII, pp. 489-550. Brooks, W. K., 1886. Report on the Stomatopoda collected by H.M.S. 'Challenger' during the years 1873-76. 'Challenger' Reports, Zool. xvi, pp. 1-116. Calman, W. T., 1916. A Nezv Species of the Crustacean Genus Squilla from West Africa. Ann. & Mag. Nat. Hist. ser. 8, xvm, pp. 373-6. Claus, C, 1871. Die Metamorphose der Squilliden. Abh. Kgl. Ges. Wiss. Gottingen, xvi, pp. n 1-63. Giesbrecht, W., 1910. Fauna und Flora des Golfes von Neapel. Stomatopoden, xxxm, pp. 1-239. Gurney, R., 1945. Notes on Stomatopod Larvae. Proc. Zool. Soc. Lond. cxvi, pp. 133-75. Jurich, B., 1904. Die Stomatopoden der deutschen Tiefsee-Expedition. Wiss. Ergeb. deutsch. Tiefsee-Exp. 'Valdivia', VII, pp. 361-408. Kemp, S., 1913. An account of the Crustacea Stomatopoda of the Indo-Pacific region. Mem. Ind. Mus. iv, 1, p. 41. [Discovery Reports. Vol. XXVII, pp. 235-278, Plate XIII, April 1955]. THE DISTRIBUTION OF SAGITTA GAZELLAE RITTER-ZAHONY By P. M. DAVID '*&■ CONTENTS Introduction . Acknowledgements Methods . Stages of maturity Synonymy Racial characteristics Horizontal distribution Vertical distribution Growth rate and life history Food .... Summary .... References Appendix: Tables 18 and 19 page 237 237 238 241 244 254 260 265 271 275 276 277 278 THE DISTRIBUTION OF SAGITTA GAZELLAE RITTER-ZAHONY By P. M. David (Plate XIII, text-figs. 1-27) INTRODUCTION ON E of the main objects of the plankton studies carried out by the Discovery Committee and the National Institute of Oceanography is to disclose the principles which underlie the distribution of oceanic organisms. Among the different species of pelagic zooplankton there are very wide contrasts in horizontal and vertical ranges, and in the relations between distribution, life cycle, and environ- ment. Yet there are few oceanic species of which even the limits of distribution have been fully delineated, and fewer still whose life cycle has been adequately related to their distribution. To under- stand the basic factors it is clearly desirable to know how much the range of habitat and the life cycle vary among planktonic organisms, and to consider whether any types of distribution are associated with certain methods of reproduction or feeding habits, or with the phylogenetic position of the species. One method of approaching these problems is to work out as fully as possible the distribution of a number of species at all stages of their life history, so that those of different habits and from different taxonomic groups can be compared; and the present paper is intended as a contribution to this end. The Discovery Committee's field of work was mainly in Antarctic and Subantarctic waters, and these regions were very widely sampled by the ' Discovery II ' in five voyages before the war. During a sixth voyage, under the National Institute of Oceanography in 1950-51, much material was collected in the winter months, which had been insufficiently represented in the pre-war plankton samples. The collections are now well suited to studies of distribution since they cover the whole range of a number of southern species at all times of year. The closing nets provide good information on vertical distribu- tion, and detailed hydrological data are available from most stations. At present, a most exhaustive study of the life history and distribution of Euphaiisia superba by Mr J. W. S. Marr is nearing completion, and it is hoped that a study of a carnivorous species such as Sagitta gazellae will be of special interest when compared with a herbivorous species such as Euphaiisia superba which has wholly different habits. Sagitta gazellae was chosen as an example of a carnivorous species for various reasons ; experience had shown that it is widely distributed ; it is easily recognized ; it occurs in large, but not inconveniently large, numbers; the animal itself is of considerable size, and is thus easily picked out of samples; it is transparent and the state of its gonads can be determined without dissection. In addition to these points, very considerable data on this species had been amassed by Mr Marr during two pre-war commissions of ' Discovery II '. Although in its distribution S. gazellae is restricted to the Southern Ocean, data from other areas concerning other species collected during the course of this work suggest that its habits are not unique and that S. lyra, a more cosmopolitan species, has, at least in the subtropics of the southern hemisphere, a very similar life history. It may well be that other species will be found to conform to a similar pattern. ACKNOWLEDGEMENTS I am particularly grateful to Mr Marr for advice and encouragement, and for permission to use the very large collection of data which he had assembled during the second and fourth commissions of R.S.S. 'Discovery II'. ?38 DISCOVERY REPORTS The present study was suggested by Dr N. A. Mackintosh, C.B.E., who has given me advice and guidance during the whole time I have been engaged on it. I should like to thank both Dr J. H. Fraser, F.R.S.E., who has read and corrected the typescript and given me a great deal of valuable advice during the preparation of this paper, and Dr H. E. Barg- mann who read the first draft and made many helpful suggestions. Mr E. Childs has assisted me in sorting a number of the samples required for this work, and my colleagues, Mr P. Foxton and Mr A. de C. Baker, have drawn my attention to any interesting specimens which they found while they were engaged upon work on the Discovery collections. I wish to thank Professor T. Tokioka of the Seto Marine Laboratory, Japan, for kind permission to reproduce two of his figures (Figs. 9 and 13); Walter de Gruyter and Co. of Berlin for permission to reproduce four of Ritter-Zahony's figures (Figs. 1 and n) from the Deutsche Siid-Polar Expedition Report; A. H. Pettifer of the Government Printing Office, Sydney, N.S.W., for permission to repro- duce one of Johnston & Taylor's figures (Fig. 2) from the Australasian Antarctic Expedition Report ; and the Council of the Linnean Society for permission to reproduce two of Fowler's figures (Figs. 3 and 4) from the Report on the Biscayan Plankton of H.M.S. 'Research'. METHODS All the material used in this investigation came from plankton samples in the Discovery collections. Most of the samples were taken from stations worked in the Southern Ocean, but in order to confirm that 758. 762> 764> 767. 770"794> 79°. 798, 800, 802, 805, 807, 809, 812, 814, 817, 821, 823, 826, 827, 832, 833, 835, 843, 846,864, 874, 875, 876, 878, 888, 896, 897, 901, 907-910, 912-918) 942-1019 (excluding 944, 952-955, 957, 963, 979-982, 984, 987, 989, 991, 993, 997, 998, 1002, 1004-1012, 1014, 1016, 1018) 2699-2704 (excluding 2701) 2726-2731, 2736-2741, 2743, 2757, 2759, 2768-2770, 2785, 2794-2797 2802-2892 (excluding 2809, 2822, 2827, 2829, 2845, 2847, 2883-2885) species. The vertical series of hauls from lines of stations which have been repeated on a number of occasions in the meridian of o° have been examined with the object of getting a picture of vertical distribution in one locality for most months of the year. These were long lines of stations covering Antarctic and Subantarctic waters. A series of deep vertical 2 m. net hauls, which extended from the Antarctic into the tropics in the meridian of 30° W., were also used to check the depth range and to examine the possibility of tropical submergence. Descriptions of the nets employed by the expedition and the manner of fishing them may be found in Kemp, Hardy & Mackintosh (1929), Ommanney (1936) and Marr (1938). The positions and hydrological information for most of the stations which will be referred to, may be found in the various Discovery Station Lists. ('Discovery' Reports, Vols, iv, xxi, xxn and xxiv), but those from St. 2699 onwards will appear in the Station List for 1950-51 which is expected to be published shortly. 240 DISCOVERY REPORTS All the specimens were fixed, and preserved in neutralized sea-water formalin, in most cases in the whole plankton sample; this seems to give the best results. Chaetognaths removed from hauls before fixation and preserved separately often fix in a bent or distorted position, and if placed upon formalin- soaked blotting-paper to keep them straight, tend to get damaged. Specimens of S. gazellae preserved in large plankton samples are usually in good condition, and samples consisting mainly of copepods such as Rhincalanus gigas or Calanus propinquns nearly always contain perfect specimens of Sagitta gazellae. The material from 70 cm. vertical nets is the most perfect, and the 70 cm. oblique nets yield specimens in better condition than those from the 1 m. nets. Specimens from 2 m. and \\ m. nets are often damaged, none being found with seminal vesicles intact, although the animals were of such Table 2. List of stations from which data were obtained during the present work Counts and measurements N70V Stations 2010, 2012, 2014, 2015, 2017, 2018, 2020, 2022, 2023, 2024, 2025, 2026, 2027 2311, 2313, 2316, 2318, 2320, 2322, 2355, 2356, 2357, 2358, 2359, 2361, 2385, 2386, 2387, 2389, 2391, 2393 2424, 2425 2492, 2494, 2495, 2496, 2498, 2501 2530, 2531, 2532, 2533, 2535, 2538, 2541, 2543, 2545, 2547 N100B Stations 414, 1774, 1775, 177° 2312, 2313, 2316, 2318, 2320, 2322 2355, 2359, 2361, 2374 2385, 2386, 2387, 2389, 2391, 2393 2424, 2425, 2426, 2428, 2430 2459, 2460, 2461, 2463, 2465 2492, 2495, 2496, 2498, 2500 2531. 2535. 2538> 2541. 2543. 2545. 2547 2606-2626 2374-2379 (excluding 2376) 1 608-1 625 (excluding 1613, 1620, 1622) 1644-1664 (excluding 1649, 1650, 1656-1658, 1661) 1687-1699 (excluding 1691, 1692, 1698) TYF and N450 oblique Stations 391, 395, 401, 405, 413 1298 1702, 1707, 1715, 1718, 1719, 1723 1871, 1876, 1917, 1919, 1944, I946- J966, 1970, 1972, 1974. !989. !99i> 1993, 1995, 2001, 2006, 2008, 2029, 2031, 2033, 2036, 2042 TYF vertical Stations 661, 663, 666, 668, 669, 671, 673, 675, 677, 679, 681 1154, 1156, 1158, 1160, 1165, 1167, 1173, 1175 a size and stage of maturity that these structures should have been present. However, for the examina- tion of hooks, teeth and ovaries, material from large towed nets is quite satisfactory. The ovaries are some- times inadequately preserved, especially in the larger specimens, and cannot be accurately measured, but this only occurs at the fourth stage of maturity (see p. 244) (when the ovaries are greatly swollen) ; the thin rod-like ovaries of the earlier stages are always adequately preserved. Thus the stage of maturity can always be determined, even if the degree of advancement of maturity within the stage cannot be. Considerable shrinkage takes place in formalin-preserved material (George, 1952) and as measure- ments of freshly preserved material from the 1931-33 and 1935-37 commissions of 'Discovery II' were available it has been possible to gain some idea of the extent. Unfortunately the degree of shrinkage is not consistent, and it has not therefore been worth while making any corrections to measurements to allow for it. Measurements of 100 specimens ranging from 20 to 105 mm. (freshly preserved) show an average shrinkage of 7-25%. In 1 95 1 a few measurements were done on board ship to determine the extent of shrinkage during SAGITTA GAZELLAE 241 fixation. The same specimens were measured in 1952 and 1954, and although the number of observa- tions is very small, the results are fairly consistent, and suggest that, disregarding the rather large difference which occurs between live and fixed animals, most of the shrinkage occurs during the first two years in preservative (Table 3). It is quite possible that anomalies in tables of head armature and tail percentages may be due to shrinkage, as the tail segment is often much less affected than the rest of the body. This is especially so in animals at stage II (those with their tail segments packed with sperms). Bollmann (1934) remarks that preservation affects the appearance of the hooks in the closely allied species S. lyra, but no difference, other than a slight increase in opacity, has been observed between the hooks of live and dead specimens of S. gazellae. Table 3. Shrinkage of a series of specimens shortly after preservation and after several years in formalin (Each column of figures refers to one specimen.) Date measured Length in mm. 25. viii. 51 Alive 93 92 92 88 — 83 — — 26. viii. 51 Preserved 89 87 87 84 — 80 — — 26. viii. 5 1 Alive — — — 89 — 86 — 79 77 27. viii. 51 Preserved — — — 85 — 82 — 77 73 i.x. 52 Preserved 86 85 84 83 83 81 77 76 7i 1. x. 54 Preserved 85-5 85 84 83 83 81 77 76 71 Hooks and teeth (in animals over about 12 mm. long) have been examined and counted under a low-power binocular microscope ; for animals below this length a monocular has been used. The head armature can best be seen if the animal is held by the pressure of a mounted needle just behind the corona; the anterior part of the head then faces the objective of the microscope, and by varying the pressure on the mounted needle the whole armature can be seen, and counted quite rapidly. Occa- sionally the use of methylene blue has helped in counting teeth, as it leaves an area free of stain at the base of each tooth. All counts of hooks include the rudimentary ones. These are often very small, and may be only just projecting through the cuticle. They are only present in animals below about 30 mm. in length, and are usually easily seen, if strong reflected light is used ; they are sometimes difficult to see by transmitted light. A weak aqueous solution of methylene blue has been used to show the corona. As Thomson (1947) has observed, methylene blue fades rapidly, and animals stained with it can be returned to samples after examination. For permanent preparations a dilute solution of aqueous haematoxylin has proved fairly satisfactory, though it does not show the corona and sensory spots as distinctly as methylene blue. The heads of some specimens have been boiled in caustic potash to verify the hook and teeth numbers counted under the binocular microscope. There were no discrepancies, but it is possible that among the large numbers of teeth counted, some immature ones may have been missed. It is unlikely, however, that the hook numbers are in error, as even the rudimentary ones are large in comparison with the smallest teeth. STAGES OF MATURITY Various systems for the classification of the stages of maturity of Chaetognaths have been proposed by Russell (1932), Kramp (1939), Thomson (1947) and Pierce (1951). Although the Chaetognatha are hermaphrodite some of these authors have based these stages only on the ovaries, but Kramp has used both testes and ovaries, and this I think provides a sounder basis; though no system of this kind seems universally applicable. Each species or closely related group of species tends to differ in the 242 DISCOVERY REPORTS development of the gonads, and in dividing up the stages of S. gazellae I have adhered as nearly as possible to the system proposed by Kramp (1939) for S. maxima, a closely related species. Specimens of S. gazellae at advanced stages of maturity are rarely taken, and there is no record of any in the literature ; this rarity is due to the breeding migration to deep water (see p. 270). Even in the extensive Discovery collections there are few perfect mature specimens, and although I have ex- amined a large number of deep hauls in the collection I have only ninety specimens in the later stages (III to V) of maturity. Previous records of S. gazellae have been only of comparatively immature animals from the upper water layers, and the total absence of really mature stages led to the view that the most advanced stages then found (stage II) were nearer full maturity than in fact they were. Ritter-Zahony (1909), in the original description of the species, stated that the ovaries were knee-shaped and compressed, and gave an impression of maturity despite their shortness; he also remarked that the seminal vesicles Table 4. Comparison of maturity stages for Sagitta maxima and S. gazellae S. maxima (from Kramp, 1939) S. gazellae Stage v* • • •• • • 00 0 se 0 » 0 9 0 oco oe 00 cgcJsaBo Oj 0 e • • * * V ••»• • • • 0 00 00 0 O Co 8 0 0 OOQCBOQ^ <$ e •• • •••#• ••• • uo 00 00 0 0 ■ • • ^ 7 0 0 Q? 00 000 •• • ••••••• • O A ^ ° 000 • o6 X 5 0 0 0 00 1 • S. GAZELLAE S.N. ° 0 00 0 O O OjQj 0 0^ 00 0 o4 2 0 S. GAZELLAE stages iht.v 0 0 0 0 0 0 3 ad 3 0 s LYRA 00 0 QjO 00 0 °0 z O 10 20 30 40 50 60 70 LENGTH MM Fig. 8. The number of hooks plotted against length for S. gazellae S.N. race and S. lyra. Johnston & Taylor (1921) record a specimen of S. lyra ( = S. gazellae) with 13 hooks on each side, but other authors have agreed with Ritter-Zahony. The larger number of hooks in Subantarctic specimens of S. gazellae is of considerable value in differentiating the species from S. lyra, and Fig. 8 gives a comparison of hook numbers and length in both species. Corona. The shape of the corona in S. lyra is simple and pear-shaped. Fig. 9 (after Tokioka, 1940) shows it in a specimen of S. lyra "lyra" -type. I have examined this feature in a fairly large series of specimens of S. lyra "gazellae" -type and find it to be similar. Though exhibiting a range of variation (Fig. 10) it does not depart from the general pear shape. Ritter-Zahony (191 1) gives two illustrations of the corona of S. gazellae (fig. 6 A, B) (Fig. 11). Fig. 1 1 b is the corona drawn directly from a specimen, and Fig. 1 1 a is a simplified representation of it, the latter was done because the author considered the sinuosities of outline to be due to contraction of the head muscles. Although the corona is often damaged or absent in preserved specimens of S. gazellae, I have been able to observe it many times among the very large number of specimens which are available in the Discovery collections, and the sinuosities of outline are always present except in very small specimens. I do not think they are due to contraction of the head muscles, for no evidence SAGITTA GAZELLAE 25i for contraction can be seen, and in specimens having the hood fully extended (Fig. 12 a) the corona still exhibits a sinuous outline. The corona of S. lyra shows no sinuous outline, yet the head of such a similar species must be equally liable to muscular contraction. On the other hand, I have not found a perfectly symmetrical corona in S. gazellae, but this structure in any specimen of any species is seldom perfect, and the complexity of the 5. gazellae corona makes it difficult to speculate as to the perfect outline. The features of the S. gazellae corona which serve to differentiate it from that of S. lyra are the marked sinuosities between the eyes (Fig. \2a-d) (this portion of the corona often 6) § % * Fig. 9. Corona of S. lyra. (From Tokioka, 1940.) Fig. 10. Variation in the shape of the corona in S. lyra "gazellae" -type. Fig. 11. Corona of S. gazellae. (From Ritter-Zahony, 191 1.) Fig. 12. Variation in the shape of the corona in S. gazellae. remaining when the rest is absent); the close proximity of the eyes to the corona; and the straight cross-portion (Fig. 12 a, b, d) at the posterior end, which is quite different from the continuous curve of the equivalent part in S. lyra. Although the difference in outline of the corona in the two species may be considered a valuable taxonomic difference, it is not a very practical one, owing to the delicacy of the structure and the consequent fact that it can so seldom be observed in specimens of S. gazellae. Width between the eyes. This feature is perhaps the most convenient and precise of all. The width between the eyes is given as a percentage of the width of the head. Table 9 gives the data from fifty specimens of S. lyra "gazellae" -type and fifty specimens of the Subantarctic race of S. gazellae, to 2S2 DISCOVERY REPORTS show the range of variation in the measurements. It can be seen that the average percentage for S. lyra "gazellae" -type is double that of S. gazellae for the same state of contraction of the head and the highest percentage for the latter species does not overlap the lowest percentage of the former. But it will be seen from Table 9 that the range for each species is considerable, and this is due to the position in which the head has been fixed. I have not found any overlap between the species; but it is quite possible that in specimens with very distorted heads this may occur. Therefore, when using this method, if specimens give a percentage between 30 and 40 the shape of the head should be noted ; if the hooks are pointing in towards the mouth, and the hood is withdrawn, the head will be much wider than normal, and the specimen will be S. lyra; if, on the other hand, the hood is forward, and the hooks point straight forwards or outwards then the specimen is 5. gazellae. As a general rule, however, specimens giving a percentage over 35 are S. lyra and under 35 are S. gazellae. Reference to Tokioka's (1940) illustration of S. lyra (" lyra" -type) shows that the percentage of eye width to head width of the specimen illustrated is 60, which fits the range of S. lyra given in Table 9. Similarly, reference to Johnston & Taylor's (1921) illustra- tion of one of the specimens which they describe as S. lyra shows that the percentage of eye width to head width is 16-7, which is within the observed range for S. gazellae given in Table 9. Mature stages. The appearance of the seminal vesicles and the receptaculum seminis serves to distinguish between the species in mature specimens only. The use of these differences is limited by the fact that such specimens occur only at depths of 750 m. or more, and are therefore not commonly taken. Both S. lyra (Ramoult and Rose, 1946 ; Ghirardelli, 1 950) and S. gazellae (see p. 270) breed in deep water. The seminal vesicles of S. gazellae occur midway between the posterior and caudal fins, whereas Fraser (1952) states that in S. lyra they lie ' nearest the posterior fin '. Those of S. gazellae appear to differ slightly in shape from those of S. lyra (Fig. 13), but are also of the ' enflata-type ' (Tokioka, 1939). Although I have examined fairly large numbers of S. lyra in varying degrees of maturity I have found none with perfect seminal vesicles, and must therefore depend upon published figures for comparison. The receptacula seminis are conspicuous white objects in S. lyra, but are not visible in S. gazellae. Two other features of difference in the mature animals may be noted. The number of hooks at these stages in S. lyra is nearly always three on each side, whereas in S. gazellae there may be from three to six (see Figs. 8 and 15). In 5. gazellae the posterior teeth are reduced to none, or occasionally one on each side, whereas in fifty specimens of mature S. lyra I have never seen less than two, and usually three on each side. Distribution. S. gazellae is confined to Antarctic and Subantarctic waters, whereas S. lyra is a species with a world-wide distribution in tropical and subtropical waters, and is found in both hemispheres. S. lyra has not been recorded from Arctic or Subarctic waters. Though Michael (191 1) records S. lyra from very cold waters, he failed to differentiate it from S. maxima, and his table of measurements of twenty specimens of S. lyra appears to contain a number of specimens of S. maxima (nos. 1, 3, 5, 8, 11, 13, 14, 17, 18 and 19). On page 38 of his report he states 'it is possible that I may have obtained both S. lyra and S. maxima and confused them ' ; I think it is just such confusion which accounts for the records of S. lyra at very low temperatures, for Fraser (personal communication) informs me that he has never observed S. lyra in Arctic or Subarctic waters. If S. gazellae is merely a cold-water form of 5. lyra (as some authors maintain) which has colonized the Antarctic and Sub- Fig. 13. Seminal vesicle of 5. lyra. (From Tokioka, 1940.) SAGITTA GAZELLAE 253 Table 9. A comparison of the width between the eyes calculated as a percentage of the zvidth of the head in Sagitta lyra "gazellae" -type and S. gazellae Total length (mm.) 5. lyra "gazellae" -type S. gazellae Eye width/head No. of Eye width/head No. of width (%) specimens width (%) specimens 56-5 — — 32-3 49 — — 25-0 1 48 — — 22-4 45 — — 25-9-33-3 44 — — 269 42 — — 28-6 39 — — 22-2 38-5 — — 28-0 38 — — 20-8 37 — — 27-3-30-3 36 — — 26-1 35 — — 25-0 34'5 — — 20-0 34 50-°-53-3 2 28-6 33-5 — — 21-7 33 — — 23-8 31 50'0-S3-3 5 29-4 3° 57- 1 1 27-9 29 — 26-9-27-8 28 5o-o-53-3 2 27-5 — — 27-7 1 27 50-0-65-3 3 24-4-25-0 26-5 46-7 1 2I-I 26 46-7-46-8 2 2f9 25-5 46-2 1 25-0 25 50-0-58-3 2 2I-I 24-5 — — 22-2 24 46-2-58-2 6 22-2 23-5 50-0 1 20-0 23 50-0 3 20-0 22-5 — — 25-0 22 42-8-50-0 2 21-4 21-5 44-5 1 — 21 — — 20-0-27-3 3 20-5 46-2 1 — — 20 60-0-62-5 2 16-7-23-8 2 19 50-0-55-5 2 20-0-27-3 2 18-5 50-0-55-6 2 — 18 45H 1 I9-2-33-3 2 17-5 — — 23-1 17 44-5-5o-o 3 20-0 16-5 42-9 1 16 50-0-62-5 3 18-2 iS'S 50-0 1 H — — 25-0 i3"5 — — ■ 22-2 13 6o-o 1 — — 12 40-0 1 28-6 1 Av. 49-1 50 Av. 24-5 50 3-2 2S4 DISCOVERY REPORTS antarctic waters in profusion, then it might be expected that a similar colonization would have taken place in the Arctic, as there is ample opportunity for specimens to be carried north by the North Atlantic drift especially since the so-called 'intermediate' S. lyra "gazellae" -type has been reported from various parts of the northern hemisphere. From the various points discussed above it may be concluded that S. gazellae Ritter-Zahony is a species in its own right, which can be distinguished from all the varieties of S. lyra Krohn. RACIAL CHARACTERISTICS The surface layers of the Southern Ocean are divided into two main areas, the Antarctic zone, and the Subantarctic zone. The line which divides these areas is the Antarctic Convergence, which is the point where the Antarctic surface water sinks beneath the Subantarctic surface water. The division is, in most areas, sharp and distinct, and is to some extent a faunistic boundary (Hart, 1934; John, 1936). Examination of a series of hauls made from the Subantarctic across the Convergence and into the Antarctic zone, reveals striking differences in the appearance of S. gazellae on each side of the Con- vergence. In the Subantarctic, specimens are found at advanced stages of maturity when 60 mm. in length, whereas specimens of the same length in the Antarctic are found to be immature; in the Subantarctic surface waters no specimens are found at a length greater than 68 mm., yet in the Antarctic surface waters specimens of more than 90 mm. are quite common. In an unpublished progress report written on board R.R.S. 'Discovery II', in 1932, Mr Marr suggested that the Antarctic and Subantarctic populations of S. gazellae are two separate races ; and by a thorough study of the state of maturity of the various size groups was able to differentiate the two races in areas where they mixed, notably in the Scotia Sea (especially south-east of the Falkland Islands) . The stages of maturity which he used prove to be subdivisions of the stages which I have called I and II (pp. 242-3). Marr's stages are distinguished from mine by the use of arabic numerals 1 and 2, and letter X, whereas I have used roman numerals 1 , 1 1 , etc. Fig. 1 4 shows these stages diagrammatically . There are three stages of maturity of the tail segment. Stage 1 includes specimens with small visible rudiments of the testes up to specimens in which these rudiments extend to the transverse septum. Stage 2 is usually characterized by the curl of the testes across the anterior end of the tail segment. Stage X is equivalent to my stage II (p. 243). The tail segment is full of sperm and is opaque. No further stage was found as no further development takes place in the surface layers (the deep-breeding habits of S. gazellae were not known at the time when this system was devised) though the presence of rudimentary seminal vesicles was noted in the most advanced specimens. Marr also noted the state of maturity of the ovaries. These stages depend on the relative sizes of the transverse to the longi- tudinal portions of the ovary. In S. gazellae the ovary opens on the dorsal surface, and its posterior part extends towards the ventral surface at right angles to the long axis of the animal before con- tinuing ventrolaterally along towards the head of the animal. This is the characteristic 'knee-shape' described by Ritter-Zahony (1909, 191 1). Marr's stage 1 included animals which had visible rudiments of ovaries, and all intermediate stages up to the stage where the longitudinal part of the ovary equalled the transverse part in length. The upper limit of his stage 2 was reached when the longitudinal part was double the length of the trans- verse part, and so on up to stage 4. In the surface waters animals were seldom found with ovaries more mature than stage 4. Marr often recorded these stages as ' weak ' or ' strong ', according to their degree of maturity within the stage. Animals in which no rudiments of the gonads were visible were recorded as immature, and given no stage number. SAGITTA GAZELLAE 255 Marr referred to the races as Large Southern (L.S.), the Antarctic form, and Small Northern (S.N.) the Subantarctic form. I have also followed this practice. Table 10 a, b shows Marr's data from two hauls, one Antarctic, and the other Subantarctic, in which the stages of maturity and total lengths are compared. It is evident that there can be no confusion between the two races when the gonads are visible. In the Subantarctic race Marr's stage 1 begins at about 25 mm. total length, and it is only in specimens smaller than this that the gonads are not visible; thus below 25 mm. it is not possible to determine the race by the method just described. 1 2-3 4 Fig. 14. Stages of maturity devised by J. W. S. Marr for differentiating the races of S. gazellae. The upper series show the stages of the testes and the lower series the stages of the ovaries. No ovaries have been drawn in the upper series, and no testes in the lower series, as no particular stage of maturity of the testes is necessarily combined with a particular stage of maturity of the ovary. In practice this does not matter, as hauls are usually composed of one race or the other, and mixing only takes place in a few areas. However, it is possible in most cases to differentiate the immature forms by the number of hooks, although the method is not infallible, as there is slight overlap in the hook numbers, but in general it is satisfactory. I have found the basic number of hooks in the S.N. race to be 14 and in the L.S. race 12 (rarely 13). Comparison of Figs. 8 and 15 shows that below 25 mm. individuals of the L.S. race are likely to have 1 1 or 12 hooks, whereas individuals of the S.N. race are likely to have 13 or 14. The overlap is evident, but not extensive. Counts of hook numbers from individuals below 12 mm. are not reliable, as even with a high-power monocular it is difficult to be certain of the numbers present. 2S6 DISCOVERY REPORTS Figs. 8 and 15 also show that the L.S. race is, in its hook numbers, intermediate between S. lyra "gazellae "-type and the S.N. race. This is an interesting point, as it shows that the tendency in S. gazellae is to have less hooks in cold water than in warm, whereas the cold water form of S. lyra (S. lyra "gazellae "-type) has more hooks than the warm water S. lyra " lyra "-type. That the L.S. race is Table 10 a. ' Small Northern ' form of Sagitta gazellae (Stages distinguished by J. W. S. Marr.) Station 839 Length (mm.) Stages Remarks 9 6 61 4 X Seminal vesicles* present but clear 53 4 X Seminal vesicles* present but clear 52 4 X Seminal vesicles* present but clear 5° 4(3) X Seminal vesicles present, ovary only just 4 47 3 2 Ovary just 3 43 2 2 — 39 2 2 — 39 2 2 — 39 2 2 — 38 2 2 — 38 2 2 — 38 2 2 — 34 1 — High limit ovary nearly 2 33 1 — — 28-5 1 — — 28 — — All immature 9-28 — * Rudimentary vesicles. Table 10b. 'Large Southern' form of Sagitta gazellae (Stages distinguished by J. W. S. Marr.) Station 825 Length (mm.) Stages Remarks $ 6" 82 3 2 Very strong 3 approaching 4 77 3 2 — 74 63 62 3 2 2 2 2 Trace (1) Strong 2 approaching 3 Strong 2 approaching 3 61 2 1 — 59 2 1 — 54 2 1 — 47 1 — — 44 1 — — 4i 22-41 — — All immature in this way apparently an intermediate form between S. lyra and S. gazellae is of no practical taxonomic significance, for the L.S. race seldom extends more than a hundred miles or so north of the Antarctic Convergence, and so could never be found mixed with S. lyra : indeed the whole Subantarctic area separates the two forms Another feature in which the L.S. race resembles S. lyra is the occasional retention of posterior teeth at stage III and IV maturity. The total loss of posterior teeth is an almost invariable feature of the O ^-k to O N cL ex? E Q. 6 o U CO CO c M -o U •4-> ■M _o J2 o o I i— O O O* 00 P- -O ■< fl CO s»ooh jo yjawnN 258 DISCOVERY REPORTS S.N. S. gazellae, and as I have remarked (p. 243) mature S. lyra retain two or three teeth on each side at full maturity. Of fifty specimens of S. lyra, all had at least two posterior teeth on each side, and most had three; of thirty specimens of L.S. S. gazellae, sixteen had no posterior teeth, and fourteen had at least one tooth on one side, and in most cases on both sides ; whereas of thirty specimens of S.N. S. gazellae twenty-five had no posterior teeth, and only five retained one tooth on one or both sides (all the specimens referred to here were at advanced stages of maturity). Fig. 16 shows a com- parison of the teeth numbers of both races throughout their size ranges, and it can be seen that the numbers of posterior teeth found in the L.S. race are slightly higher than those found in the S.N. race. In the form of the body, fins, seminal vesicles and corona, the L.S. race appears to be identical with the S.N. race. Table 11. ' Very zvarm water ' form of Sagitta gazellae (Stages distinguished by J. W. S. Marr.) Length (mm.) Stages Remarks ? 