^ GENERAL < Bulletin of the British Museum (Natural History Zoology series Vol 40 1981 British Museum (Natural History) London 1981 Dates of publication of the parts No 1 26 February 1981 No 2 30 April 1981 No 3 28 May 1981 No 4 30 July 1981 No 5 30 July 1981 ISSN 0007-1498 Printed in Great Britain by Henry Ling Ltd, at the Dorset Press, Dorchester, Dorset Contents Zoology Volume 40 Page No 1 Eugene Penard's slides of Gymnamoebia : re-examination and taxonomic evaluation By Frederick C. Page 1 No 2 Japanese earthworms : a synopsis of the Megadrile species (Oligochaeta) By E. G. Easton 33 No 3 Phylogenetic versus convergent relationship between piscivorous cichlid fishes from Lakes Malawi and Tanganyika By Melanie L. J. Stiassny 67 No 4 Miscellanea The calceolus, a sensory structure of gammaridean amphipods (Amphi- poda: Gammaridea) By R. J. Lincoln and D. E. Hurley 103 A new species of Lernaea (Copepoda : Cyclopoida) from Papua-New Guinea By G. A. Boxshall . 117 Some type specimens of Isopoda (Flabellifera) in the British Museum (Natural History), and the isopods in the Linnaean Collection By J. Ellis 121 Conchoecia hystrix n. sp. a new halocyprid ostracod for the Porcupine Bight region of the Northeastern Atlantic By M. V. Angel and C. Ellis .129 The Conchoecia skogsbergi species complex (Ostracoda, Halocy- prididae) in the Atlantic Ocean By A. J. Gooday 137 No 5 Miscellanea The larval and post-larval development of the Edible Crab, Cancer pagurus Linnaeus (Decapoda : Brachyura) ByR.W. Ingle . . . 211 A taxonomic study of the larvae of four thalassinid species (Decapoda, Thalassinidea) from the Gulf of Mexico By N. Ngoc-Ho 237 The status of Glyphocrangon rimapes Bate 1888 (Crustacea, Decapoda, Glyphocrangonidae By A. L. Rice .275 Crab zoeae and brachyuran classification : a re-appraisal By A. L. Rice 287 Bulletin of the British Museum (Natural Histofy) Eugene Penard's slides of Gymnamoebia: re-examination and taxonomic evaluation Frederick C. Page Zoology series Vol 40 No 1 26 February 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to : Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull, Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1981 ISSN 0007-1498 Zoology series Vol 40 No 1 pp 1-32 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 26 February 1981 * GENERAL -3 MARK Eugene PenarcTs slides of Gymnamoebia: re- examination and taxonomic evaluation Frederick C. Page k Culture Centre of Algae and Protozoa, Institute of Terrestrial Ecology (N.E.R.C.), 36 Storey's Way, Cambridge CB3 ODT Contents Synopsis 3 Introduction 4 Observations 7 Amoeba proteus 8 Amoeba nitida 10 Amoeba sp 11 Amoeba nobilis 12 Amoeba laureata 13 'Amoeba peritissima' 15 Amoeba terricola 16 Amoeba papyracea 17 Amoeba sphaeronucleolus 19 Amoeba fibrillosa 20 Amoeba alba 22 Amoeba granulosa 24 Amoeba lucens 25 Amoeba vespertilio 26 Amoeba muralis 27 Dinamoeba mirabilis 28 Pelomyxa palustris 29 Pelomyxafragilis 30 Pelomyxa vivipara 31 Pelomyxa belevskii 32 Pelomyxa binucleata 33 Generic diagnosis 35 Classification 35 References 36 Synopsis The slides of freshwater naked amoebae prepared early in this century by Eugene Penard have been examined with modern optical systems, and much of the material is presented in photomicrographs. The new genus Thecochaos is proposed for amoebae which resemble Thecamoeba except in being multinucleate, with T. fibrillosum (Greeff, 1891) as the type-species. Other new combinations include Chaos nobile (Penard, 1902), Thecamoeba papyracea (Penard, 1905), and Thecochaos album (Greeff, 1891). The validity of several named species of Pelomyxa is considered without a definite conclusion. The taxonomic positions of several species cannot be determined on the basis of the present material. Bull. Br. Mus. nat. Hist. (Zool.) 40( 1 ) : 1-32 Issued 26 February 1 98 1 2 F. C. PAGE Introduction Eugene Penard (1855-1954) was one of the classical students of protozoan natural history. From 1887 until his eyes failed him in 1922, he investigated diverse Sarcodina, ciliates, and flagellates, all collected from natural sources and studied alive and in fixed preparations. Most of his life was spent in his native Geneva, but his publications include material from such far places as Loch Ness, Sierra Leone, the Rocky Mountains, and the Himalayas. To the biography by Deflandre (1958) are appended a list of publications and a list of 24 genera and 426 species of Protozoa (including 3 1 species of Gymnamoebia) erected by Penard. Altogether Penard produced at least 82 protistological publications, the great majority on Sarcodina. His most important work on rhizopods, including naked amoebae, is Faune Rhizopodique du Bassin du Leman (1902), foreshadowed by his 'Etude sur les Rhizopodes d'eau douce' in 1 890. The Faune Rhizopodique is among the classical publications on naked amoebae, which also include those of Leidy (1879), Cash (1905), Cash & Wailes (1919), and Schaeffer (1926). All except the last also deal with testaceans, which indeed take up the larger part of the space. Penard was interested especially in observations of living organisms, considering such 'physiology' more interesting and more important than systematics, which he did, however, regard as essential. Amongst the characters which he could observe in naked amoebae, he considered the nucleus especially important. To preserve his organisms and facilitate observations of such characters as the nucleus, he made many permanent preparations. His standard method (Penard, 1902) was fixation with absolute alcohol, staining with borax carmine, and mounting in Canada balsam. His observations were made without an immersion objective (Deflandre, 1958), and of course he depended on drawings rather than photographs to convey visual impressions. Deflandre praised these drawings highly, though it must be said that in the case of Gymnamoebia one might often wish for more details, more depth, and especially more individual amoebae. The subsequent reproduction by later authors of a single Penard drawing per species, which even if composite in origin could only represent a single cell at a single moment, has over the years proved inadequate for identification. According to Deflandre (1958) Penard's slides are preserved in three major collections. When Deflandre was writing (he died in 1973), the smallest of these three main collections was in his own possession. It comprised 280 slides, of which 233 were of rhizopods, including 18 slides of Amoeba (in Penard's broad sense of the genus) and five of Pelomyxa. The second collection was at the Museum d'Histoire Naturelle of Geneva, containing 663 preparations, according to Deflandre. In 1952 Deflandre found these preparations in good condition but could not summarize their contents for lack of an adequate catalogue. However, Grospietsch (1975), in a publication with limited distribution, listed 791 slides in the Geneva collection. These include 48 slides of 13 species of Amoeba (in Penard's broad sense) and 1 1 slides of five species of Pelomyxa. All these species of naked amoebae except two of Amoeba are also represented in the third and largest major collection, that at the British Museum (Natural History), consisting of 950 slides. The great majority of these are testaceans. The 55 slides in this third collection bearing naked lobose amoebae (Gymnamoebia) are the object of the present study. According to Penard's nomenclature, they contain specimens of 20 species, as well as one slide with no specific identification. Besides the three major collections, Heal (1965) lists three smaller collections of Penard's slides in Britain, but these contain only three of Penard's species of Gymnamoebia, all represented in the collection at the British Museum (Natural History). There are several reasons for re-examining Penard's slides now. Free-living amoebae are much better known than they were a few decades ago. They have been investigated by more workers with a greater diversity of aims than ever before, applying modern tools. Knowledge of taxonomic value has been developed in some cases by workers who initially had no particular interest in taxonomy. It is principally large amoebae which have been much used in cell biological investigations, with the exception of Acanthamoeba, and Penard's PENARD'S SLIDES OF GYMNAMOEBIA 3 publications dealt with and his slides preserve mainly larger amoebae. Among the Penard material are represented several groups whose taxonomic status and boundaries are still matters of some uncertainty. For example, how many species of Pelomyxa did Penard actually have in front of him? There are also organisms which apparently have been seen by few other workers, so that on the basis of Penard's text and few drawings their existence and identification have been questioned: Are there, already described and identifiable, more species of Chaos than the two familiar to present-day cell biologists? Are there multinucleate amoebae which resemble Thecamoeba more than they resemble Amoeba and thus perhaps should not be classified as Chaosl With modern optical and photographic equipment, the Penard material can provide answers to some questions and give information for further consideration of those questions which remain unsettled. Bright field, phase contrast, and differential interference contrast optics have been used to examine and photograph the preparations. In most cases, it was possible to examine and photograph the slides with an objective up to x 63 (giving x 800 final magnification for direct observation, x 3 15 on the photographic negative). In a few cases the thickness of the mount made use of an objective greater than x 10 impossible, but enlargement of the negatives provided further information, such as nuclear size and structure. Most of the material on the slides was photographed, the omissions being a very small quantity of useless material and some repetitive material. The same cells were usually photographed at two magnifications and with at least two optical systems. From these hundreds of photomicrographs, the most informative are included as illustrations. The optical systems are not consistent amongst the illustrations, since the most informative photomicrographs of one species or feature were not always obtained with the same system which was most useful for another. That this inconsistency does not prevent useful comparison can be seen by examining, for example, the figures of Amoeba fibrillosa and those of Amoeba alba. All reasonably intact cells were measured. In uni- or binucleate species, all discernible nuclei were measured. In multinucleate cells, 25 nuclei were measured (if that many were present) in one or more cells. Finally, it should be noted that Penard's over-all taxonomic system for rhizopods was simple and imprecise (Penard, 1902). He recognized two 'groups', the Lobosa and the Reticulosa, with the Filosa as a subgroup of the former. Within the naked lobose amoebae he classified most in the genus Amoeba, although he also used the two supposedly anucleate genera Protamoeba Haeckel, 1866, and Glol'dium Sorskine, as well as Dinamoeba Leidy, 1874, which is still recognized. He considered a subdivision of the genus Amoeba sensu lato premature at that time. Pelomyxa he defined as 'Amibes a mouvements lents, toujours pourvues de bacteries symbiotiques'. A modern classification of as many of Penard's organisms (on these slides) as possible will be suggested. Observations These are in no sense species descriptions but only summaries of the information derived from the slides. Because of the limited taxonomic usefulness of fixed light-microscopical preparations of Gymnamoebia (greater for these larger amoebae than for smaller ones), a fuller picture of the organisms requires consultation of the descriptions of living amoebae in the Faune Rhizopodique (1902) and Penard's other publications. It will be found that the measurements given in that publication differ somewhat from those derived from the fixed preparations. For some species listed here, Page (1976, 1977) gives more complete information, under modern generic names. The headings below use Penard's generic classification; some re-classifications with gender modifications of specific epithets are proposed later in this publication. The authorships given in the headings are those required by today's nomenclatorial regulations whether or not they correspond with those used by Penard. Dates are absent from many slides. All slides bearing dates of preparation were made between 1901 and 1903, but it is almost certain that some were made later. The first two digits of the slide numbers indicate the date of deposit in the British Museum (Natural 4 F. C. PAGE History): -04= 1904, etc. All these slides bear material collected in Geneva and vicinity, including Haute Savoie. Fixation is as one would expect given the use of absolute alcohol alone, with some shrinkage of conical subpseudopodia (Amoeba vespertilio and Dinamoeba mirabilis) and blurring of the distinction between hyaloplasm and granuloplasm. The nuclear stain remains good. Most of the amoebae are located near the centres of the preparations. Some cells are broken. Only a minority of the amoebae are in the form taken during steady locomotion, with more irregular forms, e.g., producing branches in several directions or spreading on the substratum, common. The references under each species heading list the publication(s) in which Penard described each species, since it is only his material which is being considered here. Publications of other authors are cited in the remarks where relevant. Amoeba proteus Leidy, 1878 (Figs 1-6) Penard, 1902, pages 57-60 SLIDE NUMBERS. 04.5.9.19; 04.5.9.30; 20.12.8.10 (labelled Amoeba laureata with notation 'a\ec\ Amoeba proteus'); 20.12.8. 15; 20.12.8.16; 20.12.8.17; 20.12.8. 18. TOTAL NUMBER OF AMOEBAE. 13. DESCRIPTION AND REMARKS. Several of these cells had the elongated form often seen in more rapid locomotion of A. proteus, the longest (Fig. 1), 524 um long, with a second pseudopodium which may have been undergoing retraction at the moment of fixation. These preparations were examined in the light of Schaeffer's (1916) statement: 'I therefore suggest the specific name dubia for the organism named proteus by Penard.' None of the amoebae in these preparations had a cell form not reconcilable with that of A. proteus as understood since Schaeffer's ( 1 9 1 6) more precise definition of that species, though one could equally say that these shapes could also be encountered in Polychaos dubium (Schaeffer, 1916). Certainly the three locomotive forms on slide 20.12.8.15 are fully compatible with A. proteus. Schaeffer pointed out the inconsistency between the discoid nuclear shape described for A. proteus by Leidy ( 1 879) and that which Penard described as 'toujours ovoi'de' except in a variety where it was 'toujours parfaitement globuleux'. The nuclei in most of Penard's specimens could indeed be ovoid, but at least one of those on slide 20.12.8.15 is discoidal (Figs 1 , 4), and all the amoebae on that slide appear to be of the same type. Nuclear diameters in Penard's specimens are 27*5 to 54 um, with only one below 37 urn. Although it is therefore possible that some of Penard's slides labelled "Amoeba proteus' do bear members of other species, according to present specific distinctions, the amoebae on slide 20.12.8.15 certainly correspond to A. proteus. One or two amoebae on that slide appear to have the surface ridges which Schaeffer made diagnostic of A. proteus, though care is necessary in evaluating ridges or folds on these preparations. The slide 20. 1 2.8. 1 7, with the notation 'Variete', bears an amoeba which could be a P. dubium if any of these are (Fig. 3). Most of these amoebae contain ingested diatoms, and a few other algae and possible protozoa were seen in some. Amongst these slides is one, 20.12.8.18 ('avec prolongements cryptogamiques') with a single amoeba trailing a tuft of filaments (Fig. 5). These presumed hyphae (Fig. 6) are apparently nonseptate and about 2*5 um in diameter, and the longest extends about 74 um from the amoeba. Slide 04.5.9.19 is a preparation of an amoeba 'ecrasee pour montrer les parasites'. These filaments are apparently nonseptate, with a diameter up to 3 um or slightly more. In both cases, Penard's notation (not taxonomic label) refers to the "Ouramoeba of Leidy (1879), now generally acknowledged to have consisted simply of such infected members of the genus Amoeba. PENARD S SLIDES OF GYMNAMOEBIA Figs 1-4 Amoeba proteus. ( 1 ) to (3) Whole cells x 200. (4) Nucleus x 1000. N = nucleus. F. C. PAGE Figs 5, 6 Amoeba proteus with apparent fungal parasite. (5)x250. (6) x 1000. D = diatom. F = fungal hyphae. Amoeba nitida Penard, 1 902 (Figs 7-11) SLIDE NUMBERS. 04.5.9.23; 04.5.9.24; 20.12.8.12; 20.12.8.13. TOTAL NUMBER OF AMOEBAE. 8. DESCRIPTION AND REMARKS. At least four of these amoebae are in an elongate locomotive form (maximum length 408 /urn) with one main pseudopodium, though some pseudopodia are found as separate fragments. In four amoebae (one binucleate) the nucleus is distinctly discoid (Figs 9, 10). Although the angle of viewing makes the other nuclei appear roughly circular or, in one case, oval in outline, a closer examination suggests that these nuclei, too, are discoid. The thickness is about half the greatest diameter or even less. The irregular outline of the nuclei in flat view (Fig. 1 1) recalls Penard's (1902) emphasis on the foldings and invaginations to which the nuclear envelope is susceptible. The nucleolar spherules are arranged in a layer just beneath the nuclear envelope, but in several nuclei there is also a central mass, perhaps a fixation artifact. The maximum diameters of nuclei in the uninucleate amoebae were 34 to 54 //m; the maximum diameters of the nuclei in the binucleate amoeba were 35 and 3 1 //m. Ingested material included diatoms and possibly a few protozoa. At least one amoeba contained crystals, which appeared to be truncate bipyramids, though they were somewhat deteriorated. Schaeffer (1916) asserted that Penard's A. nitida was equivalent to the A. proteus of Leidy (1879). My examination of Penard's slides labelled 'Amoeba proteus' (see previous section) showed that the amoebae on at least one of them could not be Polychaos dubium, with which PENARD'S SLIDES OF GYMNAMOEBIA Figs 7-11 Amoeba nitida (=A. proteus). (7), (8) Whole cells x 200. (9) to (11) Nuclei x 1000. D = diatom. N = nucleus. Schaeffer (using the name Amoeba dubid) equated the amoebae described by Penard (1902) as A. proteus, even if some of Penard's 'Amoeba proteus^ might not belong to the latter species as now defined. However, on the basis of these preparations as well as Penard's text, I accept Schaeffer's view that A. nitida is a junior synonym of A. proteus. The deformability of the nuclear envelope in 'Amoeba nitida'' is not a strong enough character for a specific separation. To separate the species on that basis would require isolation of a strain identifiable as A. nitida and demonstration that this deformability, leading to marked infolding and invagination, rests on an ultrastructural difference from the nuclear envelope of A. proteus (Flickinger, 1974). Amoeba sp. (Figs 12-1 5) SLIDE NUMBER. 20. 1 2.8.30. TOTAL NUMBER OF AMOEBAE. 2. DESCRIPTION AND REMARKS. This slide bears Penard's comment: 'Grande amibe a noyau curieux. L'un des individus renferme un petit rhizopode encore inconnu.' These amoebae (Figs 12 and 13) are 320 and 398 //m long. The nucleus (Fig. 14) has in each case an irregular outline, with maximum diameters of 46 and 49 //m. Careful focussing showed the thickness to be about half the greatest diameter or less, but the shape appeared to be lenticular rather F. C. PAGE Figs 12-15 Amoeba sp. (12), (13) Whole cells x 200. (14) Nucleus, with rhizopod test out of focus x 1000. (1 5) Rhizopod test, with nucleus out of focus x 1000. N = nucleus. T = test. than discoid. There is a peripheral layer of granules, each slightly more than 1'5/zm in diameter, just beneath the nuclear membrane, and a less distinct, large central mass, which appears granular. Both amoebae contained ingested diatoms; one contained the test (Fig. 1 5) to which Penard's note refers (29 x 22 //m); and both contained what appeared to be small protozoa, which in one amoeba include apparent ciliates. Given the difficulty of judging from fixed preparations alone, one cannot identify these amoebae with certainty, though they seem to be either Amoeba or Polychaos. Amoeba nobilis Penard, 1 902 (Figs 16-21) SLIDE NUMBERS. 04.5.9.2 1 ; 04.5.9.22; 20. 12.8. 1 4. TOTAL NUMBER OF AMOEBAE. 