0 •o 0 • O O O 00 o 99 CO Z • • V 99 CO •o 0 £>o vi/j/ • • i &c • o < a: • •• •• • S CO o *9 • * CO • 2: • o • o Q£ •• 00 • • o 0 0 » oo o «• » V o • — 00 o • • a 8B CO • • o o • • • 1 • • ao o • o 0 o o o ■ © ® ® ® 93 Y • o or: Z Ui a UJ IT 0 • ry I z> O H i t- H o LU UJ i- < < X ul UJ 2 2 1— t- UJ 3 a. H I I o O at o UJ UJ Ul 0 UJ K i- UJ K LU l/> H U) K O Z O Z CL a < a < < • O ® ® CO V u o a ■a a o a. O "*- £ 3 C H 60 o x - h- a < o f\j — o Hi33i jo y^anriN Hi33i jo yaawriN 26o DISCOVERY REPORTS might be regarded as separate species, but the close anatomical similarity and an apparently common life history suggests that the two may be regarded as racially, rather than specifically, distinct. It is possible that the Antarctic Convergence is an isolating mechanism, and that the two races of S. gazellae represent an early stage in the formation of a new species. As individuals of both races are found, and can live, on the 'wrong ' side of the Convergence one may suppose that the irregularity and rapidity of environmental changes in the region of the Convergence, more than the widely different, but stable, conditions on each side of it represent the real barrier. Further, the distribution of the separate, though closely related, species S. lyra overlaps that of S. gazellae slightly in the region of the subtropical convergence. It might be postulated that individuals of the S. lyra "gazellae "-type crossed this convergence and successfully colonized the Subantarctic zone, giving rise to the specifically distinct Small Northern S. gazellae form, and subsequently these gazellae were equally successful in establishing themselves in the Antarctic zone, and produced the Large Southern form. The suggestion that colonization by S. gazellae of the Antarctic from the Sub- antarctic zone took place is supported by the evidence of its greater abundance in the Subantarctic (see Fig. 17). Although the colonization of the Subantarctic was sufficiently long ago to enable the Subantarctic population to become specifically distinct from the subtropical stock from which it originated, the time since the colonization of the Antarctic zone has only been sufficient for the formation of a new race. HORIZONTAL DISTRIBUTION S. gazellae is an oceanic species with a continuous circumpolar distribution in Antarctic and Sub- antarctic water; Baker (1954) records it as the most consistently occurring of the Antarctic zooplankton species which he has examined, being present in 96% of the hauls. Fig. 17, which shows the horizontal distribution of density of the species, has been compiled from data from the 1 m. oblique nets (N 100 B) fished in the surface (100-0 m.) layer. Ommanney (1936) has discussed the significance of the variations in fishing depth of these nets, and concluded that they can be disregarded for most purposes. The shallow N100B is always fished for 20 minutes at a standard speed of hauling, and thus the symbols on Fig. 17 represent numbers per 20-minute haul. It has not been necessary to take into account the time of day when the hauls were made, as Table 12 leaves little doubt that no diurnal migration takes place. Although this species has usually been regarded as an Antarctic form (Ritter-Zahony, 191 1), its main centre of abundance, as Fig. 17 shows, is in the Subantarctic zone, and no catch in excess of 200 individuals per 20-minute haul has been obtained from Antarctic waters. Many of the negative and very small hauls in Antarctic water shown on Fig. 17 were made in winter, and give the impression that the Antarctic population is sparser than it is; regrettably there are not sufficient data to make separate charts for winter and summer. There is no evidence that the species 'swarms' or forms 'shoals', and the consistency of its occurrence suggests a fairly even distribution. The northern limit of the distribution of the species is the region of the subtropical convergence ; in making Fig. 17 the mean position of this convergence has not been used, but for each line of stations which has crossed it, its position has been found by reference to the continuous thermograph records ; this convergence is a variable phenomenon, and the mean position given by Deacon (1937) is a simplified representation of it. In areas where tropical currents such as the Aghulas or the S.E. Australian current influence it, it is probably never stable, and the limits of the distribution of S. gazellae fluctuate in a similar manner. Even when a contemporary position for the subtropical convergence is used, some hauls of S. gazellae are occasionally taken north of it, for example at SAGITTA GAZELLAE 261 30° (KJ 90" 120 Fig. 17. The horizontal distribution of S. gazellae in the top 100 m. of water. Data from the shallow oblique 1 m. nets. The chart shows the Antarctic Convergence, and, closer to the continent, the normal position of the pack-ice edge in March. The subtropical convergence is shown as a short line at right angles to each line of stations which crossed it, and is the con- temporary position found by reference to the continuous thermograph, and not a mean position. 262 DISCOVERY REPORTS Station 2625 (410 497' S., 180 49-9' E.). Reference to the hydrological data for this station shows that a northward extension of Subantarctic water was present between 50 and 100 m., and the net which fished from 93 to o m. may well have taken the majority of the catch from this layer. Similar penetra- tions below subtropical water are shown at Stations 1773 and 1774, and also at Station 2036 (Fig. 22). It seemed possible that the penetration may be even farther north in the deeper waters, but examina- tion of a line of 2 m. vertical nets in subtropical waters at Stations 675 (fishing depths 2750-2000, 2000-1500, 1 500-1000, and 1000-750 m.), 677 (fishing depths 2000-1 500, 1500-1000, 1000-750 m.) and 679 (fishing depths 2000-1500, 1500-1000, 1000-750 m.), has not revealed a single specimen, and it is therefore very unlikely that tropical submergence takes place. A series of hauls from stations off the limits of the chart shown in Fig. 17 have been examined for S. gazellae, and in all cases the results have been negative. The stations were made in the Peru current, the Benguela current, and on the meridian of 900 E. in the Indian Ocean. Table 12. A comparison of day and night vertical stations from various localities, and various months; numbers corrected for a 250 m. haul; percentage is the number at a particular horizon to the total for all depths Depth (m.) Day Night No. Percentage Hauls No. Percentage Hauls 50-0 325 23-3 35 420 2I"5 35 100-50 7*5 51-4 35 1075 55-2 35 250-100 281 20-2 35 34° 17-4 35 500-250 S2 374 35 76 390 35 750-500 6 o-43 35 21 1-08 35 1000-750 10 072 35 13 o-68 33 1 5 00- 1 000 2 0-14 14 4 0-20 24 Totals I391 9993 224 1949 99-96 232 The southern limit of distribution of the species is the Antarctic continent ; and although it has not been possible owing to ice conditions to prove this in many areas, it can be said that where the Antarctic coastal waters have been reached, S. gazellae has been found. Owing to peculiarities of its life history (p. 270) it is improbable that it can breed in shallow shelf-waters, but even in the almost uniformly shallow Ross Sea it occurs in small numbers. In this area it averaged seven specimens per 20-minute haul in the month of January. In oceanic waters the average per 20-minute haul for the month was 47-7. Discovery stations in the Weddell Sea have been infrequent, and were only made in the eastern end, owing to the treacherous ice conditions often encountered in the area; but Bollmann (1934) reported that the ' Deutschland ' found S. gazellae in the Weddell Sea. Unfortunately he gives little indication of its frequency of occurrence. Previous work on the distribution of the species has been somewhat confused by the inclusion of S. gazellae in the synonymy of S. lyra, and also by the existence of a form of S. lyra which closely resembles S. gazellae. This has been discussed on pp. 245-254. Although S. gazellae was not described until 1909, there were several earlier records of 5. hexaptera which have been attributed to the species. These are discussed by Ritter-Zahony (191 1). In his paper giving the original description of the species, Ritter-Zahony observes that although his specimens came from the South Indian and South Pacific Oceans, the Berlin Museum possessed a specimen from the Tonga islands; he does not refer to this specimen again, and, bearing in mind the fact that he was unable to determine the shape of the fins in his alcohol-preserved specimens, it seems SAGITTA GAZELLAE 263 likely that the Tonga specimen was not S. gazellae. In the Revision of the Chaetognaths ( 1 9 1 1 ) he states that the distribution of S. gazellae is between 6o° and 66° S. in the Antarctic, and that the most northerly catch was 35-5° S. He also states that it occurs in deeper water farther north; but I think this last statement is a reference to the catch from 35-5° S. rather than a suggestion of tropical submergence. Subsequent records are mainly from Antarctic and Subantarctic waters, but there are a few exceptions. Burfield (1930) records S. gazellae from Rio de Janeiro (230 S.) from the 'Terra Nova' hauls, but on p. 219 of his report remarks 'The horizontal range of the stations at which this form was captured extends from lat. 230 N. to lat. 770 380 S.\ From the station numbers and distribution charts, it would appear that 23 ° N. is a misprint, and should be 23 ° S. Through the kindness of the British Museum (Natural History) I have been able to examine the original ' Terra Nova ' hauls from Rio de Janeiro (Stations 39 and 40) and, although the specimens are in a bad state of preservation, it is evident from the position and shape of the seminal vesicles that they are S. enflata and not S. gazellae. Tokioka (1939) records S. gazellae as a synonym of S. lyra in Japanese waters, but these specimens are almost certainly the cold water form of S. lyra. Thomson (1947) found S. gazellae in Tasmanian waters, but followed the synonymy proposed by Johnston & Taylor (1921) and Tokioka (1939) and described them as S. lyra. The Discovery collec- tions from Tasmanian waters contain specimens of S. gazellae, and their presence there may be due to vagaries of the subtropical convergence. Furnestin (1953) reports S. gazellae from the Mediterranean, but has informed me that, after reading a typescript copy of part of this paper, she considers them to be S. lyra "gazellae "-type. Table 13 a gives the numerical data from 496 shallow oblique hauls. The year has been divided into four equal seasons, and the table is in two parts, one for the Subantarctic, and the other for the Antarctic zone. The difference in size of the population for the two zones is immediately apparent, the average for 165 Subantarctic samples is 120-3 per 20-minute haul, as compared with 27-2 based on 331 Antarctic samples. The striking decrease in numbers in the upper layers during the winter months is evidence of a seasonal migration, comparable with that described by Mackintosh (1937) for certain other Southern Ocean plankton species. Though in the Subantarctic the total for August is rather exceptional owing to a single haul at Station 2869 of 502 individuals, three other hauls made at different times of the year in almost the same position (in the vicinity of the Crozet islands) all yielded very large numbers, and the presence of S. maxima in quite large numbers at the surface in this area suggests that some exceptional conditions exist there. The haul at Station 2869 is the only August haul in the collections which contains large numbers of S. gazellae, and if it is omitted the average for August becomes 48-0, and the winter average 66-3 ; these figures (shown in brackets in Table 13 a) give a more accurate picture of normal conditions. The Subantarctic winter average is low because a part of the population is at a depth of more than 100 m. In spring the average increases with the return of most of the population to the surface. The summer average is the highest owing to the appearance of young individuals at the surface. In autumn the average decreases for two reasons, the commencement of the winter vertical migration in April and May, and the extensive migration in March and the beginning of April in which the larger size groups of the population descend to the deeper layers to mature and breed. However this decrease is to some extent offset by the appearance in the surface layers in May of young from the March April breeding period. 264 DISCOVERY REPORTS The Antarctic population, shows a similar rise and fall in numbers, but the peak appears to be later in the year, and the summer and autumn averages are almost the same. The low average in February may be due to the comparatively small number of hauls available in this month, and if this is so then possibly the summer total is rather too low. The extent of the winter migration in the Antarctic can be seen by comparing the totals and averages for the months during which the animals are in deep water (May-September) with the totals and averages for the remainder of the year (Table 136). Table 130. Monthly and seasonal totals and averages per 20 min. haul zvith N100B in the surface layers (For explanation of figures in brackets see p. 263.) Subantarctic Antarctic Month Total Hauls Average Total Hauls Average Winter June 1,013 H 72-35 309 *7 18-17 July 461 6 76-8 18 6 3 Aug. 886 (384) 9(8) 98-4 (48-o) 179 25 7-1 2,360 (1,858) 29 (28) 813 (663) 506 48 105 Spring Sept. 3.989 28 142-4 196 37 5-3 Oct. 1. 5i9 •5 101*3 634 29 21-8 Nov. 2,916 25 Il6-6 i.43o 52 27-5 8,424 68 123-8 2,260 118 191 Summer Dec. 1. 595 7 227-8 926 31 31-0 Jan. 685 7 97-8 2,292 48 47-75 Feb. 1,887 H 1347 204 H H-5 4,167 28 1488 3.458 93 367 Autumn Mar. 1,077 5 2I5-4 1.231 17 72-4 April 616 12 5I-3 1,105 35 3!-57 May 3.217 23 139-8 459 20 22-9 4,910 40 122-7 2.795 72 388 Totals 19,861 165 9,091 331 — Table 13ft. A comparison of winter and summer totals from 1 m. oblique net hauls in Antarctic water Season Months Totals No. of hauls Average Winter Summer May-Sept. Oct.-Apr. 1,161 7,858 105 226 11-05 3476 From this it can be seen that the surface population is three times larger in summer than in winter. Although this is due in part to the presence of many young broods, it must be remembered that there are very large numbers of young hatched in the autumn, and therefore a major part of the difference between the two seasons is due to migration. The numbers of S. gazellae decrease in the vicinity of the Antarctic Convergence, and although this is not very apparent in Fig. 17, Fig. 18 shows some North to South lines with the stations plotted according to their distance in nautical miles from the Convergence, which demonstrate the point more clearly. The position of the Antarctic Convergence fluctuates slightly, and it is likely that the sudden and violent changes in physical conditions brought about by these fluctuations are unfavourable to S. gazellae (Eukrohnia hamata is numerous in this area and is presumably not so sensitive to environ- mental fluctuations). SAG ITT A GAZELLAE 265 BOO 6OO NORTH 400 200 AlC ZOO NAUTICAL MILES 6OO BOO SOUTH IOOO Fig. 18. A series of hauls from north-south lines of stations plotted by distance in nautical miles from the Antarctic Con- vergence. The position of the Antarctic Convergence has been determined by reference to the continuous thermograph record from each line. VERTICAL DISTRIBUTION Fowler's (1898, 1906) method of dividing the water column into three zones, the epiplanktonic, mesoplanktonic and hypoplanktonic is generally used in describing the vertical distribution of the chaetognatha. 5. gazellae, as I shall show, is both epiplanktonic and mesoplanktonic according to the season of the year, and its state of maturity. Fig. 19, which has been compiled from the numerical data from 129 vertical stations, shows the average depth distribution of Sagitta gazellae at all seasons in both the Subantarctic and Antarctic zones. Although these vertical hauls (made with the 70 cm. net) can be checked by the deeper oblique nets to determine whether the falling off in numbers below 100 m. is real or due to inaccurate sampling of larger size groups by a small, comparatively slow-moving net, it is not possible to check the difference between the 50-0 and 100-50 m. layers by the oblique hauls, as the shallowest layer normally fished is 100-0 m. Nevertheless, the vertical catches seem consistent. The species extends in quite considerable numbers up to the surface, for although the horizontal surface nets have not been systematically examined, occasionally they have been observed to contain z66 DISCOVERY REPORTS large numbers of S. gazellae. These are nets which have been towed at many stations at a depth of less than 5 m. below the surface. Fig. 20 shows the size and numbers of individuals taken in the 1 m. oblique nets at Station 461 (56°44' S. o° 23' W, 21.X.30), a 24-hour station in Antarctic water, and it demonstrates that size increases with depth. The animals have been plotted on Fig. 20 by their length in mm. against the % OF TOTAL 20 JO IOOO I500 J 300 4 OO 500 z 6OO I Q- UJ a 700 . .jt,'.v-MSj t at ■ vvtf • » • • •• •/ • • • tc >. /•• 90 Fig. 19 10 20 30 iO 50 60 70 8C TOTAL LENGTH MM Fig. 20 Fig. 19. The vertical distribution of S. gazellae at 129 vertical stations: the numbers at each depth are represented as a percentage of the total for all depths. Fig. 20. The vertical distribution of S. gazellae at Station 461 showing the increase in size with depth. mean depth of each haul, e.g. the mean depth for a haul from 280 to 180 m. has been assumed to be 230 m. The scatter of the observations when plotted is considerable but, although there are a number of animals of the smaller size groups present in the deeper hauls, it must be noted that these were made with flights of oblique nets, and therefore the paying out of the nets must be halted from time to time for attachment of more nets; thus the deeper ones may have sampled shallow layers for short periods of time. As the small animals are present in the upper layers in large numbers, it is likely that many of the small animals in the deep hauls are due to this source of error. SAGITTA GAZELLAE 267 It is also evident from Fig. 20 and from Table 18, Appendix, which gives the numbers plotted on Fig. 20, that there is a considerable decrease in numbers with depth. This confirms the evidence from the series of vertical stations mentioned above. The evidence from the vertical series suggests that the maximum concentration of S. gazellae is in the 100-50 m. layer, and this confirms Bollmann's (1934) statement for both S. lyra and S. gazellae (considered together). The data in Fig. 20, however, suggest that the main concentration is between 100 and 150 m. The numbers in this figure have been plotted against the mean depth of each haul and the concentrations between no and 135 m. may well have been taken at a depth of less than 100 m., as the nets which took them fished into the 100-50 m. layer. The data from the vertical nets may be considered to be more reliable in this respect. The depth range of the species obtained from the samples available extends down to 3000 m., but it is quite possible that its total range may reach to the bottom; at great depths specimens are rare. Table 12 shows the numerical data from thirty-five daylight and thirty-five night vertical stations and suggests that there is no diurnal migration in the species; Fig. 21 shows the depth distribution during the 24-hour Station 461, and confirms the absence of diurnal migration. 2230 DARK HOURS 0230 0430 IO30 Fig. 21. The vertical distribution of 5. gazellae at Station 461 showing absence of diurnal migration. Data from oblique 1 m. nets. Horizontal scale 1 mm. = 2 individuals. Fig. 22, which has been constructed by the method described by Mackintosh (1934, p. 377), shows the monthly vertical distribution of the species in the meridian of o°, and demonstrates the seasonal vertical migration in winter. Table 19, p. 278, gives the numerical data from which the figures were compiled. In the Antarctic (south of 510 on the meridian of o°) this migration is into the warm deep layer (between approximately 150 and 1000 m) (Fig. 23 a, b confirms this), but in the Subantarctic (north of 510) the migration is not so great, and only in August does the centre of maximum abundance of the population move below 100 m. However, a study of the size-group percentages in the surface 100 m. as sampled by the 1 m. oblique nets in winter (Fig. 26, see p. 271) reveals that the population is mainly composed of one brood, hatched out in late March and early April ; and as there is evidence that breeding has been taking place throughout the summer (see p. 274) one must conclude that the larger individuals have migrated to below 100 m. This is confirmed by the size of individuals taken in the N70V nets at this time of year, and also by Fig. 24 which is a comparison of size-group percentages from the shallow (100-0 m.) and deep (250-100 m.) oblique nets during the months of May and June. Fig. 23 a, b referred to earlier gives the numerical data from N 70 V nets from the o° line plotted on a temperature-salinity diagram, and shows the shift of the Antarctic population from the surface water towards the warm deep water in winter. This would be more striking if it were not for the fact that the 250-100 m. nets pass through both the warm deep and the Antarctic surface layers; and although it is probable that the animals in a haul come from the warm deep and not from the Antarctic surface water (for example when there is a large number in the 500-250 m. net, but none in the 50-0 and LATITUDE 40 SO 2494 2496 2501 2492 I 2495 I 2498 60 70 40 50 60 1772 1774 1776 1778 1781 1771 1773 1775 1777 1779 1782 70 750 750-1 250 MAR 500- APR 2386 : 23.85 2387 >389 2393 12391 1 50 IOOh ^mT K 250- 1 AUG ^ / \ 500- 1806 1808 I8IO 500- 750 750 Fig. 22. The vertical distribution of S. gazellae for 8 months of the year on the meridian of o°. The shading represents density per 250m. haul: (i) (the darkest) >4i individuals, (ii) 21-40 individuals, (iii) 11-20 individuals, (iv) (the lightest) 5-10 individuals. Hauls containing less than 5 individuals have been disregarded. The mean position of the Antarctic Convergence on this meridian is 510 S. See Table 19, p. 278. SAGITTA GAZELLAE 269 100-50 m. nets) it is still necessary to use the temperature and salinity data for the whole 250-100 m. horizon when calculating the mean for the haul, and this inevitably shifts it away from the warm deep layer. The change in position of the Subantarctic population in winter is also evident in Figs. 23 a, b, but not enough is known about the characteristics of the surface and subsurface waters in this area to draw any conclusions from these figures, other than to point out a movement down towards the Antarctic intermediate water, which bears out the data from oblique nets. JO 19 18 17 16 IS 14 I 3 12 I I 10 9 8 7 6 5 4 2 («) MARCH • 1-5 • 6-19 #20 -i- (b) JULY- AUGUST . • 33 6 7 89 34 I 234567 9 35 I 2 3 4 9 34 1 2 3 4 5 Fig. 23. The vertical distribution of S.gazellae in relation to temperature and salinity, based on three lines of stations on the meridian of o°, one in summer and two in winter. The core of the South Atlantic Central water is marked S. A. and that of the Antarctic Intermediate water A.I. Above the Antarctic Intermediate water and to the left of the South Atlantic Central water is the Subantarctic surface water. Below the Antarctic Intermediate water, enclosed by two lines is the warm deep water, while to the left of this is the Antarctic surface water. Open circles represent hauls containing no S. gazellae. Fig. 25, in which the animals at stage III and IV maturity from N70V hauls have been plotted against temperature and salinity, shows that they are concentrated in the warm deep layer. This is discussed further on pp. 270-1, but the fact must be mentioned here in order to obtain a complete picture of vertical distribution. These mature animals habitually migrate to deep water (750- 2000 m.) to breed, and animals sinking down for this purpose are taken at intermediate depths. It has been shown above that for most of its life, S. gazellae is an inhabitant of the surface waters, and is mainly concentrated in the 100-50 m. layer; but in the winter months in the Antarctic it sinks to the upper layers of the warm deep water between 250-100 and 500-250 m. At the same season in the Subantarctic, the population is about equally concentrated in the 100-50 and 250-100 m. layers, partly in Subantarctic surface water, and partly in Antarctic intermediate water. In winter the smallest individuals form that part of the population which is nearest to the surface, the larger-sized animals occurring below, and increasing in size with increasing depth. In summer the 5-z 270 DISCOVERY REPORTS larger-sized animals are mixed with the small in the surface layers, but large animals descending to breed are found in the deeper layers. The mature animals are normally only found below 750 m., in the warm deep water. /o 60 r— MAY 50 SHALLOW QOO-O) 40 30 20 IO r^ 60 MAY SO DEEP (250-IOO) 40 JO 20 \ 10 H .... 60 JUNE 50 SHALLOW (ioo-q) 40 1 1 30 20 10 1 ' - I 60 JUNE 50 DEEP (2SO-IOO) 40 30 20 10 r~J ~L- 3 r-n ™ 0 5 10 15 20 25 30 35 40 45 50 55 60 65 SIZE GROUPS Fig. 24 10 9 8 7 6 5 4 3 2 UJ 5 i < a o UJ |-' UJ 1- -2 336 7 SALINITY 9 34 I 2 -3 4 5 6 7 8 9 35 Fig- 25 Fig. Fig. 24. The vertical distribution of size groups of 5. gazellae in early winter, from shallow and intermediate oblique i m. nets. 25. The distribution of mature specimens of 5. gazellae in relation to temperature and salinity. VERTICAL DISTRIBUTION OF MATURE ANIMALS Table 14 shows the depth distribution of all mature specimens taken in closing nets (N450, TYF, N 100, N70). Here I class as 'mature' those specimens in which the seminal vesicles have grown to full size, and these include all in the stages of maturity from II/III to V. It is evident from the data that the advanced stages are very rare at depths above 750 m., and probably most common in the 1 500-1000 m. layer. The number of hauls made below 1000 m. is small compared with the number made in the shallower layers, and therefore the one stage III individual taken in the top 250 m. is the only mature specimen from approximately 35,000 immatures observed from this layer, whereas the fifteen mature specimens from the 1500 to 1000 m. layer represent a high percentage of the total numbers taken in this layer. Deep hauls have been searched for mature specimens, and Table 1 5 has SAGITTA GAZELLAE 271 been compiled from ninety-three vertical stations, the number of individuals at each stage of maturity being expressed as a percentage of the total numbers in each horizon. Both races of S. gazellae are combined and the stations are from nearly all months of the year. It can be seen that in the deepest hauls (1 500-1000 m.), the mature stages III and IV represent more than 50% of the total. Stage I is the commonest in the 100-50 m. layer, though the difference in percentage composition in the three shallow layers is small. Stage II reached a maximum in the 750-500 m. layer, falling off above and below this level. Stage III appears in the 1000-750 m. layer, being as numerous as stage II at that depth. Stage IV appears in the 1500- 1000 m. layer, but is not as numerous as stage III. Table 14. Vertical distribution of mature specimens from closing nets Depth (m.) Stages of maturity II/III III III/IV IV IV/V V 250-0 — 1 — — — — 500-250 — — I — — 750-500 — 1 — I — — 1000-750 4 5 — 4 8 i I 1 500- 1 000 2 1 1 1 4 4 — 2 >i5oo 1 1 1 2 2 — 3 (S.N. race, light type; L.S. race, heavy type.) Table 15. Vertical distribution of all stages at ninety -three stations — each stage given a percentage of the total for each depth (percentages in heavy type) Depth (m.) Stages of maturity I II III IV No. of hauls 50-0 100-50 250-100 500-250 750-500 1000-750 1 500- 1 000 1075 964 1990 985 695 974 141 893 30 639 9 474 6 300 4° 3'5 3° IS 19 26 17 107 17 361 5 263 3 i5-° 5 263 9 4S« 2 100 92 92 93 93 90 89 73 GROWTH RATE AND LIFE HISTORY The seasonal vertical migration of S. gazellae, in which part of the population sinks below the horizon of the shallow oblique nets in winter, is the main factor which makes a study of the growth rate of the species possible. In summer months the range of size groups in the population is so great that it is not possible to identify, or follow the progress of, one particular brood among all the others. In winter, however, one brood is left within the horizon of the shallow (100-0 m.) oblique nets, and its progress can be observed from May when it first appears until September and October when it becomes mixed with other broods, and its individuality is lost. This, however, applies only to the Subantarctic zone, for in the Antarctic the seasonal migration carries the population almost entirely below the horizon of the shallow oblique nets, and in addition the ice conditions prevent a study of the southern part of the Antarctic zone during the winter months. Thus the growth rate and life history of the L.S. race can only be inferred from that of the S.N. race. 272 DISCOVERY REPORTS Fig. 26 shows the data from the shallow oblique (1 m.) nets fished in the Subantarctic zone during the months of May to November. The animals have been plotted in 5 mm. size groups, the numbers of each group calculated as a percentage of the total haul or set of hauls. Each haul or set of hauls has been plotted against the day or mean of the days on which they were made. Regrettably it has not been possible to utilize data from one area alone (or from one year alone) as the numbers available in the series of repeated stations on o° have been insufficient by themselves. Thus Fig. 26 is composed of data from many localities taken in several different years. The results are, however, remarkably consistent, and this suggests that the general pattern of breeding and development is the same from year to year throughout the Subantarctic zone. The variation in the upper size limit during the winter months may well be due to the variation in depth of the nets; for example the first set of June observations made at Stations 2824 and 2825 show a greater number in the larger size groups than the other June observations, because the net at Station 2824 was fished from 163 to o m., and thus sampled the older and larger-sized broods below 100 m. The smallest size groups are affected by the occasional presence of a small brood hatched later than the main one under consideration ; this later brood seems to separate out in September. 881/92 I 2648/') 2538 1 r 2459/60 983/90 2736/36 1614 ^r 1 r Fig. 26. The growth of S. gazellae, based upon counts and measurements of specimens taken in the shallow oblique 1 m. nets. The station numbers of the hauls used are given along the top and the number of individuals in each haul or set of hauls is enclosed in each figure. The growth rate from May to October is shown by Fig. 26 to be about 5 mm. per month. This is evidently a low rate which is most probably higher in spring and summer, and which appears to be lowest during August. The normal size at maturity is 55-60 mm. and this would be reached in 12 months at a 5 mm. per month growth rate : Figs. 8 and 15 suggest that the maximum size is reached in the surface layers, for none of the mature animals (found in the deep layers) was larger than at stage II which is the most advanced stage found in the surface layers. Thus in considering the growth rate one must make allowance for both the time needed for the breeding migration, and the time taken for the maturation of the gonads. It has not been possible to discover how much time elapses between the commencement of the breeding migration and the shedding of eggs. If one assumes that for at least 1 month, and possibly 2, no growth takes place, then the summer growth rate must be somewhat higher to achieve a length of 60 mm. in 10 or 11 months. This seems a reasonable assumption, as the food supply is much greater in spring and summer. There are certain stages in the life history of S. gazellae which I have been unable to find, but in most cases they can be inferred from other evidence, and it may be possible at some future date to fill in the gaps, now that it is known when and where to look for those stages. SAGITTA GAZELLAE 273 I shall describe the life history of one particular brood, referred to as the March/April brood, which is observed during the winter months in the surface 100 m. Then having established the pattern of its life history I shall proceed to discuss that of other broods. Since the fully mature specimens are found only in deep water it can be assumed that the eggs are laid also in deep water, and that they, or the newly hatched young, rise to the surface. Certainly the youngest individuals taken in our nets occur in the surface layers. Judging by the growth rate indi- cated in Fig. 26 it seems likely that eggs are laid in about late March or early April. The youngest individuals are found between 250 m. and the surface in early May. Although at this stage they are mainly in the 10-15 mm- grouP> there are still quite a large number in the 5-10 mm. group; individuals at a length less than 5 mm. are rare at all times. It seems likely that hatching takes place in the top 250 m. as no very small individuals have been found below this depth (see Table 16), though specimens Table 16. Monthly distribution of 0-10 mm. size group, from seventy-one vertical stations Depth (m.) Months Total Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. 