7. PENARD S SLIDES OF GYMNAMOEBIA 16 Figs 16-21 Amoeba nobilis (=Chaos nobile comb, nov.) (16-18) x200. (19) x 150. (20) Nuclei x 1000. (2 1 ) Amoeba apparently infected by fungus x 200. DESCRIPTION AND REMARKS. These are multinucleate Amoebidae. The five amoebae on slide 20.12.8.14 are in the locomotive form; their lengths are from 262 to 446 /mi. The largest amoeba in these preparations, on slide 04.5.9.22 (Fig. 19) is not a locomotive form and has pseudopodia projecting in several directions, with its greatest dimension 524 /^m across, so 10 F. C. PAGE that it would be much longer in locomotion; this resembles the cell described by Penard (1902) at the top of page 66. However, none of these fixed amoebae clearly shows the distal expansion of the pseudopodia pictured by Penard (1902) in Fig. 1 on page 66, which gives it the Polychaos-\ike appearance mentioned by Page (1976). It must be kept in mind that these fixed preparations undoubtedly do not show the pseudopodial form as accurately as do observations of live amoebae. The number of nuclei counted in these seven amoebae ranged from 25 to 100 per amoeba, approximately, in the majority between 42 and 59, but these counts and those of other multinucleate species made from these slides are on the conservative side, since one cannot flatten the cells to observe all nuclei well. One hundred nuclei, 25 in each of four amoebae, had diameters ranging from 8*5 to 13'0//m; the mean diameter varied from 9'8 um in one amoeba to 12'2/im in another. The nuclei (Fig. 20) are spherical to ovoid. The apparent nucleolar material is arranged in several small, irregular bodies on the inner side of the nuclear envelope, but there are also smaller granular bodies and filamentous material, apparently in the inner part of the nucleus. Amongst the ingested material are diatoms and probably small protozoa. One preparation, with the notation 'Avec cryptogames parasites', contains an amoeba, not in the locomotive form, with hypha-like filaments extending from one side to a maximum length of 1 84 jam from the amoeba (Fig. 2 1 ). Penard ( 1 902) discussed this at some length. Von wilier (1913) isolated from an aquarium at Wurzburg an amoeba which, after comparing his material with Penard's, he concluded was A. nobilis. Siemensma (1980) has found a similar amoeba in the Netherlands. Cysts have not been observed by these workers, although cysts are known to occur in the two recognized species of multinucleate Amoebidae, Chaos carolinense (Wilson, 1900) and C. illinoisense (Kudo, 1950) (Chapman- Andresen, 1979). Amoeba nobilis, as seen in these preparations and in Penard's descriptions, is undoubtedly a member of the family Amoebidae as now defined (Page, 1976); It should therefore be known as Chaos nobile (Penard, 1902) comb. nov. • Amoeba laureata Penard. 1 902 (Figs 22-24) SLIDE NUMBERS. 20. 1 2.8.9; 20. 12.8.10 (also contains A. proteus). TOTAL NUMBER OF AMOEBAE. 2. DESCRIPTION AND REMARKS. Observations on this species were limited by the facts that only two amoebae were present and the thickness of the preparations did not permit use of an objective lens above x 10. Furthermore, neither of the amoebae is a normal locomotive form. The amoeba on slide 20.12.8.9 (Fig. 22) is made up of two thick branches and a knobby posterior end. The cell surface is separated from the cytoplasm around much of the periphery of the amoeba and is somewhat wrinkled. However, a comparison with the second amoeba (Fig. 23) suggests that this surface may not be a Thecamoeba-\ike pellicle. Possibly the fixation method is responsible for the separation. The second amoeba likewise has two arms or branches, but much longer and slender, proceeding from a main mass which includes the more or less knobby uroidal region. These amoebae do not look like the one shown by Penard (1902) in Fig. 1, page 132, which is an Amoeba proteus-\ike locomotive form, although Penard states that such a form is very rare in this species. Nor is the villous character of the uroid, described by Penard for this species, discernible in these preparations, perhaps having been distorted in fixation. The length of the thicker amoeba, from the posterior end to the tip of the main branch, is 3 10 //m. The greatest extent of the more slender amoeba, from the tip of one pseudopodium to the tip of the other, is 3 1 4 //m. Although the conditions of observation did not permit a count of the nuclei, it is obvious that there are hundreds per amoeba. (Penard said that the number sometimes exceeds 1000.) PENARD S SLIDES OF GYMNAMOEB1A 11 22 24 Figs 22-24 Amoeba laureata. (22), (23) Whole cells x 200. (24) Enlargement of amoeba in (22) to show nuclei x 500. N = nucleus. The micrographs of the nuclei (Fig. 24) do not show their structure very distinctly, but denser patches around the periphery suggest that many of them have the structure shown by Penard (1902) in his Fig. 4, page 132, with presumed nucleolar material in a few small parietal bodies. The greatest diameter of a nucleus measured on these photomicrographs was about 6'5/zm, somewhat below Penard's figure of 8-10 /urn, which appears to be derived from live material. A few diatoms and possibly a few other unicellular algae were seen, but many inclusions were not identifiable because of the conditions of observation. The taxonomic position of this species is probably not determinable from these preparations alone. The thickness of the amoeba in one preparation recalls Pelomyxa. However, the branching of these two amoebae is uncharacteristic of Pelomyxa. Furthermore, the possession of symbiotic bacteria was considered by Penard a characteristic of Pelomyxa and is so considered today (though bacteria occur in the cytoplasm of some Amoebidae). Penard explicitly mentions their absence in this species, which he would have classified as a Pelomyxa if he had found such endosymbionts. The presence of crystals, reported by Penard, also suggests that this is not a Pelomyxa (Griffin, 196 1 ). Therefore, A. laureata may well be a Chaos, -but the limitations of the available material make it advisable to reserve judgement. 'Amoeba peritissima Penard' (Figs 25-27) SLIDE NUMBER. 06.4.27.3. TOTAL NUMBER OF AMOEBAE. 2. DESCRIPTION AND REMARKS. These are thickly limax-shaped, multinucleate amoebae, if the two individuals available are representative. One cell measured 208 //m long by 68 //m broad; the other, 204 x 73 //m. 12 F. C. PAGE 25 Figs 25-27 "Amoeba peritissima'' (nomen nudurri). (25), (26) Whole cells x 200. (27) Enlargement of part of amoeba in (26), to show nuclei, x 1000. C = conidium. N = nucleus. Since the cytoplasm of one cell was less densely stained than that of the other, observations of nuclei were made on the former. This amoeba (Fig. 26) contained about 200 nuclei, on a conservative count. Observations of nuclear structure were not completely satisfactory. The nuclei (Fig. 27) appeared to contain a compact nucleolus, which at times appeared central and at other times eccentric in the nucleus. The central region of the nucleolus sometimes stained less densely than the outer part, leaving a lacuna. The diameters of 25 amoebae were 6-0 to 8*4 /zm, mostly toward the lower end of that range. One amoeba (Fig. 27) contained a multicellular conidium. The other (Fig. 25), which appeared to be ingesting an object at the time of fixation, contained several truncately bipyramidal crystals of sizes up to 12 x 9 /zm. Although this slide is labelled 'Amoeba peritissima Penard', there is in fact no such specific name in the literature, and use of the name here is not intended as a publication to make it taxonomically available. It is a nomen nudum. A full taxonomic treatment would be possible only if this organism were found again and examined in sufficient numbers. Its generic position is uncertain. The thick limax form resembles Pelomyxa. However, the presence of crystals again suggests that it is not a Pelomyxa (Griffin, 1961). I can neither confirm nor rule out the presence of symbiotic bacteria on examination of these two preserved amoebae, but Penard's use of the generic name Amoeba indicates that he found no symbiotic bacteria. Again, this carefulness of Penard, who was familiar with diverse multinucleate amoebae, contrasts with the loose usage of some recent authors, who would throw all large multinucleate lobose amoebae into the genus Pelomyxa no mather how they differ in light- and electron-microscopical structure and in such basic physiological characters as locomotion and respiration. Amoeba terricola Greeff, 1866 (Figs 28-33) Penard, 1902, pages 104-121; 1905; 1913. SLIDE NUMBERS. 20.12.8.20; 20.12.8.21; 20.12.8.22; 20.12.8.23; 20.12.8.24; 20.12.8.25; 20.12.8.26; 20.12.8.27; 20.12.8.28. TOTAL NUMBER OF AMOEBAE. 16 (not including six of 'forme papyracea'}. DESCRIPTION AND REMARKS. The slides designated as this species include one (20.12.8.25) with the notation 'forme papyracea" and another (20.12.8.28) labelled 'Variete'. As will be PENARD'S SLIDES OF GYMNAMOEBIA 13 Figs 28-33 Amoeba terricola (= Thecamoeba terricold). (28) Whole cell x 500. (29) Amoeba with Diplochlamys in food vacuole x 200. (30) Whole cell x 200. (31) Whole cell labelled 'Variete' x 200. (32) Nucleus x 1000. (33) Bodies like collapsed sporangia or cyst walls in remains of amoeba, on slide labelled 'Parasitee' x 250. DC = Diplochlamys. N = nucleus. seen below, this collection also includes one slide labelled 'Amoeba papyracecC. Penard described the latter as a separate species in 1905, but in 1913 he decided to 'renounce' it and re-unite it with A. terricola. For the sake of clarity, the amoebae on slide 20.12.8.25 will be described under Amoeba papyracea, and the status of that species will be considered there. The present description is therefore derived from the amoebae on the other 'Amoeba terricola'' slides. The maximum dimensions of these amoebae ranged from 94 to 262 //m, but most were 120//m or more, and two of the smallest had been fixed 'apres 32 jours d'isolement', in which time their size may well have decreased. The forms (Figs 28-31) were typical of Thecamoeba, though the majority did not appear to have been in locomotion when fixed, even if they were extended and flattened. The nuclei (Figs 28, 32) were the elongate ellipsoids or ovoids characteristic of the species, with a maximum length : breadth ratio of 2-3 and a mean of 1 •?. In uninucleate cells (one was binucleate) the lengths of the nuclei were 24 to 55 //m (mean 34'7//m), the majority between 24 and 38 //m. Elongate parietal nucleolar pieces, mostly at the ends of the nuclei as seen in living amoebae of the species (Page, 1977), did not stain well, in contrast to the presumed chromatin in the interior of the nucleus. Identifiable ingested material included a few protozoa, including one identified by Penard as a Diplochlamys (Fig. 29) and a few small naked amoebae. 14 F. C. PAGE 34 Figs 34-36 Amoeba papyracea (=Thecamoeba papyracea comb, nov.) (34), (35) Whole cells x 200. (36) Nucleus x 1000. N = nucleus. One slide (20. 12.8.2 1) bore the notation, 'Formation de petits kystes. Enveloppe dechiree.' The interpretation of this preparation is doubtful, which must be said also about 20.12.8.23, bearing the note, Tarasitee' and containing bodies (Fig. 33) which may be empty fungal sporangia but could also be collapsed cyst walls of smaller amoebae. The nuclei of these amoebae are somewhat larger than those found by Page (1977) in English strains of Thecamoeba terricola, but there seems little doubt that these amoebae and those investigated by other authors (comparison in Page, 1977) belong to the same species. Amoeba papyracea Penard, 1905 (Figs 34-36) Penard, 1905, 1913. SLIDE NUMBER. 06.4.27. 1 (see below). TOTAL NUMBER OF AMOEBAE, one on above slide; six of Amoeba terricola forme papyracea'. DESCRIPTION AND REMARKS. Penard's change of mind about the specific status of these organisms has been mentioned under the preceding species. The description given here is derived from both the single amoeba on slide 06.4.27.1 and the six on slide 20.12.8.25, 'forme papyracea' of" Amoeba terricola'. As Penard says in both publications on this species, the amoebae appear more hyaline and more transparent than the usual A. terricola. Both his descriptions and the appearance of the fixed amoebae suggest that they are somewhat less rigid. The lengths of these seven amoebae ranged from 192 to 233 //m; all but one were more than 200 /zm long. The nuclear structure is as shown by Penard in both illustrations. The nucleus (Fig. 36) is an elongate ovoid or ellipsoid. Although Penard describes it as broader than the nucleus of A. terricola, two of the amoebae had nuclei with a length : breadth ratio of 4*2. In the other five, the L : B was between 1-5 and 1-9. Possibly the nucleus is somewhat compressed in one direction. The lengths of the nuclei were 38 to 72 //m, mean 51'4/im. There are no elongate nucleolar bodies as in the typical A. terricola. Rather there are many small spherules, diameter about 1*5 jam, arranged in the outer region of the nucleus, with the greatest concentration toward the poles, so that the central part of the nucleus appears free of them. These spherules at the poles reach to the nuclear membrane, whereas in the typical A. terricola those poles are occupied by the elongate nucleolar bodies. PENARD'S SLIDES OF GYMNAMOEBIA 15 Figs 37-40 Amoeba sphaeronucleolm (= Thecamoeba sphaeronucleolus). (37), (38) Whole cells x 250. (39), (40) Nuclei x 1000. N = nucleus. In some of these amoebae little or no ingested material was evident. In others the food vacuoles contained bacteria, one or two diatoms, fungal conidia, and possibly one or two protozoa and algal filaments. The more hyaline appearance and apparently greater plasticity of these organisms compared with the typical A. terricola could be due partly to their not having ingested many food organisms for some time before observation and fixation. The somewhat greater size than Penard found for A. terricola might likewise be due to their form being less thick and compact because of the paucity of ingested material. The nuclear structure differs from that found in A. terricola by Penard and other workers. I am inclined to consider this a separate species, but examination of living material and possibly investigation of surface fine structure (Page & Blakey, 1979) is advisable. Amoeba sphaeronucleolus Greeff, 1891 (Figs 37^0) Penard, 1902, pages 121-125; 1905; 1913. SLIDE NUMBER. 20. 1 2.9. 19. TOTAL NUMBER OF AMOEBAE. 5. DESCRIPTION AND REMARKS. The concept of this species which we follow today is that of Penard, and there is some uncertainty whether his A. sphaeronucleolus is that of the original author (Greeff, 1891; Page, 1977). The form of Penard's specimens on this single slide agrees with the usual description of the species. The lengths of these five amoebae are 92 //m, 108//m, 143//m, 156//m, and 161 /urn, thus rather large by Penard's statements that he found large individuals to about 150//m but they are often much smaller (Penard, 1902) and that in their maximum elongation they measure 100 to 130//m (Penard, 1913). The nuclei are approximately spherical or ovoid. Four have a single, more or less spherical nucleolus, while the nucleolus of the fifth (Fig. 40) is in two large fragments accompanied by four smaller pieces which may also be nucleolar fragments. The largest dimensions of the five nuclei range from 22 to 30 //m. The nucleolus is quite smooth and more or less homogeneous except sometimes for a few small achromatic lacunae. Ingested bodies include three diatoms and a conidium in one amoeba and apparent bacteria and algae in others. 16 F. C. PAGE These details correspond fairly well with those reported by workers since Penard, though the size of the nucleus and the texture of the nucleolus differ somewhat from those reported for a North American strain by Page (1977). Even given the more homogeneous nature of the nucleolus in Penard's preparations, this species is easily distinguished from Thecamoeba quadrilineata (Carter, 1856), which is a 'smooth' Thecamoeba rather than a 'rough' one (Page, 1977). There may be some variation amongst strains of Thecamoeba from different parts of the world, since the literature suggests variation even within Europe. However, investigators should be alert to the possible existence of more than one species of 'rough' Thecamoeba with a single compact central nucleolus or endosome. Amoeba fibrillosa Greef, 1 89 1 (Figs 41-45) Penard, 1913; mentioned in Penard, 1902, pages 123, 124. SLIDE NUMBERS. 20. 12.8.7 (labelled Amoeba alba 'avec 1 Amoeba fibrillosa')', 20. 1 2.8.8. TOTAL NUMBER OF AMOEBAE. 5. DESCRIPTION AND REMARKS. These multinucleate amoebae have the wrinkled pellicle and general form of a Thecamoeba, even though the form shows a greater variety and, even in the fixed preparations, evidence of a greater fluidity than that of the more typically Thecamoeba-like Amoeba alba (next section). In this respect it may be compared with Thecamoeba proteoides Page, 1976 (Page, 1976, 1977). Long, slender forms occur, sometimes with temporary branching (which can, however, occur occasionally even in the more rigid A. alba; see Fig. 46). Undoubtedly this temporary branching is associated only with a change of direction. Further comments on this character will be found in both Greeff (1891)andPenard(1913). The largest of these amoebae is that on slide 20.12.8.7, shown in Fig. 41, which is 320 /im long though certainly not in the most extended form possible. The lengths of the other four are 228 /*m, 228 /*m, 226 /im, and 158//m, this last one an irregular form. The length : breadth ratio of the larger amoeba in Fig. 43 is 4* 1 , ignoring the lateral pseudopodium near the posterior end, which was probably being withdrawn at the time of fixation. In Fig. 42, the pseudopodium with the hyaline cap (arrow) was undoubtedly the active one, with the other branch being withdrawn in a change of direction at the time of fixation. In the large amoeba in Fig. 41 , 97 nuclei were counted, and in another amoeba 85 could be found. Both these numbers undoubtedly err on the low side. Although Penard (1913) said that the nuclei are 'normalement globuleux' though fairly often elongate, the elongated condition appears normal in these preparations (Fig. 44). Furthermore, observations while focussing suggest that many if not all the more spherical and ovoid forms (Fig. 45) are actually due to polar and oblique views of elongate nuclei. The single central nucleolus has in general the shape of the nucleus, though it often appears even more elongate (with long sides straighter) than does the nucleus. It is sometimes constricted in the middle to a dumbbell-like shape, which appears to be merely another variation and not a prelude to division as Penard (1913) thought. The measurements of 25 nuclei in the largest amoeba ranged from 7'0 x 6-2 /urn to 10*8 x 7*0 /im, with a mean greatest dimension of 8'9 /*m. A food vacuole in one amoeba contains an ingested organism which appears to be an amoeba, itself containing truncately bipyramidal crystals. Another amoeba also contains an ingested organism which appears to be a protozoon. Greeff (1891) did not publish any illustrations of this species. (See remarks on A. alba.) Although I accept that this may well be the same species which Greeff saw, it must be pointed out that Greeff did not consider the nuclei to be elongate but described them as 'in der Regel rund, zuweilen leicht oval'. However, his description of the amoeba as a whole corresponds with this material. Since Greeffs description of Amoeba fibrillosa precedes in PENARD'S SLIDES OF GYMNAMOEBIA 17 Figs 41-45 Amoeba fibrillosa (= Thecochaos fibrillosum comb. nov.). (41) to (43) Whole cells; arrow indicates hyaline cap on main pseudopodium in (42) x 250. (44), (45) Nuclei x 1000. N = nucleus. 18 F. C. PAGE Figs 46^49 Amoeba alba (= Thecochaos album comb. nov.). (46), (47) Whole cells x 250. (48) Nuclei x 2000. (49) Amoeba apparently infected by fungus x 200. the same publication his description of A. alba the former will be the type-species of the new genus being erected for the two. Penard, it will be noted, had more fixed material of A. alba, if this collection is representative. Amoeba alba Greeff, 1891 (Figs 46^9) Penard, 1902, pages 123-125; 1913. PENARD'S SLIDES OF GYMNAMOEBI A 1 9 SLIDE NUMBERS. 04.5.9.17; 20.12.8.1; 20.12.8.2; 20.12.8.3; 20.12.8.4; 20.12.8.5; 20.12.8.6; 20.12.8.7('avec 1 Amoeba fibrillosa'}. TOTAL NUMBER OF AMOEBAE. 20. DESCRIPTION AND REMARKS. This is another multinucleate species of Thecamoebidae, distinct in both locomotive form and nuclear structure from the preceding. A number of these amoebae appear to have been in locomotion when fixed, thus representing the normal locomotive form well. Only the 1 5 amoebae which were apparently fixed before they died of bursting or other causes than fixation were measured. Their lengths were 166 to 276 //m, with a mean of 210/^m; their length : breadth ratios were M to 2*2, with a mean of 1 -4, quite normal proportions for a Thecamoeba. Use of an oil immersion objective and phase contrast optics permitted a closer look at the nuclei than Penard could have and resulted in a more accurate picture of their structure. However, the same problems which Penard encountered remained in counting the nuclei because, as he said, 'la plupart ne deviennent visibles qu'apres compression de 1'Amibe' (Penard, 1902), and compression was, of course, impossible. Attempts to count the nuclei in five favourable specimens yielded results of 94, 100, 102, 145, and 185, in each case certainly below the actual number. In 1902 Penard thought that the number might reach several hundreds, but in 1 9 1 3 he said only that it often exceeded 1 00. The nuclei (Fig. 48) appeared more ovoid/ellipsoid, i.e., more elongate, in some amoebae, and more spherical to ovoid, i.e., less elongate, in others. The more elongate nuclei, 25 from each of the two amoebae, measured from 7'0x5'6/^m to 12'0x5'6//m, with a mean of 9'0 /zm for the greatest diameter. The more spherical nuclei, 25 from one amoeba, had a greatest diameter of 6'5 to 1' 5 /zm, with a mean of 7*0 jam. The presumed nucleolar material was not scattered as spherules through the nucleus, as described by Penard (1913) ('dissemines . . . dans un sue nucleaire') and shown in Fig. 2, page 123, of Penard (1902) and Fig. 7 of Penard (1913). Rather, it was arranged parietally as variously shaped bodies, some band-shaped, which may all have been lobes of one or two parietal bands in each nucleus. These photomicrographs were made with an oil-immersion lens and phase-contrast optics, not available to Penard. Ingested material included apparent algal filaments and a few diatoms. One slide (20. 1 2.8.6) bears the notation 'Parasitee par cryptogame'. This preparation (Fig. 49) contains a more or less rounded amoeba with a mass of branching, non-septate filaments coming out of an invagination. The diameter of these filaments is about 2 /zm or slightly more. Penard presumably examined this amoeba alive before fixing it; otherwise one might question whether the filaments were parasitizing the amoeba or the amoeba ingesting the filaments. With his original description of A. alba, Greeff (1891) published no illustrations, an omission which led Page (1977) to doubt whether Greeffs organism was indeed a Thecamoeba and speculate whether it might not be a Leptomyxa, a fairly common genus of multinucleate amoebae in soil. However, Penard (1902) agreed with Greeff that A. alba is very rare. I have myself never seen a multinucleate Thecamoeba-like organism in many collections from nature and do not know of any reports of them by workers other than Greeff, Penard, and Cash & Wailes (1919). The figure published by Cash & Wailes is not very informative, but their text suggests that they may have had the same species as Penard. They also described A. alba as rare. A consideration of Greeffs description in the light of the Penard slides makes it quite likely that Penard's organism is the same as GreefFs. Amoeba granulosa Gruber, 1885 (Fig. 50) Penard, 1902, pages 46^ 8. SLIDE NUMBER. 06.4.27.2. TOTAL NUMBER OF AMOEBAE. 1 . 20 F. C. PAGE Fig. 51 Fig. 50 Amoeba granulosa x 250. N? = location of possible nucleus. Amoeba lucens (= Saccamoeba lucens), not normal locomotive form, x 800. N = nucleus. DESCRIPTION AND REMARKS. The single amoeba of this species which Penard said that he found in great abundance furnishes little information. Because of the thickness of the mount, it could be examined only with the x 10 objective. At any rate, the amoeba does not appear well preserved. It is flattened but not circular in outline so presumably not dead or moribund when fixed. Inside the narrow hyaloplasmic border which occupies most of the periphery, the cytoplasm is filled with formed elements, somewhat less densely packed in the central region of the cell. These elements appear to be bipyramidal crystals, as Penard thought. A slightly stained area which may be the nucleus is indicated in Fig. 50 by an arrow. The dimensions of the cell are 142 x 88 /zm; the diameter of the possible nucleus is about 29 /zm. Although this looks like the flattened cells figured by Gruber (1885), the identification is questionable. Gruber gave the diameter as 'ungefahr 0*03 mm', which Penard (1902) mistranslated as 'de 300 /*'. Their descriptions of the nucleus do not appear to agree, although the apparent difference may be due to either optics or terminology. At any rate, one would not like to hazard a guess on the identity of this amoeba, though an amoeba with such an abundance of crystals (undoubtedly not silica, agreeing with Penard rather than Gruber) might be recognizable if found again. It might be mentioned that in this paper Gruber (1885) deplored the impossibility of identifying an amoeba with any degree of certainty. Amoeba lucens (Frenzel, 1 892) (Fig. 51) Penard, 1902, pages 55-57. SLIDE NUMBER. 04.5.9.18. TOTAL NUMBER OF AMOEBAE. 1 . DESCRIPTION AND REMARKS. Although Penard (1902) illustrated his description of this species with a drawing (Fig. 1 , page 56) of what is obviously a Saccamoeba, the single cell on this slide is not so unambiguous. In fact, it might be taken for a Cochliopodium with scales PENARD'S SLIDES OF GYMNAMOEBIA 21 52 Figs 52, 53 Amoeba vespertilio (=Mayorella vespertilio); both figures of same cell, focussed on pseudopodia in (52) and on nucleus in (53) x 1000. N = nucleus. Fig. 54 Unidentified amoeba on same slide with A. vespertilio x 800. N = nucleus. either lost or invisible to the light microscope, if it were not for Penard's label. This cell appears to be made up of a more or less discoid granular mass surrounded by a flattened hyaline border of varying width. The diameter of the granular region is 72 x 65 /^m; including the hyaline border the cell is 90 x 73 //m. The most striking feature is, of course, the truncately bipyramidal crystals, of which there are about a dozen, the largest approximately 13xlO/zm. Some crystals are slightly deteriorated. The nucleus seems poorly fixed, but using Penard's description as a guide, this appears to be a nucleus in which the diameter of the central nucleolus is only a little less than that of the nucleus. The dark area which appears to be the nucleolus has a maximum diameter about 1 5 //m. Although the nuclear membrane (usually quite distinct in Saccamoebd) is not preserved, the narrow clear halo around the nucleolus suggests a maximum nuclear diameter of 19 to 20 //m. The amoeba appears to contain at least one fungal conidium. Despite the puzzling form of this preserved amoeba, Penard's account leaves no doubt of its identity, though this slide is of value chiefly for the structure of the crystals and, to a lesser degree, that of the nucleus. Saccamoeba lucens has been re-described by Bovee (1972) and is recognized as the type-species of the genus Saccamoeba. There is some inconsistency among the descriptions of Frenzel (1892), Penard, and Bovee (1972). Amoeba vespertilio Penard, 1902 (Figs 52-54) SLIDE NUMBER. 20.12.8.29. TOTAL NUMBER OF AMOEBAE. 1 of this species. DESCRIPTION AND REMARKS. Only one of the two organisms on this slide can belong to this species or even to the genus Mayorella Schaeffer, i926, in which Amoeba vespertilio is now classified. The other (Fig. 54) is elongate, apparently with flattened hyaline borders along the sides and with a different nuclear structure. 22 I F. C. PAGE Figs 55-57 Amoeba muralis. (55), (56) Whole cells x 80. (57) Edge of cell, enlarged to show mineral grains x 250. The amoeba which can be identified as A. vespertilio is 65 //m long x 31 //m wide, not including the conical pseudopodia, of which there are six or seven, counting those which have been relegated to the sides. These pseudopodia (Fig. 52), in their fixed and probably somewhat shrunken condition, are up to 9 jum long, with a basal diameter of about 4 //m. The appearance of the uroid (posterior end) suggests that as the amoeba advances the conical pseudopodia, after passing to the posterior end, form a small clump of blunt projections at the uroid before being resorbed. The nucleus (Fig. 53) has a diameter of 12*5 //m and the central nucleolus a diameter of 7-7 //m. The amoeba contains at least one ingested algal cell. Despite the shrinkage accompanying fixation, this amoeba represents the typical Mayorella form much better than do the illustrations on page 94 in Penard (1902). It most resembles, among more recently described species, Mayorella oclawaha Bovee, 1970, and M riparia Page, 1972, though both the amoeba and the nucleus are larger than the sizes reported for those species (Page, 1976). It would not be safe to derive a specific diagnosis from this single cell, since it is not possible even to know whether its length is large, small, or average for the species. Amoeba muralis Penard, 1 909 (Figs 55-57) SLIDE NUMBER. 20. 12.8.11. TOTAL NUMBER OF AMOEBAE. 1 1 . DESCRIPTION AND REMARKS. Because Penard (19090) considered this a naked amoeba rather than one with a flexible test and accordingly placed it into the genus Amoeba, it is included in this study. This preparation contains eleven very flattened, mostly circular cells arranged in a ring. Four appear to have disintegrated, with their positions now marked chiefly by the foreign material which had covered their surfaces, though the outlines are still distinct enough for measurements of diameter. The thickness of the preparation did not permit use of an PENARD'S SLIDES OF GYMNAMOEBIA 23 objective above x 10, but it is doubtful that a higher magnification would have yielded more information. All the amoebae were covered with foreign matter, apparently mineral grains with a maximum dimension of 5 to 10 //m. There are in some cells patches of denser material, probably internal and possibly the remains of ingested algal or other plant matter. The edge of the cell (or its endogenous covering) appears as a clear border extending 5 or 10 jum beyond the mineral grains around much of the periphery of some cells, with some extraneous particles helping to mark its outer edge. I could not find any nuclei, of which Penard said there might be 40, 50, or 60. These cells are marked by multiple parallel streaks, as if scraped during preparation. Penard (19090) described this as a multinucleate amoeba which could secrete a mucilaginous envelope. According to him, this envelope kept particulate matter at a distance from the cell surface. However, when the amoeba began locomotion, the mucilage disappeared, first in the anterior region, then finally from the entire surface. One genus of amoebae with a flexible test which can be detected only with difficulty is Gocevia Valkanov, 1932. The description given by Penard suggests that A. muralis may be covered by a cuticle which may accumulate foreign matter and which may stretch and become thinner during locomotion. His description of fine, digitiform pseudopodia also recalls some recently investigated organisms classified in that genus. The characters of Gocevia with descriptions of organisms which appear to belong to that genus are discussed most recently by Page & Willumsen (1980). Gocevia belongs to the family Cochliopodiidae, in the Testacealobosia. However, all known members of the genus are normally uninucleate except one which may be normally binucleate. Dinamoeba mirabilis Leidy, 1874 (Fig 58, 59) Penard, 1902, pages 134-137; 19096; 1936. SLIDE NUMBER. 04.5.9. 1 54. TOTAL NUMBER OF AMOEBAE. 1 . DESCRIPTION AND REMARKS. This amoeba shows some signs of shrinkage in fixation in that some of the conical pseudopodia appear shrunken in diameter, though the hyaline pseudopodia of this genus are at any rate quite fine even in life (Fig. 59 A, Page, 1976). The cell is 94 /urn long, with a maximum breadth of 37 jum, neither measurement including the pseudopodia. On either side of the anterior end, which appears to have a shallow hyaline cap, is one pseudopodium, with lengths of 15*5 and \9jum. There are several single pseudopodia along the sides, as well as one broad, flat, hyaline projection bearing three short pseudopodium-like extensions. At the posterior end are several uroidal filaments which appear to have originated by adhesion to the substratum but could be pseudopodial remnants. Although Penard (1902, 1909b) emphasized that the organisms which he saw were as a rule binucleate, this cell contains only one nucleus, situated toward the posterior narrowed 'neck' and elongated by cytoplasmic movement. The diameters of the nucleus are 20 x 9 //m; of the compact central nucleolus, about 8*5 x 6*2 /^m. The cytoplasmic pigmentation, including granules, suggests an algal diet, and two or three ingested cells are distinguishable. I could not make out any of the bacterium-like objects, adherent to the surface, which are characteristic of many reported Dinamoeba, though they may also be absent from living amoebae (see Fig. 59 A, Page, 1976). The possible identity of D. mirabilis with Mastigamoeba aspera Schulze, 1875, has been discussed by Penard (19096, 1936), De Groot (1936), and Page (1970). This slide sheds no further light on that question. 24 F. C. PAGE Figs 58, 59 Dinamoeba mirabilis, both of same cell x 800. (58) Phase contrast, to show nucleus distinctly. (59) Differential interference contrast, to show pseudopodia distinctly. N = nucleus. Pelomyxa palust ris Greeff , 1874 (Figs 60, 61) Penard, 1893; 1902, pages 138-143. SLIDE NUMBERS. 04.5.9.206; 04.5.9.210; 06.4.27.7; 20.12.8.530; 20.12.8.531; 20.12.8.532; 20.12.8.533. TOTAL NUMBER OF AMOEBAE. 16. DESCRIPTION AND REMARKS. These are recognizable as the species universally designated by this name. Some of the cells are more or less rounded, others an elongated ovoid, i.e., the usual locomotive form. The longest reached 1478 jum; two 'jeunes individus' are 175 and 233 //m long. In some the mineral grains are so abundant as to hinder observations of other inclusions; some mineral grains measure more than 50 //m, but most are much smaller. Nuclei are numerous. A total of 125 nuclei, 25 in each of five amoebae, had diameters from 7'0 to 14'5 //m, with a mean of 9*2 /^m; only in one of these five amoebae did the diameters exceed 10*8 //m, however. Some of the nuclei (Fig. 61) had a rather shrivelled appearance. In the nuclei there was usually a parietal layer of small granules, sometimes a few larger, darkly staining pieces of various shapes and sizes just beneath the nuclear membrane, sometimes a small body that appeared to be near the centre of the nucleus, and often some rather indistinct filamentous material. The amoebae often contained many diatoms, occasionally filamentous ones. Rods that appeared to be the characteristic symbiotic bacteria were up to 5 jum long. The cytoplasm was often highly alveolar. Since there is no question about the identity of these amoebae and since the characters of the species are well known today, only two of the photomicrographs are reproduced here. PENARD S SLIDES OF GYMNAMOEBIA 25 60 Figs 60, 61 Pelomyxa palustris. (60) Whole cell, with anterior end at top x 100. (61) Nuclei x 1000. N = nucleus. Pelomyxa fragilis Penard, 1 904 (Figs 62-65) SLIDE NUMBERS. 04.5.9.209, 20. 12.8. 529. TOTAL NUMBER OF AMOEBAE. 4. DESCRIPTION AND REMARKS. Since the amoeba on slide 04.5.9.209 was so obscured by detritus that few useful observations could be made, this description is based entirely on the three amoebae on slide 20. 1 2.8.529. Although these amoebae certainly appear more changeable in form than the typical Pelomyxa palustris, none has pseudopodia which can be described, in Penard's term, as 'dechiquetes\ presumably referring to the form in Fig. 2 of Penard (1904). They do have secondary lateral pseudopodia, probably being retracted in a change of direction at the time of fixation. Shallow, crescent-shaped hyaline caps are distinguishable on two amoebae (Figs 62, 64). The uroidal regions of two (Figs 63, 64) appear somewhat drawn out as if by adhesion. The lengths and length : breadth ratios (not including lateral pseudopodia) are: 398 //m, L : B 3'0; 403 /zm, L : B 3-6; and 456 //m, L : B 4-3. These amoebae appear to contain hundreds of nuclei each; in one, there were at least 175 to 200. The diameters of the nuclei, 25 measured in each of the three cells, ranged from 5-4 to 7'7 jum, with a mean of 6'5 /^m. The nuclei (Fig. 65) had a ring of darkly staining material just beneath the nuclear membrane and a roughly spherical or ovoid inner body which might 26 F. C. PAGE Figs 62-65 Pelomyxafragilis. (62) to (64) Whole cells x 200. (65) Nuclei x 1 000. N = nucleus. be central or eccentric; sometimes there appeared to be two of these presumed nucleoli (see Penard, 1904). Ingested material included many diatoms, other unicellular algae, a few short algal filaments, and, in one amoeba, possibly a Colpoda. I could not identify any mineral grains PENARD S SLIDES OF GYMNAMOEBIA 27 Fig 66-69 Pelomyxa vivipara. (66), (67) Whole cells x 250. (68), (69) Nuclei x 1000. N = nucleus. with certainty within the amoebae, although a few possible mineral particles (or glass fragments?) appeared to be adherent to the outer surfaces. Nor could I identify with certainty the symbiotic bacteria, which Penard reported to be abundant. The cytoplasm appeared highly alveolar. Accepting the presence of symbiotic bacteria, these amoebae differ from the typical Pelomyxa palustris, as far as can be determined from these fixed individuals, chiefly in their greater deformability and almost certainly greater pseudopodial activity, and in the absence of ingested mineral particles. It may be that their greater motility is, in fact, due to their not being packed with those particles. The difference in nuclear structure may not be of major importance, considering the variations reported for P. palustris (Daniels & Breyer, 1967; Andresen, Chapman-Andresen & Nilsson, 1968). P.fragilis may therefore well be a synonym of P. palustris. Pelomyxa vivipara Penard, 1 902 (Figs 66-69) SLIDE NUMBER. 20. 12. 8. 534. TOTAL NUMBER OF AMOEBAE. 2. DESCRIPTION AND REMARKS. These two cells look like sacs packed with diatoms. Their measurements are 21 1 x 182//m and 187 x 127/zm, and each is approximately 70-80 /im thick. Along part of the periphery of each cell is a narrow hyaline zone (extending inward up to 12 /zm from the edge). Both are full of diatoms, and in one at least one desmid was seen. No mineral grains are present in either. Useful observations of the nuclei were possible in only one of the two amoebae, which contained well over 60 nuclei. The nuclei (Figs 68, 69) are circular to oval in outline, and much of the stained granular material is parietal in each nucleus. The diameters of 25 nuclei ranged from 7'7 to 9-2^m, with a mean of 8'6/zm. Bacteria-like rods were discernible in the cytoplasm, particularly near the nuclei, as reported by Penard. . Penard described and figured 'embryos' in these amoebae, i.e., small amoebae, which 28 F. C. PAGE might have been parasites or might have been ingested cells which were extruded before digestion. No information on that phenomenon could be gained from this preparation. Modern workers would tend to regard these amoebae as another phase of P. palustris lacking mineral grains, a matter to which reference will be made in connection with the next two species. Pelomyxa belevskii Penard, 1 893 (Figs 70-74) Penard, 1893; Penard, 1902, pages 144-146. SLIDE NUMBER. 04.5.9.207. TOTAL NUMBER OF AMOEBAE. 1 . DESCRIPTION AND REMARKS. This single cell measures 470 x 398 jum and appears considerably compressed. The first thing which strikes the eye is the ingested plant matter, all apparently derived from vascular plants, which Penard (1902) identified as decaying leaf fragments. The cell membrane is somewhat wrinkled around much of the periphery. Focussing carefully, one can find at some places on the cell fine projections, sometimes sharply pointed and single, at other times broader, irregular, and somewhat divided, if indeed these two kinds of projections are the same thing, as Penard thought (Figs. 72, 73). At one point there is also a broad, flat lobe bearing many fine projections (Fig. 71). On this preserved cell, the projections showed up best with differential interference contrast, and one must wonder at Penard's visual acuity that he could make them out with his optical system. This amoeba contained no mineral grains, and the presence of symbiotic bacteria was not confirmed, though one must accept Penard's report of their presence. Twelve nuclei (Fig. 74) were seen, with finely granular material forming a layer against the inner surface of the nuclear membrane, though an area to one side of a nucleus might appear free of it, perhaps as the result of shrinkage during fixation. The nuclear diameters were strikingly greater than in the preceding three species of Pelomyxa, ranging from 24 to 29 /zm, with a mean of 26 um. The fact that some of the tiny projections from the surface (aiguillons or asperites) are very fine suggests the possibility that these are actually non-motile flagella of the type reported by Griffin (1972, 1979) for P. palustris, but the broad lobe in Fig. 71 has the appearance of an adhesion uroid with pseudovilli, and one cannot be certain that at least some of the other projections are not such pseudovilli, though the single ones look somewhat more like flagella. P. belevskii is one of the named species which Chapman-Andresen (1978) has suggested may be a phase of P. palustris. Pelomyxa binucleata (Gruber, 1885) (Figs 75-81) Penard, 1902, pages 147, 148. SLIDE NUMBERS. 