50-0 25 20 100-50 65 30 250-100 3 2 500-250 0 0 750-500 0 0 1000-750 0 0 1500-1000 0 0 No. of 9 6 stations 10 65 3 0 0 0 0 10 0 30 95 10 0 0 0 0 13 15 90 12 0 0 0 0 6 5 5° 37 0 0 0 0 11 0 10 3° 5 0 0 0 0 6 0 0 3 0 0 0 0 6 5° 45 18 0 0 0 0 7 0 165 470 93 0 0 0 0 71 of other forms (S. maxima, S. planctonis and Eukrohnia hamata) are common in the deeper nets, at lengths from 2 to 5 mm. I have been unable to find the eggs, but judging from their size in the mature ovaries they must be very small when laid, and I do not know how to distinguish them from eggs of other species. Kramp (1939) states that some of the eggs of Sagitta maxima, a species which also breeds in deep water, rise to the surface to hatch. The ovaries of the living mature S. gazellae are pale yellow in colour, and this may be due to the presence of fat or oil which might help the eggs to rise. From May to late July the March/ April brood is concentrated in the 100-50 m. layer. In late July and August they sink to the 250-100 m. layer (Fig. 22), and the upper layers are almost completely empty of this species (the haul at Station 2869 in August is quite exceptional (see p. 263) though very fortunate, as without it there would be no data for August in Fig. 26). During the winter months the animals grow at a rate of 5 mm. a month, and by September are mainly in the 25-30 mm. group. The growth rate increases in the spring, and by the end of October specimens have reached the 35-40 mm. group, but the individuality of the brood is lost by mixture with other broods which have ascended from deeper water. The March/April brood remains in the surface until February or March, and attains stage II maturity and full size. The breeding migration (p. 270) now commences, and the animals sink, passing from stages II to III at about 750-500 m., down to the 1000-750 m. layer, where they are usually at stage III, and on to the 1 500-1000 m. layer where they are mainly at stage IV. Loss of the posterior teeth occurs between stages II and III at about 750 m., and the rest of the head armature also decreases. At the advanced stages of maturity the hooks are often damaged and broken, but the main reduction in hook numbers is effected by normal shedding from the ventral end of the hook rows. After leaving the surface the animals do not appear to increase in length; their food is presumably used almost entirely by the rapidly developing ovaries, 274 DISCOVERY REPORTS which at stage III are thin rod-shaped structures, sometimes as little as 14 % of the total length but which at stage IV are thick swollen structures occupying as much as 70% of the total length. It is probable that fertilization takes place when the ovary/body length percentage is between 55 and 60, and that the eggs are laid when this percentage is about 70. No animals with seminal vesicles intact have been taken with the ovary/body length percentage in excess of 55, and none at all with this percentage in excess of 7 1 . Spawning takes place throughout the spring, summer and autumn, and perhaps to a small extent in winter. This prolonged breeding season is evident both from the fact that small-size groups are always present in the population, and from the almost complete overlapping of all size groups which makes the interpretation of a figure such as Fig. 26 impossible during the late spring and summer months. Fig. 26, however, indicates the presence of several recognizable broods apart from the March/April one. Traces of a later brood, probably hatched in April/May, are evident at Station 965 in September, and this brood may account for the 5-10 mm. group at Station 2536 in July; the graph for Stations 2424 and 2425 is based mainly on this brood, which is still recognizable at Station 2451 in early October and at Stations 983-90 later in the same month; its individuality is lost in November. It is possible that with the advent of spring the larger animals of the March/April brood grow at a more rapid rate than do the smaller animals of the April/May brood, and thus for a month there is a noticeable gap between the peaks for the two broods. At Station 2451 in October the large size groups appear. These are from broods hatched in late spring and summer of the previous year, and have spent the winter months in deep water. It is probable that the stock which gives rise to the earliest spring broods never ascends to the surface after its winter migration. These large animals are absent from the shallow hauls in November, and evidently begin their breeding migra- tion early in that month (there are always at least a few large individuals in the shallow hauls in every month). The peaks of the November graphs are probably due to a combination of the March April and April/May broods, the point of overlap of the size ranges of each brood giving a combined total in excess of the peak of either brood separately. The areas of greatest density of population in the vertical sections (Fig. 22) for December to May are composed mainly of animals less than 20 mm. in length, and the apparent shift of the centre of density of the population southwards during the summer suggests that the earliest breeding takes place in the most northerly part of the Sub- antarctic zone, and as the summer progresses the centre of breeding moves southwards. It is not possible to say much about the life cycle of the L.S. race. The normal size at maturity is from 75 to 80 mm., and no mature animals in excess of 90 mm. have been recorded. If the growth rate is assumed to be 5 mm. per month, as in the S.N. race, then the time taken to reach full size will be 15-16 months. However, the growth rate may be expected to be considerably higher in summer, and a 12-month life cycle seems probable on these assumptions. Although most of the mature specimens of the L.S. race have been taken in March, this is probably Fig. 27. A large non-breeding specimen of S. gazellae L.S. race, 100 mm. long. SAGITTA GAZELLAE 275 due to the greater number of deep towed nets fished in this month and hence the larger quantity of specimens available. The N70V net, though it obtains specimens in perfect condition, catches only a few of the smaller Subantarctic race, and it catches still fewer specimens of the large Antarctic race. The large deep-towed nets, which are the main sources of supply of specimens of the mature L.S. race, have not been regularly fished, and so no real indication of the frequency of breeding can be obtained from the catches of mature animals. Very large individuals (>9omm.) are found in the Antarctic zone; these are apparently non- breeding (Fig. 27), and are usually found at the surface in winter and spring; they do not appear to perform either a seasonal or a breeding migration. They are never found at a more advanced stage of maturity than stage II, and the largest specimen taken to date measured 105 mm. (fresh). This is probably not the largest, for preserved specimens of 99 and 100 mm. have been examined on several occasions, and taking shrinkage into account (see p. 240) the maximum length fresh is probably 108-110 mm. It is possible that in certain cases individuals have not been able to migrate to breed because they have been in shallow water, and have therefore continued to feed and grow at the surface ; the large specimen of 105 mm., for example, was taken at Station 1007 where the sounding was only 152 mm. However, in many cases these large animals are found over deep water, and one must assume that there is another factor which operates to prevent both breeding and seasonal migration. FOOD S. gazellae is frequently taken with animals held in its jaws, but this is not necessarily any indication of its normal food, for often the animals which it has seized are far too large to be ingested — large medusae or fish for example — and not infrequently specimens are found which have bent back upon themselves and seized their own tails. While a net is being hauled the animals concentrated in the net bucket are swirled together, and evidently S. gazellae will seize anything it touches ; during fixation also, the animals make convulsive movements which may have a similar result. Observations have been made of specimens with food in the gut, and owing to the great trans- parency of the animal it has sometimes been possible to identify the species of food animal. These positive observations are however surprisingly few. In two commissions of 'Discovery II', during 1931-33 and 1950-51, only in forty-five hauls out of 667 examined were specimens of S. gazellae recorded with food in the gut; this does not mean only forty-five specimens, for in many cases there were a number of animals with food in their gut found in one haul, but even so the number is very low. Table 17 shows that Copepods and Euphausians are the commonest food animals for the species, the third group, Chaetognaths, being but 13% of the total. No mature S. gazellae has been found with food in the gut, but the total number of matures taken is small, and this may therefore be of no significance. On one occasion a large specimen of S. gazellae was kept alive aboard ship for nearly 2 hours, and was given a live Euphausian to eat. The food was touched against the mouth of the chaetognath which promptly seized it and pushed it into the mouth by alternate movements of the hooks on each side ; once in the mouth the food was passed rapidly down to the middle of the gut, this process taking about 10 minutes; from this point to the posterior end of the gut the passage of the food was much slower, and it reached the posterior end about 1 hour after ingestion. The animal died before it had defaecated. The behaviour of one specimen in far from natural conditions must of course be regarded with reserve. Chaetognaths form part of the diet of many marine animals, notably of fish. Large pelagic fish are 276 DISCOVERY REPORTS uncommon in the Southern Ocean, and so it is not likely that they represent major predators of S. gazellae. Dr T. J. Hart has observed a Cape pigeon (Daption capensis) holding a large sagittain its beak, and from the position of the observation it is fairly certain that the sagitta was S. gazellae. S. gazellae is probably eaten by most of the larger plankton-feeding organisms in the Southern Ocean, but no records are yet available. Table 17. Food organisms found in the gut of Sagitta gazellae arranged by groups in order of frequency of occurrence Species Group No. Percentage 42-2 40-0 13-3 4-5 Calanus propinquus 2 Rhincalanus gigas I Pleuromamma robusta 1 Metridia gerlachii 1 Calanus sp. 3 Other Copepoda 1 1 Thysanoessa macrura 2 Thysanoessa vicina 1 Thysanoessa sp. 5 Euphausiid larvae 3 Euphausiid spp. 7 Eiikrohnia hamata 2 Chaetognatha spp. 4 Primno macropa 1 'Large red crustacean' 1 Copepoda Euphausiaceae Chaetognatha Other Crustacea 19 18 6 2 SUMMARY i . Sagitta gazellae has been examined from a series of stations made by ships of the Discovery Committee in the Southern Ocean. 2. The species is shown to be separate from S. lyra (Krohn) with which it has been considered synonymous in recent years. 3. Five stages of maturity based upon those of Kramp (1939) for S. maxima are outlined. 4. The population of S. gazellae in the Southern Ocean consists of two races divided by the Antarctic Convergence. The Northern or Subantarctic population matures at a shorter length than does the Southern Antarctic one. 5. The limits of horizontal distribution are shown to be, to the north, the region of the subtropical Convergence, and to the south, the Antarctic Continent. The region of maximum abundance of the species is in the Subantarctic zone. 6. The species has an observed vertical range of 3000-0 m., with a layer of maximum abundance between 100 and 50 m. 7. S. gazellae performs a seasonal vertical migration to deep water in winter, but does not perform any diurnal migration. 8. The Subantarctic race is shown to have a growth rate of 5 mm. per month during the winter and early spring. The growth rate is presumed to be higher in summer. 9. The species breeds only in deep water (>750 m.). The breeding season is very extended. The smallest individuals observed were taken in the upper 250 m. 10. S. gazellae feeds mainly upon copepods and small euphausians. SAGJTTA GAZELLAE 277 REFERENCES Baker, A. de C, 1954. The Circumpolar Continuity of Antarctic plankton species. 'Discovery' Rep. XXVII, pp. 201-18. Bollmann, A., 1934. Die Chaetognathen der deutschen antarktischen Exped. auf der ' Deutschland' 1911-1912. Int. Rev. Hydrobiol. (Leipzig), xxx, pp. 251-305. Burfield, S. T., 1930. Chaetognatha. Brit. Antarct. Terra Nova Exped. (Zool.), vn(4), pp. 203-28. Deacon, G. E. R., 1937. The Hydrology of the Southern Ocean. 'Discovery' Rep. xv, pp. 1-124. Fowler, G. H., 1898. Contributions to our Knozvledge of the Faeroe Channel. Proc. Zool. Soc. Lond. 1898, p. 545. 1905. Biscayan Plankton of H. M.S. 'Research'. Part HI. Chaetognatha. Trans. Linn. Soc. Lond. (Zool.), ser. 2, X, PP- 55-87- ■ 1906. The Chaetognatha of the Siboga expedition. Siboga Exp. Reps. XXI, pp. 1-86. 1907. Chaetognatha. Nat. Antarctic Exped. 1901-1904, III. Zoology and Botany, pp. 1-6. 1908. Notes on a small collection of Plankton from New Zealand, I. Ann. Mag. Nat. Hist., ser. 8, I, pp. 240-1. Fraser, J. H., 1952. The Chaetognatha and other Zooplankton of the Scottish Area and their value as Biological Indicators of Hydrographical Conditions. Scottish Home Dep., Marine Research, no. 2, pp. 1-52. Furnestin, M. L., 1953. Chaetognathes recoltes en Mediterranee par le 'President Theodore Tissier' . Travaux publies par la Station d'Aquiculture et de Peche de Castiglione. New ser. no. 4, pp. 1-44. George, P., 1952. A systematic account of the chaetognatha of the Indian coastal zvaters zvith observations on their seasonal fluctuations along the Malabar Coast. Proc. nat. Inst. Sci. India, xvm, no. 6, pp. 657-689. Germain, L., 1913. Chetognathes. Deuxieme expedition Antarctique Francaise (1908-10). Sciences Naturelles: Documents scientifiques — zoologie, pp. 87-104. Germain, L. and Joubin, L. 1916. Chetonathes provenants des campagnes des yachts ' Hirondelle' et ' Princesse Alice', 1885-1910. Res. Camp. Sci. Monaco xlix, pp. 1-118. Ghirardelli, E., 1950. Osservazioni biologiche e sistematiche sui Chetognati della Baie du Villefranche sur mer. Boll. Pesca, Piscic. Idrobiol. xxvi, v (n.s.), Fasc. 1, pp. 1-27. 1952. Osservazioni biologiche e sistematiche sui Chetognati del Golfodi Napoli. Pubbl. Staz. Zool. Napoli,xxm, pp. 296-3 12. Hamon, M., 1952. Note complementaire sur les Chetognathes de la baie d' Alger. Bull. Soc. Hist. nat. Afr. N. xliii, pp. 50-2. Hart, T. J., 1934. On the phytoplankton of the South West Atlantic and the Bellingshausen Sea 1929-31. 'Discovery' Rep. vm, pp. 1-268. Jameson, A. P., 1914. The Chaetognatha of the Scottish National Antarctic Expedition 1902-1904. Trans, roy. Soc. Edinb. xlix, pp. 979-89. John, D. D., 1936. The Southern Species of the genus Euphasia. 'Discovery' Rep. xiv, pp. 193-324. Johnston, T. H. & Taylor, B. B., 1921. The Chaetognatha. Aust. Antarct. Exped. Ser. C, vi(2), pp. 1-17. Kemp, S., Hardy, A. C. & Mackintosh, N. A., 1929. The Discovery Investigations, Objects, Equipment and Methods. 'Discovery' Rep. I, pp. 141-232. Kramp, R. L., 1939. The Godthaab expedition 1928. Chaetognatha. Medd. Gronland, no. 80, part 5, pp. 1-44. Kuhl, W., 1938. Chaetognatha. Browns Klassen. Ord. Tierreich iv(2), pp. 1-266. Mackintosh, N. A., 1934. Distribution of the Macroplankton in the Atlantic sector of the Antarctic. 'Discovery' Rep. ix, pp. 65-160. 1937- The seasonal circulation of the Antarctic Macroplankton. 'Discovery' Rep. XVI, pp. 365-412. Marr, J. W. S., 1938. On the operation of large Plankton nets. 'Discovery' Rep. XVIII, pp. 105-20. Michael, E. L., 191 i. Classification and vertical distribution of the Chaetognatha of the San Diego Region. Univ. Calif. Publ. Zool. vm, pp. 21-186. 1919. Report on the Chaetognatha of the Albatross expedition to the Philippines. Bull. U.S. Nat. Mus. ioo, I, part 4, PP- 235-77- Ommanney, F. D., 1936. Rhincalanus gigas (Brady). A copepod of the Southern macroplankton. 'Discovery' Rep. xm, pp. 277-389. Pierce, E. L., 195 1. The Chaetognatha of the West Coast of Florida. Biol. Bull., Woods Hole, c, no. 3, pp. 206-28. Ramoult, M. and Rose, M. 1946. Recherches sur les Chetognathes de la baie d' Alger. Bull. Hist. nat. Afr. N. xxxvi, pp. 45-71. Ritter-Zahony, R., 1909. Die Chatognathen der Gazelle-Expedition. Zool. Anz. xxxiv, pp. 787-93. 1911. Revision der Chaetognathen. Deutsche Stidpolar Exped. xm(5), pp. 1-71. Russell, F. S., 1932. On the biology o/Sagitta. The breeding and grozvth o/S. elegans in the Plymouth area 1930-31. J. Mar. biol. Ass. U.K. xviu(i), pp. 131-45. Steinhaus, O., 1900. Chatognathen. Ergebn. d. Hamb. Magalhaens. Sammelreise, m, 10 pp. Thiel, M. E., 1938. Die Chaetognathen-Bevblkerung des sudatlantischcn Ozean, XIII, no. 1. Wiss. Ergebn. dtsch. atlant. Exped. 'Meteor'. Thomson, J. M., 1947. The Chaetognatha of South-eastern Australia. Bull. Coun. Sci. industr. Res. Aust. no. 222, pp. 1-43. Tokioka, T., 1939. Chaetognatha collected chiefly from the Bays of Sagami and Suruga with some notes on the shape and structure of the seminal vesicle. Rec. Oceanogr. Wks. Jap. x, no. 2, pp. 123-50. 1940. A small collection of Chaetognaths from the coast of New South Wales. Rec. Aust. Mus. xx, no. 6, pp. 367-79. 278 DISCOVERY REPORTS APPENDIX Table 18. Numbers of Sagitta gazellae taken in oblique i m. nets at a z\-hour station (461) 461,4 4615 461 C 461 D 461 £ 461 F Total 80-0 0 75-0 0 95-0 2 85-0 13 75"° 4 80-0 8 27 175-80 10 160-75 27 200-95 23 180-85 22 160-75 36 i75-8o 35 153 270-170 10 255-160 15 310-200 13 280-180 4 245-160 13 270-175 6 58 385-270 3 345-255 3 420-3 10 8 385-280 14 330-245 5 375-270 6 37 5!°-385 8 440-345 7 535-420 12 490-385 16 420-330 5 490-375 6 52 650-510 10 520-440 8 660-535 6 600-490 2 5^-420 3 615-490 14 43 Table 19. The numbers of Sagitta gazellae taken in the 70 cm. vertical nets on the meridian of o°, numbers corrected for a 250 m. haul. Fig. 22 has been compiled from these data November-December May-June 2492 2494 2495 2496 2498 2501 50-0 60 5 0 0 0 0 100-50 125 10 0 0 15 30 250-100 3 10 0 0 2 0 500-250 0 1 0 4 2 0 750-500 0 0 0 1 1 0 1000-750 0 0 0 0 0 1 1 500- 1 000 0 0 0 — 0 — 1771 1772 1773 1774 1775 1776 1777 1778 1779 1781 1782 50-0 0 0 0 0 5 10 5 0 5 0 0 100-50 0 0 0 5 0 65 15 10 70 15 5 250-100 0 0 2 10 16 20 S 5 16 2b 15 500-250 0 0 0 20 18 3 1 0 1 2 0 750-500 0 0 0 0 2 0 0 2 1 0 1 1000-750 0 0 0 0 1 0 0 0 0 0 1 1 5 00- 1 000 0 0 1 0 1 0 0 1 0 0 January July 2530 2531 2532 2533 2535 2538 2541 2543 2545 2547 50-0 15 5° 15 0 0 0 10 10 0 0 100-50 24 25 100 15 15 15 25 5 55 0 250-100 8 5 0 0 3 5 3 3 7 8 500-250 0 0 2 0 3 1 0 0 0 0 750-500 0 0 0 0 0 0 0 0 0 0 1000-750 2 1 0 0 0 1 0 0 0 0 1 500- 1 000 1 1 — 1 1 — 0 0 0 0 | 2355 2356 2357 2358 2359 2361 50-0 0 0 10 45 0 0 100-50 0 15 40 50 5 0 250-100 10 17 42 15 12 5 500-250 4 7 4 2 2 7 750-500 0 0 3 0 3 2 1000-750 0 0 0 0 3 1 1 500- 1 000 1 0 0 1 0 1 March August 2027 2026 2025 2024 2023 2022 2020 2018 2017 2015 2014 2012 2010 50-0 0 0 5 3° 5 15 0 0 0 0 0 3° 10 100-50 0 10 25 90 10 5 0 20 20 10 35 35 35 250-100 0 3 12 8 10 7 10 12 7 3 3 3 5 500-250 0 1 2 1 1 5 3 1 0 1 0 1 0 750-500 0 0 0 0 0 0 2 2 0 0 0 0 0 1000-750 0 1 1 0 0 0 0 0 0 0 0 0 0 1500-1000 I 0 — — 1 0 0 0 0 0 — 0 0 2385 2386 2387 2389 2391 2393 50-0 0 0 0 0 0 0 100-50 0 0 5 0 0 0 250-100 17 40 17 9 12 19 500-250 5 0 1 7 6 1 750-500 0 0 0 3 0 0 1000-750 0 0 1 0 0 0 1 500- 1 000 0 3 0 0 0 0 April September 2311 2313 2316 2318 2320 2322 50-0 5 0 20 10 0 0 100-50 0 0 95 15 20 10 250-100 22 5 7 7 9 17 500-250 6 9 1 4 1 3 750-500 1 5 0 1 0 0 1000-750 0 0 0 0 0 0 1 500- 1 000 1 0 0 0 0 0 1805 1806 1807 1808 1809 1810 50-0 0 0 20 60 25 10 100-50 0 5 15 60 25 10 250-100 0 0 0 15 10 5 500-250 1 0 0 0 1 0 750-500 0 0 0 0 0 0 1000-750 0 0 0 1 0 0 1 500-1000 0 0 0 0 0 1 PLATE XIII Development of the gonads in Sagitta gazellae Small Northern race. Fig. i. Late stage II, showing the opaque tail segment characteristic of the stage, rudimentary seminal vesicles and short thin ovaries. Fig. 2. Stage II/III, a rare intermediate stage, the seminal vesicles are full sized but the tail segment is still opaque, the ovaries are still short and thin. Fig. 3. Late stage III, the tail segment is empty and the seminal vesicles full, the ovaries are somewhat thickened. Fig. 4. Late stage IV, the seminal vesicles have burst, and the ovaries are nearly fully developed. DISCOVERY REPORTS, VOL. XXVII PLATE XIII [Discovery Reports. Vol. XXVII, pp. 279-291, June, 1955.] CUMACEA OF THE BENGUELA CURRENT By N. S. JONES, Ph.D. Marine Biological Station, Port Erin CUMACEA OF THE BENGUELA CURRENT By N. S. Jones, Ph.D. Marine Biological Station, Port Erin (Text-figs. 1-7) INTRODUCTION THROUGH the kindness of Dr N. A. Mackintosh, Deputy Director of the National Institute of Oceanography, the Cumacea obtained by R.R.S. 'William Scoresby ' during 1950 in the Benguela Current off the south-west coast of Africa were sent to me for examination. The ship made two surveys of the region in this year, in March and in September-October, and most of the material is from the first survey. I am much indebted to Dr T. J. Hart of the National Institute of Oceanography for information and for sorting the Cumacea from the rest of the plankton. The Cumacea described in this paper were obtained in the vertical plankton net hauls, but more specimens may be available when the horizontal catches have been sorted. Very large numbers of individuals — up to about 7700 — were caught in some of the hauls. Cumacea were taken on the continental shelf at varying distances offshore, and in depths between 64 and 204 m. The rich hauls were all from depths of less than 150 m. and at distances of less than 30 miles from the coast. Full details are available in the Station List R.R.S. 'William Scoresby' (1950), 1953. Cumacea did not occur in nets fished at greater depths off the shelf. At all stations where oxygen determinations were made the 02 concentration at the bottom was low. The presence of Cumacea in large numbers in vertical nets was unexpected, and raises certain problems since they are normally bottom-living forms. When Cumacea have been caught in tow-nets at night, adult males have usually predominated, especially when attracted by artificial light, although in some recent records newly moulted adult or ovigerous females have outnumbered the males (Hale, 19536). The hauls containing the largest numbers of specimens described here were made during the hours of darkness, but all stages and both sexes were present and there was not an unduly large proportion of adult males. It is possible that the presence of the animals in the upper layers was not due to normal vertical migration. Although the species present in these hauls may be able to live normally in the plankton, they show no special adaptations to this mode of life and their nearest relatives are coastal bottom-living forms and it may be that they have been forced up from the bottom by the low 02 concentration, as Dr Hart has suggested to me. One would hardly expect them to inhabit the azoic mud which covers so much of the sea-floor in this region, but if they do not do so the alternative habitats are either the narrow bottom zone close inshore — which may be aerated by surf action — or else isolated patches surrounded by the so-called 'azoic zone' (Marchand, 1928; Currie, 1953; Copenhagen, 1953). On the other hand, the exceptionally large March haul of about 7700 individuals, which contained ovigerous females and adult males, as well as many juveniles, may re- present a nuptial swarm (Forsman, 1938). Five species were present, of which three are previously undescribed. They are Bodotria glabra sp.n., Upselaspis gen.n. caparti (Fage), Iphinoe fagei sp.n., /. africana Zimmer, and Diastylis rufescens sp.n. Of these, the two species of Iphinoe, particularly /. fagei, were much the most abundant and widespread. 282 DISCOVERY REPORTS The genus Iphinoe is apparently richest in species round the coasts of Africa. Of the seventeen so far described1 the following occur on the west or south African coasts: zimmeri Stebbing, robusta Hansen, brevipes Hansen, africana Zimmer, crassipes Hansen, hupferi Zimmer, brevidactyla Hale and truncata Hale (Hale, 1953 a), and fagei sp.n. described below. J. trispinosa (Goodsir), inermis G. O. Sars, tenella G. O. Sars, and serrata Norman are found in the Mediterranean and north-west Euro- pean waters, and trispinosa and tenella have been taken off the Atlantic coast of Morocco (Fage, 1928). Only /. sanguined Kemp from the Chilka Lake, macrobrachium Caiman (possibly identical with I. crassipes) from the Gulf of Manaar, calmani Fage (Fage, 1945) from Assam, and pellucida Hale (Hale, 1944) from Tasmania and New South Wales have not been found near the African continent. It is probable, however, that further work will show the genus to be more widely spread in southern Asiatic waters also. Three species, /. crassipes, brevipes and tenella, have recently been reported from the coast of Travancore (Kurien, 1951). Table 1. Stations arranged from North to South, but not according to date Species WS 970 WS 971 WS 980 WS 981 WS 982 WS 988 WS 989 WS 1063 WS 990 WS 1002 WS 1050 Bodotria glabra Iphinoe africana I. fagei Upselaspis caparti Diastylis rufescens 7 1 2 1 3 33 9 42 17 1 2 631 240 7450 I 88 24 26 3 There is evidently a succession of species from north to south among the Cumacea present in this collection (Table 1). Bodotria glabra occurred only at the most northerly stations at which Cumacea were caught, at 190 44' S, 120 40' E and 200 11' S, 120 30' E. Iphinoe africana also occurred at these stations and as far south as 250 11' S, 140 39' E. /. fagei was found from 220 44' S, 140 08' E to 250 36' S, 140 24' E. Upselaspis caparti was found only in the area of Walvis Bay, 220 44' S, 140 08' E., near to its original location reported by Fage. Diastylis rufescens was found only at the most southerly station, 280 40' S, 160 14' E, and it occurred in almost the same locality 8 months later. Forsman (1938) described the stages in the life-history of Diastylis rathkei (Kroyer). The larger collections from the Benguela Current offer an abundant material for a study of this kind. I hope at a later date to compare the two Iphinoe species from this point of view with some other members of the genus in British waters. SYSTEMATIC ACCOUNT Family Bodotriidae Genus Bodotria Goodsir, 1843 Bodotria glabra sp.n. (Figs. 1 and 2). Occurrence : St. WS 970, 100-0 m., 2$? (1 ovig.), 5 juv., night station. St. WS 971, 100-0 m., 1 ovig. $, 1 juv., night station. Ovigerous female. Length 4-5 mm. Carapace about twice as long as high, \\ as long as broad; antero-lateral angle (antennal tooth) prominent and acute ; dorsal carina faintly marked on carapace and pedigerous somites ; lateral carinae slightly developed on each side of carapace, but not reaching to the posterior border; eye prominent, with lenses. 1 Refer to Zimmer, 1942, except where other references are inserted. CUMACEA 283 2 mm. -1 O J I mm. -1 Fig. 1 . Bodotria glabra sp.n. Paratype ovigerous female from the side and from above, p1, p2, p3, first, second and third peraeopods; mxp3, third maxilliped; ur, uropod. 0-2 mm.-i Fig. 2. Bodotria glabra sp.n. Paratype ovigerous female, a1, first antenna; end, endopod of uropod. 284 DISCOVERY REPORTS First pedigerous somite concealed at the sides; the second somite long. First antenna with two-jointed flagellum, bearing two aesthetascs ; accessory flagellum rudimentary. Basis of third maxilliped twice as long as the remainder of the appendage, with its distal extension reaching to the end of the merus. First peraeopod with the last two joints reaching beyond the end of the pseudorostrum ; basis about i| as long as the remaining joints together; carpus nearly twice the length of the propodus, which is about the same length as the dactylus. Second peraeopod with six joints. Exopods present only on the first pair. Uropods slender, the peduncle about z\ as long as the rami, of which the exopod is slightly the longer; endopod one-jointed; the peduncle with hyaline serrations on the inner edge, interspersed with about twelve short setae; exopod with six plumose setae on the inner edge and ending in two spines ; endopod with six spines interspersed with hyaline serrations on the inner edge, and two end spines. Colour in alcohol, yellowish-white. This species seems to be near to B. siamensis Caiman, which it resembles in the absence of lateral ridges on the pedigerous somites, but the peduncle of the uropods is much longer in proportion to the rami and the second joint of the first peraeopod is longer than the rest of the appendage. Genus Upselaspis gen.n. Carapace smooth without carinae. First pedigerous somite not visible from above, the second not much longer than the following. First antenna with accessory flagellum. Only the first peraeopod provided with an exopod (male and female). The second peraeopod with the basis and ischium fused. The inner ramus of the uropod with two joints. Upselaspis caparti (Fage). Fage, 19516, p. 5. Occurrence : St. WS 981, 50-0 m., i9c?c?, 20$$ (2 ovig.), 3 juv., day station. St. WS 982, 100-0 m., IS-, day station. This species was placed by Fage in the genus Cyclaspoides Bonnier, with the proviso that it might be necessary to propose a new genus for it when C. sarsi Bonnier became better known. I have examined three out of four specimens of C. sarsi from the British Museum (Natural History), for which I am indebted to Dr I. Gordon, and it seems to me that the differences between that species and Fage's species U. caparti are too great for them to be placed in the same genus. Fage lists a number of these differences (195 ib, p. 9), to which may be added the following: C. sarsi is much more slender when viewed dorsally; the gut is coiled anteriorly, but is straight in Upselaspis caparti; the telsonic somite in C. sarsi is produced nearly half-way along the peduncle of the uropods, but is not produced in U. caparti; the endopod of the uropod in C. sarsi is one-jointed, ending in a spine much narrower than the breadth of the joint at its end, while I consider the endopod of U. caparti to be two-jointed, as Fage thought possible, because the second joint is almost as broad at its proximal end as the distal end of the first joint ; its inner margin is finely serrated, and it ends in a spine, while in the male it bears two or three plumose setae on the inner edge proximally. Generically the important difference between the two species is that in Cyclaspoides sarsi the first three pedigerous somites are fused with the carapace in both sexes, while in Upselaspis caparti only the first pedigerous somite is fused in the female and the first and second in the male. Upselaspis is distinguished from Cyclaspis by the presence of six joints in the second peraeopod, and from Bodotria by the shorter second pedigerous somite. The species was obtained in almost the same locality as the original find reported by Fage. CUMACEA 285 Genus Iphinoe Bate, 1865 Iphinoe fagei sp.n. (Figs. 3 and 4). Occurrence : St. WS 980, 100-50 m., z$$, 5$9 (2 ovig.), 9 juv.; 50-0 m., 2 i -_ *• , CD SC SL SG FC I \ PLATE XVIII Enlarged photograph, x 4, of the core of the wax plug from Balaenoptera physalus. DISCOVERY REPORTS VOL. XXVII PLATE XVIII [Discovery Reports. Vol. XXVII, pp. 303-336, Plate XIX, August, 1955.] STUDIES OF ALEPISAUROID FISHES By N. B. MARSHALL <* H. 0 / vvoc°- CONTENTS Introduction Acknowledgements .... Part I. Classification of the Iniomi The Family Scopelarchidae . Scopelarchus cavei sp.n. Neoscopelarchoides dubius sp.n. Neoscopelarchoides elongatus (Norman) Neoscopelarchoides sp. . Relationships and synonymies of the g Key to the Scopelarchidae . The family Paralepididae Macroparalepis molestus sp.n. The Family Anotopteridae Anotopterus pharao Zugmayer Part II. Comparison of the alepisauroid and The swim-bladder . Ossification of the skeleton The setting of the pectoral fins The dorsal fin .... Intermuscular bones Distensibility of the stomach and body-wall Interrelationships of the AJepisauroidea Summary Bibliography Plate XIX myctophoid fishes page 305 3°5 306 3°7 3°7 3°9 310 311 312 3i5 3IS 3*5 3i7 3i7 322 323 324 325 328 329 33° 33i 334 334 fol hiving page 336 STUDIES OF ALEPISAUROID FISHES By N. B. Marshall (Plate XIX and Text-figs 1-9) INTRODUCTION THE Alepisauroidea are voracious, bathypelagic fishes which together with the Myctophoidea form the order Iniomi. Fishes of the order Iniomi are soft-rayed teleosts, which have evolved beyond the level of the Isospon- dyli in that the premaxillaries exclude the maxillaries from the gape. The pelvic fins are usually abdominal in position, but when they are set close to the pectorals, as in the Aulopidae, there is no association between the pelvic bones and the pectoral girdle. The swim-bladder when present is closed (physoclistous) and the blood supply to the retia mirabilia enters at the anterior end. There are usually 19 principal rays in the caudal fin and 6-1 1 rays in each pelvic fin. The dorsal fin and anal fin are never opposed and an adipose dorsal fin is usually present. There is no mesocoracoid, nor is a Weberian mechanism present. The Alepisauroidea can be defined as bathypelagic Iniomi with no swim-bladder or luminescent organs, with a single row of numerous (more than 20) small teeth on each premaxillary, one or two rows of teeth on each palatine, one to three rows of teeth on each dentary and with gill-rakers in the form of spines or teeth. The anal fin has from 12-50 rays and is usually set close to the caudal. Lastly, there are 3-4 + 4-5 branchiostegal rays.1 Compared with most other oceanic, mid-water teleosts, the Alepisauroidea are large, even very large, fishes.2 Owing to their size and swiftness, our knowledge of this suborder of the Iniomi — at least from catches of mid-water nets — is very incomplete. Almost without exception, the larger known individuals of the larger species have come either from the stomachs of whales or of oceanic fishes, such as tunny, or have been taken on long lines. Alepisaarus itself, which is caught on the tunny hooks of Madeiran fishermen, has contained some relatively large paralepidids. This report is based mainly on specimens from the Discovery Collections. The material contains some new species and is sufficiently extensive to allow of a revision of the Scopelarchidae and a review of the relationships of the alepisauroid fishes. ACKNOWLEDGEMENTS My thanks are due to Dr N. A. Mackintosh for putting the material at my disposal and to Dr W. A. Gosline, who has read the manuscript and made some useful critical comments. 1 The plus sign separates the rays on the epihyal and ceratohyal bones. 2 Alepisaurus reaches a standard length of at least 1500 mm.; the largest known Anotopterus is 855 mm. in length, while three paralepidids, Magnisudis barysoma, Notolepis coatsi and Sudis hyalina, fall within the standard length range of 400-500 mm. Certain other paralepidids, Omosudis lozuei, Evermannella balbo and Neoscopelarchoides elongatus, grow to within the range 150-300 mm. 3o6 DISCOVERY REPORTS PART I. CLASSIFICATION OF THE INIOMI In order to give systematic perspective to these studies, an outline classification of the order Iniomi is given. Those workers mainly responsible for this arrangement will become apparent during the course of this paper. I. Suborder Alepisauroidea Family Scopelarchidae Genera Scopelarchus Alcock, Neoscopelarchoides Chapman Family Evermannellidae Genus Evermannella Fowler Family Paralepididae Subfamily SUDINAE Genus Sudis Rafinesque Schmaltz Subfamily PARALEPIDINAE Genera Magnisudis Harry, Paralepis Cuvier, Notolepis Dollo, Lestidium Gilbert, Macro- paralepis Ege Family Anotopteridae Genus Anotopterus Zugmayer Family ALEPISAURIDAE Genus Alepisaurus Lowe Family Omosudidae Genus Omosudis Giinther II. Suborder Myctophoidea Family Aulopidae Genera Aulopus Cuvier, Hime Starks and Latropiscus Whitley Family Chlorophthalmidae Genera Chlorophthalmus Bonaparte, Bathysauropsis Regan Family Notosudidae1 Genera Notosudis Waite, Luciosudis Fraser-Brunner Family BATHYPTEROIDAE Genera Bathypterois Giinther, Benthosaurus Goode and Bean Family IPNOPIDAE Genera Ipnops Giinther, Bathymicrops Koefoed Family NEOSCOPELIDAE Genera Neoscopelus Johnson, Solivomer Miller, Scopelengys Alcock Family Myctophidae Genera Electrona Goode and Bean, Hygophum (Taning) Bolin, Benthosema Goode and Bean, Diogenichthys Bolin, Myctophum Rafinesque, Ctenoscopelus Fraser-Brunner, Loweina Fowler, Gonichthys Gistel, Ctenobranchus Fowler, Tarltonbeania Eigenmann and Eigenmann, Diaphus Eigenmann and Eigenmann, Notolychnus Fraser-Brunner, Lampadena Goode and Bean, Lampanyctodes Fraser-Brunner, Lampanyctus Bonaparte, Ceratoscopelus Giinther, Gymnoscopelus Giinther, Lampichthys Fraser-Brunner, Noto- scopelus Giinther, Hintonia Fraser-Brunner, Scopelopsis Brauer 1 It would appear that Scopelosaurus Bleeker (Act. Soc. Sc. Indo-Neerl, vm, 1860, Elfde Bijdr. Amboina, p. 13) should be included in this family. SCOPELARCHIDAE 3°7 Family Harpadontidae Genera Harpadon Le Sueur, Bathysaurus Giinther Family SYNODONTIDAE Genera Synodus (Gronovius) Scopoli, Trachinocephalus Gill, Saurida Cuvier and Valen- ciennes, Xystodus Ogilby. The Family SCOPELARCHIDAE The first known fish of this family was described by Alcock (1896, 1899) under the name Scope- larchus guentheri.1 It was taken by H.M. Indian Marine Survey Steamer ' Investigator ' while trawling between depths of 180 and 217 fathoms off Colombo, Ceylon. Observing that the fish '...is a remarkable generalized form of Scopeloid, showing affinities with Saurus, Chlorophthahmis, Scopelus, Odontostomus and Paralepis\ Alcock (1896) placed it in the family Scopelidae, group Scopelarchina. ' Valdivia ' took twelve specimens of scopelarchids at various stations in the Atlantic, Indian and Southern Oceans. These were described by Brauer (1902, 1906) as Dissomma atiale and placed in the family Scopelidae. Roule (191 6, 1919), having examined seven scopelarchids taken off the Azores during the cruises of ' Princesse Alice ', described them as a new species, Odontostomus perar- matus, of the family Odontostomidae. However, Odontostomus Cocco (amended to Evermannella by Fowler (1901), on account of Odontostomus Cocco, 1838, being a homonym of Odontostomus Beck, 1837, a molluscan genus) is quite distinct from Scopelarchus (see below). Regan (191 1), in setting out the classification of the Iniomi, defined the family Scopelarchidae and placed it in the suborder Alepisauroidea. Then Parr (1928) pointed out that Regan had included both Evermannella and Scopelarchus (= Dissomma) in the Scopelarchidae, and that the diagnosis of the family was based on a study of Evermannella hyalina. Parr showed clearly that Scopelarchus and Ever- mannella are so distinct as to warrant separation into different families : he was thus able to give the first critical diagnosis of the Scopelarchidae. In a later paper (Parr, 1929), this diagnosis was extended to include findings from osteological studies. Furthermore, in the 1929 paper a new genus and species, Scopelarchoides nicholsi, was described, the type and a paratype being taken by ' Pawnee ' off the west coast of Mexico. Additions to this description are to be found in a later paper (Parr, 193 1 a) in which three more specimens are recorded, one from the Gulf of California, one off the west coast of Central America and one from a locality in the Pacific Ocean. A further new species, Scopelarchus elongatus, was described by Norman (1937) from an individual taken north-east of Kerguelen Island by the B.A.N. Z. Antarctic Research Expedition. This was fol- lowed by Chapman's (1939) proposal of a new genus and species, Neoscopelarchoides dentatus, the holotype coming from the central part of the Gulf of Alaska. Lastly Mead and Bohlke (1953) have described Scopelarchus linguidens from three juvenile individuals taken off northern Japan. The Discovery material contains two new species and a large larval scopelarchid, descriptions of which, together with a fuller account of Norman's species elongatus, now follow. But to anticipate a later section, it must be added, that the study of these individuals has led me to conclude that the scopelarchids can be divided into two well-marked genera, Scopelarchus Alcock and Neoscopelarchoides Chapman. Synonymies will be discussed on p. 312-14. Scopelarchus cavei sp.n. (Fig. 1 and PL XIX, figs, 5, 6.) Holotype. A specimen 70 mm. in standard length taken by R.R.S. 