04.5.9.208; 20. 12.8.528. TOTAL NUMBER OF AMOEBAE. 13. DESCRIPTION AND REMARKS. These amoebae are generally ovoid to ellipsoid, usually so packed with algae as to resemble sacs pushed out here and there by the ends of filaments. None of the uroidal villi pictured by Penard could be seen in the fixed preparations, although the posterior end is sometimes a little morulate or shrivelled. The lengths were from 96 to 240 //m, with a mean of 1 59 urn; length : breadth ratios 1 -0-1-9, mean 1 -4. Some of the many contained algal filaments are bent, but none appear to be reflected back upon themselves. Some amoebae also contain diatoms and other unicellular algae. Apparent symbiotic PENARD S SLIDES OF GYMNAMOEBIA 29 Figs 70-74 Pelomyxa belevskii (70) Whole cell, showing ingested remains of vascular plants and several nuclei, two of which are in focus at right of picture x 150. (71) Fine projections on both broad lobe (upper right of picture) and main cell body x 1000. (72), (73) Single and more complex projections from cell surface x 1000. (74) Two nuclei x 1000. N = nucleus. bacteria could be distinguished with differential interference contrast. No mineral grains were identified with certainty. All the amoebae are binucleate, with the two nuclei fairly close together in some cells but widely separated in others. Generally the diameters of the two nuclei in a given cell are similar, and the differences are slight enough to be accounted for by angle of viewing. The 26 nuclei had maximum diameters ranging from 19 to 34 //m, with a mean of 24'3 /zm. Some nuclei were isodiametric, but in others the two diameters measured differed slightly. The 30 F. C. PAGE 78 Figs 75-81 Pelomyxa binudeata. (75) Whole cell showing both nuclei x 250. (76), (77) Cells distorted by ingested algal filaments x 250. (78), (79) Same nucleus at different levels to show both parietal arrangement and reticulate form of nucleolar material x 1000. (80), (81) Another nucleus at two levels to show nucleolar structure x 1 000. N = nucleus. presumed nucleolar material was generally parietal, sometimes appearing to exist as fragmented bodies, sometimes as granules or larger clumps. However, in the best-preserved nuclei, what appeared to be small nucleolar pieces in one optical plane could be seen on focussing into another plane to be part of a reticulum just inside the nucleus, as shown in Figs. 78-81. Again, one cannot say definitely whether P. binudeata is a distinct species or fits into the cycle of phenotypic change in P. palustris (Chapman-Andresen, 1978). The nuclei appear distinctive, but one could explain the absence of mineral grains associated with a shape differing from that of P. palustris as a stage on the way to maturation. Generic diagnosis Phylum SARCOMASTIGOPHORA Subphylum SARCODINA Superclass RHIZOPODA Class LOBOSEA Subclass GYMNAMOEBIA Order AMOEBIDA Family THECAMOEBIDAE Genus THECOCHAOSnov. PENARD'S SLIDES OF GYMNAMOEBIA 31 DERIVATION OF NAME. From Theco- + Chaos because Thecochaos (multinucleate) bears, a relationship to Thecamoeba (uninucleate) in the family Thecamoebidae similar to that of Chaos (multinucleate) to Amoeba (uninucleate) in the family Amoebidae. DIAGNOSIS. Broad, flattened, often irregularly oval to oblong in outline but sometimes more elongate, always with length greater than breadth in locomotion, with surface folds and wrinkles and light-microscopical appearance of a thickened pellicle; hyaloplasm a more or less crescentic cap at anterior end, sometimes with slender lateral extensions; branching usually only when changing direction; multinucleate. Essentially a multinucleate Thecamoeba. TYPE-SPECIES. Thecochaos fibrillosum (Greeff , 1891). Classification Subclass GYMNAMOEBIA Order AMOEBIDA Family AMOEBIDAE Amoeba proteus Leidy, 1878 (including A nitida, junior synonym) Amoeba sp. Chaos nobile (Penard, 1902) comb. nov. Family THECAMOEBIDAE Thecamoeba terricola (Greeff, 1866) Thecamoeba papyracea (Penard, 1905) comb. nov. Thecamoeba sphaeronucleolus (Greeff, 1891) Thecochaos fibrillosum (Greeff, 1 89 1 ) comb. nov. Thecochaos album (Greeff, 1891) comb. nov. Family HARTMANNELLIDAE Saccamoeba lucens Frenzel, 1892 Family PARAMOEBIDAE Mayorella vespertilio (Penard, 1902) (Dinamoeba mirabilis Leidy, 1874?) Order PELOBIONTIDA Family PELOMYXIDAE Pelomyxa palustris Greeff, 1 874 (Validity of other species of Pelomyxa questionable.) Incertae. sedis: Amoeba granulosa, Amoeba laureata, Amoeba muralis, 'Amoeba peritissima' (nomen nudum). References Andresen, N., Chapman- Andre-sen, C. & Nilsson, J. R. 1968. The fine structure of Pelomyxa palustris. C. R. Trav. Lab. Carlsberg36 : 286-317. Bovee, E. C. 1972. The lobose amebas. IV. A key to the order Granulopodida Bovee & Jahn, 1966, and descriptions of some new and little-known species in this order. Arch. Protistenk. 114 : 371-403. Cash, J. 1905. The British freshwater Rhizopoda and Heliozoa 1. London. & Wailes, G. H. 1919. The British freshwater Rhizopoda and Heliozoa 4. London. Chapman-Andresen, C. 1978. The life cycle of Pelomyxa palustris. J. Protozool. 25 : 42 A (Abstract only.) 1979. Comparative cytology of the genera Amoeba, Chaos and Pelomyxa. In S. H. Hutner, Protozoological actualities — 7977. Lawrence, Kansas. Daniels, E. W. & Breyer, E. P. 1967. Ultrastructure of the giant amoeba Pelomyxa palustris. J. Protozool. 14: 167-179. 32 F. C. PAGE Deflandre, G. 1958. Eugene Penard (1855-1954). Correspondance et souvenirs. Bibliographic et bilan systematique de son oeuvre. Hydrobiologia 10 : 2-37. De Groot, A. A. 1936. Einige Beobachtungen an Dinamoeba mirabilis Leidy. Arch. Protistenk. 87 : 427^36. Flickinger, C. J. 1974. The fine structure of four 'species' of Amoeba. J. Protozool. 21 : 59-68. Frenzel, J. 1892. Untersuchungen liber die mikroskopische Fauna Argentiniens. Bibliotheca Zoologica 4(12): 1-82. Greeff, R. 1981. Uber den Organismus der Amoben. Biol. Centralbl. 1 1 : 633-640. Griffin, J. L. 1961 . Identification of ameba crystals. II. Triuret in two crystal forms. Biochim. Biophys. Ada 46 : 433-439. 1972. Microfilaments, microtubules, pseudopods, villi, cilia, and cytoplasm of various amoebae. J. Cell Biol. 55(2, part 2) : 96a (Abstract only.) 1979. Flagellar and other ultrastructure of Pelomyxa palustris, the giant herbivorous amoeboflagellate: more evidence for evolutionary distance from carnivores. Trans. Amer. Micros. Soc. 98 : 1 57, 1 58. (Abstract only.) Grospietsch, T. 1975. List of the slide collection made by E. Penard, which is deposited in the Museum d'histoire naturelle at Geneva. Informative Bulletin of Testacean Workers, No. 3 : appendix (unpaginated). Gruber, A. 1885. Studien iiber Amoben. Zeitschr. wiss. Zool. 41 : 186-225. Heal, O. W. 1965. Some slide collections of Protozoa, especially Testacea. J. Quekett Micros. Club 30 : 1-6. Leidy, J. 1 879. Fresh-water rhizopods of North America. Washington. Page, F. C. 1970. Mastigamoeba aspera from estuarine tidal pools in Maine. Trans. Amer. Micros. Soc. 89: 197-200. 1976. An illustrated key to freshwater and soil amoebae. Ambleside. 1977. The genus Thecamoeba (Protozoa, Gymnamoebia). Species distinctions, locomotive morphology, and protozoan prey. J. nat. Hist. 11 : 25-63. & Blakey, S. M. 1979. Cell surface structure as a taxonomic character in the Thecamoebidae (Protozoa: Gymnamoebia). Zool. J. Linn. Soc. 66 : 1 13-135. & Willumsen, N. B. S. 1980. Some observations on Gocevia placopus (Hiilsmann, 1974), an amoeba with a flexible test, and on Gocevia-\ike organisms from Denmark, with comments on the genera Gocevia and Hyalodiscus. J. nat. Hist. 14:41 3-43 1 . Penard, E. 1890. Etudes sur les Rhizopodes d'eau douce. Mem. Soc. Phys. Hist. nat. Geneve 31 : 1-230. 1893. Pelomyxa palustris et quelques organismes inferieurs. Arch. Sci. phys. nat. Geneve (3rd series) 29: 165-1 84. — 1902. Faune rhizopodique du bassin du Leman. Geneve. 1904. Quelques nouveaux Rhizopodes d'eau douce. Arch. Protistenk. 3 : 391-422. 1905. Observations sur les Amibes a pellicule. Arch. Protistenk. 6 : 1 75-206. 1909a. Sur quelques Rhizopodes des mousses. Arch. Protistenk. 17 : 258-296. 19096. Sur quelques Mastigamibes des environs de Geneve. Rev. suisse zool. 17 : 405-438. 1913. Nouvelles recherches sur les Amibes du groupe Terricola. Arch. Protistenk. 28 : 78-140. 1936. Rhizopode ou flagellate? Quelques reflexions a propos de la Dinamoeba mirabilis. Bull. Soc. franc. Micros. 5 : 136-140. Schaeffer, A. A. 1916. Notes on the specific and other characters of Amoeba proteus Pallas (Leidy), A. discoides spec, nov., and A. dubia soec. nov. Arch. Protistenk. 37 : 204-228. 1926. Taxonomy of the amebas. Washington. Siemensma, F. J. 1980. Amoeben. Natura 77 : 62-72. Vonwiller, P. 1913. Uber den Bau der Amoben. Arch. Protistenk. 28 : 389-410. Manuscript accepted for publication 2 September 1980. British Museum (Natural History) An Atlas of Freshwater Testate Amoebae C. G. Ogden & R. H. Hedley 1980, Hardcovers, 222pp, £17.50 (£18.00 by post). Co-published by British Museum (Natural History) and Oxford University Press. This book illustrates, using scanning electron micrographs, most of the common species of testate amoebae that are found in freshwater habitats. Information on the biology, ecology, geographical distribution and a classification are followed by descriptions of ninety-five species. Each of these is illustrated by several views of the shell. The text is designed not only to enable biologists to identify species of testate amoebae, but to serve as an introduction to students interested in the taxonomy and biology of these freshwater protozoa. It will be of special interest to protozoologists, ecologists, limnologists, water treatment specialists and micropalaeontologists interested in recent sediments. British Museum (Natural History) Publication Sales, Cromwell Road, London SW7 5BD. Titles to be published in Volume 40 Eugene Penard's Slides of Gymnamoebia : re-examination and taxonomic evaluation. By F. C. Page Japanese earthworms: a synopsis of the Megadrile species (Oligochaeta). By E. G. Easton Phylogenetic versus convergent relationship between biscivorous cichlid fishes from Lakes Malawi and Tanganyika. By M. L. J. Stiassny Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural History) Japanese earthworms: a synopsis of the Megadrile species (Oligochaeta) E. G. Easton Zoology series Vol 40 No 2 30 April 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries sould be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull, Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1981 ISSN 0007-1498 Zoology series Vol 40 No 2 pp 33-65 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 30 April 1981 - 5 MAV ti Japanese earthworms: a synopsis of the Megadrile species (Oligochaeta) E. G. Easton British Museum (Natural History), London SW7 5BD Contents Introduction 33 Geographical affinities of the Japanese earthworm fauna 34 Classification and Checklist of Japanese earthworms 35 Taxonomy 36 Moniligastridae 37 Biwadrilidae 39 Lumbricidae 40 Ocnerodrilidae 44 Acanthodrilidae 45 Octochaetidae 46 Megascolecidae 46 References 61 Introduction Earthworms play an important part in the soil ecosystem where they participate in organic matter cycles and improve soil structure. They make nitrogen available for plant growth by feeding on organic material in the soil and voiding casts which have a low C/N ratio; in addition the casts contain fragmented litter which is readily broken down by micro- organisms to produce further nitrogen for plant growth. During ingestion the soil particles are ground down in size while the subsequent casts produce a turnover of the soil. Their burrows improve soil aeration and drainage. Although the activities of earthworms are beneficial to man, the worms may be vectors of protozoan, cestode or nematode parasites of mammals and birds which commonly infest pigs and poultry. Earthworms may be very numerous in suitable habitats. One of the highest populations records was that of 845 individuals per square metre (total live weight 245 gm) from an orchard in Europe (Raw, 1959) and populations of 250 per square metre are frequently encountered in grasslands. Over 70 species of earthworms have been recorded from Japan and represent seven of the seventeen currently recognized families of the 'Megadrilacea'. In view of the importance of the group in soil fertility and their common occurrence, it is surprising that no comprehensive study has been published on the rich earthworm fauna of Japan. This situation has compelled the student needing to identify specimens, laboriously to search for matching descriptions scattered throughout the scientific literature beginning with Michaelsen's monograph (1900). A task made all the more difficult by numerous nomenclative changes which have taken place during the last 80 years. It is intended that the present work should go some way to make good this omission and provide a preliminary guide to the many species which have been reported. The Japanese species of the families Biwadrilidae, Lumbricidae, Ocnerodrilidae, Acanthodrilidae and Octochaetidae have been studied extensively but not so the species of the genus Drawida (family Moniligastridae) and of the Pheretima group of genera (Amynthas, Metaphire, Pheretima, Pithemera and Poly- Bull. Br. Mus. nal. Hist. (Zool.) 40 (2) : 33-65 Issued 30 April 1 980 34 E. G. EASTON pheretima in the family Megascolecidae) which together constitute the majority of the Japanese earthworm fauna. The present synopsis brings together the records of all earth- worms reported from Japan under their currently accepted names. Distributions within Japan and where appropriate, elsewhere in the world are also given. Diagnoses and keys are provided for the identification of all indigenous and introduced earthworms with the excep- tion of the poorly known species of the genus Drawida for which only the diagnostic characters are tabulated. Information on species of the Pheretima group of genera is derived from current revisionary studies (Easton, 1979 and in preparation). Geographical affinities of the earthworm fauna of Japan The autochthonous (indigenous) species of the Japanese earthworm fauna have diverse origins and geographical affinities. Although Japan forms part of the Holoarctic geographical region, only one indigenous earthworm family, the Lumbricidae, has a holoarctic distribu- tion. The Lumbricidae occur naturally in the eastern parts of North America and throughout the Palaearctic. The majority of species of this family have been recorded from the western Palaearctic; a few are known to be indigenous to Siberia and only a single species, Eisenia japonica, occurs naturally in Japan at the easternmost limit of the family range. The presence of indigenous taxa in both North America and the Palaearctic suggests that the family is of considerable antiquity, predating the formation of the North Atlantic during the Eocene. Several species of Lumbricidae are allochthonous (peregrine) and some, listed below (p. 35), have successfully colonized Japan, presumably after introduction by man. Elements of the fauna of the Oriental Region are represented in Japan by eight species of the genus Drawida, family Moniligastridae, which is also present in Korea, Manchuria and eastern Siberia as well as most of the Oriental Region. The family Moniligastridae contains four other genera, these inhabit India and southeast Asia including the Philippines and the islands to the west of Wallace's line. It is apparently of recent Indian origin and has affinities with the African family Alluroididae (Jamieson, 1978). Possibly the family invaded Asia after the collision of the Indian and Asian plates during the Tertiary period. The majority of the earthworms in Japan belong to the Pheretima group of genera (family Megascolecidae), the dominant earthworm group throughout southern mainland Asia, the Indo-Australasian Archipelago and the islands of the south-western Pacific. Most other Megascolecid genera occur in India or Australia and it has been suggested that the Pheretima group originated in the New Guinea/North Australia area and invaded Asia by way of the Indo-Australasian Archipelago during the Miocene or Oligocene (Easton, 1979). Several species of the Pheretima group are allochthonous, for example the. Indonesian Poly- pheretima elongata, and have been introduced into Japan by man. The family Biwadrilidae is of uncertain zoogeographical provenance being known from the single species Biwadrilus bathybates which is restricted to Japan and has the distinction of being the only member of the superfamily Biwadriloidea. This species has several primi- tive characters (male pores on segment 13; an unspecialized morphology and lateral lines) and is aquatic. Its Japanese distribution is considered to be a relict of a once, more wide- spread range (Sims, 1980). Other Japanese species cannot be included in the groups discussed above. Pontodrilus matsushimensis (family Acanthodrilidae) is littoral and has also been recorded from New Caledonia and Chatham Island; it may occur on many other beaches of the Pacific. The four species included in the genus are found on beaches of the tropical and warmer temperate regions of the world. Three other genera recorded from Japan, Ocnerodrilus (Ocnerodrilidae), Microscolex (Acanthodrilidae) and Dichogaster (Octochaetidae) are allochthonous (not indigenous to the region) and have been introduced from tropical countries through the agency of man. JAPANESE EARTHWORMS 35 Classification and checklist of Japanese earthworms (After Sims, in press) Order Moniligastrida Family Moniligastridae Drawida hattamimizu Hatai, 1930 D.japonica Michaelsen, 1892 D. keikiensis Kobayashi, 1938 D. koreana Kobayashi, 1938 D. moriokaensis Ohfuchi, 1938 D. nemora Kobayashi, 1936 D. onfunatoensis Ohfuchi, 1938 D. tairaensis Ohfuchi, 1938 Order Haplotaxida Suborder Lumbricina Superfamily Biwadriloidea Family Biwadrilidae Biwadrilus bathybates (Stephenson, 1917) Superfamily Lumbricoidea Family Lumbricidae Aporrectodea trapezoides species complex A. caliginosa (Savigny, 1826) A. trapezoides (Duges, 1828) Bimastos parvus Eisen, 1874 Dendrobaena octaedra (Savigny, 1 826) Dendrodrilus rubidus (Savigny, 1 826) Eiseniafetida (Savigny, 1826) E.japonica (Michaelsen, 1892) E. rosea (Savigny, 1 826) Lumbricus sp. Superfamily Megascolecoidea Family Ocnerodrilidae Ocnerodrilus occidentalis Eisen, 1 878 Family Acanthodrilidae Microscolex phosphoreus (Duges, 1837) Pontodrilus matsushimensis lizuka, 1898 Family Octochaetidae Dichogaster bolaui (Michaelsen, 1891) D. saliens (Beddard, 1893) Family Megascolecidae Amynthas acinctus (Goto & Hatai, 1 899) A. corticus (Kinberg, 1867) A.flavescens (Goto & Hatai, 1 898) A. glabrus (Gates, 1932) A. gracilis (Kinberg, 1867) A. habereri (Cognetti, 1906) A. hilgendorfi species-complex A. agrestis (Goto & Hatai, 1 899) A. ambiguus (Cognetti, 1906) A. communissimus (Goto & Hatai, 1899) A. glandularis (Goto & Hatai, 1899) A. gomejimensis (Ohfuchi, 1937) A. hilgendorfi (Michaelsen, 1892) 36 E. G. EASTON A. irregularis (Goto & Hatai, 1 899) A. levis (Goto & Hatai, 1 899) A. rokugo (Beddard, 1 892) A. schizoporus (Goto & Hatai, 1898) A. sieboldi lenzi (Michaelsen, 1 899) A. tappensis (Ohfuchi, 1935) A. tokioensis (Beddard, 1892) A. vittatus (Goto & Hatai, 1898) A. yunoshimensis (Hatai, 1930) A. hupiensis (Michaelsen, 1895) A. illotus species-group A. 'illotus' (Gates, 1932) 'Pheretima ' oyuensis Ohfuchi, 1937 Amynthas pusillus (Ohfuchi, 1956) A.japonicus(HoTst, 1883) A. megascolidioides (Goto & Hatai, 1899) A. micronarius (Goto & Hatai, 1898) A. minimus (Horst, 1893) A. morrisi (Beddard, 1892) A. obscurus (Goto & Hatai, 1898) A. papulosus (Rosa, 1 896) A. parvicystis (Goto & Hatai, 1 899) A. robustus (Perrier, 1872) A. scholasticus (Goto & Hatai, 1 898) Metaphire californica (Kinberg, 1 867) M.fuscata (Goto & Hatai, 1 898) M.hataii (Ohfuchi, 1937) M. parvula (Ohfuchi, 1956) M. peguana (Rosa, 1 890) M. riukiuensis (Ohfuchi, 1957) M. schmardae (Horst, 1883) M. servina (Hatai & Ohfuchi, 1937) M. sieboldi (Horst, 1883) M. tosaensis (Ohfuchi, 1938) M. yamardai (Hatai, 1930) M. yezoensis (Kobayashi, 1938) Pheretima (Parapheretima) koellikeri Michaelsen, 1928 Pithemera bicincta (Perrier, 1875) Polypheretima elongata (Perrier, 1872) P. iizukai (Goto & Hatai, 1899) Taxonomy Key to the genera of earthworms recorded from Japan 1 Male pores in front of or at anterior margin of clitellum (Intestinal gizzards usually present) Male pores behind or at posterior margin of clitellum (Intestinal gizzards absent) ... 3 2 Male pores on segment 10 or in furrow 10/11 (First of several intestinal gizzard in or before segment 13). . . DRA WIDA (family Moniligastridae) Male pores on segment 1 3* (Intestinal gizzards absent) .... BIWADR1LUS (family Biwadrilidae) "The species Eiseniella letraedra (Lumbricidae) may be confused with Biwadrilus since it has male pores on segment 13. Although it is a widespread peregrine species it has not yet been recorded from Japan. It may be distinguished from Biwadrilus by its more posteriorly placed clitellum (segments 22-37) and the possession of an intestinal gizzard and calciferous glands. JAPANESE EARTHWORMS 37 Male pores on segment 1 5 (A single intestinal gizzard in segments 1 7 or 17-1 8. . . APORRECTODEA,BIMASTOS,DENDROBAENA,DENDRODRILUS, E1SENIA and LUMBR1CUS (family Lumbricidae) 3(1) Setae, 8 on each segment 4 Setae, more than 20 on each segment .... PHERETIMA group of genera (family Megascolecidae) 4 Calciferous glands present 5 Calciferous glands absent 6 5 Calciferous glands in segment 9 or 9 and 1 0 . OCNERODRILUS (family Ocnerodrilidae) Calciferous glands in segments 15,16 and 17. . DICHOGASTER (family Octochaetidae) 6(4) Male pores on segment 17 MICROSCOLEX (family Acanthodrilidae) Male pores on segment 18 PONTODRIL US family Acanthodrilidae) Family MONILIGASTRIDAE DRAWIDA Michaelsen, 1900 DIAGNOSIS. Setae lumbricine. Dorsal pores absent. Clitellum includes segments 10-13. Male pores in or near intersegmental furrow 10/11. Prostates absent. Oesophageal gizzards, calciferous glands and intestinal caeca absent. Several intestinal gizzards in segments 12-27. Excretory system holonephric. INDIGENOUS DISTRIBUTION. Japan, Korea, Manchuria, Siberia, China, India, ?Ceylon, Burma, Thailand, Indo-China, Malaya, Philippines, Borneo. REMARKS. Eight species of Drawida have been recorded from Japan. All are either restricted to Japan or also occur in Korea. [D. japonica has been recorded from outside the Japan/ Korea area but Gates (1935: 3) is of the opinion that these records represent another species.] Oishi (1932: 18) listed five new species of Drawida but omitted to characterize them. A survey of the literature failed to reveal any subsequent descriptions so they are nomina nuda which are therefore outside of nomenclature. None of the eight species considered here is particularly well known and the specific status of each requires closer investigation. Since this appraisal would require consideration of their affinities with non-Japanese species, a project beyond the scope of the present work, only the principal morphological characters of these species are tabulated (Table 1). Drawida hattamimizu Hatai, 1930 hattamimizu Hatai, 1930a:485. Other Japanese records: Ohfuchi, 1938^:1993; Kobayashi, 19416:263. JAPANESE RECORDS. Hokkaido, ISHIKAI (Ohfuchi, 1938). Honshu (Chubu-Chiho) (Kobayashi, 19416), ISHIKAWA-KEN Hatta & Kanazawa (Hatai, 1930): (Kinki-Chiho) (Kobayashi, 19416). DISTRIBUTION. Japan. Drawida japonica Michaelsen, 1892 japonica Michaelsen, 1892 : 232. Other Japanese records: Kobayashi, 19416 : 263, d: 458,6-: 515. JAPANESE RECORDS. 'Japan' (Michaelsen, 1892). Honshu (Kanto-Chiho) (Kobayashi, 19416); TOCHII-KEN Utsunomiya (Kobayashi, 194 Id): (Chubu-Chiho) (Kobayashi, 19416): (Kinki-Chiho) (Kobayashi, 19416): (Chugoku-Chiho) (Kobayashi, 19416). Shikoku (Kobayashi, 19416). Kyushu (Kobayashi, 19416 & e)\ NAGASAKI-KEN Iki (Kobayashi, 19416). DISTRIBUTION. Japan and Korea. 38 5 a a -s: -i \ a ex o • — i :3 o «oo o ^ I c w I s «*- O C at o ke koreana — 0 ^ .52 . .a :a o -~ ^ <5 — ex o ex 2 & -H ci — o. o o. i m >> ja oil « O '•3 «j •- 2 S'G .Is. S2 cs H • 7 ell moriokaen — . O. 0 CX 1 £ia> 00> rs given above ecal pores slig •2 $ o a Cha . 6 -ts 1 • = > o § S"0 o ju Jd, ^ c JAPANESE EARTHWORMS 39 Drawida keikiensis Kobayashi, 1938 keikiensis Kobayashi, 1938a : 107. Japanese records: Kobayashi, 19416 : 263. JAPANESE RECORDS. Honshu (Chubu-Chiho) (Kobayashi, 19416): (Kinki-Chiho) (Kobay- ashi, 19416): (Chugoku-Chiho) (Kobayashi, 19416). Shikoku (Kobayashi, 19416). Kyushu (Kobayashi, 19416). DISTRIBUTION. Japan and Korea. Drawida koreana Kobayashi, 1938 koreana Kobayashi, 1938a : 102. Japanese records: Kobayashi, \94le : 515. JAPANESE RECORDS. Kyushu (Kobayashi, 1 94 1 e). DISTRIBUTION. Japan and Korea. Drawida moriokaensis Ohfuchi, 1938 moriokaensis Ohfuchi, 19386 : 44. JAPANESE RECORDS. Honshu (Ou-Chiho) IWATE-KEN Morioka (Ohfuchi, 1938); MIYAGI-KEN Tsukinoki (Ohfuchi, 1938). DISTRIBUTION. Japan. Drawida nemora Kobayashi, 1936 nemora Kobayashi, 1936 : 141. Japanese records: Kobayashi, 19416 : 263. JAPANESE RECORDS. Honshu (Chubu-Chiho) (Kobayashi, 1 94 1 6). DISTRIBUTION. Japan and Korea. Drawida ofunatoensis Ohfuchi, 1938 ofunatoensis Ohfuchi, 19386 : 33. JAPANESE RECORDS. Honshu (Ou-Chiho) IWATE-KEN coast & islands of Sanriku (Ohfuchi, 19386); MIYAGI-KEN some regions of coast (Ohfuchi, 19386). DISTRIBUTION. Japan. Drawida tairaensis Ohfuchi, 1938 tairaensis Ohfuchi, 19386 : 39. JAPANESE RECORDS. Honshu (Ou-Chiho) AKITA-KEN Honjo (Ohfuchi, 19386); IWATE-KEN Morioka (Ohfuchi, 19386); MIYAGI-KEN Tsukinoki (Ohfuchi, 19386); FUKUSHIMA-KEN Taira machi (Ohfuchi, 19386). DISTRIBUTION. Japan. Family BIWADRILIDAE BIWADRILUS Jamieson, 1971 DIAGNOSIS. Setae lumbricine. Dorsal pores absent. Clitellum includes segments 15-31. Prostates absent. Male pores on segment 13. Oesophageal gizzards, intestinal gizzards, calci- ferous glands and intestinal caeca absent. Excretory system holonephric. INDIGENOUS RANGE- Japan. 40 E. G. EASTON Biwadrilus bathybates (Stephenson, 1917) Criodrilus bathybates Stephenson, 1917 : 96. Other Japanese records: (syn. miyashitai) Yamaguchi, 1953:309. C. miyashitai Nagase & Nomura, 1937 : 36 1 . DIAGNOSIS. As for the genus. JAPANESE RECORDS. Honshu (Ou-Chiho) YAMAGATA-KEN Tsuruoka (Nagase & Nomura, 1937): (Chubu-Chiho) KYOTO-KEN Biwa-ko (Stephenson, 1917; Yamaguchi, 1953); KYOTO- FU Komori (Nagase & Nomura, 1937); OSAKA-FU Minamitoyoshima (Yamaguchi, 1953): (Kinki-Chiho) HYOGO-KEN Muko (Yamaguchi, 1953). DISTRIBUTION. Japan. Family LUMBRICIDAE DIAGNOSIS. Setae lumbricine. Dorsal pores present. Clitellum usually begins behind segment 22, rarely (Eisenielld) segment 15. Male pores usually on segment 15, rarely (Eisenielld) segments 12, 13 or 14. Prostates, oesophageal gizzards and intestinal caeca absent. A single intestinal gizzard in segment 17 or 17 and 18. Excretory system holonephric. INDIGENOUS RANGE. Palaearctic and eastern North America. REMARKS. The currently recognized Lumbricid genera are defined principally on difficult to observe somatic structures. Since the Japanese species can be readily identified without recourse to these characters, generic diagnoses are not provided. Of the eight Lumbricid species recorded from Japan, seven are widespread allochthonous forms which may have been introduced into Japan through the agency of man. Only one species, Eiseniajaponica is thought to be indigenous but even this species has been recorded in Europe (Graff, 1954). Full descriptions and distributions of the allochthonous species were provided by Gates (1972a). For a detailed description ofE.japonica see Gates (1975). The names Bimastos, E. fetida were emended to Bimastus and E. foetida respectively by Michaelsen (1900) and used by many subsequent authors. Under the articles of the Inter- national Code of Zoological Nomenclature such emendations are invalid and the original orthography is employed here. Several recent revisions of the family Lumbricidae (Omodeo, 1956; Bouche, 1972, Perel, 1976) include taxonomic changes which effect the Japanese fauna. The genus Allolobophora has been restricted to include only the type species (A. chloroticd); excluded species being accommodated in either Eisenia or Aporrectodea (syn. Nicodrilus Bouche). The species Bimastos tenuis has been placed within the synonymy of rubida which itself has been transferred from Dendrobaena to Dendrodrilus. Some of the results of these revisions are incompatible with one another and with the results of other workers. Often a species is consigned to different genera by different workers. Of the Japanese species, rosea is included in Eisenia for convenience although Perel (1974) excluded it from this genus but did not indicate to which genus it should be assigned. Key to the species of Lumbricidae of Japan 1 Prostomium tanylobic Lumbricus Prostomium epilobic 2 Setae closely paired (aa = 3ab) Setae widely paired (aa= \-2ab) 4 3 Tubercula pubertatis on segments 24, 25, 26-30 or absent Bimastos parvus Tubercula pubertatis on segments 27-29 Eiseniajaponica Tubercula pubertatis on segments 28-30, 31 Eisenia fetida Tubercula pubertatis on segments 29, 30-31 . . Eisenia rosea Tubercula pubertatis on segments 3 1-33 . . Aporrectodea trapezoides species-complex JAPANESE EARTHWORMS 41 4(2) Tubercula pubertatis on segments 28, 29-30, 3 1 (tail cylindrical) . Dendrodrilus rubidus Tubercula pubertatis on segments 3 1-33 (tail octogonal) . . . Dendrobaena octaedra Aporrectodea trapezoides species-complex trapezoides species-group, Gates, 19726 : 1. DIAGNOSIS. Length 60-140 mm. Body cylindrical or flattened posteriorly to form rectangular cross section with setal pairs at corners. Prostomium epilobic. Clitellum on segments 27, 28, 29-33, 34, 35. Tubercula pubertatis on segments 31-33. Male pores on segment 15. Spermathecal pores in furrows 9/10/11 in setal line c. First dorsal pore in a furrow between 6/7 and 13/14. Setae closely paired, setal ratio (caliginosd) 30aa=\Qab = 2Qbc = lcd=\00dd. REMARKS. The close affinities of the component taxa of this complex were first recognized by Gates (19726) when he erected the species-complex to accommodate caliginosa Savigny, 1826 (under the name turgida Eisen, 1873), trapezoides Duges, 1828 and six other species. Aporrectodea caliginosa (Savigny, 1 826) caliginosa Savigny, 1826:180. Japanese records: Kobayashi, 1941a:52, 6:264, c:378, e:5\5; Nakamura, 1967: 164; Nakamura, 1972 : 18; Nakamura, 1973a: 199,6:210. DIAGNOSIS. Length 60-85 mm. Body cylindrical, unpigmented, anterior segments flesh pink, rest of body pale grey. Clitellum on segments 27, 28, 29-34, 35. Genital tumescences incorporating setae a and b on segments 9-1 1 , 30, 32-34 and frequently 27. First dorsal pore in furrow 12/1 3 or 13/1 4. JAPANESE RECORDS. Hokkaido (Nakamura, 1972); ISHIKAI Sapporo (Nakamura, 1967; 1973a & b). Honshu (Ou-Chiho) (Kobayashi, 1941c): (Chubu-Chiho) (Kobayashi, 19416 & c): (Kinki-Chiho) (Kob_ayashi, 19416 & c): (Chugoku-Chiho) (Kobayashi, 19416 & c); SHIMANE-KEN Oki-gunto (Kobayashi, 194 la). Shikoku (Kobayashi, 19416 & c). Kyushu (Kobayashi, 1 94 1 6, c & e); KAGOSHIMA-KEN Shibushi & Yanakawa (Kobayashi, 1 94 1 c). DISTRIBUTION. Cosmopolitan (indigenous to Palaearctic). Aporrectodea trapezoides (Duges, 1828) trapezoides Duges, 1828:289. Japanese records: Michaelsen, 1892:230; Kobayashi, 19386:414; Kobayashi, 1 94 1 6 : 264, d : 459, e : 5 1 5. DIAGNOSIS. Length 80-140 mm. Body flattened posteriorly to form rectangular cross section with setal pairs at corners, slate, brown, reddish brown, often paler ventrally. Clitellum on segments 27, 28-33, 34. Genital tumescences incorporating setae a and 6 on segments 9-1 1 , 32-34 often 27 and occasionally 26, 28, 29. First dorsal pore in a furrow between 6/7 and 13/14. JAPANESE RECORDS. 'Japan' (Michaelsen, 1 892). Hokkaido (Kobayashi, 1 94 1 6 & c); OSHIMA Hakodate (Kobayashi, 1938c). Honshu (Kanto-Chiho) TOCHII-KEN Utsunomiya (Kobay- ashi, 194 Id): (Kinki-Chiho) (Kobayashi, 19416): (Chugoku-Chiho) (Kobayashi, 19416). Shikoku (Kobayashi, 19416). Tsushima (Kobayashi, 19416). Kyushu (Kobay- ashi, 19416&?). DISTRIBUTION. Palaearctic (indigenous), Nearctic, Oriental, Australasian and Neotropical regions. Bimast os parvus Eisen, 1874 parvus Eisen, 1874:46. Japanese records: Michaelsen, 1910:64; Kobayashi, 1941a:52, 6:264, c : 378, e : 515; Yamaguchi, 1953:310; Gates, 1972a : 87. beddardi Michaelsen, 1894 : 182. Japanese records: Kobayashi, 19416 : 264, e : 515. 42 E. G. EASTON DIAGNOSIS. Length 17-65 mm. Body cylindrical, reddish dorsally, yellow ventrally. Pro- stomium epilobic. Clitellum on segments 23, 24, 25-31, 32. Tubercula pubertatis absent or in the form of indistinct paired longitudinal ridges along the ventral margins of the clitellum on segments 24, 25, 26-30. Male pores on segment 15. Spermathecal pores absent. First dorsal pore in furrow 5/6. Setae closely paired, setal ratio (postclitellar segments) 3Qaa = \Qab = 25bc = 9cd = \00dd. JAPANESE RECORDS. 'Japan' (Michaelsen, 1910 — intercepted at Hamburg; Gates, 1972 — intercepted at American port). Hokkaido (Kobayashi, 1941c); ISHIKAI Chitose (Yamaguchi, 1953). Honshu (Ou-Chiho) (Kobayashi, 19416): (Chubu-Chiho) (Kobayashi, 19416 & c): (Kinki-Chiho) (Kobayashi, 19416 & c): (Chugoku-Chiho) (Kobayashi, 19416 & c); SHIMANE-KEN Oki-gunto (Kobayashi, 194 la & 6). Shikoku (Kobayashi, 19416 & c). Tsushima (Kobayashi, 19416). Kyushu (Kobayashi, 19416, c & e); NAGASAKI-KEN Gotto- retto & Iki (Kobayashi, 19416); KAGOSHIMA-KEN Kagoshima & Shibushi (Kobayashi, 1941c). DISTRIBUTION. Cosmopolitan (indigenous to eastern North America). Dendrobaena octaedm (Savigny, 1 826) octaedra Savigny, 1826:183; Gates, 1974a:15. Japanese records: Nakamura, 1967:164; Stop- Bowitz, 1969:214;Tamuraetal, 1969 : 26; Nakamura, 1972: 18; Nakamura, 1973a: 199,6:210. DIAGNOSIS. Length 17^40 mm. Body cylindrical, octagonal posteriorly, red, yellowish, brown, violet1 or copper coloured. Prostomium epilobic. Clitellum on segments 27, 28, 29-33, 34. Tubercula pubertatis in form of a longitudinal ridge on segments 31-33. Male pores on segment 15. Spermathecal pores in furrows 9/10/1 1/12 at seta d. First dorsal pore in a furrow between 4/5 and 6/7. Setae widely paired, aa-ab-bc- cd, dd slightly greater than aa. JAPANESE RECORDS. Hokkaido Stop-Bowitz, 1969; Nakamura 1972); ISHIKAI Sapporo (Nakamura, 1967; 1973a & 6); HIDAKA Hidaka-Mombetsu (Tamura et al, 1969). DISTRIBUTION. Palaearctic (indigenous), Nearctic and Oriental regions. Dendrodrilus rubidus (Savigny, 1 826) rubida Savigny, 1826 : 182. tenuis Eisen, 1874:44. Japanese records: Kobayashi, 19386:415; Kobayashi, 19416:264; Stop- Bowitz, 1969 : 227; Tamura et al, 1969 : 26; Nakamura, 1972 : 18; Nakamura, 19730: 199,6 : 210. DIAGNOSIS. Length 20-90 mm. Body cylindrical, dark red dorsally, pale ventrally and dorsally in the intersegmental furrows. Prostomium epilobic. Clitellum on segments 24, 25, 26, 27-31, 32. Tubercula pubertatis absent or in the form of indistinct paired longitudinal ridges on segments 28, 29-30, 31. Male pores on segment 15. Spermathecal pores absent or in furrows 9/10/11 in setal line c. First dorsal pore in furrow 4/5 or 5/6. Setae widely paired, 1 8aa = 1 Qab = 246c = \2cd = 5Qdd. JAPANESE RECORDS. Hokkaido (Kobayashi, 19416; Stop-Bowitz, 1969; Nakamura, 1972); ISHIKAI Sapporo (Nakamura, 1973a & 6); HIDAKA Hidaka-Mombetsu (Tamura et al, 1969); OSHIMA Hakodate & Nanayehama (Kobayashi, 1938c). Honshu (Chubu-Chiho) (Kobay- ashi, 194 16): (Chugoku-Chiho) (Kobayashi, 19416). DISTRIBUTION. Cosmopolitan (indigenous to Palaearctic). Eiseniafetida (Savigny, 1826) fetida Savigny, 1826: 182. Japanese records: Michaelsen, 1892:230; Sasaki, 1924:89; Kobayashi, 1928 : 468; Kobayashi, 19386 : 415; Ohfuchi, 193&/: 1992; Kobayashi, 19416 : 263, c : 378, d: 459, e: 5 15; Stop-Bowitz, 1969 : 210; Nakamura, 1972: 18; Nakamura, 1973a: 199,6:210. JAPANESE EARTHWORMS 43 DIAGNOSIS. Length 35-130 mm. Body cylindrical, reddish purple or brown dorsally, yellow in intersegmental furrows and ventrally. Prostomium epilobic. Clitellum on segments 24, 25, 26-32. Tubercula pubertatis in the form of paired longitudinal ridges on segments 28-30, 3 1 . Male pores on segment 1 5. Spermathecal pores in furrows 9/10/11, near the dorsal line. First dorsal pores in furrow 5/6. Setae closely paired, 4aa = ab = 4bc = cd=\ 6dd. JAPANESE RECORDS. 'Japan' (Michaelsen, 1892). Hokkaido (Kobayashi, 19416 & c; Stop- Bowitz, 1969; Nakamura, 1972); ISHIKAI (Ohfuchi, 1938) Sapporo (Nakamura, 1973a & 6); OSHIMA Hakodate (Kobayashi, 1938c). Honshu (Ou-Chiho) (Kobayashi, 19416 & c); AOMORI-KEN Aomori (Sasaki, 1924); IWATE-KEN Morioka (Sasaki, 1924); MIYAGI-KEN Sendai (Sasaki, 1924; Kobayashi, 1928): (Kanto-Chiho) (Kobayashi, 19416); TOCHII-KEN Utsunomiya (Kobayashi, 194 Id); TOKYO- TO Tokyo (Sasaki, 1924): (Chubu-Chiho) (Kobayashi, 19416 & c): (Kinki-Chiho) (Kobayashi, 19416 & c): (Chugoku-Chiho) (Kobayashi, 19416 & c). Shikoku (Kobayashi, 19416 & c). Tsushima (Kobayashi, 19416). Kyushu (Kobayashi, 19416, c & e}\ NAGASAKI-KEN Iki (Kobayashi, 19416); KAGOSHIMA-KEN Kagoshima, Shibushi & Yanakawa (Kobayashi, 1941c). Osumi-Gunto YAKU-SHIMA (Kobayashi, 19416&c). DISTRIBUTION. Cosmopolitan (indigenous to Palaearctic) Eiseniajaponica (Michaelsen, 1892) japonica Michaelsen, 1892:230. Other Japanese records: Michaelsen, 1900:481; Michaelsen, 1910:62; Kobayashi, 19386:414; Oishi, 1934:134; Kobayashi, 1941^:52, 6:264, c : 378, d:46Q, e: 515; Nakamura, 1967 : 164; Tamura et al, 1969 : 26; Nakamura, 1971 : 347; Nakamura, 1972: 1 0; Nakamura, 1973a: 199, 6 : 210; Gates, 1975 : 1. japonica f. gigantica Oishi, 1934 : 134. Other Japanese records: Kobayashi, 19416 : 264. japonica f. minuta Oishi, 1934 : 134. DIAGNOSIS. Length 24-1 75 mm. Body cylindrical, whitish grey or dark reddish brown. Prostomium epilobic. Clitellum on segments 23, 24-31. Tubercula pubertatis in the form of paired longitudinal ridges on segments 27-29, often disrupted and on segments 27 and 29 only. Male pores on segment 15. Spermathecal pores in furrows 9/10/11 in setal line c. First dorsal pore in furrow 4/5. Setae closely paired, setal ratio 93aa = 10a6 = 466c = lcd= 3 IQdd. REMARKS. Some authors (Beddard, 1895; Oishi, 1934; Kobayashi, 1941; Nakamura, 1972) have recognized varieties of this species. Their taxonomic validities are dubious nevertheless the diagnostic features tabulated by Kobayashi ( 1 94 1 0 are given below (Table 2). JAPANESE RECORDS. Hokkaido (Kobayashi, 19416 & c; Nakamura, 1972); ABASHIRI Oketo (Nakamura, 1971); KAMIKAWA Furano & Mitsumata (Gates, 1975), Monomanai & Nishi- Shibetsu (Nakamura, 1971); ISHIKAI Misumai, near Sapporo (Nakamura, 1973a), Sapporo (Nakamura, 1967, 19730 & 6); HIDAKA Hidaka-Mombetsu (Tamura et al, 1969); SHIRIBESHI Kutchan & Yotei-zan (Gates, 1975); osHiMA_Hakodate (Michaelsen, 1892); Nanayehama & Hakodate (Kobayashi, 1938c). Honshu (Ou-Chiho) (Kobayashi, 19416 & c): (Kanto- Chiho) (Kobayashi, 19416); TOCHII-KEN Utsunomiya (Kobayashi, 194 Id); KANAGAWA-KEN Eno-shima (Michaelsen, 1892*; Rosa, 1893); Yokohama (Michaelsen, 1910): (Chubu- Chiho) (Kobayashi, 19416 & c); YAMANASHI-KEN Fuji-san (Michaelsen, 1900): (Kinki-Chiho) (Kobayashi, 19416 & c): (Chugoku-Chiho) (Kobayashi, 19416 & c); SHIMANE-KEN Oki-gunto (Kobayashi, 194 la & 6). Shikoku (Kobayashi, 19416 & c). Tsushima (Kobayashi, 19416). Kyushu (Kobayashi, 19416, c & e); FUKUOKA-KEN Moji (Michaelsen, 1910); NAGASAKI-KEN Iki (Kobayashi, 19416); MIYAZAKI-KEN Aoi-dake (Kobayashi, 1941c); KAGOSHIMA-KEN Kagoshima (Kobayashi, 1941c). DISTRIBUTION. Palaearctic (Japan, Korea and Europe). *Michaelsen's original reference to this record was 'Enoshima, Japan'. It is uncertain which of the localities named Enoshima was referred to but Rosa (1893) noted that it was near Tokyo. 44 E. G. EASTON Table 2 Eisenia japonica: marker characters of varieties. (After Kobayashi, 1 94 1/) minuta typica gigantica Length (mm) 24-55 42-102 139-175 Diameter (mm) 1-75-2-80 2-5-5-0 3-5-7-2 Segment number 85-110 96-140 125-151 Colour uniformly anterior pink, dark reddish whitish grey posterior brown whitish grey Shape of tubercula pubertatis round triangular intermediate Size of seta (u) 290 x 24 390 x 36 540 x 47 Eisenia rosea Savigny, 1 826) rosea Savigny, 1826: 182; Gates, 19746:9. Japanese records: Nakamura, 1971 : 347; Nakamura, 1972: 18; Nakamura, 1973a: 199,6:210. DIAGNOSIS. Length 25-85 mm. Body cylindrical, unpigmented, anterior segments dark red, rest of body pink or pinkish grey. Prostomium epilobic. Clitellum on segments 24, 25, 26-32, 33. Tubercula pubertatis on segments 29-30, 31. Male pores on segment 15. Spermathecal pores in furrows 9/10/11, near dorsal line or halfway between lateral line and setal line d. First dorsal pore in furrow 4/5. Setae closely paired, aa greater than be, ab greater than cd, dd= 1/3-1/2 body circumference. JAPANESE RECORDS. Hokkaido (Nakamura, 1972); KAMIKAWA Monomanai & Nishi- Shibetsu (Nakamura, 1971); ISHIKAI Hiroshima (Nakamura, 1973a); Sapporo (Nakamura, 19736). DISTRIBUTION. Cosmopolitan (indigenous to Palaearctic) LUMBRICUS Linnaeus, 1758 Lumbricussp. Ohfuchi, 1941 : 255. DIAGNOSIS. Length 20-300 mm. Body cylindrical, trapezoidal posteriorly, purplish red or purplish brown dorsally, paler ventrally. Prostomium prolobic. Clitellum begins between segment 26 and 39 and occupies 5-1 5 segments. Tubercula pubertatis usually occupies more than 4 clitellar segments. Male pores on segment 1 5. Spermathecal pores in furrows 9/10/11 between setae c and d. First dorsal pore between furrows 5/6 and 9/10. Setae closely paired. JAPANESE RECORDS. Honshu (Chugoku-Chiho) YAMAGUCHI-KEN Shuhodo Akigoshi (Ohfuchi, 1941). DISTRIBUTION. Cosmopolitan (indigenous to Holarctic). Family OCNERODRILIDAE OCNERODRILUSEisen, 1878 DIAGNOSIS. Setae lumbricine. Dorsal pores absent. Clitellum includes segments 14-19. Prostates tubular, discharging through combined male and prostatic pores on segment 17. Oesophageal and intestinal gizzards absent. Calciferous glands in segment 9 or segments 9 and 10. Intestinal caeca absent. Excreatory system holonephric. INDIGENOUS RANGE. Tropical and subtropical America, tropical Africa. Several species are allochthonous, one, O. occidentalis Eisen, 1878 has been recorded from Japan. JAPANESE EARTHWORMS 45 Ocnerodrilus occidental™ Eisen, 1 878 occidentalis Eisen, 1878 : 10. Japanese records: Kobayashi, 1941a : 52, b : 263, c • 378 e • 515- Gates 1973: 16. DIAGNOSIS. As for the genus. JAPANESE RECORDS. 