'Discovery II' in a young fish trawl, hauled obliquely between 500 m. and the surface, at station 1585: oo° 06' S, 490 45-4' E., Indian Ocean : about 400 miles north-west of the Seychelles. 1 In the 1899 publication Alcock gave the specific name as giintheri. 308 DISCOVERY REPORTS Body proportions in hundredths of the standard length (70 mm.). Greatest depth of body 24-3 ; length of head 257; horizontal diameter of eye 8-6; length of snout 7-1 ; width of bony interorbital 2-6; length of lower jaw 21-4; maxilla 17-8; length from tip of snout to origin of dorsal fin 40 ; snout to adipose dorsal fin 757; snout to pectoral fin 28-6; snout to pelvic fin 47-1; snout to anal fin 61-4; length of pectoral fin 257; length of pelvic fin 12-1 ; length of caudal peduncle 14-3 ; depth of caudal peduncle 9-3. Body compressed; greatest breadth (at pectoral girdle) 14-3. Fin-rays. D. 9, A. 21, Pect. 21, Pv. 9. Scaling. Large cycloid scales on opercula; probably also present on cheeks. Scales over the trunk and tail well ossified, with a shiny, armoured appearance : cycloid. Scale rows between lateral line and origin of dorsal fin 4; between lateral line and origin of anal fin 7. Number of scales along lateral line (to caudal flexure) 43. Lateral line scales only slightly enlarged, about 1-2 times as deep as those immediately adjacent. Vertebrae. 41. 1 Dentition. About fifty-five small, pointed retrorse teeth on premaxilla. Dentary with about thirty outer, smallish, pointed teeth and 9-10 inner, larger, and depressible teeth of which the second and third are the largest. The first six of these teeth are barbed. Palatine teeth arranged in two closely adjacent rows, each consisting of 5-6 teeth. Lingual teeth ten, the first longest, the rest gradually decreasing in size. Text-fig. 1. Scopelarchus cavei. Holotype (x 1-3). Colour (in spirit). General body colour yellowish brown, the dorsal surface darker, with a fairly dense peppering of small melanophores covering the upper half of the flank-area above the lateral line. Immediately behind each eye is a black lunate area. Dorsal, anal, pelvic and caudal fins trans- parent or whitish. The pectoral fin is black, except for the two uppermost and five of the lowermost fin rays, which are whitish. Underlying the scales over the lower half of the abdomen is a layer of silvery pigment, which extends from the isthmus to the anus. (Is it possible that luminescent tissue may be associated with such a light-reflecting layer?) Above this silvery area, the black peritoneum of the body cavity shows through the body wall. This species is most closely related to Scopelarchus guentheri Alcock, but has fewer anal rays (21, cf. 26), fewer lateral-line scales (43, cf. c. 50) and these scales are not enlarged as in S. guentheri. Perhaps there is also a difference in the interorbital width, which in S. cavei is relatively broad, whereas in S. guentheri Alcock (1899) described the eyes as being separated by a 'mere linear space '. I have much pleasure in naming this species after A. J. E. Cave, Professor of Anatomy at St Bartholomew's Hospital Medical College. 1 In all counts of vertebral numbers the last, upturned, caudal element has been included. SCOPELARCHIDAE 3°9 Neoscopelarchoides dubius sp.n. (Text-fig. 2 and PI. XIX, figs. 3, 4.) Holotype. A female, 138 mm. in standard length, taken in a 450 cm. closing net hauled obliquely between 1800 and 1150 m., at station 1759: 310 54-5' S, 510 27-9' E., Indian Ocean: about 500 miles south-east of Madagascar. Body proportions in hundredths of the standard length (138 mm.). Greatest depth of body 17-5; length of head 22-8 ; horizontal diameter of eye 5-8 ; length of snout 8-7 ; width of bony interorbital 0-7 ; length of lower jaw 20-3 ; length of maxilla 15-2; length from tip of snout to origin of dorsal fin 38-4; snout to adipose dorsal fin 81-2; snout to pectoral fin 25-5; snout to pelvic fin 40-9; snout to anal fin 61-5; length of pectoral fin 21-0; length of pelvic fin 29-0; length of caudal peduncle 11-2; depth of caudal peduncle 6-2. Anus situated a little in front of the origin of the anal fin. Trunk not much deeper than wide at mid-point of body. Fin-rays. D. 9, A. 21, Pect. 27, Pv. 9. oxoxomoococcoxcoxo^^ * V] ^^? — ^^T" ^LL^n4^A{ff]r^1^ Text-fig. 2. Neoscopelarchoides dubius. Holotype ( x 1) : a, palatine teeth ( x 2-5) ; b, mandibular teeth ( x 2-5) ; c, lingual teeth ( x 2-5) ; d, lateral line scale ( x 6). Scaling. Rather small, thin cycloid scales on cheeks and gill covers and over the body, extending a short way in a B-shaped area over the bases of the principal caudal rays. Scales of lateral line much enlarged. Scales along lateral line 59. Transverse rows of scales between origin of dorsal fin and lateral line 8; between origin of anal fin and lateral line 10-11. Vertebrae. 57. Dentition. Premaxilla with about 80 small, pointed, retrorse teeth. Dentary with about thirty outer, smaller pointed teeth and nine inner, long, barbed, depressible teeth, the second and third being the largest. Left palatine with three large, barbed, depressible teeth, these being set somewhat inward from a row of eight small, pointed teeth. Lingual teeth fourteen, moderately compressed, the first the largest, the remainder gradually decreasing in size. Gill-teeth. The first gill arch bears a series of closely set, small, pointed gill-teeth. Colour (in spirit). Dark brown above the lateral line, medium brown below, except where the black peritoneum of the body cavity shows through the abdominal walls. The fins are more or less hyaline, except for the caudal, which has a fairly dense covering of small melanophores. There is a kidney-shaped iridescent area on the outer walls of the optic cup of the eye, close to the lens. This has been regarded as a luminescent organ, but it must be the elliptical mass of fibrous tissue 3IO DISCOVERY REPORTS which Brauer (1908) called a ' linsen polster ' and which is shown on his pi. 40, figs. 2 and 5, of Dissomma a nale. Table 1 shows the number of fin-rays in Neoscopelarchoides dubias compared with the other known species of this genus. N. dubius is also readily distinguishable from the other species by the highly developed pelvic fins. The proportions of the latter expressed in hundredths of the standard length are : N. dubias, 29-0 ; N. elongatus, 13-3-14-4; N. linguidens, 5-3-5-6; N. dentatus, c. 17-0 (measured from Chapman's (1939) figure). Note. I have called this species dubius, because it is possible that future work may reveal it to be the adult of the larval form, Benthalbella infans Zugmayer (see p. 312). Table 1 Number of fin-rays Dorsal Anal Pectoral Pelvic N. dubius sp.n. N. elongatus (Norman) N. dentatus Chapman N. linguidens (Mead and Bohlke) 9 8-9 6-7 9-10 21 26-27 17-21 28 27 21-22 22-25 25-27 9 9 9 9 Neoscopelarchoides elongatus (Norman) (Text-fig. 3b, V and PI. XIX, figs. 1, 2.) Scopelarchus elongatus Norman 1937. Rep. B.A.N.Z. Antarct. Res. Exped. 1929-193 1 (B), 1 (2), p. 86. As the original description by Norman (1937) is rather brief, an expanded one is given here. It is based on the type specimen (standard length 198 mm.) which was taken east of Kerguelen Island (45° 53' S, 84° 33' E) and on two individuals taken by R.R.S. ' Discovery II '. The first (s.l. 104 mm.) is from station 391 (550 48^' S, 520 35' W), about 300 miles south-east of the Falkland Islands, in a 450 cm. net hauled between 1200 m. and 1300 m.; date, 18. iv. 30 (Text-fig. 3). The second (s.l. 1 20 mm.) is from station 725 (530 23-6' S, 740 57-8' W), about 50 miles south of the western end of the Magellan Strait, in a 100 cm. net hauled obliquely between 250 m. and 196 m.; date 17. xi. 31 (PI. XIX, figs. 1, 2). In the following description the data for the type1 are given first, followed (in parenthesis) by the data for the Discovery specimens, in the above-mentioned order. Body proportions in hundredths of the standard length 198 mm. (104 and 120 mm.). Greatest depth of body 14-1 (n-i, io-8); length of head 19-8 (22-1, 22-5); horizontal diameter of eye 5-6 (5-3, 5-8); length of snout 6-i (7-2, 7-9); width of bony interorbital i-6 (i-6, 1-5); length of lower jaw (20-2, 20-0); maxilla (17-3, 16-6); length from tip of snout to origin of dorsal fin 39-1 (42-8, 41-7); snout to adipose dorsal fin 79-9 (74-0, 75-0); snout to pectoral fin 21-7 (24-0, 23-3); snout to pelvic fin 37-9 (40-4, 38-3); snout to anal fin 64-6 (65-4, 64-2); length of pectoral fin (14-4, 13-3); length of pelvic fin (14-4, 13-3); depth of caudal peduncle 7-1 (6-2, 6-o); length of caudal peduncle 12-6 (12-5, 12-8). Fin-rays. D. 9 (9, 8), A. 26 (27, 27) Pect. 22 (21, 21), Pv. 9 (9, 9). Scaling. (Based on the Discovery specimen from station 725.) Small, thin cycloid scales on cheeks and gill covers and over the rest of the body. Scales of lateral line much enlarged. Scales along lateral line 64. Transverse rows of scales between origin of dorsal fin and lateral line 8 ; between origin of anal fin and lateral line 10- 11 (the type specimen has sixty-two scale pockets along the lateral line). Vertebrae. In the type specimen: 61. 1 Owing to damage, certain measurements for the type specimen had to be omitted. SCOPELARCHIDAE 311 Dentition (based on Discovery specimens). Premaxillary teeth c. 70, pointed, retrorse, small. Dentary with c. thirty outer, pointed, retrorse teeth; eleven long, depressible, barbed, inner teeth, the 3rd~5th being the largest. Palatine teeth in two closely adjacent rows, 8-10 smaller, pointed teeth in outer row, 5-6 longer, barbed teeth in inner row. Lingual teeth eleven, barbed, decreasing in size from before backward (see Text-fig. 36'). (ukL-M. a' V Text-fig. 3. a, Neoscopelarchoides sp. late larval stage ( = Benthalbella infans) (X2): a', lingual teeth (xo/3); b, Neoscopelarchoides elongates (Norman). Specimen from station 391 (x 1-3); b', lingual teeth (xg-j). Neoscopelarchoides sp. (Text-fig. 3 a, a'.) This specimen of a late larval stage was taken by R.R.S. 'Discovery II' off the Azores. Station 3051 (lat. 390 29' N, long. 90 50' W), 26. viii. 52. Net TYF (young fish trawl), hauled obliquely between 700 m. and the surface. Body proportions in hundredths of the standard length (60-5 mm.). Greatest depth of body 13-2; length of head 15-4; diameter of forward lens-containing part of eye 4-1 ; length of snout 6-6; width of bony interorbital i-o; length of lower jaw 15-3; length from tip of snout to origin of dorsal fin 38-8; snout to origin of adipose dorsal fin 57-8; snout to anterior ray of pectoral fin 18-2; snout to anterior ray of pelvic fin 33-5; snout to origin of anal fin 64-4; depth of caudal peduncle 4-5; length of caudal peduncle 12-4. Fin-rays. D. 9, A. 22, Pect. 24, Pv. 9. As shown in Text-fig. 3 a the eyes are tubular and are directed forward and upward. Dentition. About thirty small, pointed, retrorse teeth on premaxilla. Mandible with about twenty outer, small, upright pointed teeth ; seven inner, long and slender teeth, all but the first with a tiny wooos HOLE. v MASS. / 3i2 DISCOVERY REPORTS barb and all but the first, which is more or less upright, antrorse. Palatine with about twelve smallish, pointed teeth. The lingual teeth (see text-fig. 3 a') more than any other feature, show that this must be a young scopelarchid. When caught the fish was almost transparent, the only colour being the black pigment of the eye tube. The body-wall is quite transparent, the internal organs, gut, liver and mesenteries, showing clearly. The muscles associated with the base of the pelvic fins are moderately dense, but over the rest of the body wall there is but a thin muscle-layer, which becomes less and less easy to distinguish as one works upwards from the pelvics. Along the mid-ventral line from the head to the pelvic fins, there is a sharp hyaline division between the musculature of each side of the body wall. Relationships. Following on this section, reasons will be given for recognizing two genera of scopel- archids, Scopelarchns Alcock and Neoscopelarchoid.es Chapman. One of the features distinguishing them is the relative position of the dorsal and pelvic fins. In the larval specimen from station 3051, the origin of the dorsal fin lies behind that of the pelvic fins, a characteristic of the genus Neo- scopelarchoides; hence the larva may be presumed to belong to this genus. It also has the full comple- ment of principal caudal and pelvic rays (and presumably of dorsal, anal and pectoral rays), and at this stage there is unlikely to be much shift in the relative positions of the dorsal and pelvic fins. But it cannot be fitted with certainty into any of the four known species. In numbers of fin-rays it is closest to N. dubius and it also agrees with this species in the number of body segments before the origin of the anal fin (the remaining myotomes cannot be counted with certainty). The larva has thirty-five pre-anal fin myotomes, and N. dubius has thirty-five pre-anal fin lateral line scales and vertebrae. N. elongatus has 33-35 pre-anal fin segments but the anal rays (26-27) are more numerous than those of Benthalbella. I have already drawn attention to the fact that N. dubius sp.n. may prove to be the adult of a larval form, Benthalbella infans described by Zugmayer in 191 1. Zugmayer's young fish and the Discovery larva are very similar, the only difference being in the number of anal rays (17 compared with 22). A number of other specimens, previously recorded, also nearly resemble them : (1) The fish figured in Murray and Hjort (1912) on p. 746 and labelled 'New fish resembling Dysomma ' (sic). The drawing shows nine dorsal and twenty or twenty-one anal rays. (2) Two Benthalbella larvae from the Bay of Cadiz (Schmidt, 1918). Fin-ray formula: D. 9-10, A. 21-22, Pect. 26, Pv. 8-9. (3) Two Benthalbella larvae from Madeira and one from the Azores (Roule and Angel, 1930) which were regarded as young stages of Omosudis lowei. Their Plate III, figs. 75 and 76 show a fin-ray complement of D. 8-9, A. 20. (4) Three post-larval Benthalbella from the Strait of Gibraltar (Nybelin, 1948). Fin-ray formula: D. 9-10, A. 20-22, Pect. 27, Pv. 9. The distribution of these larvae may be seen in the chart, fig. 4. It would thus seem probable that the Discovery larva and all those listed above belong to one species of Neoscopelarchoides. If this is so, the species has a fin-ray complement of D. 9-10, A. 17-22, Pect. 22-27, Pv. 8-9. N. dubius comes very close to this; and it may be concluded that either these remarkably large larvae are the young of N. dubius or of an unknown, closely-related species. RELATIONSHIPS AND SYNONYMIES OF THE GENERA Considering in chronological order the genera that have been included in the Scopelarchidae, Dissomma Brauer (1902) is clearly synonymous with Scopelarchus Alcock (1896), for S. guentheri Alcock and Dissomma anale Brauer are either very closely related or belong to the same species. 313 o B o -*_> o j~ •M M 5^. 11 -^ ro V ON d o c r8 — £ a o n _■ « T3 n d to -t- ►H ON .. j_ N £ f/j u T3 d d j n. o c_ <-> -n ~^ d » Hi re r — A 3 re re jr. — a -= 3 JJ ■z- u ■a CD o J3 .- -*-1 d <-i- o O d 0 - Q 3 J3 60 X o 314 DISCOVERY REPORTS Regan (191 1) thought that Neosudis Castelnau (1873) might be a scopelarchid ; but Neosudis has a dorsal fin of sixteen rays placed two thirds of the way back along the body, whereas the dorsal fin in the Scopelarchidae has 6-10 rays and is placed about half-way down the body. As Parr (1928) has pointed out, the genus is evidently not related either to the Scopelarchidae or Evermannellidae. Benthalbella Zugmayer (191 1) has already been considered: its status will be discussed below. Turning to Promacheon Weber (1913), which was provisionally included by Parr (1928) in the Scopelarchidae, this genus cannot be fitted into this family. The peculiar triangular tooth on each premaxilla, the number, form and arrangement of the other teeth, the absence of teeth on the palatines, the fin-ray formula (D. 14, A. 17-18, Pect. 15, Pv. 10) and the small number of lateral line scales (c. 32), show that Promacheon has a character complex quite unlike that of the scopelarchids. Scopelarchoides Parr (1929) is defined has having two areas of abdominal muscles, one surrounding the base of the pelvic fin, the other being an upper lateral musculature with a limited ventral extension. Between these two areas the body-wall is transparent, for the peritoneum is only separated by connec- tive tissue from the outer skin. The division of these muscles is very like that of Benthalbella larvae. Chapman ( 1 939) noted that small specimens of Neoscopelarchoides dentatus have a translucent abdominal cavity and that the abdominal musculature is fully developed only in the larger individuals. This sug- gests that the fishes described as Scopelarchoides nicholsi were not completely developed. If this is admitted, the differences between Scopelarchus and Scopelarchoides are clearly no more than dif- ferences between different phases of development. The other characters which Parr (1929) has used to define these two genera are nearly all osteological, involving the lesser development of various bones (prootics, opisthotics, preorbitals and subopercula) in Scopelarchoides. Evidence from these characters also indicates that Scopelarchoides Parr (1929) cannot be retained as a separate genus, and that it is synonymous with Scopelarchus Alcock (1896). It is clear from this study that the scopelarchids can be divided into two genera only: (1) Scopelarchus Alcock (1896) comprising the short-bodied species, in which the origin of the dorsal fin is in front of the insertion of the pelvic fins (S. guentheri Alcock, S. nicholsi (Parr), and S. cavei sp.n.). (2) Neo- scopelarchoides, comprising the long-bodied species having the origin of the dorsal fin behind the in- sertion of the pelvics (Scopelarchus elongatus Norman, Neoscopelarchoides dentatus Chapman, N. dubius sp.n. Benthalbella infans Zugmayer, and Scopelarchus linguidens Mead and Bohlke). Two names are available for the second genus, Benthalbella Zugmayer and Neoscopelarchoides Chapman. Zugmayer (191 1) described Benthalbella infans in an Appendix to his report, since he was unable to determine its systematic position, yet he did point out that there were certain resemblances between his fish and Brauer's Dissomma anale. Remarking that, ' Ce poisson est tres jeune ou plutot larvaire ', he none the less gave it a name because it was then unknown. Article 30 of the International Rules of Zoological Nomenclature, Section He, /?, states that: ' Species which were species inquirendae from the standpoint of the author of the generic name at the time of its publication ','... are excluded from consideration in determining the types of genera '. In view of Zugmayer's uncertainty as to its systematic position and his admission of its larval character, Benthalbella infans cannot be considered as a genotype. Neoscopelarchoides Chapman must thus be the name for the second genus, genotype N. dentatus. The characters of the two genera and the main characters of the species are given in the following key. The distribution of the better known species is also given. See also the chart on p. 313. There is a full description of the family characters in Parr's papers (1928 and 1929). SCOPELARCHIDAE— PARALEPIDIDAE 315 KEY TO THE SCOPELARCHIDAE I. Origin of dorsal fin in front of insertions of pelvic fins. Scopelarchus Alcock Other generic characters: Short bodied, greatest height of body from 17 to 25 per cent of standard length ; number of vertebrae from 41 to 48 (41 in S. cavei; 46 in S. guentheri and 48 in S. nicliolsi) ; number of scales in lateral line 43-50. * Pectoral rays 19-23. A. Pelvic fins longer than pectorals and extending beyond origin of anal fin. Fin-rays: D. 6-7, A. 21-23, Pect. 20-22. Pv. 9 S. nicholsi (Parr) Distribution. Gulf of California, off west coasts of Mexico and Central America. B. Pelvic fins shorter than pectorals and not reaching origin of anal fin b 1. Number of scales along lateral line 43. Fin-rays: D. 9, A. 21, Pect. 21, Pv. 9 S. cavei sp.n. b 2. Number of scales along lateral line 48-50. Fin-rays: D. 7-9, A. 23-26, Pect. 19-21, Pv. 9 S. guentheri Alcock Synonyms:. Dissomma anale Brauer (1902 and 1908), Odontostomus perarmatus Roule, (1916) Distribution. Atlantic Ocean, east of Bouvet Island, Indian Ocean, East Indian Archipelago. II. Origin of dorsal fin behind insertions of pelvic fins. Neoscopelarchoides Chapman Other generic characters: long-bodied, greatest height of body from 10 to 17-5 per cent of standard length (number of vertebrae 57 in N. dubius, 61 in N. elongatus). Number of scales in lateral line 56-64 (scales not formed in the holotype and two paratypes of N. linguidens). Pectoral rays 21-27. A. Pectoral fins less than half-length of pelvic fins ; caudal peduncle from 20 to 25 per cent of standard length. Fin-rays: D. 6-7, A. 17-21, Pect. 22-25, ^v- 9 N- dentatus Chapman Distribution : Gulf of Alaska ; off Central California. B. Pectoral fins equal to, or considerably more than, half the length of pelvic fins (these fins about equal in length in N. elongatus, while in linguidens and dubius the pectorals are about three-quarters the length of the pelvics). Caudal peduncle from 11 to 15 per cent of standard length. b 1. Length of pelvic fin more than one-quarter of standard length. Fin-rays: D. 9, A. 21, Pect. 27, Pv. 9 N. dubius sp.n. b 2. Length of pelvic fins considerably less than one-quarter of standard length. (i) Pectoral fins with 21-22 rays. Fin-rays: D. 8-9, A. 26-27, Pect. 21-22, Pv. 9 Distribution. Off Kerguelen Island; Falklands-Magellan Strait region. N. elongatus (Norman) (ii) Pectoral fins with 25-27 rays. Fin-rays: D. 9-10, A. 28, Pect. 25-27, Pv. 9-10 N. linguidens (Mead and Bohlke) The Family PARALEPIDIDAE Macroparalepis molestus sp.n. (Text-fig. 5.) Holotype. An individual 146 mm. in standard length, taken in a 2 m. net during an experimental haul near station 2209 (530 077' S, 1680 56-4' E), near Campbell Island, south of New Zealand. 27. i. 38. 2000 m. wire. Proportions in hundredths of the standard length (146 mm.). Greatest depth of body 4-1 ; length of head 16-1 ; length of snout 8-2; horizontal diameter of eye 2-7 ; width of bony interorbital i-8; length of lower jaw 9-5; length from tip of snout to origin of dorsal fin 63-0; snout to pelvic fin 497; snout to anal fin 787 ; length of pectoral fins 7-5 ; length of pelvic fin 4-8 ; depth of caudal peduncle 2-0 ; length of caudal peduncle 5-1. Fin-rays. D. 12 (last ray bifid), A. 30, Pect. 12, Pv. 9. 1 Counted from upper angle of operculum to caudal flexure. 3i6 DISCOVERY REPORTS Scales. Absent, but there are ossifications along the lateral line, which may be modified scales (Harry, 1953 a) see Text-fig. 5 c and d and below. Lateral line. Roofing the gutter-like lateral line along the trunk is a fragile skin in which is set a series of dumb-bell shaped ossifications, one per segment. Above and below each ossification are two pores, the anterior larger than the posterior. Immediately at the end of the ossification comes a small median pore (see Text-fig. 5 c). About ninety ossifications occur along the lateral line, the series on the right flank ending opposite the origin of the adipose dorsal fin. The last few, posterior ossifications are not fully developed (presumably the complete series is developed at a later stage). Number of vertebrae. 1 o 1 . Ventral carina. Distinctly developed between the pectoral and pelvic fins. Dentition. About twenty teeth on the premaxillae, the first being the longest, then come about fourteen retrorse, thorn-like teeth followed by eight or so, small antrorse teeth. Dentary with six long pointed teeth, each of which is flanked by smaller pointed teeth, usually one on each side. Palatines with about 10-12 teeth, these being six hooked outer teeth, and four to six larger, depressible, pointed inner teeth (see Text-fig. 56). fflf Text-fig. 5. Macroparalepis molestus sp.n. Holotype ( x 1) : a, maxillary and mandibular teeth ( x 3) ; b, palatine teeth ( x 3) : c, part of lateral line to show pores and bony segments ( x 25); d, transverse section through lateral line. Gill-teeth. Rather widely spaced, in a single row, on the upper part of the lower limb of the first gill arch. Colour (in spirit). Light brown with a peppering of small melanophores along the dorsal surface. Three dorsal ' saddles ' of more concentrated pigmentation ; the first along the base of the dorsal fin ; the second just above the origin of the anal fin, and the third just behind the adipose dorsal fin (see Text- % 5). The study of the above fish has thrown some light on the status of Macroparalepis Ege. As a result of his investigations, Harry (1951, 1953a) quite reasonably decided to divide Macroparalepis Ege into two genera, these corresponding to divisions I and II of Ege's (1933) synopsis of the species. Macro- paralepis was restricted to division I of the key, while division II was placed under another genus, Stemonosudis. Macroparalepis molestus described above is awkwardly intermediate between these two genera. Like Macroparalepis, as defined by Harry (1953), the nostrils of M. molestus come before the tip of the maxillary; like Stemonosudis the body is slightly compressed and very elongate, the tip of the lower jaw is not elevated and the end of the upper jaw is about 1 orbital diameter from the anterior margin of the eye. ANOTOPTERIDAE In fin-ray counts Macroparalepis molestus is intermediate between the two genera. 317 Dorsal Pectoral Anal Macroparalepis M. molestus Stemonosudis 11-14 12 9 IO-II 12 10-13 24-28 3° 37-5° Lastly, in colour pattern Macroparalepis molestus partly resembles Stemonosudis in having saddles of pigment along the back, but there are no alternating blotches on the base of the anal fin. Furthermore, it is interesting that a pigmentation very similar to that described for Stemonosudis has been found by Harry (19536) in a fish that he identifies as a post-larval Macroparalepis. In the present state of our knowledge, it would thus appear advisable to retain Macroparalepis as defined by Ege (1933). The Family ANOTOPTERIDAE Anotopterus pharao Zugmayer. Anotopterus pharao Zugmayer, 191 1. Result. Camp. sci. Monaco, p. 138, pi. 4, fig. 4 (for a full synonymy see Hubbs, Mead and Wilimovsky (1953)). During the 1949-50 Antarctic whaling season, the captain of one of the whale-catchers attached to F/F ' Balaena ' of United Whalers Limited, obtained a large fish belonging to the genus Anotopterus: it was swimming at the surface, the position being approximately lat. 690 S, long. 1700 E (Ross Sea area). Our knowledge of this genus is limited and a search was made through the unnamed Discovery Collections in the hope of finding other specimens. Eventually a small individual of standard length 100 mm. was found: it was taken in the South Atlantic (230 16' S, 01 ° 51-7' W) in a 450 cm. net hauled obliquely between 600 m. and the surface. The study of these two individuals, as well as of the type of Eugnathosaurus vorax Regan and of a specimen of Anotopterus pharao from Madeira, forms the basis of this section. Zugmayer (191 1), after examining a fish of total length 165 mm. taken by ' Princesse-Alice ' at the North Atlantic station 3028 (360 54' 30" N, n° 49' W), described a new genus and species A. pharao and furthermore proposed that it should be assigned to a new family, the Anotopteridae. A second fish of 291 mm. taken off Madeira was described by Roule (1927) and compared with the type speci- men. Legendre (1934) found five individuals of this species (from 127 mm. to 144 mm.) in the stomachs of Thunnus alalunga, caught by fishermen in the Bay of Biscay. These specimens were examined by Roule (1935). More recently Maul (1946) and Nybelin (1946) have written on Anotopterus. Maul obtained speci- mens of A. pharao from the stomachs of Alepisaurus ferox, caught off Madeira by local fishermen. Nybelin studied a head (length 150-160 mm.) obtained from the stomach of a halibut (caught by M/S 'Arctica' at 640 25' N, 530 30' W), and proposed the species Anotopterus arcticus. He also examined a complete specimen, of standard length 650 mm., found in the stomach of a fin whale (harpooned in the Weddell Sea at 620 02' S, 380 25' W) which he regarded as another new species, A. antarcticus. Nybelin (1946) also reviewed Eugnathosaurus vorax, which was described by Regan (1913) from a head trawled off Coats Land by the Scottish National Antarctic Expedition (station 417 at 71 ° 22' S, 1 6° 34' W), and he rightly saw that this was a species of Anotopterus, but concluded that owing to certain differences in the dentition it was distinct from A. antarcticus. Young A. pharao have been recorded from the eastern North Atlantic by Nybelin (1948) and Maul (1952). Nybelin's fish is about 14 mm. long and has 4 palatine teeth: Maul found 14 palatine teeth in a specimen of standard length 47-5 mm. 318 DISCOVERY REPORTS The foregoing papers were based on fishes from the Atlantic and Antarctic areas, but quite recently Abe (1952) and Hubbs, Mead and Wilimovsky (1953) described specimens from the north-east and north-west Pacific. The latter workers also made a close study of the literature and concluded that there is no good evidence for recognizing more than one species, pharao, of the genus Anotopterus. The study of the type of Eugnathosaurus vorax and of the two specimens mentioned earlier, together with a specimen from Madeira and published papers, has led me, independently of these ichthyologists, to much the same conclusion. But there is some indication that this one species may be differentiated into a northern and southern form ; further material may even show that these ' forms ' might better be re- garded as separate species. However this may be, my present intention is to bring forward new data on Anotopterus and then to point out the differences between fishes from the northern and southern seas. Much of the ground I had intended to cover regarding the status of the species of Anotopterus has already been thoroughly dealt with by Hubbs et al. (1953). Comparison of the type specimen of Eugnathosaurus vorax Regan (borrowed from the Royal Scottish Museum) with the large Antarctic specimen taken by the whale-catcher has shown them to be closely similar. The measurements in millimetres are given below, those for the type coming first, followed in parentheses by those of the complete fish. Length of head 150 (147); length of snout 89 (89); horizontal diameter of bony orbit 15*5 (i4*5); width of bony interorbital 8-o (8-o); postorbital length of head 46 (46); length of mandible in (108); length of premaxillae 90 (89); length of largest palatine tooth 9-5 (10*5). Dentition. The type specimen has no teeth on the first fifth of the length of the left mandible, then come twelve small, pointed and more or less upright teeth, then seven larger teeth inclined backwards and then two squat teeth, shaped rather like a rose thorn. Between the second to fifth mandibular teeth three smaller ones are inserted. In the complete Antarctic specimen, the arrangement of teeth on the left mandible is very similar: there is a short, foremost toothless part, followed by seventeen small, pointed teeth (the first seven of which are inclined forward), then by seven large retrorse teeth and finally by two small thorn-like teeth. The palatine dentition of the type is well-illustrated in fig. 1 (p. 234) of Regan's (1913) paper. In addition to the teeth shown in this figure, I have found indications of others. There are certainly the remains of a tooth base between the second and third teeth of the right palatine and there are indica- tions of another base in front of the first tooth. On the left palatine there is a tooth base between the second and third teeth ; this base is a clear-cut oval with a brownish centre, perhaps an indication of a resorbed tooth. Thus the type had thirteen functional teeth, two of these having been detached at some time. The palatine dentition of the complete individual is shown in Text-fig. 6b. As well as the eleven upright, functional teeth, there are a number of recumbent teeth lying in the dental integument and one obliquely set tooth. The latter is quite hard and is evidently moving upwards into position. The recumbent teeth can be divided into two types: (1) moderately ossified teeth (shown in the figure by cross hatching), and (2) soft teeth (shown dotted). On the right palatine it will be seen that the last two teeth have soft replacement teeth, while on the left palatine two of the moderately ossified re- cumbent teeth, soon to become functional, are closely associated with what must also be replacement teeth. The tooth pattern suggests that at any time there may be from 11 to 14 palatine teeth in use. Presumably tooth succession goes on all the time as in Alepisaurus. It was of such fishes that Owen (1840-5) wrote: ' . . .the succession of teeth is uninterrupted, the pulps of the new teeth are developed in most of the species in the soft gum or integument covering the dentigerous margins of the bones, and the calcification of the pulp is completed as it lies recumbent and buried loosely in the substance ANOTOPTERIDAE 3>9 of the gum. The point of the new tooth, which, in this state, is directed backwards, is then exposed by a gradual rotatory movement of the tooth from the horizontal to the vertical position. The small individual of standard length ioo mm. taken by R.R.S. 