'Japan' (Gates, 1973 interception at American port). Honshu (Kanto- Chiho) KANAGAWA-KEN_O-shima (Kobayashi, 19416 & c): (Chubu-Chiho) (Kobayashi, 19416 & c): (Kinki-ChihoMKobayashi, 19416 & c): (Chugoku-Chiho) (Kobayashi, 19416 & c); SHIMANE-KEN Oki-gunto (Kobayashi, 194 la & b). Shikoku (Kobayashi, 19416 & c). Kyushu (Kobayashi, 19416, c & e); NAGASAKI-KEN Gotto-retto (Kobayashi, 19416); KAGOSHIMA-KEN Kagoshima & Yanakawa (Kobayashi, 1941c). Okinawa-gunto OKINAWA- JIMA (Kobayashi, 1 94 1 6 & c). DISTRIBUTION. Palaearctic, Nearctic, Oriental and Ethiopian regions. Family ACANTHODRILIDAE MICROSCOLEXRosa, 1887 DIAGNOSIS- Setae lumbricine. Dorsal pores absent. Clitellum includes segments 14-16. Prostates tubular, discharging near male pores on segment 17. A weak oesophageal gizzard in segment 5. Calciferous glands, intestinal caeca and intestinal gizzards absent. Excretory system holonephric. INDIGENOUS RANGE. Southern South America, South Africa, Sub-Antarctic Islands. Two species, M. dubius (Fletcher, 1887) and M. phosphoreus (Duges, 1837) are allochthonous, the latter has been recorded from Japan. Microscolex phosphoreus (Duges, 1837) phosphoreus Duges, 1837 : 17. Japanese records: Yamaguchi, 1935 : 200; Kobayashi, 19416 : 263. DIAGNOSIS. As for the genus. JAPANESE RECORDS. Honshu (Kanto-Chiho) (Kobayashi, 19416): (Chugoku-Chiho) (Kobayashi 19416). Shikoku (Kobayashi, 19416). Kyushu KAGOSHIMA-KEN Osio (Yamaguchi, 1935). DISTRIBUTION. Cosmopolitan (? indigenous in South America). PONTODRILUSPerrier, 1874 DIAGNOSIS. Setae lumbricine. Dorsal pores absent. Clitellum begins on segment 14 and occupies 5 or 6 segments. Prostates tubular discharging through combined male and prostatic pores on segment 18. Oesophageal gizzard rudimentary or absent. Calciferous glands, intestinal caeca and intestinal gizzards absent. Excretory system holonephric. DISTRIBUTION. Circum-mundane, on sea shores throughout the tropics and warmer areas of the temperate zones. A single species has been recorded from Japan. Pontodrilus matsushimensis lizuka, 1898 matsushimensis lizuka, 1898 : 21. Other Japanese records: Yamaguchi, 1953 : 309. DIAGNOSIS. As for the genus. JAPANESE RECORDS. Honshu (Ou-chiho) MIYAGI-KEN Matsushima-wan (lizuka, 1898); Miyakojima (Yamaguchi, 1953): (Kinki-Chiho) HYOGO-KEN Akashi (Yamaguchi, 1953). Kyushu FUKUOKA-KEN Fukuoshima (Yamaguchi, 1953); 'shore of Ranshima near Ogura' (Yamaguchi, 1953). 46 E. G. EASTON DISTRIBUTION. Japan, New Caledonia (Beddard, 1 899) and Chatham Island, New Zealand (as var. chathamianus Michaelsen, 1 899). Family OCTOCHAETIDAE DICHOGASTER Bzddard, 1888 DIAGNOSIS. Setae lumbricine. Dorsal pores present. Clitellum includes segments 14-18. Prostates tubular discharging on segments 17 and 19 or 17 only. Male pores on segment 18. Penial setae present. Two well developed oesophageal gizzards anterior to septum 8/9. Paired calciferous glands in segments 15, 16 and 17, intestinal caeca and gizzards absent. Excretory system meronephric. INDIGENOUS RANGE. Tropical Americas, Africa. Several species are allochthonous of which two have been recorded from Japan. Key to species recorded from Japan 1 One pair of prostates which discharge onto segment 17 saliens Two pairs of prostates which discharge onto segments 17 and 19 bolaui Dichogaster bolaui Michaelsen, 1 89 1 bolaui Michaelsen, 189 la : 9. Japanese records: Kobayashi, 1941c : 379. DIAGNOSIS. Length 20-40 mm. Two pairs of prostates discharging on segments 17 and 19. JAPANESE RECORDS. Okinawa-Gunto OKINAWA-JIMA (Kobayashi, 1 94 1 c). DISTRIBUTION. Cosmopolitan (indigenous range unknown). Dichogaster saliens (Beddard, 1893) saliens Beddard, 1893 : 683. hatomaana Ohfuchi, 1957 : 259. DIAGNOSIS. Length 1 7-70 mm. One pair of prostates discharging onto segment 1 7 only. REMARKS. The description of D. hatomaana provided by Ohfuchi (1957) is indistinguishable from that of saliens. JAPANESE RECORDS. Sakishima-Gunto IRIOMOTE-JIMA Hatoma-jima (Ohfuchi, 1957). DISTRIBUTION. Cosmopolitan (indigenous range unknown). Family MEGASCOLECIDAE PHERETIMA group of genera DIAGNOSIS. Setae perichaetine. Dorsal pores present. Clitellum including segments 14-16. Prostates racemose discharging through combined male and prostatic pores on segment 1 8 or rarely 19. Oesophageal gizzard single, well developed in segment 8. Calciferous glands absent. Intestinal caeca usually present. Excretory system meronephric. INDIGENOUS RANGE. Japan, Korea, China, Burma (east of the Chinwin/Irrawaddy axis), Thailand, Vietnam, Malaysia, Indonesia, Philippines, Papua New Guinea, New Britain, New Hebrides, New Caledonia, northern Queensland, islands of the western Pacific (Easton, 1979). REMARKS. Nine closely related genera comprising about 760 nominal species, are included in the Pheretima group of genera. Five have been recorded from Japan: Amynthas Kinberg, 1867; Metaphire Sims & Easton, 1972; Pithemera Sims & Easton, 1972; Pheretima Kinberg, 1867; and Polypheretima Michaelsen, 1934. JAPANESE EARTHWORMS 47 GENERIC DIAGNOSES Polypheretima — Intestinal caeca absent. Pithemem — One pair of intestinal caeca present originating in or near segment 22. Amynthas — One pair of intestinal caeca present originating in or near segment 27; male pores superficial. Metaphire — One pair of intestinal caeca present originating in or near segment 27; male pores in copulatory pouches; no nephridia present on spermathecal ducts. Pheretima — One pair of intestinal caeca present originating in or near segment 27; male pores in copulatory pouches; nephridia present on spermathecal ducts. For full generic descriptions and keys to nominal species and species-groups see Sims & Easton, 1972 (also Easton, 1979 for keys and descriptions of the species of Polypheretima). Key to the species of Megascolecidae of Japan 1 Intestinal caeca absent 2 Intestinal caeca present 3 2 Spermathecal pores absent or in furrows 5/6/7, 5/6 or 6/7; genital markings presetal Polypheretima elongata Spermathecal pores in furrows 5/6/7/8/9; genital markings postsetal . Polypheretima iizukaii 3(1) Intestinal caeca originate in segment 22 Pithemera bicincta Intestinal caeca originate in segments 2 5-2 7 4 4 Spermathecal pores absent 5 Spermathecal pores present 6 5 Intestinal caeca simple Amynthas illotus species-group Intestinal caeca manicate Amynthas hilgendorfi species-complex (part) 6(4) Spermathecal pores segmental 7 Spermathecal pores intersegmental 9 7 Spermathecal pores presetal Amynthas parvicystis Spermathecal pores postsetal 8 8 Spermathecal pores on segment 6 Amynthas glabrus Spermathecal pores on segments 6, 7 & 8 Amynthas obscurus 9(6) First spermathecal pores in furrow 4/5 10 First spermathecal pores in furrow 5/6 11 First spermathecal pores in furrow 6/7 First spermathecal pores in furrow 7/8 29 10 Four pairs of spermathecal pores; male pores on segment 1 8 . Amynthas scholasticus Five pairs of spermathecal pores; male pores on segment 19 .... Amynthas megascolidioides 11(9) One pair of spermathecal pores Amynthas minimus Two pairs of spermathecal pores Amynthas morrisi Three pairs of spermathecal pores Four pairs of spermathecal pores 17 12 Intestinal caeca simple Intestinal caeca manicate 16 1 3 Male pores superficial Male pores in copulatory pouches Metaphire yezoensis 14 Male porophores small; genital markings present 15 Male porophores large: genital markings absent Amynthas acinctus 48 E. G. EASTON 1 5 Genital markings in transverse rows (body length less than 80 mm) Amynthas papulosus Genital markings in clusters associated with male pores (body length up to 200 mm) Amynthas gracilis 16(12) Male pores superficial or absent . . . Amynthas hilgendorfi species-complex (part) Male pores in copulatory pouches Metaphire hataii 17(11) Intestinal caeca simple 18 Intestinal caeca manicate Amynthas habereri 1 8 Male pores superficial 19 Male pores in copulatory pouches 20 Male pores in seminal grooves Metaphire riukiuensis1 (part) 19 Genital markings small, segmental Amynthas corticus Genital markings large, intersegmental Amynthus micronarius 20(18) Genital markings present Metaphire fuscata Genital markings absent 21 2 1 Copulatory pouches restricted to segment 18 Metaphire tosaensis Copulatory pouches extending onto segments 1 7 and 1 9 . Metaphire riukiuensis^ (part) 22(9) One thecal segment Amynthas hilgendorfi species-complex (part) Two thecal segments 23 Three thecal segments 25 23 Male pores in seminal grooves Amynthas japonicus Male pores simple or absent .... Amynthas hilgendorfi species-complex (part) Male pores in copulatory pouches 24 24 Intestinal caeca simple Metaphire parvula Intestinal caeca manicate Pheretima koellikeri 25(22) Intestinal caeca simple 26 Intestinal caeca manicate 28 26 Male pores superficial 27 Male pores in copulatory pouches Metaphire peguana 27 Genital markings segmental Amynthas flavescens Genital markings intersegmental at 17/18 and 18/19 .... Amynthas hupiensis 28(25) Genital markings paired, median to male pores Metaphire servina Genital markings numerous, within copulatory pouches . . . Metaphire yamardai Genital markings absent Metaphire sieboldi 29(9) One thecal segment Amynthas hilgendorfi species-complex (part) Two thecal segments 30 30 Male pores superficial; genital markings present Amynthas robustus Male pores in copulatory pouches; genital markings absent ....... 31 3 1 Intestinal caeca simple Metaphire californica Intestinal caeca manicate Metaphire schmardae Amynthas acinctus (Goto & Hatai, 1 899) acincta Goto & Hatai, 1899:16. Other Japanese records: Yamaguchi, 1930a:52; Ohfuchi, 1938^: 1933;Kobayashi, 19416 : 260; Yamaguchi, 19620: 10. ?phaselus Hatai, 19306:659. Other Japanese records: Kobayashi, 19386:410; Kobayashi, 1 9416: 260, e: 513; Yamaguchi, 1962a: 11. ?maculosus Hatai, 19306 : 661 . 'It is uncertain whether the male pores of M. riukiuensis are in copulatory pouches or in seminal grooves. The species has been keyed out twice to allow for either condition. JAPANESE EARTHWORMS 49 ?kamitai Kobayashi, 1934:5. Other Japanese records: Kobayashi, 19386:411- Kobayashi 1941n Cognetti, 1906 : 777. DIAGNOSIS. Spermathecal pores closely paired in furrows 5/6/7/8/9. Male pores superficial on large porophores on segment 18. Genital markings small, paired, pre and postsetal, in line with the male pores on segments 19 and 20. Intestinal caeca manicate each with about 10 diverticula originating in segment 26. JAPANESE RECORDS. Honshu (Kanto-Chiho) KANAGAWA-KEN, Yokohama (Cognetti, 1906). DISTRIBUTION. Japan only. Amynthas hilgendorfi species-complex INCLUDED SPECIES hilgendorfi Michaelsen, 1892 : 235; (syn. rokugo, irregularis, schizopord) Beddard, 1900 : 633. Other Japanese records: Michaelsen, 1899 : 9; Michaelsen, 1916 : 11; Michaelsen, 1923 : 237; Yamaguchi, 1930a:50; 19306:89; Kobayashi, 19386:407; Ohfuchi, 1938^:1994; Kobayashi, 1941a:51, 6:260, c:378, d:459, oro (Nakamura, 1967); OSHIMA Hachimano-cho & Maksukage-cho in Hakodate & Ono (Yamaguchi, 1962). Honshu (Ou-Chiho) (Kobayashi, 19416 & c): (Kanto-Chiho) (Kobayashi, 19416); TOCHII-KEN Utsunomiya (Kobayashi, 1941J): (Chubu-Chiho) (Kobayashi, 19416 & c): (Kinki-Chiho) (Kobayashi, 19416 & c); HYOGO-KEN Nakahama (Michaelsen, 1899): (Chugoku-Chiho) (Kobayashi, 1 94 1 6 & c). Shikoku (Kobayashi, 1 94 1 6 & c). Kyushu (Kobayashi, 1 94 1 b, c & e); NAGASAKI-KEN Iki (Kobayashi, 19416); KAGOSHIMA-KEN Kagoshima & Shibushi (Kobayashi, 1941c). Okinawa-Gunto OKINAWA-JIMA (Kobayashi, 19416 &c). DISTRIBUTION. China and Japan, introduced into North America and New Zealand. Amynthas illotus species-group illotus species-group Sims & Easton, 1972 : 236. DIAGNOSIS. Spermathecal pores absent. Male pores superficial. Intestinal caeca simple, originating in or near segment 27. REMARKS. This species-group was recognized by Sims & Easton (1972) to accommodate several poorly described species which lack Spermathecal pores. These species do not have any special affinities with one another, instead the group is one of convenience. It is highly probable that when more data become available most of the included species will be found to be synonymous with other thecate species. Two species have been recorded from Japan: A. illotus and A. pusillus. For convenience the athecate species 'Pheretima' oyuensis Ohfuchi, 1956 is also considered here although it is not a member of the illotus species-group. This species was considered incertae sedis by Sims & Easton (1972) since it lacks male pores, the structure of which are diagnostic of the caecate genera Amynthas, Metaphire and Pheretima. Amynthas 'illotus' (Gates, 1932) illota: Ohfuchi, 1956 : 136 (non Gates, 1932 : 397). DIAGNOSIS. Length 125-155 mm, 125-144 segments. Genital markings absent. 109-114 setae on segment vii. REMARKS. The diagnosis of A. illotus was restricted by Gates (\912a : 196) to exclude the Japanese specimens identified by Ohfuchi. They cannot be assigned to another species but it is not proposed to recognize a new species to accommodate them since it is probable that 54 E. G. EASTON when more data become available, they will be found to belong to a previously described species. JAPANESE RECORDS. Sakishima-Gunto ISHIGAKI-SHIMA Ibaruma (Ohfuchi, 1956); IRIOMOTE- JIMA Hatoma-jima & Hoshitate (Ohfuchi, 1956). DISTRIBUTION. Japan. Amynthas pusillus (Ohfuchi, 1956) pusilla Ohfuchi, 1956 : 136 [non Ude, 1893 : 63 (= Amynthas minimus)]. DIAGNOSIS. Length 44-50 mm, 75-93 segments. Genital markings absent. 33-37 setae on vii. JAPANESE RECORDS. Sakishima-Gunto IRIOMOTE-JIMA Sonai (Ohfuchi, 1956). DISTRIBUTION. Japan. 'Pheretima' oyuemis Ohfuchi, 1937 oyuensis Ohfuchi, 1937« : 24. DIAGNOSIS. Length 50-55 mm, 87-94 segments. Male pores and prostates absent. Genital markings absent. C. 38 setae on vii. JAPANESE RECORDS. Honshu (Ou-Chiho) AKITA-KEN Inariyama, Oyu Katstuno district (Ohfuchi, 1937). DISTRIBUTION. Japan. Amynthas japomcus (Horst, 1883) japonica Horst, 1883 : 192. DIAGNOSIS. Spermathecal pores paired, c. 0-43 body circumference apart in furrows 6/7/8. Male pores superficial on segment 18 in seminal grooves which extend onto segment 17. Genital markings absent. Intestinal caeca manicate, each with about 8 diverticula, originating in segment 26. JAPANESE RECORDS. 'Japan' (Horst, 1883). DISTRIBUTION. Japan. Amynthas megascolidioides (Goto & Hatai, 1 899) megascolidioides Goto & Hatai, 1 899 : 2 1 . Other Japanese records: Kobayashi, 1 94 1 a : 5 1 , b : 260. DIAGNOSIS. Spermathecal pores paired, in furrows 4/5/6/7/8/9. Male pores superficial on segment 19. Genital markings small, paired, postsetal in line with the male pores on segments 17,18 and 20. Intestinal caeca simple, originating in segment 27. JAPANESE RECORDS. Honshu (Ou-Chiho) (Kobayashi, 19416): (Kanto-Chiho) TOKYO- TO Tokyo (Goto & Hatai, 1_899): (Chubu-Chiho) (Kobayahi, 19416): (Kinki-Chiho) (Kobayashi, 19416): (Chugoku-Chiho) (Kobayashi, 19416); SHIMANE-KEN Oki-gunto (Kobayashi, 194 la & 6). Shikoku (Kobayashi, 1 94 1 6). Tsushima (Kobayashi, 1 94 1 6). Kyushu (Kobayashi, 19416). DISTRIBUTION. Korea and Japan. Amynthas micronarius (Goto & Hatai, 1 898) micronaria Goto & Hatai, 1898:74. Other Japanese records: Ohfuchi, 19376:50; Kobayashi, 1 9416: 260, c : 378,.Diegeographische Verbreitung der Oligochaeten, pp 1-186. Berlin: Friedlander&Sohn. 1910. Zur Kenntnis der Lumbriciden und ihrer Verbreitung. Ezheg. zool. Muz. 15 : 1-74. 1916. Oligochaten aus dem Naturhistorischen Reichsmuseum zu Stockholm. Ark. Zool. 10 (9): 1-21. 1922. Oligochaten aus dem Rijks-Museum van Natuurlijke Historic zu Leiden. Capita Zool. 1 (3) : 1-67. 1923. Oligochaten von Neuseeland und den Auckland-Campbell-Inseln einigen anderen Pacifischen Formen. Vidensk. Medder dansk naturh. Foren. 15 : 197-240. — 1928. Miscellanea Oligochaetologica. Ark. zool. 20 (2) : 1-15. 1934. Oligochaeta from Sarawak. Quart. Jl microsc. Sci. 77 : 1-47. Nagase, I. & Nomura, E. 1937. On the Japanese aquatic Oligochaeta Criodrilus miyashitai n. sp. Sci. Rep. Tohoku Univ. 11 : 361-402. Nakamura, Y. 1967. [Population density and biomass of the terrestrial earthworm in the grasslands of three different soil types near Sapporo.] Jap. J. appl. Ent. Zool. 11 : 164-168. [In Japanese] 1971. Distributions of soil animals in three forests of northern Hokkaido (III). Lumbricids (Oligochaeta : Lumbricidae). J. Jap. for. Soc. 53 (1 1) : 347-349. 1972. [Ecological studies on the family Lumbricidae from Hokkaido I. Ecological distribution.] Jap. J. appl. Ent. Zool. 16 : 18-23. [In Japanese] 1973a. On the number of worms emerging from individual cocoons of several lumbricid species in Japan (Oligochaeta : Lumbricidae). Appl. Ent. Zool. Tokyo % (3) : 199-200. 19736. [Ecological studies on the family Lumbricidae from Hokkaido II. Environmental factors, especially vegetation and soil properties.] Jap. J. appl. Ent. Zool. 17 : 210-214. [In Japanese] 64 E. G. EASTON Ohfuchi, S. 1935. On some new species of earthworms from northeastern Hondo, Japan. Sci. Rep. Tohoku Univ. 10 : 409^415. - 1937a. Descriptions of three new species of the genus Pheretima from northeastern Honshu, Japan. Res. Bull. Saito Ho-on Kai. Mus. 12 : 13-29. - 1937/>. On the species possessing four pairs of spermathecae in the genus Pheretima, together with the variability of some external and internal characters. Res. Bull. Saito Ho-on Kai Mus. 12:31-136. - 1938a. On the variability of the opening and the structure of the spermatheca and the male organ in Pheretima irregular is. Res. Bull. Saito Ho-on Kai Mus. 15 : 1-31. - 19386. New species of earthworms from northeastern Honshu, Japan. Res. Bull. Saito Ho-on Kai Mus. 15 : 33-52. - 1938c. New and little known forms of earthworms, Pheretima from Nippon. Res. Bull. Saito Ho-on Kai Mus. 15 1:53-66. - \938d. [Zoological observations of the earthworms of the genus Pheretima from the ricefield in Ishikai, Hokkaido.] Botany Zool. Tokyo 6 (12): 1991-1998. [In Japanese] — 1939. Further studies of the variability in the position and number of male and spermathecal pores in the case of Pheretima irregularis based on local analysis. Sci. Rep. Tohoku Univ. 14 : 81-1 17. - 1941. The cavernicolous Oligochaeta of Japan. Sci. Rep. Tohoku Univ. 16 : 243-256. - 1956. On a collection of the terrestrial Oligochaeta obtained from the various localities in the Riu-kiu Islands, together with the consideration of their geographical distribution (Part I). J. Agric. Sci. Tokyo 3: 131-176. 1957. On a collection of terrestrial Oligochaeta obtained from the Riu-Kiu Islands, together with a note on their geographical distribution (Part II). J. Agric. Sci. Tokyo 3 : 243-26 1 . Oishi, 1932. Zool. Mag. Tokyo 44 (5 \9-52Q): 17-18. [In Japanese] - 1934. Zool. Mag. Tokyo 46 (545) : 133-134. [In Japanese] O mod to, P. 1956. Contribute alia revisionedei Lumbricidae. Arch. Zool. Ital. 41 : 129-212. Perel, T. S. 1974. 'Revision of the genus Eisenia Malm, 1877, em. Michaelsen, 1900 (Lumbricidae, Oligochaeta)' [Russian; English summary]. Zool. Zh. 53 : 1604-1615. - 1976. A critical analysis of the Lumbricidae genera system (with key to the USSR fauna genera). Rev. Ecol. Biol. Sol 13: 635-643. Perrier, E. 1 872. Recherches pour servir a 1'histoire des Lombriciens terrestres. Nouv. Archs Mus. Hist. naf. Paris 8: 5-1 98. 1874. Sur un nouveau genre indigene des Lombriciens terrestres (Pontodrilus marionis E.P.). C. r. hebd. seanc. Acad. Sci., Paris 78 : 1 582-1 586. 1875. Sur les Vers de terre des ties Philippines et de la Cochinchine. C. r. hebd. Seanc. Acad. Sci., Paris (D)81 : 1043-1046. Rosa, D. 1887. Microscolex modestus (n. gen., n. sp.). Boll. Mus. Zool. Anal. comp. R. Univ. Torino 2(19): 1-2. — 1890. Viaggo di Leonardo Fea in Birmanica e regioni vicine, XXVI. Perichaetidi. Annali Mus. civ. Stor. nat. Giacomo Doria 10 : 107-122. 1891. Die exotischen Terricolen des k.k. naturhistorischen Hofmuseums. Annln naturh. Mus. Wein 6: 319-406. 1893. Revisionedei Lumbricidi. Memorie Acad. Sci. Tori no (series 2) 43 : 1-50. 1 896. 