'Discovery II' in the South Atlantic (station 2048, lat. 23° 16' S, long. oi°5i-7'W.; 25. iv. 37, N. 450B: 600-0 m.) has six functional teeth on the left palatine and seven on the right. The dentition of the right palatine is shown in Text-fig. 6 a. On the right mandible there are seventeen teeth: one retrorse tooth near the symphysis + 5 antrorse teeth +11 upright teeth. b Text-fig. 6. Palatine teeth of Anotopteras pharao: a, right palatine of a specimen 100 mm. in standard length ( x 6-5) ; b and c diagramatic representations of the palatine dentitionsof specimens732mm.and278mm.instandardlength(6, x 1 ;c, x 1-5). In b and c the functional teeth are shown in black : the moderately developed replacement teeth are cross hatched; the soft, relatively undeveloped replacement teeth are dotted. Compared with the type specimen of Engnathosaurus vorax, the proportions of the different parts of the head of the whole fish are as follows. (These figures are in hundredths of the head length, those for the type coming first, followed in brackets by those of the Discovery specimen) : Length of head 150 (24) mm. Proportions per cent: length of snout 59-3 (58-3), width of bony interorbital 5-3 (67), horizontal diameter of bony orbit 10-3 (13-3); postorbital length 307 (27-1); length of mandible 74-0 (70-8). The palatine dentition of an individual from Madeira of standard length 278 mm. is shown in Text- fig. 6c. There are six completely erect, functional teeth, together with two moderately ossified, obliquely set teeth, one on the left palatine between the first and second upright teeth, the other situated towards the posterior end of the right palatine. Opposite the latter tooth is a fairly well-ossified recumbent tooth, together with a soft replacement tooth. The tooth pattern suggests that, as the fish grows, the complement of these stabbing palatine teeth increases, probably to between eight and ten. 32o DISCOVERY REPORTS The body proportions and meristic characters of the three complete specimens dealt with above may be seen in Table 2. For other data the reader is referred to Nybelin (1945) and Abe (1952) and the comprehensive tables given by Hubbs, Mead and Wilimovsky (1953). There are no marked differences in body proportions between all the known individuals of Anoto- pterus pharao. Hubbs et al. have remarked that the differences in proportions between A. pharao and A. antarcticus considered by Nybelin (the pelvic-anal distance, the diameter of the eye, and the depth of the caudal peduncle) are likely to be an expression of age rather than of specific distinctness. Comparison of the 100 mm. and 732 mm. individuals in Table 2 shows them to be closely similar, apart from the relative size of the characters mentioned above. There can be little doubt that these characters are correlated with age. But there seems to be some differentiation within the species in respect of the number of functional palatine teeth and perhaps in the number of vertebrae. (See Table 3.) Table 2. Body proportions (expressed in hundredths of the standard length) and meristic features of three individuals of Anotopterus pharao Region Madeira South Atlantic (230 16' S, oi° 517' W) Antarctic (690 S, 1700 E) Standard length in mm. 278 100 732 Body proportions : Head length Diameter of orbit 20-0 2-2 24-0 2-8 23-0 1-9 Interorbital width 1-25 i-8 1-2 Length of snout Greatest depth of body Depth of caudal peduncle Length of pectoral fin Length of pelvic fin 12-2 5-3 i-6 Damaged Damaged 14-5 5"5 i-6 Damaged Damaged 12-0 3"3 1-0 7-4 3-3 Length between: Snout tip — adipose dorsal Snout — origin of pectoral fin Origins of pectoral — pelvic fins Pelvic — anal fins Pelvic fin — base of caudal fin 91-7 23-0 38-8 277 38-5 90-0 23-0 36-0 25-0 36-0 91-2 20-2 36-2 33-i 43-8 Meristic characters: Number of palatine teeth 6 13 11 Number of vertebrae 80 83 Number of lateral line 'scales' — — 83 Number of fin rays: Pectoral fin 12 14 14 Pelvic fin Anal fin 9 J5 9 15 9 H In general the table indicates that fishes from the North Atlantic and North Pacific tend to have fewer (4-10) palatine teeth than those (11-14) m South Atlantic and Antarctic waters, and although there are certain anomalies, the data also suggest that in fishes from northern waters there is a tendency for the number of palatine teeth to increase with age. The number of vertebrae [(79 ±1/2 to 81) in three northern fishes and 83 in two southern fishes] perhaps lends some support to the suggestion that within this wide-ranging species there is a northern and a southern form. However, Maul's (1952) record from Madeira of a 47-5 mm. fish with 14 palatine teeth has to be considered. All that can be done at present is to draw attention to the possibility con- sidered by Hubbs et al. (1953) that Anotopterus, like certain ceratoid angler fishes, '. . .occurs chiefly ANOTOPTERIDAE 321 as young to small adults in the warmer waters and as large adults only in colder waters towards the poles '. Perhaps there is some overlap between the spawning areas of northern and southern forms ; perhaps the idea of these two forms is an over-simplification. But these questions must obviously await further data. Table 3. Numbers of palatine teeth and vertebrae in Anotopterus pharao Authority Locality Size of fish (s.l. mm.) Number of pala- tine teeth Number of vertebrae North Pacific Hubbs et al. (1953) Kamchatka 861 10 79 ± 1 or 2 centra Abe 1952 Kamchatka or Hokkaido 855 10 — Hubbs et al. (1953) British Columbia 87-3 (head) 5 — Hubbs et al. (1953) Hokkaido 640-0 10 — Hubbs et al. (1953) 3S°35'N, 1220 25' W 299-0 4 — Hubbs et al. (1953) 3i0S4-3'N, i52°2i-6'W 112-5 8 — North Atlantic Nybelin (1946) Davis Strait, 640 25' N, 53° 30' W c 150 (head) 10 — . Collett(i896) Off Cape Finisterre 253 (head) 7 — Roule (1935) Bay of Biscay 126-144(5) 6? 81 Nybelin (1948) 380 25' N, 10° 23' W H 4 — Roule (1927) Madeira 276 6 — Maul (1946) Madeira 265-342 6 — Maul (1952) Madeira 47-5 14 — Hubbs et al. (1953) Madeira 247 8 — Hubbs et al. (1953) Madeira 260 8 — Marshall Madeira 278 6 80 South Atlantic and Antarctic Marshall 230 16' S, oi° 517' W 100 13 — Nyeblin (1946) Weddell Sea, 620 02' S, 380 25' W 650 H 83 Marshall Ross Sea area, 69 °S, 170° E. 732 11 83 Regan (1913) and Marshall Coats Land, 710 22' S, 160 34' W 150 (head) 11 +2 — Lastly, it should be mentioned that the large Antarctic fish bears, on either side of the tail, two basicaudal dermal keels very similar to those described by Hubbs et al. (1953) for their large Pacific fishes. Distribution. It will be apparent from the quotation above, that Hubbs et al. (1953) concluded that adult Anotopterus probably have an antitropical (bipolar) distribution. The facts presented here — the presence of a young fish in the South Atlantic at latitude 230 S a>nd of a 732 mm. individual in Antarctic waters — does not conflict with this conception. Concerning the vertical distribution there is definite support for the contention of the foregoing workers that Anotopterus, at least when adult, is a surface-dwelling fish. The fish caught at the surface in the Ross Sea area by the whale-catcher had a very distended abdomen. Soon after capture, this individual vomited up two moderately large Notolepis coatsi Dollo of standard lengths 180 mm. and 270 mm. The stomach of each of these fishes was bloated with krill (Euphausia superba). N. coatsi are often to be found in the stomachs of whale-bone whales : in fact nearly all the sizeable specimens in the museum collections have been obtained from whales (see also Clarke, 1950). Examination of the stomach contents of twelve Notolepis has invariably shown them to have been eating adult krill. As Notolepis feeds on the krill of the surface waters, it is evidently preyed on by Anotopterus and in turn both may be fortuitously swallowed by the krill-engulfing whales. It may be 3-2 322 DISCOVERY REPORTS concluded that in Antarctic waters Anotopterus lives, at least for part of the time, in the upper ioo m. But little is known of the vertical limits of distribution and the possibility of tropical submergence needs to be borne in mind. PART II. COMPARISON OF ALEPISAUROID AND MYCTOPHOID FISHES Although the Iniomi have been divided into the suborders Myctophoidea and Alepisauroidea, it is by no means easy to assign clear-cut diagnostic features to these groups. Regan's (191 1) definition of the Alepisauroidea was mainly based on the osteology of Alepisauras and Evermannella, and he placed the paralepidid fishes and Sudis among the Myctophoidea and considered that Ateleopus should be put into a third and separate suborder. Regan's use of the forward attachments of the palatines as a character for separating the alepisau- roids and myctophoids cannot be sustained, as a study of Parr's (1929) paper soon shows. The degree of development of the interoperculum (the smallness of this bone in the alepisauroids contrasting with its more normal development in the myctophoids), is a trenchant distinguishing feature, as long as Regan's inclusions within the Alepisauroidea of Scopelarchus, Evermannella, Omosudis, Alepisaurits are accepted. But Harry (1951, 1953) has cogently argued that the Paralepididae show a closer rela- tionship to the alepisauroids than to the myctophoids ; and in this family the interoperculum is well developed (Parr, 1929). Furthermore, Anotopterus, which has many features in common with Alepisaurus, differs from the latter in possessing a relatively large interoperculum. Similarly, the preoperculum, which is narrow and almost vertically placed in the Scopelarchidae, Evermannellidae, Omusodidae and Alepisauridae, is broad and boomerang-shaped in the Paralepididae and fairly wide in the Anotopteridae. In both these latter families the preoperculum is obliquely inclined, following the line of the forwardly directed suspensorium. Regan also based his classification on the design of the pectoral girdle, remarking that in the Mycto- phoidea the cleithrum is attached to the lower end of the supracleithrum, while in the Alepisauroidea the attachment comes at the upper end of the latter bone. As far as I have been able to extend this observation, it seems true for the myctophoids in which a more 'solid' pectoral girdle is developed, there being a firm join between the supracleithrum and cleithrum. In the myctophoids (Neoscopelus and Lampanyctus crocodilus were examined) there is a wide area of overlap between these two bones, the cleithrum extending rather beyond the midpoint of the supracleithrum. In the alepisauroids, Omosudis, Evermannella and Alepisaurits the cleithrum is attached to the upper part of the supraclei- thrum (Regan, 191 1 ; Parr, 1929), but in the paralepidids, although the above attachment is found in Paralepis speciosa, Parr (1929) remarks that in Lestidium intermedium on the other hand the cleithrum only extends somewhat above the middle of the supracleithrum. Again, in the figure of the pectoral girdle of Scopelarchus anale (Parr, 1929), the cleithrum is shown to be attached to the lower end of the supracleithrum. Clearly the type of linkage between these two bones cannot be used as a subordinal diagnostic. The position of the pectoral fins was also listed as a distinguishing character by Regan, who pointed out that these fins were laterallyplaced in the myctophoids but were low in position in the alepisauroids. There are a number of exceptions to this generalization and the whole question of the position of the pectoral fins will be more fully discussed in a separate section (pp. 325-28). Lastly, Regan (191 1) said that the Alepisauroidea had '. . .strong pointed canines in the lower jaw and on the palatines'. This is true for his inclusions, Scopelarchus, Evermannella, Omosudis and Alepisaurits and for Anotopterus, but this description does not cover certain of the paralepidids, such as Paralepis and Magnisudis (Harry, 1953). However, there are certain features of dentition possessed ALEPISAUROID AND MYCTOPHOID FISHES 323 by all alepisauroids, namely, the premaxillae each carry a single row of teeth and there are one or two rows of teeth on each palatine and 1-3 rows on each dentary. Turning now to more recent studies, Harry (1953 a) stated that the symphysis of the lower jaw is more or less elevated in the Paralepididae, fitting into a corresponding, arched toothless emargination on the upper jaw. As Harry has pointed out elsewhere, the symphysis of the upper jaw is usually toothless in the Alepisauroidea and may be arched. (A specimen of Evermannella balbo which I ex- amined had an arched symphysis provided with teeth.) It is also of interest that certain of the Mycto- phoidea have a bony boss at the symphysis of the lower jaw, which fits into a toothless emargination of the upper jaw (Chlorophthalmidae, Bathypteroidae, Ipnops, Bathymicrops, Notosudis, Luriosudis and Neoscopehis). In many of the Myctophidae there is certainly a small bony boss at the symphysis of the lower jaw, but this fits against a toothless, slightly depressed area between the two premaxillae, but such close-fitting structures are absent in the jaws of the Aulopidae, Harpadontidae and Syno- dontidae. Harry (1953 a) has also drawn attention to the form of the gill-rakers in the Alepisauroidea. In Magnisudis each gill-raker consists of a basal part bearing four long, stiff, depressible, filamentous parts. The rakers of Alepisaurus ferox are somewhat similar, each consisting of a basal part supporting a number of spines, of which one to four are longer than the others. Needle-like or spinous gill-rakers are also found in Paralepis, Oinosudis and Evermannella.1 In Scopelarelius, Notolepis, Lestidinm, Macroparalepis and Sudis the rakers are tooth-like. All the gill-arches in Anotopterus are without rakers. More normal lath-like gill-rakers are characteristic of all families of the Myctophoidea, except the Harpadontidae and Synodontidae, these having tooth-like or spinous rakers. There is therefore much justification for the opening remarks concerning the difficulty of defining the suborders Myctophoidea and Alepisauroidea, but before summarizing this discussion in a section on alepisauroid relationships, there are certain interesting morphological trends within this group which will now be considered. Points of contrast with the myctophoids will then be seen. The swim-bladder. A survey of the literature and a number of dissections {Alepisaurus ferox, Anotopterus pharao, Omosudis lowei, Evermannella atrata, Scopelarelius guentheri and Notolepis coatsi) show that a swim-bladder is absent in the Alepisauroidea. In the Myctophoidea this organ is only found in the Myctophidae, Marshall (1951 and unpublished) finding a closed swimbladder in twenty- three of the twenty-six species studied. The absence of the swim-bladder in the Aulopidae, Chloro- phthalmidae, Synodontidae, Harpadontidae, Ipnopidae and Bathypteroidae was recorded by Gunther (1864, 1887). Correlations between environment and the degree of development of the swim-bladder (including its loss) in teleosts have already been considered by Marshall (1950, 1951) and Jones and Marshall (1953), but in these papers little reference was made to the alepisauroids. The Alepisauroidea are active, voracious, mid-water fishes, while the Myctophoidea, other than the Myctophidae and Notosudidae seem to be predominantly bottom-dwellers. Many benthic fishes lack a swim-bladder, but the absence of this organ in a mid-water group, such as the Alepisauroidea, is of interest and will be further considered. Marshall (1951 and unpublished) has found a well-developed swim-bladder in many of the bathypelagic fishes, which have centres of abundance in the uppermost 1000 m. of the ocean. But notable exceptions to this rule are found in the Melanostomiatidae, as well as in the Alepisauroidea. (A consideration of the reports by Ege, 1930, Legendre, 1934, and Maul, 1945, supports the view that many paralepidids occur in the uppermost 500 m. and the same would 1 In the specimen of Evermannella balbo examined, there are no rakers on the first gill arch but some are present on the second arch. 55. 324 DISCOVERY REPORTS also appear to be true for Alepisaurus and Anotopterus. Our knowledge of the vertical distribution of the Scopelarchidae, Evermannella and Omosudis is less complete.) Fishes without a swim-bladder are heavier than their environment (Jones and Marshall (1953) have reviewed the data on densities). The only pelagic fish that has been well studied is the mackerel, Scomber scombrus with a density of 1-071 and a sinking-factor of 1043. * Very probably other pelagic fishes without a swim-bladder are also heavier than the surrounding water, and, like the mackerel keep their level by unceasing activity. A study of their appearance and food suggests that the Alepisauroidea are active swimmers and their lightly ossified skeletons may point to their having lower sinking-factors than the mackerel. Moreover, Taylor's ( 1 92 1 ) work on the density of fish tissues would indicate that the sinking-factor of any alepisauroid is likely to be more than 1000. While there are no observations of swimming in alepisauroids which bear on this problem, it is in- teresting that Skowron (1928), who studied the luminescence of Chauliodus shard, saw that these fishes at first swam vigorously, but on ceasing active movements sank to the bottom of the container. This, as I have seen in the Red Sea, is also true of Stomias affinis. Evidently both fishes are heavier than the surrounding water, and it is significant that, like the alepisauroids, both are without a swim- bladder. Ossification of the skeleton. Hubbs, Mead and Wilimovsky (1953) found no reason to accept Roule's chimerical hypothesis that Ajiotopterus is a ' monstre normalise ' or ' teratobionte '. Nor is there evidence that this fish or any other alepisauroid is rachitic. Their skeletons are rather delicately but not abnormally built. A five foot Alepisaurus ferox weighs about 4 lb. (Lowe, 1835), a striking indication of its lightly ossified skeleton as well as of its compressed and rather lean-looking body. Harry (1953 a) has drawn attention to the peculiar cartilaginous development of the jaws in the Paralepididae and Anotopteridae. Certainly, the long time of exposure required to get tolerably clear hard-tube X-ray photographs of these fishes provides further evidence of the lack of calcification in the skeleton ; and this is also true of Omosudis lowei (s.L. 60 mm.) Evermannella hyalinus (s.l. 153 mm.), Neoscopelarchoides elongatns (s.l. 198 mm.) and N. dubius (s.l. 138 mm.). Only in Scopelarchus cavei sp.n. (s.l. 70 mm.) does the degree of ossification of scales and skeleton compare with that, say, in a myctophid. Moreover, in some of the alepisauroids the calcification of the skeleton seems to be a slow, irregular and protracted process. Beebe (1932) remarks that a post-larval Omosudis lowei of total length 1 1-45 mm. did not have ' . . . a particle of bone in the entire body, fins, fin-rays or appendicular skeleton except for a faint trace in the centre of the supracleithrum '. The head bones, particularly the pre- maxillae and dentary were well ossified. Similarly, in the large ' Benthalbella ' larvae of the scopel- archids, the skeleton, except for the skull, has a soft, cartilaginous texture. In Anotopterus the thin, scale-like segments along the lateral line canal (see Text-fig. 7) are only plainly visible in larger indi- viduals. Maul (1946), who studied specimens from 265 mm. to 342 mm. in standard length, remarks that, along the anterior half of one individual, they only became visible after alizarin staining. Lastly, in certain paralepidids (Ege, 1930; Parr, 193 1 ; Harry, 1953 a), the scaling is not completed until very late in development. I have also found this feature in Notolepis coatsi. Parr (1937) has suggested that the amount of available vitamin D falls off with increasing depth, and that fishes living at deep levels in the ocean have developed lower requirements for this vitamin, this resulting in less extensive ossification of the skeleton. But assuming vitamin D to be necessary for the normal formation of bone in fishes, Anotopterus, Alepisaurus and the paralepidids live at levels not very far removed from the phytoplankton-bearing, vitamin D-producing, upper layer of the ocean. , 0. . . r density of fish 1 binking-factor = -j : ~ ; — - x 1000. density of environment ALEPISAUROID AND MYCTOPHOID FISHES 325 On the other hand, abyssal benthic fishes might be expected to have poorly ossified skeletons, yet in the Bathypteroidae, and Ipnopidae (to consider only the Iniomi), the skeleton is well formed and much denser than in the alepisauroid fishes. (This is well shown in X-ray photographs.) A more solid skeleton is no handicap to a fish which lives on the deep-sea floor, whereas reduced ossification would appear to be an advantage to the bathypelagic Alepisauroidea. This fact compensates these active predacious fishes for the absence of a swim-bladder, for it must result in the lowering of the sinking- factor, and hence less energy is required to maintain the fish at a particular feeding-level in the ocean. The correlation between the lack of a swim-bladder and the laying-down of less skeletal material in bathypelagic fishes will be considered more fully in a later paper, but one comparison may be made now. Gonostoma denudatum has a large, well-developed swim-bladder, while in G. bathyphilum there is no trace of this organ. A specimen of denudatum (131 mm.) was placed alongside one of bathyphilum (122 mm.) and an X-ray photograph obtained. This strikingly confirmed the impression formed by handling these fishes, that denudatum had a better ossified skeleton than bathyphilum.1 If this interpre- tation comes near the truth, more striking corroboration would be hard to find than these structural differences between the two species. Text-fig. 7. Part of the lateral line (near the origin of the adipose dorsal fin) of Anotopterus pharao, showing two of the scale-like ossifications and the lateral line pores ( x 6-6). The setting of the pectoral fins. In defining the two suborders of the Iniomi, Regan ( 1 9 1 1 ) stated that the pectoral fins were lateral in the Myctophoidea, but low in position in the Alepisauroidea. Certainly, the pectorals of Alepisaurus, Anotopterus, Omosudis and Evermannella are placed very low on the shoulders, while those of the Aulopidae, Chlorophthalmidae, Synodontidae, Harpadontidae, Ipnopi- dae and Bathypteroidae are laterally placed. But a number of exceptions to these generalizations are to be found in the Myctophidae, Scopelarchidae and Paralepididae. Measurements of the angle between the horizontal axis of the body and the axis of the pectoral fin have indicated that such an index comes close to being a diagnostic feature of each suborder. In most alepisauroids this angle is below 450 : in most myctophoids it is above 450. Data, showing what may be called the pectoral angle, are listed in Table 4. It is interesting to consider these data in the light of the ideas of Harris (1952) on the function and the evolution of the fins in fishes. Harris contended that, with the evolution of a swim-bladder, there was no longer any tendency for the body to sink or even tilt, because the centres of gravity of the body and the swim-bladder are usually either close together, or else coincident. Pectoral fins set as hydro- planes and an asymmetrical tail, both giving a lift-force (as in sharks) were thus no longer necessary. So the pectorals were freed to become limbs in the Choanichthyes, and brakes (or oars) in the Teleostii, while the caudal fin attained a symmetrical neutral shape, but with a different structural basis, in the Dipnoi, Crossopterygii and Actinopterygii. With increasing specialization of the pectoral fins as brakes, there appears to have been an upward movement of the fins along the sides of the fish and a rotation of the axis of each fin so as to bring the axis near to a vertical position. Since the Alepisauroidea have no swim-bladder they might be expected to have a specific gravity somewhat greater than their environment (see p. 324). In order to remain at one level, an alepisauroid 1 Fishes of this size may contain well-developed ova, showing that they are nearing, or have attained, the adult phase. Thus the differences in ossification are unlikely to be due to differences in the phase of life. 3z6 DISCOVERY REPORTS must presumably make continual compensating movements. Now the setting of the pectoral fins in alepisauroids gives the strong impression that one of their functions is to produce an upward lift during forward motion. Moreover, during the swinging of the tail, the combined caudal and anal fins, which are set close together, may act rather like a heterocercal caudal fin also giving an upward lift. If this is so, the alepisauroids have certain dynamic parallels with the sharks. Table 4. Pectoral fin 'angles' in some iniomous fishes. Alepisauroidea Scopelarchus guentheri 20° Paralepis brevis 35-40° S. cavei sp.n. 30° P. coregonoides 25-30° Neoscopelarchoides elongatus 20° P. sphyraenoides 30° N. dubius sp.n. 25° Anotopterus pharao 10° Evermannella balbo 15-20° Alepisauriis ferox 5-io° Sudis hyalina 45° Omosudis lowei 20-250 Myctophoidea Auloptis filamentosus 50° Gonichthys cocco 70° Chlorophthalmus agassizii 45° Ctenobranchns nigro-ocellatus 75° Bathysauropsis gracilis 70° Diaphus coeruleus 60° Bathypterois filifer 70-80° D. lutkeni 55° Ipnops murrayi 70° D. ostenfeldi 45° Bathymicrops regis 65° D. lucidus 40° Harpadon nehereus 40-45° D. garmani 6o° H. squamosus 45° Notolychnus valdiviae 75° Synodus similis 45° Ctenoscopelus phengodes 45° Saurida elongata 40° Lampadena nitida 70° Electrona antarctica 40-45° L. chavesi 60-65° E. rissoi 6o° Ceratoscopelus townsendi 55° Myctophum affine 55° Notoscopelus elongatus 75° M. humboldti 50° Yet this is to regard the pectorals merely as relatively passive hydroplanes. A study of their structure soon shows that the pectorals of any alepisauroid must have greater mobility than those of a shark. No observations on alepisauroids are available, but Breder and Krumholz (1943) have analysed the functions of the pectoral fins in Anchoa mitchilli, which like most clupeids and alepisauroids has low set pectorals, each with an axis making less than 450 with the horizontal axis. (In A. mitchilli this angle is about 300). In spite of a well-developed swim-bladder, an adult A. mitchilli is evidently a little heavier than the surrounding water. During swimming there is apparently a slight tendency for the fish to roll, this being counteracted by the constant, irregular beating-down movement of the pectorals. Between these movements, the pectorals also move together in simultaneous down-beats so as to raise the head for swimming. Perhaps the pectorals of the Alepisauroidea have rather similar functions. At all events it is clear that pectorals, with the axis nearer to the horizontal than to the vertical, may act not only as hydroplanes but also as roll-compensating and elevating devices. But there is another aspect to these considerations. While the evolution of a swim-bladder made possible the evolution of pectoral fins as brakes, the inference that all teleosts with well-developed swim-bladders should have pectorals with this function is only partly true. The remarks above on A. mitchelli show this very clearly. When discussing locomotion in the Isospondyli, Breder (1926) remarked that their pectoral fins ' . . . while capable of being folded back do not have the flexibility to be seen in many of the higher teleosts and are not often capable of being used effectively to back water. Their use is more nearly like that of the shark's pectorals. . . '. ALEPISAUROID AND MYCTOPHOID FISHES 327 It is significant that the little available evidence suggests that certain isospondylous fishes are some- what heavier than the environment. This is borne out by the observations of Breder and Krumholz (1943) on Harengula pensacolae and Anchoa mitchilli, and by measurements of the specific gravity of Clupea harengus (Magnan, 1929). Moreover, the studies of the first two authors on Megalops atlanticus suggest that the simple, unspecialized type of open swim-bladder, characteristic of most Isospondyli, has a limited function in regulating the specific gravity. After remarking that the pectoral movements of tarpon are not unlike those of Anchoa mitchilli, Breder and Krumholz (1943) continue thus: ' Observations in an aquarium of tarpon ranging from 6 cm. to 100 cm. indicate that they seem to be- come heavier as the time for them to rise for breath approaches. Their pectorals typically work harder and finally with a burst of tail effort they rush to the surface and gulp .... After the ingestion of air they are usually lighter than water and frequently have difficulty descending, until they emit small bubbles by way of the gill clefts, after which they reach a state of approximate balance and from then on become heavier again.' It would appear therefore, that as far as present knowledge goes, the development of a capacious swim-bladder in the Isospondyli has not always led to a close correspondence between the specific gravity of the body and that of the environment. Correlated with a slight tendency of the fish — at least for part of the time — to sink, the pectorals are set in such a way that they can act as hydroplanes and beat downwards to raise the head. The angle between the axis of the pectoral fin and the horizontal axis of the body in a number of isospondylous fishes is given below. Elops saurus (400) Megalops atlanticus (450) Chanos chanos (15-200) Clupea harengus (25 °) Sardinops sagax (300) Harengula pensacolae (20°) Ilisha filigera (45°) Pristigaster cayanus (30-35°) Odontognathus compressus (45-50°) Salmo salar (40°) Coregonus sp. (45°) Salvelinus alpinus (45°) Osmerus eperlanus (40-45°) Retropinna oameroides (40°) Plecoglossus altirelis (25°) Galaxias fasciatus (55°) But not all Isospondyli have an open swim-bladder (see Jones and Marshall (1953)). A closed swim- bladder is commonly found in the Gonostomatidae and Sternoptychidae (Marshall, 1951) and it is of interest to measure the pectoral angle in representatives of these families. Maurolicus muelleri (6o°) Vinciguerria attenuata (40°) Gonostoma denudatum (45-50°) Ichthyococcus ovatus (45°) Argyropelecus aculeatus (65°) Sternoptyx diaphana (60°) Polyipnus nuttingi (50°) A comparison of the pectoral angle in Gonostoma denudatum with that in G. elongatum and G. bathyphilum (20-25°) is revealing, when considered in the light of the foregoing discussion. G. denudatum has a well-developed gas-filled swim-bladder with a powerful gas-secreting complex; G. elongatum has a fat-filled swim-bladder with a degenerate complex; G. bathyphilum has no swim- bladder. Since the two latter species are probably heavier than their environment, the presence of pectoral fins with a hydroplane-like setting is not surprising. And as G. denudatum can presumably bring its weight in water to vanishing point, the possession of pectorals which are beginning to look more like brakes than hydroplanes is more comprehensible. From these considerations we may return to the Myctophidae, most of which have a swim-bladder. In all investigated species (Marshall, 1951), this swim-bladder is closed and is usually provided with a highly developed gas-secreting complex. Reference to the list of pectoral angles of myctophids (see p. 326) shows that most species have pectorals with a setting more efficient for braking (or paddling) than for hydroplaning. 328 DISCOVERY REPORTS It would appear therefore that in the Isospondyli and the Iniomi there is some correlation between the possession of a closed swim-bladder and pectorals having angles greater than 450.1 But the specialization of pectoral fins as brakes is best seen in the higher groups such as the Percomorphi. It must also be remembered that many of the isospondylous and iniomous fishes with closed swim- bladders undertake considerable diurnal vertical migrations. At the end of a descent, compression of the swim-bladder may leave the fish somewhat heavier than the sea-water, and, until the volume can be restored by gas secretion, the fish will need to make compensating movements. Perhaps this is why the pectorals are not so highly specialized as in the Percomorphi and are set so as to allow some lift- force. While a closed swim-bladder means that the fish is independent of the surface for replenishing the gases and is able to achieve a finer control of specific gravity, there are certain limitations. Moreover, this discussion does not imply that all fishes with open swim-bladders have only a rough control of their specific gravity. Measurements of their specific gravity (Jones and Marshall, 1953) and observa- tions of their behaviour indicate that cyprinid and esocid fishes are perfectly able to reduce their weight in water to vanishing point. Such fishes are able to manoeuvre more easily than, say, clupeids. Lastly, the pectorals tend to be lateral and ' brake-set ' not only in the Myctophidae but also in other myctophoid fishes, the Aulopidae, Chlorophthalmidae, Synodontidae, Harpadontidae, Ipnopidae and Bathypteroidae. As already mentioned these are predominantly bottom-dwelling fishes and have no swim-bladders, and it is hardly surprising that the pectoral fins should have this setting, for it is a common feature in benthic fishes. Considering only those without a swim-bladder, pectorals with an upright or nearly upright axis are found in the scorpaenids, cottids, nototheniids, gobiids and gobie- socids. Certainly the pectorals in these fishes are used in locomotion and, being so placed, may allow of better control of body movements as the fish grubs for food. The similarity in the setting of these fins in the pelagic Myctophidae on the one hand and the remaining benthic myctophoids on the other is therefore probably an instance of an unusual type of convergence (the independent acquisition of like characters). The dorsal fin. Except for Alepisaurus, the dorsal fin of alepisauroids has 6-15 rays and is set about half-way down the body length. In Anotopterus there is no trace of a dorsal fin, while Alepisaurus has a long sail-like dorsal, very like that of the sail-fishes {Istiophorus). This convergence is, indeed, so striking that it led Dollo (1909) to conclude — on very slender evidence — that Alepisaurus uses its dorsal fin as a sail, much in the manner sometimes seen in Istiophorus. In Alepisaurus ferox the dorsal fin, which has about 40 rays, originates just behind the head and ends above the origin of the anal fin, i.e. the base of the dorsal extends over more than two-thirds of the standard length. The longest of the anterior rays may measure at least twice the greatest depth of the body. The muscles which move the dorsal fin rays are shown in Text-fig. 8 a, the upper figure being of the first three rays, the lower of the 14th and 15th rays. It will be seen that the erector muscles are power- fully developed, while the depressors and particularly the inclinators are a good deal smaller. The basal supports for the dorsal fin are also of interest. At the base of each ray is a distal radial, but it would appear that the intermediate and proximal elements have fused to form a long continuous rod, running the entire length of the dorsal fin, an unusual feature in fishes (Text-fig. 8 a, b). But such a development is hardly surprising in view of the strains imposed on the high dorsal as it is hauled up L Two myctophids listed on p. 326, Gonichthys cocco and Ctenobranchus nigro-ocellatus have no swim-bladder yet have pec- toral angles of 70-750. But both species have a more stream-lined, slender body than other myctophids and both may well keep to a particular level by vigorous swimming movements (rather like Scomber). Beebe and Vander Pyl (1944) have remarked on the vigorous unceasing swimming movements of Gonichthys cocco as compared with other lantern-fishes. ALEPISAUROID AND MYCTOPHOID FISHES 329 during active movements. Perhaps the sudden raising of the dorsal gives stability during turns and enables turning to be more rapidly effected, a capacity that may well be important in a fish which preys on such large active fishes as the paralepidids. EM Text-fig. 8. Dorsal fin of Alepisaimis ferox: a, the basal elements supporting the first three dorsal rays. DM, EM, IM, depressor, erector and inclinator muscles of fin-ray (X4/5); b, the basal elements supporting the 14th and 15th dorsal rays ( x 7-5). Intermuscular bones In his Classification of Fishes, Berg (1947) remarks that in the lower teleosts true intermuscular bones (epineurals, sometimes also epicentrals) are usually present. In defining the larger groups, Berg states that intermuscular bones are present in the Isospondyli, Haplomi, Ostariophysi (Cypri- noidea) and Apodes. (Definite absence is recorded for the gadoids, Solenichthyes, Microcyprini and Percomorphi.) Intermuscular bones are also present in certain of the alepisauroid fishes. Hubbs, Mead and Wilimovsky (1953) observed the prominent development of these bones in Anotopterus and Alepi- saurus and described the complex as consisting of a series of thick dorsal elements running hori- zontally, together with more slender median bones stretching backward and slightly downward. The positions of the intermuscular bones in a transverse section across a medium-sized individual of Anotopterus pharao may be seen in Text-fig. 9 a. The pattern of intermuscular bones in Alepisaurus is remarkably similar to that found in Anotopterus. Certain of the Paralepididae and Omosudis lowei also have intermuscular bones, which are developed in both the epaxial and hypaxial parts of the myotomes. A drawing of the more conspicuous elements associated with the first three vertebrae of Omosudis may be seen in Text-fig. gb. The extraordinary 4-2 33o DISCOVERY REPORTS interneural, median, ossified body connected with the first vertebra is also shown. Parr (1929) has already described and figured this ossification, but it may be added that the median dorsal ligament, which runs down the trunk, is attached to its posterior end. In the Paralepididae, I have found inter- muscular bones in Paralepis speciosa, P. brevis, P. brevirostris and P. coregonoides, but none were found in Notolepis coatsi, N. rissoi kroyeri, Lestidiwn sphyraenoides and Sudis hyalina. Lastly, X-ray photographs of Scopelarclms cavei sp.n., S. guentheri, Neoscopelarchoides dubius sp.n., N. elongatus and Evermannella balbo have not revealed the presence of intermuscular bones. (These show up quite clearly in X-ray photographs of paralepidids, Anotopterus and Alepisaurus.) Text-fig. 9. a, transverse section through the trunk muscles of Anotopterus pharao, showing the intermuscular bones (imb.) (x9'3); b, some of the larger intermuscular bones associated with the first three vertebrae of Omosudis lowei (x io-6) d.l. = dorsal ligament. Distensibility of stomach and body-zuall Evermannella, Omosudis, Anotopterus and Alepisaurus are among the alepisauroids capable of swallowing very large prey. Alcock (1899) has a drawing of Evermannella atrata very much distended with a large squid, while in the Discovery Collections there is a specimen of E. indica with a gono- stomatid fish folded up in its stomach, the length of the prey being appreciably longer than the length of the predator's abdomen. The capacity of Omosudis for swallowing large prey is well known since Gunther's (1887) account of a Sternoptyx in the stomach of one individual, victim and predator being equal in bulk. Anotopterus also has a highly distensible stomach and body-wall, as the presence of two large Notolepis coatsi in the specimen described earlier (p. 321) must indicate. Alepisaurus is a highly voracious fish with a comparable capacity for dealing with large prey : in the collections of the Natural History Museum there is an A. ferox considerably distended, having swallowed one of its own kind. There is no record of a scopelarchid or a paralepidid containing large prey (although I have seen many Notolepis coatsi, each somewhat distended by a stomach crammed with Euphausia superba) and it is interesting that of all the families of alepisauroid fishes, the scopelarchids and paralepidids are the only ones containing species which are fully scaled. Most probably all the scopelarchids are fully scaled, while among the paralepidids this condition obtains in Magnisudis, Paralepis and Notolepis only (Harry, 1953 a, b). INTERRELATIONSHIPS OF THE ALEPISAUROIDEA 33> Considering all fishes known to swallow relatively large prey, the striking fact emerges that all are scaleless or have a very reduced squamation. Omosudis, Alepisaurus and Evermannella are completely naked, while Anotopterus has scale-like segments only along the lateral line. Other fishes with dis- tensible stomachs and body-walls (the Melanostomiatidae, Idiacanthidae, Astronesthidae, Malacos- teidae, Lyomeri, Giganturoidea, Chiasmodus niger and Melanocetus have no scales). The only known scaled fishes which can engulf large prey are Stomias and Chauliodus. In Chauliodus — and probably in Stomias — the scales, which are very thin and do not overlap, lie between the epidermis and the pig- mented layer of the corium (Brauer, 1908). Moreover, the scales are not held in pockets, which sug- gests that fishes cannot have both a normal overlapping scale system and a highly distensible body- wall. Presumably the connective tissue pattern associated with an imbricating squamation sets a limit to the degree of tissue distensibility. However this may be, there is no record of a fish with well- developed, overlapping scales being able to cope with large prey. It is interesting that the scaleless (or virtually scaleless) paralepidids, Lestidium, Macroparalepis and Sudis have relatively larger teeth on the palatines and mandibles than the fully scaled Magnisudis, Paralepis and Notolepis. Is it pos- sible that the larger toothed genera are able to capture (and retain) larger prey and that the loss of scales is an adaptation towards distensibility of the stomach and body wall? On the other hand the scopelarchids with a formidable dentition of stabbing, barbed teeth are completely scaled. A study of the food of these fishes would be of particular interest. (In the stomach of Scopelarchus gnentheri of standard length 65 mm. I found the remains of copepods and ostracods and a mass of nematocysts, probably of Siphonophora.) The alepisauroids, which are all pelagic, may be compared and contrasted with the pelagic members of the Myctophoidea, the Myctophidae, Luciosudis and Notosudis. These latter fishes are fully scaled, the myctophids feeding particularly on copepods and smallish members of the zooplankton (Beebe and Vander Pyl, 1944). There is no evidence that any myctophid fish is able to tackle large prey. Harpadon, which of all myctophoid fishes has a dentition most like an alepisauroid, feeds on small fishes and shrimps (Hora, 1934). To summarize : among the pelagic families of the Iniomi the predominant evolutionary trend in the alepisauroids has been towards adaptations for dealing with nektonic prey, extraordinary elasticity of the tissues of the stomach and body wall being associated with a loss or marked reduction in the scaling. By contrast, the Myctophidae, which feed on planktonic animals, have a complete scaling. INTERRELATIONSHIPS OF THE ALEPISAUROIDEA Considerations of the interrelationships within the Alepisauroidea are inseparable from the question whether the suborder represents a natural group, but before attempting to deal with this problem the similarities and differences between the families must be elaborated. Parr (1929) considered that the Scopelarchidae and Evermannellidae were fairly closely related. Moreover, the other four families, the Alepisauridae, Anotopteridae, Omosudidae and Paralepididae have certain common features. The characters of these two groups can be listed as follows : Group 1 . Scopelarchidae and Evermannellidae Intermuscular bones absent : parietals fused with the frontals ; anus near to origin of anal fin or mid- way between pelvic and anal fins;1 body short to moderately elongate; definite trend towards develop- ment of tubular eyes. 1 In larval or juvenile scopelarchids, the anus is closer to the pelvics than to the origin of the anal fin, e.g. in Neoscopel- archoides linguidens (Mead and Bohlke, 1953) and in Benthalbella larvae. 332 DISCOVERY REPORTS Group 2. Alepisauridae, Anotopteridae, Omosudidae and Paralepididae Intermuscular bones present; parietals not fused with frontals (no information on Anotopterus) ; anus much nearer to insertion of pelvic fins than to origin of anal fin ; eyes normally formed ; definite trend towards development of elongate to very elongate body forms. It will be noticed that there are certain morphological trends within these two groups. In addition to those listed under the group headings, mention has been made already (pp. 322-31) to others, which can be briefly summarized here. There are tendencies to the development of a lightly ossified skeleton, to the loss of scales, and to the acquisition of means for dealing with very large prey (certain of the mandibular and palatine teeth have become large stabbing canines, and the tissues of the stomach and body wall are distensible). As in the more generalized Isospondyli, the pectoral fins are also set low down on the shoulders and make an angle of less than 45 ° with the horizontal axis of the body. These trends lead to more specialized forms, but within each of the two groups there is one family with a more generalized character complex. For example, in Group 1 the Scopelarchidae are fully scaled and have lingual teeth, whereas both these features are absent in the Evermannellidae. In Group 2, some of the Paralepididae are also fully scaled, in contrast to the other three families, the Alepi- sauridae, Anotopteridae and Omosudidae. In discussing the relationships of the Paralepididae, Harry (1953 a) pointed out that Magnisudis has many generalized features in common with the Aulopidae and with another myctophoid family, the Chlorophthalmidae. He also considered that the ' Paralepididae are most closely related to the Anotopteridae. They both have the same general proportions, essen- tially similar osteology, the same peculiar cartilaginous development of the jaws, which is found in these two families alone in the order, and a good number of other similarities.' These two families also show striking resemblances in the development of scale-like ossifications along the lateral line (Text-figs. 5 c and 7) and in certain skull features already mentioned on p. 322. The remaining two families in Group 2, the Alepisauridae and Omosudidae, are perhaps the most nearly related of all. Regan (191 1) observed that Omosudis has a head, mouth and teeth very much like Alepisaurns, and both genera have completely lost the scaling, even along the lateral line. The alepisauroid fishes are difficult to define, lacking diagnostic characters and it might be argued that they are not a monophyletic group. But Simpson (1953), after expressing his belief that the fissipede carnivores are monophyletic, remarked that he was unable to find a single character that occurred in all fissipedes and in no other mammals. Again, Berg (1940), after defining the Clupeiformes ( = Isospondyli), concluded by remarking that 'this order, from which a series of higher orders has arisen, represents an artificial assemblage, its separate members, as may be seen from the diagnosis, greatly differing from one another. In time, the Clupeiformes will doubtless be divided in many orders.' If the Isospondyli are considered to be merely a convenient grouping of fishes of a 'certain level of organization ', then the same may be true of the Iniomi. It has been pointed out on p. 305, that the Iniomi have evolved beyond the ' isospondylous level ' mainly in that the premaxillaries exclude the maxillaries from the gape, and it is possible that more than one group of isospondylous — or more likely pre-isospondylous — fishes may ' have tried this experiment ' with these jaw-bones. There may well be some parallel with the early evolutionary history of mammals. 'Palaeontologists use an arbi- trary criterion that a reptile became a mammal when the dentary-squamosal joint developed and the functional jaw-movement ceased to be on the articular-quadrate joint. This line was probably crossed separately by at least five different lineages. . .' (Simpson, 1953). It has also been stressed by Simpson that the development of a higher group of animals seems always to be bound up with definitive adaptive features correlated with a spread into some major living INTERRELATIONSHIPS OF THE ALEPISAUROIDEA 333 space. 'The adaptive characters involved may be quite broad and varied, as in carnivores and ungu- lates, or may be quite specific, virtually "single characters" in a taxonomic sense, as in rodents and bats.' As already stated, the alepisauroids are bathypelagic fishes and consideration of the group in- volves features such as the dentition, type of gill-raker, loss of scales, acquisition of distensible tissues etc., all of which would seem to be correlated with a predacious, large-prey-seizing habit. Such characters remind one more of the broad and varied features of the mammalian carnivores than of those of the bats. (Perhaps in the type of character-complex, the ceratioid angler fishes are the nearest oceanic equivalents to the bats.) Bearing these theorizings in mind, these studies have done little to dispose of the (partly intuitive) idea that the Iniomi may be no more than an assemblage of lineages, each having attained certain morphological features. When the great differences between, say, a myctophid and Anotopterus are considered, the value of keeping the order Iniomi seems questionable. Yet if this is not a compact ' monophyletic ' order, it is at the very least a convenient gathering of those ' isospondyloid ' fishes with more specialized jaws. Furthermore, there is little reason for discarding the two suborders. A consideration of the position of the Notosudidae (Notosudis Waite, 1916, and Luciosudis Fraser- Brunner, 193 1), provides something of a test of the usefulness of the classification used in this report (p. 306). The dentition of both genera is somewhat similar to that described for the alepisauroids (p. 305), while the general body-form is reminiscent of certain paralepidids. (Fraser-Brunner, 193 1, regarded Luciosudis as a paralepidid, while Mead and Taylor, 1953, have listed their new species Luciosudis harryi under the Paralepididae.) But the combination of certain features (lath-shaped gill- rakers, anal fins with relatively few rays, lateral pectoral fins with 'axial angles' of more than 450, pelvic fins set well forward of the mid-standard length and a well-developed scaling) indicates clearly that the Notosudidae have closer affinities with the Myctophoidea. Similarly there are certain iniomous fishes (not yet described) in the Discovery Collections with no scales and a single row of teeth on the premaxillae, dentaries and palatines — all alepisauroid features — but with an underlying myctophoid character-complex quite like that described for the Notosudidae. But unlike the Notosudidae, one of these new species appears to have a peculiar system of light organs over the flanks. These new fishes must undoubtedly be placed in a new family and if closer examina- tion confirms the presence of light-organs, then the contrast between the Myctophoidea, having three families in which photophores of peculiar structure have arisen, and the Alepisauroidea, in which photophores have not been evolved, will be further emphasized.1 But it is also of interest that in number and arrangement of branchiostegal rays (3-4 + 4-5) the alepisauroids are a more homogeneous suborder than the myctophoids, in which there are three main groups: (1) The Chlorophthalmidae, Ipnopidae, Bathypteroidae, Neoscopelidae and Luciosudis most nearly resemble the alepisauroids in having 3-5+4-8 branchiostegal rays. (2) The Myctophidae have 2 + 5-6 + 0-2 rays (numbers on the epihyal, ceratohyal and hypohyal). (3) The third group, which consists of the Aulopidae, Synodontidae and Harpadontidae, has more numerous branchiostegal rays. (Aulopus filamentosus 8 + 7, Hime japonica 7 + 7, Synodontidae 12-16; Synodus lucioceps 8+10, Saurida undosquamis 6+10; Harpadontidae 17-25 : Harpadon nehereus 9+14. One specimen of each species examined.) Furthermore, the orbitosphenoid bone seems to be consistently absent in the alepisauroids, but may be present or absent in the myctophoids (Regan, 191 1 ; Parr, 1929). However, these aspects of the character-complex of the Myctophoidea will be considered in a later paper. It is reasonable to conclude that the Myctophoidea and the Alepisauroidea are, at least, useful taxonomic groupings. Each can be adequately defined by a rather loose assemblage of characters and by certain morphological trends. But further discussion of the classification in terms of evolution had 1 Apart from the deep-sea Isospondyli, no predacious, bathypelagic fish has an elaborate system of light-organs. 334 DISCOVERY REPORTS best be postponed until we know more of the functional morphology and ecology of these fishes. Perhaps study of the Notosudidae and the new fishes mentioned on p. 333 will throw some light on these interesting but somewhat intractable problems. SUMMARY This report deals with certain aspects of the classification and functional morphology of the alepi- sauroid fishes. These are voracious, bathypelagic forms, which, together with the sub-order Mycto- phoidea form the order Iniomi. Part I contains a revision of the Scopelarchidae, which are shown to fall into two well-marked genera, Scopelarchus Alcock and Neoscopelarchoides Chapman. Two new species, Scopelarchus cavei and Neoscopelarchoides dubius are described and N. elongatus (Norman) is redescribed. The remarkably large larval form, Benthalbella infans Zugmayer is shown to be a young stage of a species of Neoscopel- archoides. The remainder of the first part is taken up with a description of Macroparalepis molestus sp.n. and a review of the family Anotopteridae. While there is no good evidence for regarding Anotopterus as consisting of more than one species, A. pharao Zugmayer, there is some indication that this species may be differentiated into northern and southern forms. Some discussion of the classification of the Iniomi, leading to certain considerations of functional morphology, forms the second part of the report. Alepisauroids are compared and contrasted with myctophoids and possible correlations between the presence and absence of a swim-bladder, the degree of ossification of the skeleton and the setting of the pectoral fins are traced. Adaptations for dealing with large prey and the development of intermuscular bones are also reviewed. In the concluding section, relationships within the Alepisauroidea are considered ; and finally there is some discussion on the status of the order Iniomi. BIBLIOGRAPHY Abe, T., 1952. Records of the ' Mizu-uo-Damahi' (New Japanese name), Anotopterus pharao, and a record of the 'Etchiopia', Brama raii, from near the surface of the North-western Pacific. Jap. J. Ichth. 11 (4/5), pp. 230-8. Alcock, A., 1896. Natural history notes from H.M. Indian Marine Survey Steamer ' Investigator', Series II, No. 23. A supple- mentary list of the marine fishes of India, with descriptions of 2 new genera and 8 new species. J. Asiat. Soc. Beng. lxv (2), no. 3, pp. 301-8. 1899. A descriptive catalogue of the Indian deep-sea fishes in the Indian Museum. . .collected by the Royal Indian Marine Survey Ship 'Investigator'. Calcutta. Beebe, W., 1932. Nineteen new species and four post-larval deep-sea fish. Zoologica, N.Y. xm (4), pp. 47-107. Beebe, W. and Vander Pyl, M., 1944. Eastern Pacific Expeditions of the New York Zoological Society. XXXIII. Pacific Mycto- phidae (Fishes). Zoologica, N.Y. xxix (2), pp. 59-95. Berg, L. S. 1947. Classification of fishes, both recent and fossil. Trav. Inst. zool. Acad. Sci. U.R.S.S. v (2), pp. 1-517. (Trans- lation in Ann Arbor, Michigan.) Breder, C. M. 1926. The locomotion of fishes. Zoologica, N.Y. IV (5), pp. 159-297. Breder, C. M. and Krumholz, L. A., 1943. On the locomotion and feeding behaviour of certain post larval Clupeoidea. Zoologica, N.Y. xxviii (2), pp. 61-7. Brauer, A., 1902. Diagnosen von neuen Tiefseefischen, welche von der Valdivia-Expedition gesammelt sind. Zool. Ariz, xxv (668), pp. 277-98. 1906. Die Tiefsee-fische. I. Systematischer Teil. Wiss. Ergebn. 'Valdivia', xv, pp. 1-420. 1908. Die Tiefsee-fische. II. Anatomischer Teil. Wiss. Ergebn. 'Valdivia', xv (2), pp. 1-266. Castelnau, F. L. de Laparte de, 1873. Contributions to the ichthyology of Australia. VII. Fishes of New Caledonia. Proc. Zool. Acclim. Soc. Victoria, 1872, pp. 110-22. Chapman, W. M., 1939. Eleven new fishes and three new genera of oceanic fishes collected by the International Fisheries Com- mission from the north-eastern Pacific. Proc. U.S. nat. Mus. lxxxvi (3062), pp. 501-42. BIBLIOGRAPHY 335 Clarke, R., 1950. The bathypelagic angler fish Ceratias holbolli Krbyer. 'Discovery' Rep. xxvi, pp. 1-32. Cocco, A., 1838. Su di alcuni Salmonidi del mare di Messina; lettera al Ch.d. Carlo Luciano Bonaparte. Nuovi Annali delle Scienze Naturali, Bologna, 11, pp. 161-94. Collett, R., 1896. Poissons provenant des campagnes du yacht THirondelle' (1885-1888). Result. Camp. sci. Monaco, x, pp. 1-198. Dollo, L., 1909. Les poissons voiliers. Zool. Jb. xxvn (5), pp. 419-38. Ege, V., 1930. The North Atlantic and the Mediterranean species of the genus Paralepis Cuv. A systematical and biological investi- gation. Reprinted from Rep. Danish oceanogr. Exped. Medit. 11, A. 13, pp. 1-201. 1933- On sotne new fishes of the families Sudidae and Stomiatidae. Vidensk. Medd. dansk. naturh. Foren. Kbh. XCiv, pp. 223-36. Fowler, H. W., 1901. Note on the Odontostomidae. Proc. Acad. nat. Sci. Philad. liii, pp. 211-12. Fraser-Brunner, A., 193 i. Some interesting West African fishes and descriptions of a new genus and two new species. Ann. Mag. nat. Hist. Ser. 10, vm, pp. 217-25. Gunther, A., 1864. Catalogue of the fishes in the British Museum, v, pp. 1-455. 1887. Report on the deep-sea fishes collected by H.M.S. ' Challenger' during the years 1873-1876. Rep. Sci. Res. Challenger. Zool. xxii, pp. 1-335. Harris, J. E., 1952. Fin patterns and mode of life of fishes. Essays in marine biology. Oliver and Boyd, Edinburgh, pp. 17-28. Harry, R. R., 195 1. Deep-sea fishes of the Bermuda Oceanographic Expeditions. Family Paralepididae. Zoologica, N.Y. xxxvi (O. PP- !7-35- 1953 a. Studies on the bathypelagic fishes of the family Paralepididae 1. Survey of the Genera. Pacif. Sci. vn(2), pp. 219-49. 1953 b. Studies on the bathypelagic fishes of the family Paralepididae (Order Iniomi). 2. A revision of the North Pacific species. Proc. Acad. nat. Sci. Philad. cv, pp. 169-230. Hora, S. L., 1934. Wanderings of the bombay-duck, Harpodon nehereus {Hamilton Buchanan) in Indian waters. J. Bombay nat. Hist. Soc. xxxvn (3), pp. 640-54. Hubbs, C. L., Mead, G. W. and Wilimovsky, N. J., 1953. The widespread, probably antitropical distribution and the relation- ship of the bathypelagic iniomous fish Anotopterus pharao. Bull. Scripps. Instn. Oceanogr. tech. VI (5), pp. 173-98. Jones, F. R. H. and Marshall, N. B., 1953. The structure and functions of the teleostean swimbladder. Biol. Rev. xxvm, pp. 16-83. Koumans, F. P., 1953. Biological results of the Snellius Expedition. XVI. The Pisces and Leptocardii of the Snellius Expedition. Temminckia, ix, pp. 177-275. Legendre, R., 1934. Lafaune pelagique de V Atlantique au large du Golfe de Gascogne recueillie dans les estomacs de Germons. Premiere Partie: Poissons. Ann. Inst, oceanogr. Monaco, xiv (6), pp. 249-418. Lowe, R. T., 1835. Additional observations on Alepisaurus ferox. Trans, zool. Soc. Lond. I, pp. 395-400. Magnan, A., 1929. Les caracteristiques geometriques et physiques des poissons avec contributions a V etude de leur equilibre statique et dynamique. Ann. Sci. nat. (Zool.) (10), XII, pp. 1— 133. Marshall, N. B., 1950. Air bladder structure and vertical distribution in deep-sea fishes. Ann. Rep. Challenger Soc. in, No. 11, p. 26. 1951- Bathypelagic fishes as sound scatterers in the ocean. J. Mar. Res. x (1), pp. 1-17. Maul, G. E., 1945. Monografia dos peixes do Museu Municipal do Funchal. Familia Sudidae. Bol. Mus. Funchal No. 1(1), PP- 1-37- 1946. Monografia dos peixes do Museu do Funchal. Ordem. Iniomi. Bol. Mus. Funchal No. 2 (2), pp. 5-61. 1952. Additions to previously revised families. Bol. Mus. Funchal No. 6 (16), pp. 51-62. Mead, G. W. and Bohlke, J., 1953. Scopelarchus linguidens, a new bathypelagic fish from off northern Japan. Jap. J. Ichth. 11 (6), pp. 241-45. Mead, G. W. and Taylor, F. H. C, 1953. A collection of oceanic fishes from off north-eastern Japan. J. Fish. Res. Bd. Can. x (8), pp. 560-82. Murray, J. and Hjort, J., 1912. The depths of the ocean. London: Macmillan. Norman, J. R., 1937. Fishes. Rep. B.A.N.Z. Antarct. Res. Exped. 1929-1931, (B) 1 (2), pp. 49-88. Nybelin, O., 1946. Sur quelques especes du genre Anotopterus Zugmayer et sur la position systematique de la famille des Anoto- pteridae. Goteborgs VetenskSamh. Handl. (6) vb (1), pp. 3-16. 1948. Fishes collected by the ' Skagerak' Expedition in the eastern Atlantic 1946. Goteborgs VetenskSamh. Handl. (6) vb (16), pp. 1-95. Owen, R., 1840-5. Odontography ; or a treatise on the comparative anatomy of the teeth; their physiological relations, mode of development, and microscopic structure in the vertebrate animals. Vols. I and II. London : Hippolyte Bailliere. Pappenheim, P., 1914. Die Fische der deutschen Sildpolar -Expedition 1901-1903. II. Die Tiefseefische. Dtsch. Siidpol Exped. 15, Zool. vii, pp. 163-200. 336 DISCOVERY REPORTS Parr, A. E. 1928. Deep sea fishes of the order Iniomi from the waters around the Bahama and Bermuda Islands. Bull. Bingham oceanogr. Coll. in, Art. 3, pp. 1— 193. 1929. A contribution to the osteology and classification of the orders Iniomi and Xenoberyces. Occ. Pap. Bingham oceanogr. Coll. No. 2, pp. 1-45. 193 1 a. Deep-sea fishes from off the western coast of North and Central America. Bull. Bingham, oceanogr. Coll. II, Art. 4, PP- !-53- 19316. On the genera Paralepis and Lestidium and the taxonomic status of some of their species. Copeia, No. 4, pp. 152-8. 1937. Concluding report on fishes. Bull. Bingham oceanogr. Coll. HI, Art. 7, pp. 1-79. Regan, C. T., 191 i. The anatomy and classification of the teleostean fishes of the order Iniomi. Ann. Mag. nat. Hist. Ser. 8, vii, pp. 120-33. 1913. The Antarctic fishes of the Scottish National Antarctic Expedition. Trans, roy. Soc. Edinb. xlix, pp. 229-92. Roule, L., 1916. Notice preliminaire sur quelques especes nouvelles ou rares des poissons provenant des croisieres de S.A.S. le Prince de Monaco. Bull. Inst, oceanogr. Monaco, No. 320, pp. 1-32. 1919- Poissons provenant des campagnes du yacht 'Princesse Alice' et du yacht 'Hirondelle II'. Result. Camp. sci. Monaco, lii, pp. 1-190. 1927. Considerations sur deux especes abyssales de poissons recemment entrees au Musee Oceanographique, et sur la valeur teratologique possible de I'une d'elles. Bull. Inst, oceanogr. Monaco, No. 497, pp. 1-12. 1935- Nouvelles observations sur unpoisson abyssal ( Anotopterus pharao Zugm.). Bull. Soc. zool. Fr. lx (3-4), pp. 324-30. Roule, L. and Angel, F., 1930. Larves et alevins des poissons provenants des croisieres du Prince Albert Ier de Monaco. Result. Camp. sci. Monaco, lxxix, pp. 1-148. Schmidt, J. 1918. Argentinidae, Microstomidae, Opisthoproctidae, Mediterranean Odontostomidae. Rep. Danish oceanogr. Exped. Medit. 11, Biol. A 5, pp. 1-40. Simpson, G. G. 1953. The major features of evolution. New York: Columbia University Press. Skowron, S., 1928. tjber das Leuchten des Tiefseefisches Chauliodus sloanii. Biol. Zbl. xlviii, pp. 680-5. Taylor, H. F., 1921. Air bladder and specific gravity. Bull. U.S. Bur. Fish, xxxviii, pp. 121-6. Waite, E. R., 1916. Australian Antarctic Expedition, ign-igi4. Under the leadership of Sir Douglas Mawson, D.sc. B.E. Fishes. Austral. Antarct. Exped. Sci. Rep., Ser. C, ill (1), pp. 1-92. Weber, Max, 1913. Die Fische der Siboga-Expedition. Siboga Exped. Monog. 57, Livr. lxv, pp. 1-710. Leiden. Zugmayer, E., 191 1. Poissons provenant des campagnes du Yacht ' Princesse- Alice ' (1901-1910). Result. Camp. Sci. Monaco. xxxv, pp. 1-174. PLATE XIX Fig. i. Neoscopelarchoides elongatus (Norman), lateral view x i Fig. 2. Neoscopelarchoides elongatus, dorsal view of head x i Fig. 3. Neoscopelarchoides dubius sp.n., dorsal view of head x 1 Fig. 4. Neoscopelarchoides dubius sp.n., lateral view x 1 Fig. 5. Scopelarchus cavei sp.n., dorsal view of head x 2 Fig. 6. Scopelarchus cavei sp.n., lateral view x i\. DISCOVERY REPORTS, VOL. XXVII PLATE XIX [Discovery Reports. Vol. XXVII, pp. 337-376, August 1955.] EUPHAUSIACEA OF THE BENGUELA CURRENT FIRST SURVEY, R.R.S. 'WILLIAM SCORESBY', MARCH 1950 By BRIAN P. BODEN CONTENTS Acknowledgements ....... Introduction ........ Distribution of the Euphausiacea in the Benguela Current Remarks Developmental phases Nyctiphanes capensis Euphausia lucens . Eitphausia tenera . Nematoscelis megalops Bibliography 339 339 34° 345 346 346 359 365 37° 375 EUPHAUSIACEA OF THE BENGUELA CURRENT FIRST SURVEY, R.R.S. 'WILLIAM SCORESBY', MARCH 1950 By Brian P. Boden (Text-figs. 1-24) ACKNOWLEDGEMENTS TH I s report was started while I was oceanographer on the staff of the Bermuda Biological Station for Research Inc. My thanks go to the succession of directors the station had during this short period for encouraging me to continue with the work at the expense of station duties. Dr T. lohn Hart, National Institute of Oceanography, performed the onerous task of preliminary plankton sorting with his usual skill and patience and sent the euphausiids to me in Bermuda. I am particularly indebted to several of the staff of the Scripps Institution of Oceanography for their critical reading of the manuscript and their numerous suggestions. 25° 25° WS997 WS999^SI001 WS996* « #•••"* -WS 1002 WS998 WSI000 .UNION. OF SOUTI 30° INTRODUCTION The material upon which this report is based is drawn from the collections made by R.R.S. 'William Scoresby ' during her eighth commission and her first survey of the Benguela Current off the coast of South-West Africa. The stations occupied during this cruise are shown in Fig. 1. The collections were made in March 1950. The hauls were vertical and made with an N 70 closing-net of plankton silk of 70 meshes to the inch and with a mouth diameter of 70 cm. Fourteen species were represented at these stations; their occurrence is given on pp. 340-45. An overall picture of distribution based on the results of a single cruise may very well be misleading, and therefore the counts are presented simply in tabular form and efforts at interpretation are postponed until the results of the second survey are available. Full station data may be obtained from the Discovery Reports, Station List, vol. xxvi. Descriptions of the South African species as adults are to be found in Boden (1954). 10° EAST WS978. 20° VVS976« T WS977 WS979 WS980 a i«WS98l WS987 WS986« 4 ••WS989;: WS988 20° Fig. 1 . Plan of stations at which euphausiids were taken during the survey of the Benguela Current by R.R.S. 'William Scoresby', March 1950. 34° DISCOVERY REPORTS DISTRIBUTION OF THE EUPHAUSIACEA IN THE BENGUELA CURRENT Thysanopoda microphthalma G. O. Sars Thysanopoda microphthalma G. O. Sars, 1885, p. 106; T. distinguenda Hansen, 1905a, p. 17. One adult female was taken from between 500 m. and 250 m. at station WS 977. Illig (1930) has previously reported the species from the coast of South-West Africa. Nyctiphanes capensis Hansen Nyctiphanes capensis Hansen, 1911, p. 19; N. australis Stebbing, 1905, p. 113 {nee Sars, 1885); JV. australis Stebbing, 1910, p. 299; N. africana Zimmer, 1912, p. 3. Table 1 . Occurrence of Nyctiphanes capensis Station no. Depth (m.) Adult males Adult females Larvae WS979 50-0 6 12 4071 100-50 4 5 1 1 1 1 WS980 50-0 — 2 5« 100-50 1 3 r73 WS981 50-0 — — 5 WS988 50-0 — — 267 100-50 — 1 452 WS 989 50-0 — — 454 WS997 50-0 — — — 100-50 — — — 250-100 — — — 500-250 — — — 750-500 — — — 1000-750 — — 1 WS998 50-0 — — i57 100-50 — — 24 175-100 — — 7 WS 999 50-0 — 3 952 100-50 — — 19 150-100 — — 5 WS 1000 50-0 3 13 62 100-50 — 2 10 150-100 — — 52 WS 1001 50-0 9 22 651 100-50 — 1 1 17 WS 1002 50-0 — 1 1 The stations at which N. capensis was taken are shelf stations. Like other members of the genus, the species is characteristically an inhabitant of shallow, coastal waters. Good collections have been made previously in the Sturrock Dock in Capetown (Boden, 1954). N. africana Zimmer (19 12) is included here as a synonym. Zimmer's rather vague descriptions and drawings of two specimens taken from the stomach contents of the Cape snoek, Thyrsites atun, almost certainly refer to Nyctiphanes capensis and the fish were taken well within the distribution range of N. capensis. DISTRIBUTION OF THE EUPHAUSIACEA IN THE BENGUELA CURRENT 34i Euphausia lucens Hansen Euphausia splendens Dana sens G. O. Sars, 1885, p. 80; E. lucens Hansen, 19056, p. 9; E. uncinata Colosi, 1917, p. 186. Table 2. Occurrence of Euphausia lucens Station no. Depth (m.) Adult males Adult females Larvae WS976 50-0 — — ■ — 100-50 — — — 250-100 — — — 500-250 — 4 5 750-500 — 5 4 1000-750 — — — WS 977 50-0 — 1 — 100-50 — — — 250-100 — — — 500-250 — — — 750-500 — — — 1000-750 — — — WS 978 50-0 — — 1 100-50 — — — 250-100 — — — 500-250 — 1 — 750-500 — — — WS 979 50-0 — — H 100-50 — — — WS 986 50-0 — — s5 100-50 — — — 250-100 2 6 1 500-250 — 2 1 750-500 — — 1 1000-750 — — 3 WS996 50-0 — — 6S 100-50 — 1 6 250-100 — — 37 500-250 — — 21 750-500 — — 5 1000-750 — — 8 WS 997 50-0 — 3 H 100-50 — — 16 250-100 — 1 2 500-250 — — — 750-500 — — 1 1000-750 — — 10 WS 1000 50-0 36 237 — 100-50 — 5 — 150-100 — — — WS 1 00 1 50-0 2 18 3 100-50 — 14 3 WS 1002 50-0 — 6 37 E. lucens is an oceanic species which flourishes in the Pan-Antarctic (Rustadt, 1930) waters. It is probable that in the waters of the Benguela Current it is approaching the limit of its range (John, 1936; Boden, 1954). 