1 Lumbrichi raccolti a Sumatra dal dott Elia Modiglani. Annali Mus. civ. Stor. nat. Giacomo Doria 16 : 502-532. Sasaki, K. 1924. Allolobophorafoetida (Sav.) in north Japan. Sci. Rep. Tohoku Univ. 1 : 89-90. Savigny, J. C. 1826. Analyse d'un Memoire sur les Lombrics par Cuvier. Mem. Acad. Sci. Inst. Fr. (Hist.) 5: 176-184. Sims, R. W. 1980. A classification and the distribution of earthworms, suborder Lumbricina (Haplotaxida : Oligochaeta). Bull. Br. Mus. nat. Hist. (Zool.) 39 (2) : 103-124. - (in press). Oligochaeta : Haplotaxida. In Barnes, R. D. Invertebrates Taxonomy and classifi- cation of living organisms. New York: McGraw-Hill. Sims, R. W. & Easton, E. G. 1972. A numerical revision of the earthworm genus Pheretima auct. (Megascolecidae : Oligochaeta) with the recognition of new genera and an appendix on the earth- worms collected by the Royal Society North Borneo Expedition. Biol. J. Linn. Soc. 4(3): 1 69-268. Song, M. J. & Paik, K. Y. 1970. On a small collection of earthworms from Geo-je Isl., Korea. Korean J. Zool. 13: 101-111. - 1971. Earthworms of MtJiri, Korea. Korean J. Zool. 14 : 192-198. Stephenson, J. 1917. Aquatic Oligochaeta from Japan and China. Mem. Asiat. Soc. Beng. 6 : 85-99. JAPANESE EARTHWORMS 65 Stop-Bowitz, C. 1969. A contribution to our knowledge of the systematics and zoogeography of Norwegian earthworms. Nytt. Mag. Zool. 17 : 169-280. Takahashi, S. 1932. Zool. Mag. Tokyo 44 (527) : 343-360. [In Japanese] Takahashi, S. & Yamaguchi, H. 1961. Occurrence of Phereti ma communissima (Goto et Hatai) in Hokkaido. J. Hokkaido Gakugei Univ. 12 : 1-3. Tamura, H., Nakamura, Y., Katsusuke, Y. & Fujikawa, T. 1969. An ecological survey of soil fauna in Hidaka-Mombetsu, South Hokkaido. J. Fac. Sci. Hokkaido 17 : 17-57. Ude, H. 1893. Beitrage zur Kenntnis auslandischer Regenwiirmer. Z. weiss. Zool. 57 : 57-75. 1905. Terricole Oligochaten von den Inseln der Siidsee und von verschiedenen andern Gebieten der Erde. Z. wiss. Zool. 83 : 405-501 . Yamaguchi, H. 1930a. [Some species of earthworms from Sapporo, Hokkaido (Preliminary report).] Zool. Mag. Tokyo 42 : 49-58. [In Japanese] 19306. On the variability of the capsulogenous glands in the earthworm (Pheretima hilgendorfi Michaelsen). Trans. Sapporo nat. Hist. Soc. 11 : 89-95. 1935. Occurrence of the luminous Oligochaete, Microscolex phosphoreus (Dug) in Japan. Annotnes zool. Jap. 15 : 200-202. 1953. Studies on the aquatic oligochaeta of Japan. VI. A systematic report with some remarks on the classification and phylogeny of the Oligochaeta. J. Fac. Sci. Hokkaido Univ. 11 : 277-342. 1962. On earthworms belonging to the genus Pheretima collected from the southern part of Hokkaido. J. Hokkaido Gakugei Univ. 13 : 1-21. Manuscript accepted for publication 8 September 1980 British Museum (Natural History) 1881-1981 Centenary Publications Alfred Waterhouse and the Natural History Museum Mark Girouard During his long life (1660-1753) Sir Hans Sloane amassed a vast collection of books, manuscripts and natural history specimens. These were the foundation of the British Museum which was established in the year of his death. By 1881 the natural history collections had grown to such an extent that they had to be moved from Bloomsbury, to a new, purpose-built museum in South Kensington. There, in some five acres of land, Alfred Waterhouse had designed what is still one of London's most outstanding pieces of architecture — The British Museum (Natural History), popularly known as the Natural History Museum. In this book Mark Girouard, traces the chequered development of the design, describes the influences of the key personalities involved and highlights some of the buildings most interesting features. Written by one of Britain's leading architectural historians the lively text, supported by an attractive design and lavish illustration, will appeal to all those interested in architecture and natural history. 218 x 230 mm, 64 pages, 8 pages of full colour, illustrations, many black and white and 2 colour line illustrations. £1-75 (paper); £4-95 (hardback). Co-published with Yale University Press. Titles to be published in Volume 40 Eugene Penard's Slides of Gymnamoebia : re-examination and taxonomic evaluation. By F. C. Page Japanese earthworms: a synopsis of the Megadrile species (Oligochaeta). By E. G. Easton Phylogenetic versus convergent relationship between bisci vorous cichlid fishes from Lakes Malawi and Tanganyika. By M. L. J. Stiassny Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester N GENERAL Bulletin of the ^ British Museum (Natural History) Phylogenetic versus convergent relationshi] between piscivorous cichlid fishes from Lakes Malawi and Tanganyika Melanie L. J. Stiassny Zoology series Vol 40 No 3 28 May 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England. World List abbreviation: Bull, Br. Mus. nat. Hist. (Zool.) Trustees of the British Museum (Natural History), 1981 ISSN 0007-1498 Zoology series Vol 40 No 3 pp 67-101 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 28 May 1981 Phylogenetic versus convergent relationship between piscivorous cichlid fishes from Lakes Malawi and Tanganyika Melanie L. J. Stiassny Department of Morphology, Zoological Laboratory, University of Leiden, The Netherlands'. Contents Synopsis Introduction Methods, nomenclature and materials . Methods Nomenclature Materials Abbreviations used in text figures . Anatomical description Ethmovomerine region of the neurocranium Cephalic muscles Pharyngeal jaw apparatus .... Discussion . Acknowledgements References 67 67 68 68 69 70 70 73 73 77 87 98 99 100 - f JUN i USftARY Synopsis The anatomy and phylogenetic relationships of two genera of African cichlid fishes, Rhamphochromis from Lake Malawi and Bathybates from Lake Tanganyika, are investigated. In accordance with the current methods of cladistic analysis data from representatives of a wide range of cichlid taxa are included for outgroup comparison. Particular emphasis is placed upon the anatomy of the ethmovo- merine region of the neurocranium, the cheek musculature, and the pharyngeal jaw apparatus. Based upon a number of synapomorphic characters an hypothesis of a sistergrbup relationship between the monophyletic genus, Bathybates, and a monophyletic assemblage consisting of the genera Hemibates and Trematocara is formulated. A similar resolution of the relationships of Rhampho- chromis has not been achieved; the differential success of the study is discussed and additional data relevant to unravelling the status and relationships of Rhamphochromis are introduced. Introduction The similarities that exist between individual species and whole communities of cichlid fishes in Lakes Malawi and Tanganyika have often been remarked upon (eg. Pellegrin, 1903; Regan, 1921, 1922; Fryer, 1959; Fryer & lies, 1972; Galis & Barel, 1980). Regan (1921) felt that the majority of the Malawian genera were phyletically distinct from any found else- where and that they formed a 'natural group'. Since that time, because it has been assumed that the cichlid flock of each lake has had a separate ancestry, their similarities have been interpreted as examples of convergent evolution (Fryer, 1969; Fryer & lies, 1972). The recent papers of Greenwood (1978, 1979, 1980) cast considerable doubt upon existing ideas about the phylogeny and interrelationships of the lacustrine Cichlidae. With reference 'This research was carried out in the Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD and was submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy in the Faculty of Science, London University. The author is at present holding a European Science Exchange Programme fellowship awarded by the Royal Society. Bull. Br. Mm. Nat. Hist. (Zool)40(3): 67-101 67 Issued 28 May 1981 68 MELANIE L. J. STIASSNY to the Lake Malawi haplochromine species group Greenwood (1979) suggests that the prevalent idea of the Malawi group being entirely derived from one or a few anatomically generalized fluviatile haplochromine cichlids (see Regan, 1921; Trewavas, 1935, 1949; Greenwood, 1974) is an oversimplification. He now suspects that lineages related to Thoracochromis, Astatotilapia, and even to Serranochromis and Chetia may have contributed to the flock. Clearly there is a need to reconsider the phyletic relationships of these lacustrine Cichlidae. Apart from the large number of species the problem of determining the interrelationships of the lacustrine Cichlidae is intensified by the fact that few characters have been found to be of use in phylogenetic analyses. The common occurrence of a range of morphological features confers a rather characteristic 'facies' upon many trophic groups (Greenwood, 1974). The problem of determining monophyletic assemblages, on the basis of shared derived characters, in the face of an apparent 'web of parallelisms' (Liem, 1978) is therefore particularly acute amongst these lacustrine Cichlidae. The two genera that form the subject of the present investigation, Bathybates Boulenger, 1898 from Lake Tanganyika and Rhamphochromis Regan, 1921 from Lake Malawi, have been selected for a number of reasons. Superficially, at least, Bathybates and Rhamphochromis are rather similar in external appearance, and Boulenger (1898) suggested that Bathybates ferox and Paratilapia longiceps (=Rhamphochromis longiceps) were closely related. Later, because Regan (1921) felt that the Malawian flock was a natural group, Rhamphochromis and Bathybates were considered to be an example of convergent evolution (Fryer & lies, 1972; Lowe-McConnell, 1975). Although data on the habits and ecology of these fishes are sparse it would appear that both occupy broadly similar biological positions in their respective lakes (Poll, 1956; Coulter, 1966, 1967; Fryer & lies, 1972; Lowe-McConnell, 1975). Both are essentially off- shore, open-water predators feeding upon cichlid as well as non-cichlid members of the pelagic communities. Species in both genera are also known to frequent the benthic zone where they are able to adapt to the prevailing near anoxic conditions (Coulter, 1967; Lowe-McConnell, 1975). The fact that both Bathybates and Rhamphochromis are exclusively piscivorous has also been a factor contributing to their selection for a comparative study. Amongst the trophic groups represented in the lacustrine flocks the piscivorous predatores comprise one of the largest groups (c. 30-40% of total species number). The large number of piscivorous species seems to be attributable both to the large biomass of other cichlid species (Fryer & lies, 1972; Witte, in press) and to the apparently minor anatomical modification necessary to facilitate the capture and ingestion of this food source (Fryer & lies, 1972; Liem, 1978). Methods, nomenclature and materials Methods In order to determine the phylogenetic relationships of Bathybates and Rhamphochromis a cladistic approach has been adopted. Monophyletic groups are defined on the basis of shared derived characters (synapomorphies), and in estimating the relative plesiomorph (primitive) or apomorph (derived) nature of various character states the 'commonality principle' (Schaefler, Hecht and Eldredge, 1972) is applied. A number of problems arise when cladistic principles are applied to an analysis of cichlid interrelationships (see Greenwood, 1979: 270). Apart from the phylogenetic breakdowns provided by Greenwood (1979, 1980), Liem and Stewart (1976) and Liem (1979) little guidance is available to aid the selection of appropriate cichlid outgroup taxa. Highly tenta- tive and noncladistic phylogenies for members of the Lake Malawi and Lake Tanganyika flocks are to be found in Fryer & lies (1972) (see also Regan, 1920, 1921, 1922; Trewavas, 1949). In the course of this investigation those taxa which have been thought by these authors to be 'close to' or 'implicated in the ancestry' of Bathybates or Rhamphochromis PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 69 have received particular attention. I have also placed an emphasis upon the examination of other lacustrine taxa although some riverine forms have been included. Otherwise my selection of cichlid outgroup taxa has been somewhat arbitrary. The following review is a partial account of the anatomy of selected cichlid taxa. Only those structures found to yield characters suitable for phylogenetic analysis at the level of universality under consideration have been selected for description here. Characteristics of the ethmovomerine region of the neurocranium, the cheek musculature and the pharyngeal jaw apparatus (PJA sensu Hoogerhoud & Barel, 1978) receive particular attention. Descriptions are based upon the type species, Rhamphochromis longiceps and Bathybates ferox, of the two major genera and where possible also of the type species of outgroup genera. For more detailed and comprehensive accounts of cichlid anatomy see Goedel (19740, b) and Anker (1978) for myology, and Barel et al. (1976) for osteology. Nomenclature The nomenclature of muscles follows that of Winterbottom (1974) and Anker (1978). Topographical and skeletal nomenclature is based upon that of Nelson (1969), Rosen (1973), Patterson (1975) and Barely al. (1976). To investigate myological structures specimens were dissected under a Wild M-7 stereo- microscope. Osteological study specimens were cleared in buffered trypsin solution and double stained following the procedure of Dingerkus & Uhler (1977). This material was supplemented by reference to the extensive osteological collections in the British Museum (Natural History). Taxonomic nomenclature Greenwood (1979) divides the polyphyletic genus Haplochromis into a number of monophyletic lineages (=genera) restricting Haplochromis to five species. Difficulties arise when reference is made either to species formerly included within the genus but have yet to be assigned to other genera, or to the former concept of the genus Haplochromis. To avoid confusion Greenwood (1979) suggests adopting a convention proposed by Patterson & Rosen (1977). Thus the specific names of species formerly placed in the genus Haplochromis and not allocated to other genera will be prefixed with the name Haplochromis cited between quotation marks. When referred to collectively all cichlid fishes with an Haplochromis type of pharyngeal apophysis (Greenwood, 1978) are termed haplochromine cichlids. Taxonomic note on Bathybates and Rhamphochromis The genus Bathybates is entirely restricted to Lake Tanganyika and was originally described by Boulenger (1898; type species Bathybates ferox Boulenger, 1898). Boulenger (1898) believed that Bathybates was closely related to Paratilapia with which it was connected by Paratilapia longiceps Giinther (=Rhamphochromis longiceps) of Lake Malawi. He believed that the more formidable dentition coupled with characters of the body scales warranted the establishment of a new genus. Since that time six more Bathybates species have been described; Bathybates fasciatus Boulenger, 1901; Bathybates vittatus Boulenger, 1914; Bathybates minor Boulenger, 1906; Bathybates graueri Steindachner, 1911; Bathybates horni Steindachner, 1911; Bathybates leoPoU, 1956. A key to the species of Bathybates can be found in Poll (1956). In 1915 Boulenger brought together Paratilapia caerulea Boulenger, 1908, Paratilapia esox Boulenger, 1908 and Hemichromis longiceps Giinther, 1864 into a new genus, Champsochromis. Regan (1921) was of the opinion that the type species of that genus, Champ sochromis caeruleus, was not generically distinct from Haplochromis as then defined and therefore placed caeruleus in Haplochromis. For the remaining species he established the genus Rhamphochromis, and designated Rhamphochromis longiceps the type species. He added four additional species to his genus: Rhamphochromis macrophthalmus Regan, 1921; 70 MELANIE L. J. STIASSNY Rhamphochromis ferox Regan, 192 1 ; Rhamphochromis woodi Regan, 1 92 1 , and Rhampho- chromis leptosoma Regan, 1921. Since that time two more species have been described: Rhamphochromis lucius Ahl, 1926 and Rhamphochromis brevis Trewavas, 1935. A key to the species of Rhamphochromis can be found in Trewavas (1935). Unlike Bathybates, Rhamphochromis is not exclusively lacustrine and specimens have been collected in the Upper Shire River (Boulenger, 1915; Ricardo-Bertram et al., 1946; pers. obs.). More species of Rhamphochromis from Lake Malawi, particularly from the Nkata Bay region, have yet to be described (pers. obs.). Materials Material representative of the following cichlid genera has been examined. A complete list of specimens used in this study is deposited in the fish section of the British Museum (Natural History). The number in brackets following each generic name indicates the number of species examined. South American genera: Acaronia(l) Astatotilapia(5) Lichnochromis (1) Aequidens ( 1 ) A ulonocara (2) Limnochromis (2) Apistogramma ( 1 ) A ulonocranus ( 1 ) Limnotilapia ( 1 ) Cichla ( [ ) Bathybates (7) Neotilapia ( 1 ) Cichlasoma (4) Boulengerochromis ( 1 ) Orthochromis ( 1 ) Crenicichla (3) Callochromis ( 1 ) Perissodus (3) Geophagus (3) Chetia(l) Pharyngochromis (1) Petenia ( 1 ) Ctenochromis ( 1 ) Prognathochromis (3) Diplotaxodon ( 1 ) Rhamphochromis (8) Asian genera: Ectodus ( 1 ) Sarotherodon ( 1 ) Etroplus (2) Haplochromis ( 1 ) Serranochromis (5) "Haplochr amis' (25) Teleogramma ( 1 ) Madagascan genera: Haplotaxodon ( 1 ) Telmatochromis (2) Paratilapia ( 1 ) Hemibates ( 1 ) Tilapia ( 1 ) Paretroplus ( 1 ) Hemichromis ( 1 ) Trematocara (8) Ptychochromis ( 1 ) Hemitilapia ( 1 ) Trematocranus ( 1 ) Lamprologus ( 1 0) Tylochromis (2) African genera: Lethrinops (2) Xenotilapia (2) Aristochromis(l) Abbreviations used in the text figures A,,A2, A3, Aw parts of the adductor man- dibulae muscle aa anguloarticular aap adductor arcus palatini muscle ad5 5th adductor muscle ad-fos adductor fossa art.fc-pb3 articulatory facet of pharyn- gobranchial 3 asp-aa ascending process of the anguloarticular asp-aa. f flange on ascending process of the anguloarticular ca central aponeurosis cart.ext.ep2 cartilaginous extension of epibranchial 2 PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 7 1 cc cranial condyle con. tiss. tract connective tissue tract cp coronoid process of the dentary d-bb dorsal bony bridge dent dentary dl-f dorsolateral fenestra do dilatator operculi muscle ect ectopterygoid end endopterygoid epl-4 epibranchials 1-4 ex.hd-ep4 expanded head of epibranchial 4 exs extrascapula fr frontal h-lpe horn of lower pharyngeal element hym hyomandibula im intermandibularis muscle intorb-s interorbital septum lac lachrymal l.ang-dent angulodentale ligament lap levator arcus palatini muscle le lateral ethmoid le-b body of lateral ethmoid le-e anterior extension of lateral ethmoid le-p lateral ethmoid process lev-ext 1-4 levatores externi muscles 1-4 lev-int 1-2 levatores interni muscles 1-2 lev-post levator posterior muscle 1-fos lateral fossa lo levator operculi muscle l.pal-le palatine-lateral ethmoid ligament l.pal-mes palatine-mesethmoid ligament Ipe lower pharyngeal element me MeckeFs cartilage m.c-p2 musculus cranio- pharyngobranchialis 2 mes mesethmoid mes-a arm of mesethmoid mes-p plate of mesethmoid mes-w wing of mesethmoid met metapterygoid m.t-e2 musculus transversus epibranchialis 2 m.t-p2 musculus transversus pharyn- gobranchialis 2 mx-f maxillary flange MELANIE L. J. STIASSNY nlf 0-5 neurocranial lateral line foramina 0-5 nip-pr nipple process od-a obliquus dorsalis anterior muscle od-p obliquus dorsalis posterior muscle oes oesophagus op operculum ov-f ovoid fenestra pal palatine pal-fc palatine facet pb 1-3 pharyngobranchials 1-3 pmx-art articular process of the premaxilla pmx-asp ascending process of the premaxilla pop preoperculum postorb postorbital process preorb preorbital process psh parasphenoid pt posttemporal quad quadrate r-d retractor dorsal is muscle r-f rostral fenestra rost-cart rostral cartilage sc supracleithrum shank-s shank spine spin. oes sphincter oesophagi sym symplectic tA,,^ tendons of A, tA2 tendon of A2 tA3 tendon of A3 tAw tendon of Aw td-p transversus dorsalis posterior muscle t.lev-ext3 tendon of levator externus 3 muscle t.lev-int 1 tendon of levator internus 1 muscle up-4 fourth upper tooth-plate V ramus mandibularis V vl-f ventrolateral fenestra vl-bb ventrolateral bony bridge vo-fc vomerine facet vo-fos vomerine fossa vo-h head of the vomer v-s stalk of the vomer vo-wing wing of the vomer Note on the figures The scale on all figures indicates 5 mm PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 73 Anatomical description The ethmovomerine region of the neurocranium Because of the position of the ethmovomer it acts as a base upon which the premaxillae move and are supported (Alexander, 1967; Rosen & Patterson, 1969). It is thus of importance in feeding and hence trophic adaptation, one of the key elements in the ecological success of lacustrine Cichlidae (Fryer & lies, 1972; Greenwood, 1974). As in most higher teleosts, the ethmovomer of the Cichlidae is composed of four bones: the paired lateral ethmoids, the mesethmoid and the vomer. Both the mesethmoid and the vomer are compound elements incorporating a number of dermal components (Patterson, 1975). The modal arrangement of these bones and associated cartilage in the African Cichlidae is illustrated in Fig. 1 . nlf2 nlfl dl-f intorb-s mes-w mes-p mes-a preorb vo-s Fig. 1 Semidiagrammatic drawing of the modal arrangement of the bones of the ethmovomer in African cichlids. A. Lateral view. B. Dorsal view. C. Lateral view (exploded). Regan (1920) was the first to attribute any phylogenetic significance to the particular associations of the bones in this region of the skull in cichlid fishes. He divided the genus Tilapia into a number of subgenera using the presence or absence of a sutural connection between the mesethmoid (=ethmoid of Regan, 1920) and the vomer as one of the characters for so doing. Trewavas (1973), Liem & Stewart (1976) and Liem (1979) also utilize ethmovomer characteristics in phylogenetic analyses of cichlid fishes. None of these authors has discussed the relationship of the lateral ethmoid bones to the other bones of that region. 