342 DISCOVERY REPORTS Euphausia tenera Hansen Euphausia gracilis G. O. Sars, 1885, p. 89 (nee E. gracilis Dana, 1852, p. 644); E. tenera Hansen, 19056, p. 9. Table 3 . Occurrence of Euphausia tenera Station no. Depth (m.) Adult males Adult females Larvae WS996 50-0 100-50 250-100 500-250 750-500 1000-750 — 2 5 2 6 1 6 This tropical to subtropical species occurred at only one station. Since this station is a southern one, it is probable that many of the unidentified pre-furcilia larvae encountered at other stations are really E. tenera. Euphausia recurva Hansen Euphausia recurva Hansen, 19056, p. 13. Table 4. Occurrence of Euphausia recurva Station no. Depth (m.) Adult males Adult females Larvae WS976 50-0 — — — 100-50 1 — — 250-100 1 2 — 500-250 6 4 4 750-500 7 7 — 1000-750 — — — WS977 50-0 4 5 100-50 1 — — 250-100 — — — 500-250 — — — 75°-500 — — — 1000-750 — — — WS996 50-0 — — — 100-50 1 — — 250-100 1 2 — 500-250 7 9 — 750-500 — — — 1000-750 — — — WS 997 50-0 7 13 — 100-50 2 2 — 250-100 3 2 — 500-250 — — — 750-500 — — — 1000-750 — — — Evidence of diurnal vertical migration may be discerned in these data. At the day stations WS 976 and WS 996 the animals are absent from the surface layers while at the night stations WS 977 and WS 997 they are concentrated in the upper layers and absent from the lower layers. DISTRIBUTION OF THE EUPHAUSIACEA IN THE BENGUELA CURRENT 343 Euphausia hanseni Zimmer Euphausia hanseni Zimmer, 191 5, p. 180. Four adult females of this species were taken, one from station WS 978 and three from station WS 987. All were taken between 250 m. and 100 m. Euphausia similis var. armata Hansen Euphausia similis var. armata Hansen, 191 1, p. 24; E. similis var. lobata Zimmer, 1914, p. 425. This species occurred at station WS 997. One adult male was taken between 100 m. and 50 m. and one female between 1000 m. and 750 m. Nematoscelis megalops G. O. Sars Nematoscelis megalops G. O. Sars, 1885, p. 127. Table 5. Occurrence of Nematoscelis megalops Station no. Depth (m.) Adult males Adult females Larvae WS976 50-0 — — 10 100-50 — — 2 250-100 — — 2 500-250 — — 750-500 — 1 4 1000-750 — — ■ WS 977 50-0 1 4 108 100-50 — — — 250-100 2 2 — 500-250 — — 7 750-500 — — 1000-750 — — WS978 50-0 — — — 100-50 — — 250-100 — — 37 500-250 — — 350 500-350 — — 44 750-500 — — WS986 50-0 — — 37 100-50 — — 37 250-100 — — 135 500-250 — — 23 750-500 — — — 1000-750 — — 2 WS997 50-0 — — 1 100-50 — — — 250-100 — — 9 500-250 — — 750-500 — — 1000-750 — — 2 Interesting indications of diurnal vertical migration are evident in these data. Station WS 977 is a night station (21.03-03.45 hr.) and the larvae are concentrated in the upper 50 m. At station WS 978 (10.00-13.04 hr.) a premature closing of the net at 350 m. and subsequent repetition of the 500-250-m. haul reveals a concentration of larvae between 250 m. and 350 m. This, in conjunction 344 DISCOVERY REPORTS with the layering between 250 m. and 100 m. at station WS 986 (08.05-14.07 hr.), indicates that the animals concentrate around 250 m. in the daytime. Evidence of the vertical migration of a very closely allied species, N. difficilis, and its possible relation to sonic scattering layers in the Pacific has been presented previously (Boden, 1950a). Thysanoessa gregaria G. O. Sars Thysanoessa gregaria G. O. Sars, 1883, p. 26. Table 6. Occurrence of Thysanoessa gregaria Station no. Depth (m.) Adult males Adult females Larvae WS976 50-0 — — 4 100-50 — — — 250-100 — — — 500-250 — — — 750-500 — — 1 1000-750 — — — WS996 50-0 — — 3 100-50 — — — 250-100 — — — 500-250 — — — 750-500 — — — 1000-750 1 — — WS 997 50-0 — — 1 100-50 — — 2 250-100 1 — 2 500-250 — — — 750-500 — — — 1000-750 — — — This species is usually quite common in this area (Boden, 1954). Its worldwide distribution is of an antitropical nature (Hubbs, 1950, 1952), although Tattersall (1926) claims that it is common in the deeper waters in the tropical oceanic area. This leads to an attractive speculation that the apparent tropical discontinuity of distribution may be associated with a submergence along temperate iso- therms. Actually Tattersall's and Moore's (1952) records are not below 280 N. Brinton (Scripps Inst, of Oceanography — personal communication) finds that the species actually does seek deeper waters toward the tropical limits of its range. The limited number of deep hauls which he has made in the equatorial regions of the Pacific, however, have not yielded any specimens of T. gregaria and he is of the opinion that it is a true antitropical species. Nematobrachion boopis Caiman Nematodactylus boopis Caiman, 1896, p. 17; Nematobrachion boopis Caiman, 1905, p. 153. Two adult females were captured between 750 m. and 500 m. at station WS 976. Stylocheiron maximum Hansen Stylocheiron maximum Hansen, 1908a, p. 92. Two larvae of this species were taken, one at the 250-100-m. level at station WS 976 and the other in the upper 50 m. at station WS 986. At station WS 997 an adult female was taken between 100 m. and 50 m. and another between 500 m. and 250 m. DISTRIBUTION OF THE EUPHAUSIACEA IN THE BENGUELA CURRENT 345 Stylocheiron elongatum G. O. Sars Stylocheiron elongatum G. O. Sars, 1883, p. 32. Eleven larvae were encountered between 1000 m. and 500 m. at station WS 976. One was taken in the upper 50 m. at station WS 997. Stylocheiron longicorne G. O. Sars Stylocheiron longicorne G. O. Sars, 1883, p. 32; S. mastigophorum (partim) Chun, 1887, p 30. One larva was taken between 500 m. and 250 m. at station WS 997. Stylocheiron afhne Hansen Stylocheiron affine Hansen, 1910, p. 118. Five larvae were taken in the upper 250 m. at station WS 976. A small number of unidentifiable larvae occurred at stations WS 976, 977, 978, 988, 997, 998, 999 and eggs were present in the upper layers at stations WS 976, 977, 978, 986, 996, 997, 998 and 999. REMARKS Physical oceanographic conditions in the Benguela Current (Currie, 1953) are somewhat similar to those found in the California Current, and a very close similarity in phytoplankton populations and successions in the two currents has been stressed by Hart (1953). In general, the distribution of comparable species of euphausiids substantiates this similarity in the two regions. Nyctiphanes capensis occupies the same inshore niche in South African waters that N. simplex does in Californian waters. Euphausia lucens, reaching the northern limit of its range in the Benguela, may be compared with E. pacifica, which reaches its southern limit in the California Current. The distribution of Nyctiphanes megalops in this survey is very similar to that of N. difficilis in California, and that of Euphausia recurva and Thysanoessa gregaria is the same in both regions. Curiously enough, species with comparable distributions in the two currents show remarkable morphological similarities. The appearance of Nyctiphanes capensis closely resembles that of N. sim- plex. This is true of all four species of Nyctiphanes. These species all occur in areas of upwelling where conditions are roughly the same. Euphausia lucens and E. pacifica are undoubtedly the most closely morphologically allied species in this large genus. Even the copulatory organs of the males show many points of similarity. Nematoscelis difficilis and N. megalops are so closely similar that the females are inseparable. On these grounds Einarsson (1942) cancelled N. difficilis. The males are separable on small but consistent differences in the copulatory organs, however, and the species has, therefore, been revived (Boden, 1954)- The distribution of euphausiids in the Pacific is now fairly well understood (Brinton, unpublished) and it may become possible to draw further analogies when the results of the second survey of the Benguela Current have been evaluated. 346 DISCOVERY REPORTS DEVELOPMENTAL PHASES It was possible to glean some information on the developmental phases of four of the species encountered. This is presented below. Nyctiphanes capensis Hansen Egg. Size range 0-34-0-38 mm. in formalin. The eggs are carried in two pear-shaped sacs attached by their anterior broad sides to the endopodites of the sixth and seventh cormopods, and by their narrow posterior ends to the exopodites of the eighth cormopods (precisely as in N. simplex). Only one female with intact egg-sacs was found in the collection (Fig. 2). There were fourteen eggs in one sac and sixteen in the other. The perivitelline space of the eggs is small (Fig. 3 a). Fig. 2. Nyctiphanes capensis. Ovigerous female. Nauplius. The nauplius stage (Fig. 36, c) is passed within the egg-sac. The specimens shown in the figure were removed from the unruptured egg-sac of the female. The length of the two specimens examined was 0-42 mm. each. The larva is simple and egg-shaped. It has two small processes at the posterior end, each flanked by one smaller process. The first antenna is uniramous with three terminal setae. The second antenna is biramous, the rami with four and five terminal setae. The mandible is bud-like and naked. Pseudometanauplius. Only one specimen of this stage was found (Fig. 3 d, e). It was removed from the ruptured egg-sac of a female. It is probable that the rest had already hatched out. Length of specimen 0-48 mm. The carapace is oval in dorsal view. The first antenna is unchanged. The second antenna is the same as in the nauplius stage. A median eye has appeared. The anterior lip and buds of three metanaupliar limbs have appeared. There is now a short tail with one pair of postero-lateral and two pairs of terminal spines. The carapace, mandibles, anterior lip, limb-buds and tail are covered by a thin skin. DEVELOPMENTAL PHASES 347 Metanauplhis. No specimens in this stage were encountered. The metanaupliar stage is very transitory and the larva begins to eat only in the next stage. It is probably almost identical with the metanauplii of other species of this genus (cf. Lebour, 1925; Boden, 195 1). First calyptopis. Description based on examination of six specimens (Fig. ^a-c). Length range 0-98-1-10 mm.; average length 1-03 mm. Carapace length 0.60-0-63 mm.; average length 0-62 mm. The carapace is a long, high, narrow dome covering the entire anterior part of the body, and it has no projections of any sort along its inferior margin. OSmtn 0-5 mm H Fig. 3. Nyctiphanes capensis. a, egg; b, nauplius, lateral view; c, nauplius, dorsal view; d, pseudo- metanauplius, lateral view ; e, pseudometanauplius, ventral view. The single median eye is visible with the ocular plates on either side of it bearing the rudimentary ocular luminous organ. The first antennae project forward and downward and are to all intents uniramous. The rudiments of the inner ramus can be seen, however, bearing three setae. The outer ramus carries five terminal setae. The setae are naked. The second antennae project outward and downward and are carried on a peduncle having two segments. The anterior and longer ramus of the antennae is unsegmented and is armed terminally 348 DISCOVERY REPORTS with four strong setose spines, and laterally with two spines. The second ramus is segmented distally and carries six terminal setose spines. The mandible is naked and provided with strong biting jaws. The first and second maxillae and first thoracic limb are present and very strongly resemble the same appendages in N. australis as figured by Sars (1885) and in N. couchii as figured by Lebour 0-25 mm Fig. 4. Nyctiphanes capensis. a, first calyptopis, lateral view ; b, telson of same ; c, first calyptopis, dorsal view. (1924). They are directed forward, are setose, and are used in feeding. The larva first begins eating in this stage. The thoracic segments without limbs are distinct as narrow encircling wrinkles in the integument. The abdomen is smooth and unsegmented. The telson has six, serrate, terminal spines decreasing in size from the outer to the inner, three pairs of serrate, postero-lateral spines increasing in size from the outer to the inner, and one pair of small ventro-lateral spines. Second calyptopis. Six specimens examined (Fig. $a-c). Length range 1-40-1-52 mm.; average length 1-46 mm. Carapace length ±0-85 mm. DEVELOPMENTAL PHASES 349 The carapace is a little longer than half the total length, and is smooth along the inferior margin (cf. N. austral is, smooth; A7, simplex, smooth; and N. couchii, denticles). The median eye is lost and the compound eyes are now well defined but not pigmented. The ocular luminous organ is visible. There are three segments to the peduncle of the first antenna, which is biramous, the outer ramus carrying four terminal spines and the inner, three. 0'5mm 0-25 mm Fig. 5. Nyctiphanes capensis. a, second calyptopis, lateral view; b, telson of same; c, second calyptopis, dorsal view. The remaining thoracic appendages are larger than in the preceding stage but are otherwise un- changed. Occasionally the rudiments of the second thoracic limb may be discerned as a small naked process. The segmentation of the thorax is clearly visible. The abdomen now has five segments and a telson. The developing uropods can be seen within the integument of the telson. The telson retains the three pairs of postero-lateral and one pair of ventro- lateral spines of the previous stage, but now has seven terminal, serrated spines. 35° DISCOVERY REPORTS Third calyptopis. Six specimens examined (Fig. 6a-c). Length range 179-2-22 mm.; average length 1-90 mm. Carapace length ±0-97 mm. The carapace is a little longer than half the total length of the larva. At this stage a small but dis- tinct tooth appears on the inferior margin of the carapace toward its posterior end. The carapace covers the compound eyes which are now globular and distinctly pigmented. 0-25mm 0-5mm Fig. 6. Nyctiphanes capensis. a, third calyptopis, lateral view ; b, telson of same ; c, third calyptopis, dorsal view. The outer, distal margin of the basal segment of the first antennal peduncle is produced as a strong, toothed spine which reaches to about the distal end of the third and last segment of the peduncle. The two rami of the first antennae remain unaltered. The second antennae and the other thoracic appendages are much the same as in the preceding stage. DEVELOPMENTAL PHASES 3Si The abdomen now has six segments and a telson. There are no abdominal appendages. The two rami of the lateral uropods are now visible, the outer ramus being somewhat longer than the inner and having its outer margin produced as a strong spine. The remaining armature of the telson is unchanged. Dominance in the furcilia stages For reasons which have been adequately discussed before (Fraser, 1936; John, 1936; Einarsson, 1945; Boden, 19506, 1951) the term ' cyrtopia ' for the later larval stages has been discarded, and these stages are now included in the furcilia series. Thus the furcilia stages in this report include all the instars, from the one in which the carapace first withdraws from over the eyes and there are no pleopods at all, to that in which there are five setose pleopods, the second antennal endopod is segmented and there is a single terminal and three pairs of postero-lateral spines on the telson. After one pair of postero-lateral spines has been lost the animal is regarded as 'juvenile ' or post-larval, until it reaches adult size. This is done because the ecdyses in the juvenile condition (apparently 9 or 10) are orderly and there is no evidence of dominance in any of the stages. In addition the animal may be sexually mature before attaining fully the adult form. There are twenty-five possible instars in this series. Six are numerically dominant and are regarded as the actual furcilia stages (Fig. 7). This scheme has been followed by most modern workers since it was first suggested by Fraser (1936). Recently, however, Sheard (1953) has proposed an alternative scheme, based mainly on an exhaustive investigation of Nyctiphanes australis, in which the number of furcilia stages is reduced to three and the dominants are disregarded. The stages he proposes are as follows: Furcilia stage 1 (F I). Eyes developed and free of the carapace, pleopods absent or present as non-setose rudiments. Terminal spines (of telson) generally seven. Lateral spines (postero-lateral of telson) three in number, the central spine always present. Furcilia stage 2 (F II). Some, or all, pleopods setose. Terminal spines sometimes reduced in number, the central spine always present. The pair of long lateral spines unaltered at the base. Furcilia stage 3 (F III). All pleopods setose and functional. The pair of long lateral spines are altered at the base. The terminal spines are progressively reduced to one (the median spine). The stage terminates with the end of the phase. The outer limits of Sheard's series are actually the same as those proposed herein for N. capensis, or those of N. simplex (Boden, 1951), and his F I and F II stages are the same as the first and second furcilia stages presented here. His F III stage, however, embraces the final four furcilia stages of N. capensis. Sheard's F II and Fill stages are differentiated mainly on the alteration in the shape of the base of the long lateral spines on the telson. This is a somewhat nebulous character, frequently rather difficult to detect, and is not as reliable an indicator as a definite reduction in the numbers of terminal spines on the telson or the sudden change in both form and function of the second antennal endopod. Sheard 'lumps' the final furcilia stages into his F III stage and throws considerable doubt on the validity of regarding the dominant instars as actual stages and the remainder as variants, on the grounds that the dominants will change from time to time and from place to place. This concept of the dominance of certain instars he regards as a result of the inadequacy of plankton sampling. His arguments are based almost exclusively on the data he obtained for N. australis and are applied in an inferential manner to the other species whose development he describes. It is perfectly true that most plankton sampling is pitifully inadequate, but this cannot be said of Fraser 's (1936) work, in which the concept of the dominants constituting the actual stages was origi- nally expressed. His material was spread adequately through time and space and a very large number of individuals was examined. It is also untrue of Einarsson's (1945) work and most of John's (1936). 352 DISCOVERY REPORTS While the dominants in Nyctiphanes may differ (i.e. the modal position may shift) from time to time or place to place they do not appear to move out of the stage boundaries. For example, in N. capensis (Fig. 7) it is possible that Fill may at times be expressed as a form having seven terminal spines on the telson instead of five (those with six are abnormalities), but the mode will not be absent altogether. Thus, while a false impression may be registered by dogmatically regarding F III as typified by a form having five terminal spines, this impression will probably be less false than that obtained by disregarding all the modes occurring after F II, and considering all the lateral individuals as belonging to one stage. Sheard states that it is highly probable that any individual will pass through only a limited number of the many possible instars in the stages he proposes. This is undoubtedly true, and it seems only logical that the modes represent the limited number of instars through which the majority of individuals passes and that they can therefore be expressed as stages, and the remainder as variants. The fact that in Nyctiphanes the modes will shift and that by lumping several series of furcilia taken at different times or places one can hide these modes can, perhaps, be explained without doubting that the dominants may be regarded as valid stages. The genus Nyctiphanes is peculiar among the Euphausiacea in that all four known species are shallow-water, neritic forms with limited geographical ranges. N. simplex occurs in Californian waters, N. couchii in the North Atlantic, N. australis in southern Australian and northern New Zealand waters, and N. capensis in South African waters. All these species have a great number of variants in their furcilia series and, therefore, substantiate Frost's (1935) claim that species inhabiting coastal waters regularly exhibit greater larval variation than oceanic species. As Sheard points out, the considerable variation exhibited by Euphansia superba, generally regarded as an oceanic species, may not controvert this claim, since it inhabits the region of pack ice, which simulates neritic conditions in some respects. The marked reduction in the number of instars in oceanic species has led to some speculation that the Euphausiacea represent arthropods in which gradual development is giving way to metamorphosis. On this premise, the genus Nyctiphanes with its great larval variation could be regarded as somewhat more primitive than other genera. It is probable, however, that the large number of variants is not a primitive, but a secondary feature developed during invasion of coastal waters. The genus has several rather specialized features, including an egg-sac in which the first, more vulnerable, larval stages are passed. This, too, is probably a secondary adaptation to the more varied coastal conditions. Thus the genus, in which variation is most marked and the dominant stages are least ' fixed ', is one which is specially adapted to a variety of coastal conditions. The fact that dominant instars do occur in any particular place or time can, perhaps, represent the basis for an argument that these instars are an expression of development under the particular environmental conditions obtaining. In other words, such a genus may have slightly different developmental stages under different environmental conditions. I agree with Sheard that much experimental work with live animals is desirable before the question can be examined much further, but feel that under the present circumstances it is less useful to ' lump ' the furcilia stages than to 'split' them. It is possible, unless one is as careful as Sheard has been in giving full codes and counts for all the variants, that much valuable information on development as an expression of environment may be obscured by disregarding dominance. Full data on geographical position, temperature, salinity, oxygen and phosphate content for the stations at which the species reported on herein were taken are available in the Discovery Report, vol. xxvi, pp. 21 1-58. DEVELOPMENTAL PHASES 353 Furcilia stages Twenty-five furcilia instars were encountered. If the same terminology is adopted for all four species of Nyctiphanes as for N. simplex (Boden, 1951), the total number of furcilia types in each species is as follows: N. couchii 15, N. simplex 16, N. australis 24, N. capensis 25. For N. simplex it was sus- pected that three other types may occur. Frequency distributions of the N. capensis instars are given in Fig. 7. In the case of N. simplex and N. capensis six of these types were dominant and are regarded as stages or successive instars, while the remainder are considered variants. According to Gurney (1942) it seems highly probable that there are six stages in N. couchii too. While Sheard recognizes only three stages in N. australis, it is possible to interpret his data differently. Table 7 shows the course of development of three species of Nyctiphanes and it can be seen that they are fairly similar. Of very great interest is the fact that two of the species do not conform to the rule first shown by Fraser (1936) that the pleopods first appear as non-setose rudiments and these rudiments become setose at the next moult. N. capensis and N. simplex develop setose pleopods directly in stage II on somites which bore no non-setose rudiments in stage I. The six stages of N. capensis encountered are as follows : First furcilia. Pleopods non-setose. Second antennal endopod simple. Second furcilia. Pleopods mixed. Second antennal endopod simple. Third furcilia. Pleopods setose. Five terminal spines on telson. Second antennal endopod simple. Fourth furcilia. Pleopods setose. Three terminal spines on telson. Second antennal endopod simple. Fifth furcilia. Pleopods setose. Three terminal spines on telson. Second antennal endopod segmented. Sixth furcilia. Pleopods setose. One terminal spine on telson. Second antennal endopod segmented. First furcilia. Twenty-two specimens examined. Twelve measured (Fig. 8a-g). Length range 2-2-2-8 mm. ; average length 2-5 mm. The following types of furcilia are included in this stage. Those with (a) no pleopods, (b) one pair, (c) two pairs, (d) three pairs, (e) four pairs of non-setose pleopods. No specimens with five pairs of non-setose pleopods were encountered. The occurrence of the above types was as follows: (a) 199, (b) 35, (c) 34, (d) 18, (e) 6. There is no rostrum. The frontal plate is broader than long with a small, distinct, emargination of the anterior border. There is a small, but definite, denticle on the inferior margin of the carapace toward the posterior end. The eyes are heavily pigmented and project beyond the edge of the carapace. The first antennal peduncle has three segments and the basal segment bears a strong spine, reaching just beyond the middle of the third segment, on its anterior outer margin. The rudiments of the two flagella are present on the third segment. The second antenna is biramous, natatory, and its endopod is simple. The usual condition of the thoracic appendages is that the first and second maxillae are clearly visible, the first cormopod is present and bifurcate, and the bud of the second cormopod is free but naked. The remaining cormopods are contained in the integument and difficult to discern. Individuals in which the pleopods are developed frequently have the second cormopod present, bifurcate and with bilobed gill rudiments. 354 PLEOPOO DEVELOPMENT TERMINAL SPINES ON TELSON :OVERY REPORTS FREQUENCY OF OCCURRENCE STAGE NUMBERS II II II 7 7 7 7 1 1 1 1 1 1 1 \ 1 1 1 1 1 1 \\ 1 1 1 1 1 1 \\\ 1 1 1 1 1 1 WW 1 1 1 1 1 1 7 7 7 7 7 7 7 7 1 1 1 ■ w 1 1 1 1 1 1 \w 1 1 1 1 1 1 o LJ w\ Z 1 1 1 1 1 1 O WW CO 1 1 1 1 1 1 O WW O 1 1 1 1 1 1 o o WW 111 1 1 1 1 II _l < z w\w £ 1 1 1 1 1 1 £ < ww\ 1 1 1 1 1 1 CVJ s\w 5 1 1 1 1 1 II ww\ 1 1 1 II 1 7 6 5 1 \\\ w 1 1 1 1 1 1 wwv 1 1 1 1 1 1 ■ w\w 1 1 1 1 1 1 4 3 2 ■ 1 wvw 1 1 1 1 1 1 4 wwv 1 1 1 1 1 1 \ w w o 1 1 1 1 1 1 5 3 1 Q- O wvw 5 Q 5 ^ 1 1 1 1 1 1 5 " 1- _l z \ww i S 1 1 1 1 1 1 2 1 1 Z CO 51 %s. Sars, G. O., 1883. Preliminary notices on the Schizopoda of H. M.S. 'Challenger' Expedition. Forh. VidenskSelsk., Oslo, 7, pp. 1-43. 1885. Report on the Schizopoda collected by H.M.S. 'Challenger' during the years 1873-1S76. Rep. Sci. Res. Voy. H.M.S. 'Challenger'. Zool. 13, pp. 1-228, 38 pis., 4 figs. Sheard, K., 1953. Taxonomy, distribution and development of the Euphausiacea (Crustacea). B.A.N.Z.A.R.E. Rept. (Series B), 8(i),pp. 1-72. Stebbing, T. R. R., 1905. General catalogue of South African Crustacea III. Marine Invest. South Africa, 4 (2), pp. 21- 213, 9 pis. 1910. General catalogue of South African Crustacea V. Ann. S. Afr. Mus. 6 (6), pp. 281-599, 7 pis. Tattersall, W. M., 1926. Crustacea of the Order Euphausiacea and Mysidacea from the Western Atlantic. Proc. U.S. nat. Mus. 69(8), pp. 1-31. Zimmer, C, 1912. Siidwestafrikanische Schizopoden. Schultze, Forschungsreise in Siidafrika 5. Denkschr. med.-naturw. Ges. Jena, 17, pp. 1-11. 1914. Die Schizopoden der Deutschen Sudpolar Expedition, igoi-j. Deutsch. Siidpol. -Exped. 15, pp. 377-445, 3 pis. 1915. Schizopoden des Hamburger Naturhistorischen (Zoologischen) Museums. Mitt, naturh. Mus., Hamburg, 32, pp. 159-82, 4 figs. [Discovery Reports. Vol. XXVII, pp. 377-395, Plates XX, XXI, September 1955.] CESTODES OF WHALES AND DOLPHINS FROM THE DISCOVERY COLLECTIONS By S. MARKOWSKI CONTENTS Material and methods 379 Abbreviations used in the figures 379 General discussion 380 List of cestodes examined 380 Summary of discussion 382 Synonymy 382 Systematic notes 382 Tetrabothrius affinis .......... 382 T. wilsoni ............ 384 T. ruudi 385 T. schaeferi sp.n. ........... 386 Trigonocotyle globicephalae 388 Tr. prudhoei sp.n. ........... 389 Priapocephalus grandis 390 Diplogonoporas balaenoptcrae ......... 392 Phyllobothrium larva 393 Tetraphyllidean larva .......... 394 Relationships between the host and the parasite 394 References 394 Plates XX, XXI following page 395 CESTODES OF WHALES AND DOLPHINS FROM THE DISCOVERY COLLECTIONS By S. Markowski (Plates XX, XXI and Text-figs. 1-45) MATERIAL AND METHODS HE material for this study was collected mainly at whaling stations and on board factory ships by T various members of the staff of the Discovery Investigations.1 It consists of eighty-eight samples taken from the intestines and blubber of forty-six Blue Whales (Balaenoptera musculus), eight Fin Whales (B. physalus), thirteen Sei Whales (B. borealis), fourteen Sperm Whales (Physeter catodon) and seven dolphins, namely: two Globicephala edwardi, two Lagenorhynchus obscurus, one L. australis and two Steno bredanensis. Of these, four samples of Tetrabothrins affinis from Blue Whales belong to the collections of the British Museum (Natural History). The material comprises intestinal forms of cestodes, as well as several examples of larval stages en- cysted in the blubber. The specimens are on the whole in very good condition, except for a few cases where slight maceration has occurred. In some samples the cestodes have been fixed and preserved attached to portions of the gut. The bulk of the collection was preserved in 4% formalin, although some of the worms were fixed in Bouin's solution. About 600 slides of serial sections and whole mounts have been made, stained in Ehrlichs' haema- toxylin and counter-stained with erythrosin. The whole mounts, stained in Mayer's paracarmine were cleared in benzyl-alcohol with very good results, even for the larger portions of the strobila. The serial sections have been cut at 10-15// thick. The hosts were caught in South Africa (Saldanha Bay, Durban, Cape Town and Simon's Town), the Azores, the Falkland Islands, South Georgia, the South Shetlands and at various positions in the Southern Ocean. The localities are given with the description of each species. I have the greatest pleasure in expressing my gratitude to The Royal Society for advancing me a grant from the Browne Research Fund which enabled me to complete this investigation. I also take this opportunity of expressing my thanks to Dr N. A. Mackintosh, C.B.E., of the National Institute of Oceanography and to Dr H. E. Bargmann of the same Institute for their kind assistance in the course of this work, and to Mr S. Prudhoe of the British Museum (Natural History). My thanks are also extended to Dr Gwendolen Rees who kindly lent whole mounts of Tetrabothrius from whales. ABBREVIATIONS USED IN THE FIGURES c.s., cirrus-sac. u., uterus. e., excretory system. u.o., uterine opening. l.m., longitudinal muscles. v., vagina. o.v., ovary. v.d., vas deferens. t., testis. v.g., vitelline gland. 1 Now incorporated with the National Institute of Oceanography. 38o DISCOVERY REPORTS GENERAL DISCUSSION The cestodes collected represent ten different species belonging to six genera. Of these species, two appear to be new to science and two represent larval stages, the specific identification of which does not seem possible at present, although it can be seen that they belong to two different genera. The list given below shows the names of the cestodes examined and Table 2 (p. 394) their occurrence in particular hosts. LIST OF CESTODES EXAMINED 1. Tetrabothrius affinis (Loennberg, 1891). 2. T. wilsoni (Leiper & Atkinson, 1914). 3. T. ruudi Nybelin, 1928. 4. T. schaeferi sp.n. 5. Trigonocotyle globicephalae Baer, 1954. 6. Tr. prudhoei sp.n. 7. Priapocephalus grandis Nybelin, 1922. 8. Diplogonoporus balaenopterae Loennberg, 1892. 9. Phyllobothrium larva. 10. Tetraphyllidean larva. With the exception of the two larval forms and Diplogonoporus balaenopterae, the rest of the above mentioned cestodes belong to the family Tetrabothriidae, which appears to contain the bulk of the cestodes found in Cetacean hosts. Because of their great similarity the identification of certain species presents certain difficulties. The scolex in the majority of these cestodes bears four suckers, which are variable in shape and size. These suckers are usually provided with modifications called ' epaulettes ' or with fleshy protuberances such as occur in the Genus Trigonocotyle. The scolex is subject to contraction and consequently its size and shape give no help in specific determination. In the genus Priapocephalus, however, this organ is acorn-like and easy to distinguish, though some variations in its contour do occur, as was mentioned by Nybelin (1928) for P. minor. The strobila in the Tetrabothriids is either cylindrical and worm-like, or flattened, as in the other groups of cestodes. The cylindrical strobila appears to be typical of Tetrabothrius affinis, though in some cases, a short portion showing flattened segments occurs in the hind part of the body (PI. XX, fig. 1). In Trigonocotyle globicephalae a short anterior portion of the body is cylindrical, while the remainder of the strobila is flattened. Modifications of the anterior part of the body have also been noticed in the only example of the Pseudophyllidea, present in this collection, namely Diplogonoporus balaenopterae, where that portion of the strobila is markedly undulate dorso-ventrally (PI. XXI, fig. 6). It is difficult to prove at the moment whether these modifications represent natural features or are caused by fixation. The genital organs in Tetrabothriids are also very similar and uniform in structure. Although there is much variation in the number of the testes in immature proglottids and their size and arrangement in the medullary parenchyma is difficult to determine, in the fully mature segments where egg pro- duction has not yet begun, it is possible to ascertain their number and arrangement. In some forms, as in Trigonocotyle globicephalae , the testes, when examined in the cylindrical part of the body, are distributed in many layers. In the flattened portion of the body, however, they are arranged in a single layer. CESTODES EXAMINED 381 In the species of Tetrabothrius examined during the present work, the testes are situated dorsally, except in T. affinis, where they are distributed dorso-ventrally. This arrangement, however, may be altered and affected by the growth of the uterus, particularly when filled with eggs. The testes seem also to decrease rapidly in number in the gravid segments, being either partially spent or else crowded nearer to the surface of the medullary parenchyma by the uterus. Some differences in the size of the testes in mature segments (measured in transverse sections of the body) are notice- able, and this may be regarded as a valuable feature in specific differentiation (Table 1, p. 387). The cirrus-sac, although similar in structure, is also subject to much contraction and its size and shape depend greatly on this factor. In some cases, as in Trigonocotyle globicephalae, and 7V. prudhoei sp.n., some muscular modification occurs in the area of the genital atrium, forming a kind of sucker. This, however, is also met with in other species of Tetrabothriids not found in whales. In cetacean Tetrabothriids, the ovary and vitelline glands are situated in the ventral half of the segment. In some of the species examined, the ovary contains fairly large egg-cells. The full develop- ment of the ovary is reached in the portion of the strobila occupied by the mature proglottids. From this point onwards towards the posterior end, the ovary gradually undergoes a process of deterioration, caused by the overgrowth of the uterus. In fully gravid segments the uterus is so strongly developed that it practically replaces the rest of the genital system. It is a centrally situated, sac-shaped organ extending laterally and gradually occupying the whole of the medullary parenchyma. There is no proper uterine pore in cetacean Tetrabothriids. Instead, the uterus opens through the body-wall, when the eggs are ready to be discharged from the segment. The rupture through which the eggs are passed appears on the dorsal surface of the body, at one definite point, and the path in the cortical parenchyma, along which the distending uterus is pushed, is surrounded by a distinct mass of eosinophil tissue. In Priapocephalus grandis, however, many such ruptures on the dorsal surface have been observed. No uterine openings or ducts have been found in the two species of Trigonocotyle from the present collection. It seems that a kind of ' cement M protects the newly formed uterine opening from the effects of the gastric juices in the host's intestine and which closes the opening after the eggs have been discharged. It is well known that the eggs of cestodes lose much of their characteristic structure after fixation and extraction from the uterus. Nevertheless, specific differences seem to occur between the eggs of particular species found in the Discovery Collection. Their size appears to be related to the thickness of the outer and inner membranes in the various species, and in some cases to the number of membranes as well. The embryos also seem to vary in size with species, and, with the exception of Tetrabothrius ruudi (Fig. 15), the three pairs of hooks with which they are all provided exhibit specific differences (Table 1). The vagina runs ventrally to the cirrus-sac in all examined species of Tetrabothrius. The most striking differences among particular species seem to occur in the structure and arrange- ment of longitudinal muscles, as seen in transverse section. In the present material, the longitudinal muscles of Tetrabothriids undergo morphological changes within the individual strobila itself. These changes are associated with the development of the genital organs and it appears that in the immature portion of the strobila, where the genital organs are either non-existent or rudimentary, the longitudinal muscles are differentiated into bundles, but in the mature portion of the strobila, these bundles have taken on a very definite arrangement. As the mature proglot- tids ripen, and the uterus becomes increasingly distended, the muscle bundles tend to atrophy, although their characteristic arrangement can still be distinguished. This series of changes is shown in sections of Trigonocotyle globicephalae (Figs. 24, 25) and Priapocephalus grandis (Figs. 34-36). 1 This 'cement' is probably produced by gland-cells in the modified tissue. 382 DISCOVERY REPORTS SUMMARY OF DISCUSSION To sum up : Specific differences are to be found in the size and the arrangement of testes, in the size and structure of eggs, embryos and embryonic hooks, and in the arrangement and shape of the longitudinal muscles in the fully mature, but not gravid segments. SYNONYMY The history and the nomenclature of some of the Tetrabothriids occurring in Cetacea have been given by Rees (1953). Recently, Baer (1954) has published a list of synonymous names of Tetrabothriids and their specific classification. From the present work it seems, however, that Tetrabothrius zvilsoni, which he considers to be synonymous with T. affinis, actually represents a distinct species (Table 1, p. 387). Baer's (1954) criticism of Baylis's (1926) description of T. affinis appears to be justified, as the material examined by Baylis seems to represent another species, described in this paper as T. schaeferi sp.n. The form mentioned by Rees (1953) as Tetrabothrius sp. also appears to represent the same species. The two new species described in this paper, namely T. schaeferi and Trigonocotyle prudhoei have been named after Miss F. H. A. Schaefer who, during my recent illness nursed me in hospital, and after my friend Mr S. Prudhoe of the British Museum (Natural History). SYSTEMATIC NOTES Genus Tetrabothrius Rudolphi, 1819 This genus is represented in the present material by four species. Of these, one seems to be new. Tetrabothrius affinis (Loennberg, 1892). (Figs. 1-6, PI. XX, figs. 1-2.) Diplobothrium affine Loennberg, in Jagerskiold, 1891. Tetrabothrium {Diplobothrium) affine. Loennberg, 1892. Hosts: Blue Whale (Balaenoptera musculus); Sperm Whale (Physeter catodon). Localities: Saldanha Bay and Durban (South Africa); South Georgia; and 580 32' S, 340 52' E. The specimens examined are about 20 cm. long and about 5 mm. wide. The strobila is cylindrical with very short segments. In some specimens the posterior end of the body is modified into flattened segments, markedly separated from each other (PI. XX, fig. 1). This region is the spent part of the body containing the degenerate genital organs. The scolex, about 2 mm. long and 2 mm. broad, is variable in shape and bears typical 'epaulettes'. A neck does not seem to be present. The musculature is strongly developed in this species. Cross-sections of the body show that the longitudinal muscles form two distinct rings. The internal ring is composed of strongly developed bundles of a very characteristic shape shown in Fig. 4. The circular and dorso-ventral muscles are also very well developed. The testes are dorso-ventrally elongate in younger segments and about 90 x 53 ^ in diameter, but in the more developed segments they are irregularly spherical and seem to be arranged in several layers, which are distributed dorso-ventrally. The thick-walled, coiled vas deferens is well developed. The pear-shaped or rounded cirrus-sac possesses thick muscular walls and a strongly developed cirrus. The lobate ovary extends transversely in the ventral part of the segment and is composed of dis- tinct egg-cells. The vitelline gland is situated ventrally to the ovary. CESTODES EXAMINED 383 8 : \ ■ ? I'm'/- 1 1 I B i ■= ? * 1 1 1 II I I Ss I B SS Sin i f ll I i Figs. 1-6. Tetrabothrins affinis from Blue Whale. 1, cross-section through mature segment; 2, formation of uterine opening and mode of discharging eggs ; 3, eggs being discharged from uterus ; 4, transverse section of a portion of longitudinal muscle ; 5, transverse section of segment showing cirrus-sac and vagina; 6, egg. 384 DISCOVERY REPORTS The uterus is a well-developed sac. There is no special uterine pore, but dorsally in the central part of the segment, the parenchyma undergoes some modification. Transverse sections of this part of the segment show differentiated tissues which appear to form a cone. Gradually the distending uterus, filled with eggs, breaks through this part of the segment and the eggs are discharged through the rup- ture in the body-wall. The diameter of the egg is about (66-106x33-83)^ with the embryonic membrane measuring 51 x66it. The embryo measures about 52x2611 and is provided with three pairs of hooks, of which the lateral pairs measure 17 /a in length, and the central pair 20 11. The excretory system, composed of two main canals with transverse vessels, is very well developed. The ventral canal is relatively wider in diameter than the dorsal one, but the latter possesses thick walls. Tetrabothrius wilsoni (Leiper & Atkinson, 1914). (Figs. 7-10.) Oriana wilsoni Leiper & Atkinson, 1914. Tetrabothrius wilsoni Baylis, 1926. Tetrabothrius affinis Baer, 1954. Hosts: Sei Whale (Balaenoptera borealis); Blue whale (B. musculus). Localities : Durban ; South Africa ; South Georgia. Of the thirty-four specimens examined several appear to be gravid, a condition not hitherto reported. The length of the worms varies from 6 to 14 cm. and the width from 2 to 3 mm. The body is dorso-ventrally flattened. The scolex bearing four suckers with weakly developed 'epaulettes' is about 3 mm. broad. The longitudinal muscles are arranged in two very indistinct rings, as shown in Fig. 9. Individual bundles are not well defined and are composed of rather thick muscle fibres. Their arrangement appears diffuse and differs considerably from that found in Tetrabothrius affinis. The circular muscles, though present, are poorly developed. The testes are situated in the dorsal part of the segment, being 99 x 120 11 in transverse section. They seem to be arranged in an irregular layer. The vas deferens is not very strongly developed, while the cirrus-sac is pear-shaped or more or less spherical in transverse sections of the strobila. The ovary is extremely well developed, lobate, extending transversely in the ventral part of the segment, and occupying a considerable area of the proglottid. The egg-cells are fairly large, being about 1514 in diameter. The vitelline gland, situated ventrally beneath the ovary, is well developed. The uterus, in fully gravid segments, occupies the whole of the medullary parenchyma of the seg- ment, considerably displacing the internal organs. The walls of the segments are much thinner than in T. affinis. The eggs are discharged through a small opening in the dorsal face of the segment. They measure about 90 x 60 ix. The embryonic membrane is about 45 x 30/x and the embryo 36 x 21 /x in diameter. It seems that the latter is attached by its membrane to the external egg-capsule, as shown in Fig. 10. The external and internal pairs of embryonic hooks are of similar length, i.e. about 12 /a. The excretory system appears to be weakly developed and both the ventral and the dorsal trunks are of similar diameter. In the younger segments, as well as in the gravid, the ventral trunk of the excretory system seems to have a greater diameter. This, however, may be caused by the fixation and con- traction of the tissues. This form, described by Leiper & Atkinson in 1914, by Baylis in 1926 and recently by Baer (1954), represents a valid species, distinguished by differences in the arrangement and structure of the longi- SYSTEMATIC NOTES 385 tudinal muscles, the arrangement of the testes, the development of the ovary, and the structure of the eggs. The above-mentioned authors appear to have dealt only with immature specimens, which of course are not satisfactory for comparative purposes. ^WmMM m %» M Figs. 7-10. Tetrabothrius wihoni from Sei Whale. 7, cross section through mature segment ; 8, transverse section of segment showing cirrus-sac and vagina; 9, cross-section through portion of longitudinal muscle; 10, egg. Tetrabothrius ruudi Nybelin, 1928. (Figs. 11-15.) Host: Fin Whale (Balaenoptera physalus). Locality: South Georgia. The length of the specimens examined is about 9 cm. and the width about 2 mm. The scolex, pro- vided with four suckers and ' epaulettes ' is about 2 mm. broad. The longitudinal muscles form two more or less distinct rings. The inner one is formed by well accentuated bundles, composed of well developed fibres. The circular muscles are not well developed. The testes, situated dorsally, are about 42 X 45^ in transverse section. They are not very numerous. The vas deferens is well developed. The cirrus-sac is pear-shaped, as depicted by Nybelin (1928). The ovary, extending horizontally in the ventral part of the segment is well developed. The vitelline gland lies ventrally to the ovary. The very well developed uterus possesses the same kind of opening as in previous forms and similar modified parenchymatous tissue. The formation of the uterine duct commences in the mature segments where the uterus has not yet reached full development. In the early stages the modified tissue surrounding the duct is markedly visible. The eggs are 40-50 ^ in diameter, the embryonic membrane 30 x 30/Li and the embryo 17 x i8/l*. In eggs cleared in glycerine, embryonic hooks have not been observed. The excretory system shows the ventral vessel to be relatively greater in diameter than the dorsal one, the latter possessing thick muscular walls. 386 DISCOVERY REPORTS 14 Figs. n-15. Tetrabothrius ruudi from Fin Whale. 11, scolex in lateral view; 12, cross-section of mature segment; 13, cross- section through gravid segment showing uterus and uterine opening situated dorsally; 14, cross-section through portion of longitudinal muscle; 15, egg. Tetrabothrius schaeferi sp.n. (Figs. 16-20). Tetrabothrius afflne. Baylis, 1926. Tetrabothrius sp. Rees, 1953. Host: Blue Whale (Balaenoptera muscuhis). Localities: South Georgia, Southern Ocean. The body is rather slender and flattened, varying from about 7 to 14 cm. in length and about 2 to 3 mm. in breadth. The segmentation is distinct. The scolex, provided with four suckers and ' epaulettes ' is 3 mm. broad. A neck is present and well defined. The longitudinal muscles are seen to form two rings, if the segment is examined in cross-section. The inner ring is composed of distinctly separated bundles, irregularly oval in outline and well developed. Immediately, above the inner ring lies the external ring of longitudinal muscles which are weakly developed, as are the circular and dorso-ventral muscles. The testes, about fifty in number, are about 90 x 90/x when measured in transverse section. They occupy the dorsal part of the segment, being arranged in a single or many layers in fields extending laterally and antero-posteriorly. The cirrus-sac, spherical or pear-shaped is about 30 n long and 150M wide. The vas deferens is a much coiled, fairly well-developed organ. The well-developed ovary lies in the ventral parenchyma. It is lobate and in transverse section more or less oval in outline. The egg- cells are fairly large, about 21 /x in diameter. The vitelline gland is ventral to the ovary. The uterus is a sac-like organ which as it matures gradually fills the medullary parenchyma. The uterine duct appears fairly early in the strobila as a differentiation of parenchymatous tissue. The uterine opening is spherical when observed in horizontal section. It is well marked, and surrounded with modified cells which appear to absorb more haematoxylin than the counter-stain erythrosin. The egg is about 90 x 75 it in diameter, or more or less spherical. The inner membrane containing SYSTEMATIC NOTES 387 an embryo is about 45 x 36/* in diameter, while the embryo is about 24 x 45^. The lateral pairs of hooks are 12 it in length and the central pair 18//,. The proportions of these hooks is not represented accurately in Fig. 20. The ventral and dorsal trunks of the excretory system differ in diameter and have rather thin walls. The differences between this new form and Tetrabothrius affinis lie in the shape of the body and in the arrangement of the longitudinal muscles. These characters also differ from those found in other species of Tetrabothrius. Further differences in structure and in the eggs and the embryos have been given in Table 1. ' ' '.**>*< ft H -ft ?«f * 3 ,-w. - v^vS^it'L,yi:iU;4f-'i,''(au;'' Figs. 1 6-20. Tetrabothrius schaeferi sp.n. from Blue Whale. 16, Cross-section of semi-gravid segment with eggs; 17, cross- section through longitudinal muscle; 18, cross-section of segment showing cirrus-sac and vagina; 19, cross-section showing cirrus-sac and everted cirrus; 20, egg. Table 1 . Measurements of Tetrabothriids occurring in the Cetacea from the Discovery Collections Species Body Body Testis Egg Embryonic Embryo Lateral Central length width diameter shell diameter hooks hooks Tetrabothrius affinis 20 cm. 5 mm. 9°><53M 106 x 83/4 51 x66fj. 52 x 26/x 17/* 20/x T. wilsoni 14 cm. 3 mm. 99 x 120/x 90 x 60/i 45X3°M 36X 21 /j. I2/z 12/z T. ruudi 9 cm. 2 mm. 42X45/X 40 x 50/i 30x30/* 17X 18/x — — T. schaeferi sp.n. 14 cm. 3 mm. 90 x 9011 90x75,* 45x36/* 24x451* 1211 i8fi Trigonocotyle 21 cm. 3/2 mm. I20X 135^ 75x45/* 60 x 40/1 54X3°/x 18/n 12/i globicephalae Tr. prudhoei sp.n. 14-5 cm. 1 mm. 120 x 6oti 60 x 84 fl 45 x 30/* 27 x 3011 9/* 6fJL Priapocephalus grandis 15 m. 24 cm. 12 mm. 68 x 60 ft 99 x 66/1 33 x 50/* 17/* 17/* 388 DISCOVERY REPORTS Trigonocotyle Baer, 1932 Trigonocotyle globicephalae Baer, 1954. (Figs. 21-26.) Prosthecocotyle monticelli Linton 1923; nee Furhmann, 1899. Trigonocotyle monticelli Linton, 1923; Baer 1932. Trigonocotyle lintoni Yamaguti, 1942; nee Guiart, 1935. Host : Globicephala edwardi. Locality: Off Cape Town (South Africa). 23 Figs. 21-26. Trigonocotyle globicephalae from Globicephala edwardi. 2 1 , scolex in lateral view ; 22, cross-section through mature segment in anterior cylindrical part of strobila; 23, cross-section through segment in flattened part of strobila, showing cirrus- sac, vagina and longitudinal muscles; 24, cross-section through mature segment in anterior cylindrical part of strobila showing longitudinal muscles; 25, cross-section through gravid, flattened part of strobila showing longitudinal muscles; 26, egg. The length of the body in the specimens examined is about 21 cm. The strobila appears to be divided into two portions ; the anterior part is cylindrical and about 2 cm. in length, composed of very short segments, while the rest of the body is flattened. It seems that the anterior portion, which bears a resemblance to the cylindrical strobila in Tetrabothriiis affinis, contains mature segments and the flattened part gravid segments with eggs. The scolex, 1-5 mm. broad is provided with four suckers, bearing fleshy protuberances in the form of horns. SYSTEMATIC NOTES 389 The longitudinal muscles are well developed and composed of thick fibres. In transverse section they are collected in elongate bundles, separated from each other by muscular septa. The longitudinal muscles in the cylindrical portion of the strobila are far better developed than those in the gravid segments. The bundles in the anterior portion are thicker than in the gravid segments as shown in Figs. 24 and 25. This is probably owing to the pressure of the uterus as it becomes distended with eggs. The circular muscles are exceedingly well developed, noticeably in the anterior portion of the body. The testes are about 120 x 135 /x in transverse diameter and arranged in many layers in the cylindrical part of the body. In the flattened portion they seem to be arranged in a single layer. The vas deferens is well developed. The pear-shaped cirrus-sac is about 186X90/X in diameter. The genital atrium possesses a muscular modification in the form of a sucker. The ovary extends ventrally in the medullary parenchyma and is well developed. The vitelline gland situated ventrally under the ovary is composed of compactly arranged cells. Neither the uterine opening nor the rudiment of its duct has been found. The uterus is very well developed and gradually pushes away other parts of the genital complex. The eggs are 75 X45//. in diameter and the embryo 54x36/^. The lateral hooks measure i8|a and the central 12/x. Yamaguti (1942) could not find any hooks in his material. The excretory system is well developed and has transverse vessels connecting the two main trunks. Trigonocotyle prudhoei sp.n. (Figs. 27-32). Hosts: Steno bredanensis, Lagenorhynchus obscurus, L. australis. Locality: 140 45' N, 180 34' W; East Falkland Islands (southernmost point of Bay of Harbours). The body is about 14-5 cm. in length and 1 mm. in width and composed of slightly overlapping segments. The scolex, measuring from 1-5-2 mm. in width, is provided with four suckers bearing fleshy protuberances. The longitudinal muscles are collected into bundles, with some indication of two layers dorsally, when examined in transverse section. They seem to be separated by muscular septa. The circular muscles are well developed. There are also dorso-ventral muscle fibres. The testes are not very numerous and occur mainly in the dorsal part of the segment. They are about 120 ■: 60/x in transverse section, the longer axis being in the dorso-ventral plane. The cirrus-sac is flask-shaped, about 32 x 15/tt in diameter, with thick muscular walls and a well-developed cirrus. The vas deferens is well developed. Radiate muscles forming a sucker occur in the genital atrium. The ovary is situated ventrally and extends laterally on both sides of the median line. The well- developed vitelline gland is situated ventrally beneath the ovary. Neither a uterine pore nor the rudiment of a uterine duct have been observed. The eggs are 60 x 84/x. The inner membrane is 45 :■; 30 ju and the embryo 27 x 30 /x. The lateral pairs of embryonic hooks measure 9^ and the central 6/x in length. The excretory system is well developed. This species differs mainly from other species of Trigonocotyle occurring in Cetacea in the form of the serrate strobila and in the arrangement of the longitudinal muscles. It also differs in the number and arrangement of the testes, these being less numerous and distributed dorsally. There are also differences in the size of the eggs and in the length of the embryonic hooks as shown in Table 1 . 39° DISCOVERY REPORTS O 05mm Figs. 27-32. Trigonocotyle prudhoei sp.n. from Lagenorhynchus obscurus. 27, the scolex; 28, part of the strobila; 29, cross- section through part of longitudinal muscles; 30, cross-section through mature segment; 31, cross-section showing cirrus-sac and vagina; 32, egg. Priapocephalus grandis Nybelin, 1922. (Figs. 33-41 ; PI. XX, 3-4). Hosts: Blue Whale {Balaenoptera musculus), Sei Whale (B. borealis), Fin Whale (B. physahis), Sperm Whale (Physeter catodon). Localities : Saldanha Bay and Durban (South Africa) ; South Georgia ; Southern Ocean (560 40' 32' S, 24° 33' W); Porto Pirn, Azores. The size of the specimens examined varies from 4 cm to 1 5 m. 24 cm. in length with a width varying from 2 to 12 mm. The worms are in various stages of development, from immature to full-grown individuals. There are also fragments of worms, measuring from 12 cm. to 5 m. 75 cm. in length and 2-12 mm. in width; some of these have scolices intact. These are typically acorn-like and are deeply embedded in the mucous which exudes from the gut of the host. Fig. 33 shows the mode of attachment. It seems that there is a free space between the scolex and the surrounding intestinal tissues of the host. The scolex bears a kind of collar, which possibly serves as a supplementary organ of attachment. The maximum length of the scolex amounts to 7 mm., including the collar; the corresponding width is 4 mm. The segments are wider than long (the length being about 1-5 mm.), and not craspedote. The body is ivory white or yellowish in preserved specimens. The longitudinal muscles are very well developed. They are arranged in two rings, the outer one being the less well developed. The inner one is composed of strong bundles, oval in transverse section. The longitudinal muscles are variously developed in the same individual. In the juvenile forms or in SYSTEMATIC NOTES 39i Figs. 33-41. Priapocephalus grandis from Blue Whale. 33, longitudinal section through scolex of immature specimen embed- ded in mucous ; 34, cross-section through part of longitudinal muscles ; 35, cross-section through part of longitudinal muscles ; in segment with rudimentary genital organs; 36, cross-section of longitudinal muscles in mature segment; 37, sagittal section of body showing uterine ducts and openings for discharge of eggs; 38, mode of egg-discharge; 39, cross-section of segment showing uterine pores; 40, cross-section of cirrus-sac and vagina; 41, egg. 392 DISCOVERY REPORTS the anterior portion of the body the bundles are arranged very densely. They seem to form definitely shaped bundles in the mature and gravid segments. In the young specimens their arrangement is radiate. The differences in the arrangement of the longitudinal muscles in the various segments are shown in Figs. 34-6. The circular muscles are well developed. The testes are numerous and arranged in many layers. They measure about x 68(53-60)^ in transverse section and about 45 - 53^ in horizontal section. The vas deferens is well developed and arranged in coils. The cirrus-sac is about 135 /x long and about 60 /j, wide. The fairly large ovary is composed of large egg-cells. The uterus, filled with eggs, occupies most of the medullary parenchyma. The eggs, after the uterus has broken through the cuticle, are discharged by numerous pores situated in the dorsal part of the segment. No specially defined uterine pore is present, as in the species of Tetrabothrius described above. There is no difference in the formation of the uterine ducts, but instead of one, several openings are formed. Figs. 37-39 show the mode of dis- charging the eggs. The egg is about (83-99 :< 66)/u, and contains the embryo, 33x50^ in diameter. The embryonic hooks are about ij/j. in length. The excretory system is well marked. Genus Diplogonoporus Loennberg, 1892 Diplogonoporus balaenopterae Loennberg, 1892. (Figs. 42-43; PL XXI, figs. 5-7.) Hosts: Blue Whale (Balaenoptera musculus); Fin Whale (B. physalus); Sei Whale (B. borealis). Locality: South Georgia. The body is flat and tape-like. In all the specimens with a scolex, irrespective of size, the anterior part of the body is undulate, as shown in PI. XXI, fig. 6. Though this modification of the body may be caused by fixation, it seems nevertheless to be constant. Although the only complete worms occurring Figs. 42, 43. Diplogonoporus balaenopterae from Sei Whale. 42, scolex; 43, transverse section showing arrangement of longitudinal muscles. in the collection measure up to about 13 cm. in length, there are many fragments, one of which is over 8 m. long. The scolex is typical, with two grooves from 1 to 1-5 mm. broad and from 1 to 2 mm. long. The bothria are deep and well pronounced. There is no neck. The longitudinal muscles, as examined in transverse section, form irregular groups of muscle bundles composed of loosely arranged fibres. Individual groups are separated by muscle fibres, run- ning dorso-ventrally. The circular muscles are well developed. 393 SYSTEMATIC NOTES The genital organs occur as separate units in both lateral fields of the segment. In some specimens, however, there are three sets of organs on one side, and two on the other. Two parallel furrows run along the strobila, marking the position of the genital organs. The testes are about 120 X 150/A in transverse diameter and are arranged in a single layer. Their number between the right and left genital sets seems to be from about thirty-seven to forty. It seems also that the number of testes occurring in each of the genital sets varies greatly when examined in transverse sections, the left-hand side of the segment may possess a few testes or none, while on the right side sixteen or more testes may occur. The cirrus-sac is about 345x300/*. The vas deferens is well developed and much coiled. The cirrus is fairly large. The ovary forms an arch in the ventral half of the segment. The uterus, forming a ' rosette ' is filled with operculate eggs of about 66X48/X in diameter. The egg-shell is 3/x thick. The vitelline glands are situated dorsally and ventrally, in the cortical parenchyma, leaving a free field around the genital organs. They are more or less spherical in transverse section. LARVAE Two types of larval stages have been found in the present material. All specimens examined seem to belong to the Genus Phyllobothrium, except one specimen which probably belongs to the Tetraphyllidea. Specific identification is not possible owing to the complete lack of any morphological characters. Phyllobothrium (sensu lato) sp. larva. (Fig. 44; PI. XXI, fig. 8.) Hosts: Sperm Whale (Physeter catodon) and (Lagenorhynchus obscurus). Localities : Durban, Simon's Town, South Africa and South Georgia. This larval stage is found in cysts in the blubber and may be regarded as a cysticercus. The body is composed of a bladder with an invaginated scolex and neck. The size of the cyst is variable, measuring from 3-20 x 2-9 mm. The scolex is about 2 mm. broad and is similar to that of Phyllobothrium. The evaginated neck is 6 mm. long and 3 mm. wide in the larger specimens. Fig. 44. Phyllobothrium sp. from Sperm Whale. 44. larval stage extracted from blubber. Fig. 45. Tetraphyllidean larva from blubber of Sperm Whale. 394 DISCOVERY REPORTS Tetraphyllidean larva. (Fig. 45.) Host: Sperm Whale {Physeter catodon). Locality: South Georgia. Only one specimen has been found; it occurred in the same host together with the previously described form. The total length of the evaginated cyst is about 9 x 3 mm. The scolex is 1 mm. wide, provided with four discoid suckers, which measure about 420 x 525/x in diameter. RELATIONSHIPS BETWEEN THE HOST AND THE PARASITE From the material examined it seems that host-specificity does not exist. The only exception is Tetra- bothrius ruudi, which at present has been recorded only in the Fin Whale. This, however, does not mean that this species may not occur in other species of whales, and it is probably a matter of the number of the hosts examined. This appears to be true as far as the present material is concerned. Table 2. Infection of hosts with particular species of cestodes Host No. of hosts examined Tetra- bothrius affmis T. wilsoni Species of cestodes Trigono- cotyle globi- cephalae T. prudhoei sp.n. Priapo- cephalus grandis Diplogono - porus balaeno- pterae Phyllo- bothrium sp. larva Tetra- phyllidean larva T. ruudi T. schae- feri sp.n. Balaenoptera musculus B. borealis B. physalus Physeter catodon Lagenorhynchus obscurus L. australis Steno bredanensis Globicephala edwardi 46 13 8 14 2 1 2 2 11 1 2 10 4 15 1 1 1 2 3i 1 2 1 1 3 2 12 1 1 The Blue Whale has been found to harbour five different species of Cestodes, while in the seven other hosts only one to three different species occurred. Blue Whales formed by far the greater number of the hosts examined (Table 2) and this undoubtedly accounts for the greater variety of worms from this species of whale. Hitherto Diplogonoporus balaenopterae has been found only in the Sei Whale. In the present material this cestode has been found in two other hosts, the Fin Whale and the Blue Whale. The latter appears also to be a new host for Tetrabothrius wilsotii. The Fin Whale is a new host for Priapocephalus grandis and Tetrabothrius affinis. Two species of tapeworms may occur together in the same gut, for in Blue Whale, Priapocephalus grandis was found either with Diplogonoporus balaenopterae or with Tetrabothrius affinis, or with T. schaeferi sp.n. Triple infestation has not, however, been noted in any of the hosts examined. REFERENCES Baer, J. G., 1932. Contribution a I'etude des Cestodes de Cetaces. Rev. Rev. Suisse Zool. xxxix, 3, pp. 195-228. Baer, J. G., 1954. Revision taxinomique et etude biologique des Cestodes de la famille des Tetrabothriidae, parasites d'oiseaux de haute mer et de Mammiferes marins. Mem. Univ. Neuchatel, I, pp. 1-121. Baylis, H. A., 1926. Some Tetrabothriid Cestodes from Whales of the genus Balaenoptera. Journ. Linn. Soc. London, xxxvi, pp. 161-72. Baylis, H. A., 1932. A list of Worms parasitic in Cetacea. Discovery Rep. vi, pp. 395-418. Fuhrmann, O., 1899. Das Genus Prosthecocotyle. Zool. Ang. xxn, pp. 180-3. SYSTEMATIC NOTES 395 Fuhrmann, 0., 1903. Die Tetrabothrien der Sdugetiere. Zbl. Bakt. xxxv, pp. 744-52. Guiart, J., 193 s. Cestodes parasites provenant des Campagnes scientifiques du Prince Albert ier de Monaco. Res. Camp. Sci. Albert ier, fasc. 91, pp. 68-74. Jagerskiold, L. A., 1891. Einiges tiber die Schmarotzer der nordatlantischen Balaenopteriden. Biol. Foren. Forhandl., Stock- holm, hi, pp. 127-34. Leiper, R. T. & Atkinson, E.L., 1914. Helminths of the British Antarctic Expedition, 1910-1913. Proc. Zool Soc. Lond. pp. 222-6. Leiper, R. T. & Atkinson, E. L., 19 15. Parasitic worms with a note on free-living nematoda. Brit. Antarctic ('Terra Nova') Exped. 1910. Nat. Hist. Rep. Zool. 11 (3), pp. 19-60. Linton, E., 1923. A new Cetacean Cestode (Prosthecocotyle monticelli spec, nov.) with a note on the genus Tetrabothnus Rudolphi. J. Parasit. x, pp. 51-5. Loennberg, E., 1892. Anatomische Studien tiber skandinavische Cestoden, II Kgl. Sven. Vetensk.-Akad. Handl. xxiv, pp. 1-30. Nybelin, O., 1922. Anatomisch-systematische Studien tiber Pseudophyllideen. Goteb. Kungl. Vetensk. Handl. xxvi, pp. 1-228. Nybelin, O., 1928. Zweineue Cestoden aus Bartenwalen. Zool. Anz. lxxviii, pp. 309-14. Rees, G., 1953. A record of some parasitic worms from whales in the Ross Sea Area. Parasit. xliii, i, pp. 27-34. Yamaguti, S., 1942. Studies on the Helminth Fauna of Japan. Part 42. Cestodes of Mammals. XI, 18 pp. 4 pis. Privately printed, 20 June. Kyoto, Japan. PLATE XX Figs. 1-2. Tetrabothrius affinis from Blue Whale, i, cylindrical body with portion of spent segments; 2, strobila. Figs- 3~4- Priapocephalus grandis from Blue Whale. 3, strobila; 4, anterior part of strobila and scolex. DISCOVERY REPORTS VOL. XXVII PLATE XX PLATE XXI Figs. 5-7. Diplogonopoms balaenopterae from Fin Whale. 5, strobila; 6, anterior part of body; 7, part of the strobila. Fig. 8. Portion of Sperm Whale blubber infected with Phyllobothrium sp. DISCOVERY REPORTS VOL. XXVII PLATE XXI J 10 II 12 DISCOVERY REPORTS Vol. XXVII, pp. 1-162, plates I-XII Issued by the National Institute of Oceanography SIPHONOPHORA OF THE INDIAN OCEAN TOGETHER WITH SYSTEMATIC AND BIOLOGICAL NOTES ON RELATED SPECIMENS FROM OTHER OCEANS by A. K. Totton CAMBRIDGE AT THE UNIVERSITY PRESS 1954 Price sixty shillings net PUBLISHED BY THE SYNDICS OF THE CAMBRIDGE UNIVERSITY PRESS London Office : Bentley House, n.w. i American Branch : New York Agents for Canada, India, and Pakistan: Macmillan Printed in Great Britain at the University Press, Cambridge (Brooke Crutchley, University Printer)