74 MELANIE L. J. STIASSNY Rhamphochromis The lateral ethmoid bones (Figs 2 & 1 3 A) The elongate lateral ethmoids form the posterolateral part of the ethmovomer. The two bones are separated posteromedially by a cartilaginous interorbital septum and antero- medially by the mesethmoid and the ethmovomerine cartilage. Each bone may be described in two parts; the body and the anterior extension. Dorsally the body contacts the frontal and ventrally it forms the floor, anterior wall and side of the anterior myodrome. Medially it bears a large well developed process (the lateral ethmoid process) which abuts against the ethmovomerine cartilage. The anterior extension of the lateral ethmoid is suturally united with the vomer at two separate points. The dorsal sutural contact (the dorsal bony bridge) is separated from the lateroventral contact (the lateroventral bony bridge) so that the ethmovomerine cartilage is bridged dorsally and ventrally to expose an ovoid region of cartilage (the mesethmoid palatinad articulation facet of Barel et al., 1976). A broad strap-like ligament, the palatine-lateral ethmoid ligament, originates from the ventral face of the anterior extension of the lateral ethmoid and attaches to the laterodorsal ridge on the palatine. fr mes-a vo-h psh dl-f ov-f Fig. 2 Rhamphochromis longiceps, ethmovomerine region. A. Lateral view. B. Lateral view (suspensorium removed). C. Dorsal view. The mesethmoid (Figs 2 & 1 3 A) The mesethmoid is a single, elongate bone situated medially and forming the posterodorsal part of the ethmovomer. The dorsal plate of the mesethmoid underlies the frontal bones forming the floor of the large frontal fossa. Anteriorly the dorsal plate is bifurcated and each PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 75 arm of bone is suturally united with the vomer medial to its dorsal suture with the lateral ethmoid. Between the two arms of the mesethmoid the ethmovomerine cartilage is exposed; the anterior margin of the mesethmoid forms the posterior margin of the rostral fenestra. A pair of medial wings are borne on the ventrolateral face of the dorsal plate of the mesethmoid. In lateral view the mesethmoid wings are hemispherical and are separated medially by the ethmovomerine cartilage. The vomer (Figs 2 & 1 3 A) The vomer is a large strongly ossified element that forms the anterodorsal and ventral parts of the ethmovomer. It is most conveniently described in two parts: the head and ventral stalk. The stalk of the vomer is slender and its posterior extremity is enclosed within a channel in the parasphenoid. Dorsally a wing of the stalk is produced above the lateral fossa and becomes suturally united with the lateral ethmoid (the lateroventral bony bridge). Above the lateroventral bony bridge the head of the vomer is suturally united with the dorsal part of the lateral ethmoid anterior extension (the dorsal bony bridge) and medial to this it is suturally united with the mesethmoid. Dorsomedially the head of the vomer is divided by the rostral fenestra. The vomer bears a pair of well developed facets for articulation with the palatine. d-bb vo-h A vl-f vl-bb ov-f -bb vo-h vo-h vl-bb ov-f Fig. 3 Ethmovomer (lateral view). A. Tylochromis lateralis. B. Geophagus brasiliensis. C. Petenia splendida. Hathy bates and other cichlid taxa Outgroup comparisons amongst other cichlid taxa indicate that the arrangement of the bones of the ethmovomer of Rhamphochromis represents the modal (ie. plesiomorphic) type found in the majority of African Cichlidae. 76 MELANIE L. J. STIASSNY Interestingly, in the majority of South American taxa examined the anterior extension of the lateral ethmoid only contacts the vomer dorsally; the ventrolateral bony bridge is absent (compare Fig. 3B of the South American Geophagus and Fig. 3A of the African Tylochromis). However, in the South American Petenia (Fig. 3C) a ventrolateral bony bridge is present, but unlike the 'African' arrangement, in Petenia it is the vomer, rather than the vomer and the lateral ethmoid, that is produced to form the bridge. The ethmovomerine region of Bathybates differs from that of Rhamphochromis in two salient features: (i) the absence of a palatine-lateral ethmoid ligament (Figs 4 A & 1 3B), (ii) the reduction or loss of a dorsal bony bridge (Figs 4 & 1 3B). The palatine-lateral ethmoid ligament is present in all the other cichlid taxa examined with the notable exceptions of Hemibates and Trematocara (Figs 5 & 14). Because of the widespread intrafamilial occurrence of this ligament its absence is interpreted as a synapomorphy uniting Bathybates, Hemibates and Trematocara. A well developed dorsal bony bridge, formed by the sutural union of the anterior extension of the lateral ethmoid and the head of the vomer, is also present in all other taxa except Bathybates, Hemibates and Trematocara. In a few specimens of Bathybates and Hemibates a narrow splinter of bone has been found linking the lateral ethmoid and the vomer, but in no instance was a well developed dorsal sutural union found. The reduction or loss of the dorsal bony bridge is interpreted as a further synapomorphy uniting the genera Bathybates, Hemibates and Trematocara. Hemibates and Trematocara share a further synapomorphy of the ethmovomerine region; the presence of a palatine-mesethmoid ligament (Fig. 5). In all of the cichlids examined, the dorsal articulatory facet of the palatine (the meseth- moidad process of Barel et a/., 1976) abuts against the lateral aspect of the ethmovomer. nlf2 nlfl mes-w mes-a -rvo-h fr C le dl-f Fig. 4 Bathybates ferox, ethmovomerine region. A. Lateral view. B. Lateral view (suspensorium removed). C. Dorsal view. PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 77 Whilst in other taxa the facet of the palatine is attached to the ethmovomer by a few strands of connective tissue, in Hemibates and Trematocara the facet is also attached to the meseth- moid by a palatine-mesethmoid ligament. This ligament arises from the dorsal aspect of the palatine facet and passes posterodorsally to the dorsal face of the mesethmoid. Since it has only been found in Hemibates and Trematocara it is interpreted as a synapomorphy indicative of a sistergroup relationship between the two genera. Further evidence of a sistergroup relationship between these two genera is found in the form of the rostral cartilage. In the majority of cichlids each premaxilla has ascending and articular processes and the two ascending processes are closely apposed and firmly bound by connective tissue. The rostral cartilage is a small nubble of cartilage situated on the postero- dorsal face of the ascending processes, just above the articular processes. In Trematocara and Hemibates (Fig. 6C, D), although well developed articular and ascending processes are present, the rostral cartilage is not restricted by them and extends about halfway along the long ascending process. A similar elongation of the rostral cartilage has not been found in any other cichlid taxa. The premaxillae ofBathybates is exceptional since there are no distinct articular processes (Fig. 6B). The Cephalic muscles On the basis of their shared innervation the muscles of the cheek may be divided into two groups. The first (group one muscles) includes the adductor mandibulae, the levator arcus palatini, and the dilatator operculi muscles. All are innervated by branches of the fifth nlf2 nlfl dl-f l.pal-mes pal-fc preorb B .pal-mes pal-fc le-p vl-f dl-f l.pal-mes vo-h nlfO le-e mes-p Fig. 5 Ethmovomerine region. Left, Trematocara marginatum A. Lateral view. C. Dorsal view. Right, Hemibates stenosoma B. Lateral view. D. Dorsal view. 78 MELANIEL. J. STIASSNY cranial nerve. Group two muscles include the adductor arcus palatini, the adductor operculi and the levator operculi muscles; this group is innervated by branches of the seventh cranial nerve. pmx-asp rost-cart pmx-asp pmx-art rost-cart Fig. 6 Right premaxilla (lateral view). A. Rhamphochromis longiceps. B. Bathybates ferox. C. Trematocara marginatum. D. Hemibates stenosoma. Group one muscles These develop from the masticatory muscle plate which divides early in ontogeny into a dorsal and ventral section. The former gives rise to the constrictor dorsalis and the latter develops into the adductor mandibulae (Edgeworth, 1935). The adductor mandibulae muscles. This is the largest muscle complex of the head; it occupies the lateral face of the suspensorium and inserts onto both upper and lower buccal jaws. In the Cichlidae four subdivisions are recognized. Following Anker (1978) a distinction is made between a part and a section of a muscle. A part is a subdivision having no, or hardly any, anatomical connection with other subdivisions of the same muscle. A section remains anatomically continuous with the rest of the muscle but may be distinguished from it on the basis of some other criterion such as fibre direction. Hardly any connection exists between the four subdivisions of the adductor in cichlid fishes. Each part is well defined and easily separable from the others, and each is inserted via a separate tendon. Although, as compared with some other families, the range of morphological variation of the adductor within the Cichlidae is narrow, numerous small differences are detectable. pal-le l.ang-dent Fig. 7 Rhamphochromis longiceps. A. Lateral view of the adductor mandibulae muscle complex after removal of the lachrymal and circumorbital bones and the eyeball. B. After removal of A2 to expose A3. C. After the removal of A2, and with A, cut away. 80 MELANIE L. J. STIASSNY Rhamphochromis INNERVATION. The path of the ramus mandibularis V is constant. It passes internal to A,, external and dorsal to A3, and divides, medial to A2, sending a minor branch to the anteromedial surface of A2. The main trunk enters the lower jaw dorsal to the tendon of A3. On the medial face of the lower jaw the nerve branches, innervating Aw. For a precise description of the course of the ramus mandibularis see Goedel ( 1 974a, b) and Anker ( 1978). PART A, (Figs 7 & 8). This is the dorsal, superficial part of the adductor defined by its dorsal position and insertion onto the maxilla (Vetter, 1874). Part A, is a complex elongate muscle which connects the suspensorium with the maxilla and mandible. Origin is from the dorsal face of the vertical limb of the preoperculum and from the hyomandibular transverse zone. Dorsal muscle fibres pass anteroventrally and converge upon a well developed sheet-like aponeurosis situated beneath the eye. The fibres which originate from the middle section of the vertical limb of the preoperculum pass forward, bypassing the dorsal aponeurotic sheet, and intercalate with fibres that have originated from the anterior region of the aponeurosis. Together these fibres converge upon an anteroventral aponeurosis from which two tendons arise. The dorsal tendon, tAla, passes forward, medial to the maxillary shaft, and inserts on the anterior border of that bone just below the cranial condyle of the maxillary head. The second tendon, tA,b, passes ventrally, medial to tA3 and the ramus mandibularis V. On the inner aspect of the lower jaw a portion of tAlb intergrades with tAw, the remainder passes lateral to Aw and inserts onto the nipple process on the anteromedial face of the anguloarticular. PART A2 (Fig. 7 A). This is the largest component of the adductor. Part A2 is a parallel-fibred muscle occupying the ventrolateral region of the cheek, connecting the ventral elements of the suspensorium with the lower jaw. Dorsally A2 overlies A, and it conceals the anteroventral aponeurosis of the latter. Origin is from the crescentic zone and horizontal limb of the preoperculum, the symplectic, and the ventrolateral ridge of the quadrate. Part A2 is composed of two sections distinguished by their differing sites of insertion. The dorsal section (A2a) is composed of fibres originating from the crescentic zone of the preoperculum and inserting onto the coronoid process of the dentary. The fibres contact the coronoid process on its lateral face via an association with the angulodentale ligament and, on its medial face, via a strap-like tendon, tA2. In the ventral section (A2^) the fibres originating from the horizontal limb of the preoperculum, the cc asp-aa B tAw ca im nip-pr tAs mc Fig. 8 Rhamphochromis longiceps. A. Buccal jaws and anterior suspensorial elements (medial view). B. Dentary, A2 and Aw removed (medial view). PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 8 1 symplectic, and the quadrate insert musculously onto the posterior border of the ascending process of the anguloarticular. Part A2 has no connection with part Aw. PART A3 (Fig. 7C). This is the medial part of the adductor complex and it underlies both A, and A2. Part A3 is a small, flat muscle which is roughly triangular in outline with the apex directed rostroventrally. Origin is from the lateral face of the metapterygoid and the dorsal border of the symplectic. Fibres converge upon a well developed strap-like tendon, tA3 which runs anteroventrally and enters the lower jaw lateral to tAlb. On the inner aspect of the lower jaw tA3 passes lateral to Aw and inserts onto a small sesamoid ossification (the coronomeckelian ridge of Barel et al., 1976). PART Aw (Fig. 8 A). Part Aw lies on the medial face of the mandible and connects it with the medial face of the suspensorium. Fibres fan out from a central sheet-like aponeurosis, and attach musculously to the medial face of the coronoid process, the inner and outer walls of the Meckelian fossa and the ventral face of the anguloarticular. Towards the quadrato-mandibular articulation the aponeurosis becomes consolidated into a tendon, tAw, which is firmly inserted onto the medial face of the quadrate and preoperculum. THE INTERMANDIBULARIS (Fig. 8 A). According to Edgeworth (1935) the mandibular muscle plate, in the early stages of its development, extends from the Gasserian ganglion to the anterior edge of the pericardium or to the midventral line. With the development of Meckel's cartilage the mandibular muscle plate separates into a masticatory muscle plate and the intermandibularis. The latter is innervated by either the ramus mandibularis V (Kesteven, 1943) or it may also receive innervation from the VII cranial nerve. In the Cichlidae a single intermandibularis is present just caudal to the symphyseal facets of the dentaries. The intermandibularis passes transversely between left and right rami, attaching musculously to their medial faces. Bathybates The adductor in Bathybates is essentially similar to that of Rhamphochromis and a separate description is not warranted. However, certain differences in detail do exist and these are described and illustrated: (i) The tAla is considerably reduced in length. The anterior muscle fibres of A, have apparently encroached upon the tendon so that the anterior portion of A, comes to lie medial to the maxillary shaft (Fig. 9A). (ii) The short tAla inserts onto a distinct flange developed on the anterior border of the maxilla, just below the cranial condyle of the maxillary head (Fig. 9 A, B). Fig. 9 Bathybates ferox. A. Buccal jaws and anterior suspensorial elements (medial view). B. Left maxilla (medial view). 82 MELANIE L. J. STIASSNY (iii) The posterior face of the ascending process of the anguloarticular is expanded laterally to accommodate a large insertion of section A2yj(Fig. 10A). (iv) Part A3 is large and well developed. Other cichlid taxa In the majority of cichlid taxa examined the mode of A, insertion is as described in Rhamphochromis. The presence of an elongate tAla is therefore interpreted as a plesio- morphic character of cichlid fishes. The reduction of the length of tAla in Bathybates is, on the basis of its limited distribution, interpreted as a derived character. Liem (1978: Fig. 6) illustrates the medial aspect of the mandible, maxilla, and the anterior portion of the suspensorium and associated muscles, tendons, and ligaments of Hemibates stenosoma. He represents tAla (tarn,) as a short tendon of the Bathybates-type. This contradicts my own observations of tAla in a range of//, stenosoma specimens in which I found tAla to be of the standard Rhamphochromis (i.e. modal) type. One species of Lamprologus (a large genus, most of whose species are endemic to Lake Tanganyika, see Poll, 1978) has a tAla approaching the Bathybates type. This species, Lamprologus elongatus, is united with the other Lamprologus, none of which has a short tAla, by the possession of a suite of characters that on the basis of their restricted distribution, are assessed to be apomorphic. For example all Lamprologus possess a hyomandibula in which the head is deeply notched, widely spaced extrascapular bones, a prominent flange on the prootic, and the infraorbital bones reduced or absent. All these characters are restricted to Lamprologus, or, in the case of infraorbital reduction, also occur in Telmatochromis, Julidochromis, Chalinochromis, and Teleogramma. In view of the number of synapo- morphies uniting L. elongatus with the remaining Lamprologus species, the presence of a reduced tAla in L. elongatus is interpreted as a convergent development rather than a synapomorphy shared with Bathybates. In all species of Bathybates the maxilla bears a distinct flange on its anterior border just below the cranial condyle of the maxillary head. This maxillary flange forms a platform onto which tAla is firmly inserted. An incipient flange is sometimes present in individual speci- mens of Trematocara and Hemibates. The presence of a well developed maxillary flange is interpreted as an apomorphic character within the Cichlidae. asp-aa.f Fig. 10 Left anguloarticular bone. A. Bathybates ferox. B. Trematocara marginatum. C. Rhamphochromis longiceps. D. Callochromis melanosligma. PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 83 In all species of Bathybates (Fig. 10 A), Trematocara (Fig. 10B), and Hemibates the ascending process of the anguloarticular has the posterior border laterally expanded into a ledge on which A2^ inserts. The presence of this expansion of the anguloarticular is inter- preted as a synapomorphy uniting these three genera. Liem and Osse (1975) describe an adductor fossa in a number of Lake Tanganyika genera. This fossa is an indentation on the lateral face of the anguloarticular (Fig. 10D) which also serves to increase the insertion area of A2o, and it is of a form quite distinct from that of Bathybates, Trematocara and Hemibates; it therefore is assessed as an independent character state. The relatively large A3 of Bathybates is difficult to interpret phylogenetically as there is considerable variation in A3 development throughout the family. The A, of all Trematocara species displays a novel tendinous connection with the lachrymal bone (Fig. 1 1). In Trematocara a third tendon (tAlc) arises from the anteroventral aponeurosis of A, and passes forward to insert on the medial face of the large lachrymal bone. A similar connection between the lachrymal and A, was not found in any other cichlid taxon examined, and its presence in Trematocara is interpreted as a synapomorphy uniting the species of that genus. dent Fig. 11 Trematocara marginatum, buccal jaws, lachrymal bone and anterior portion of A, (lateral view). The outline of the lachrymal bone is indicated with a broken line. The dilatator operculi. Rhamphochromis (Fig. 12 A) The muscle lies behind and above the levator arcus palatini but there is no intercalation of their fibres and the two muscles are distinct. The main muscle mass of the dilatator occupies the well developed dilatator fossa, which is formed anteriorly by the sphenotic and posteriorly by the pterotic bones. The fossa lies between the crest bearing neurocranial lateral line foramina 4, 5 and 6 (Barel et al., 1976) and the hyomandibular articulation sockets. Anteriorly the dilatator is bounded by the postorbital process of the sphenotic bone. The dilatator origin is extensive, occupying the entire surface of the fossa. Fibres converge ventrocaudally and merge into a tendon-like aponeurosis inserting firmly on the dorsal surface of the dilatator process of the operculum (the dorsal process of Winterbottom, 1974). Bathybates (Fig. 12B) The dilatator of Bathybates differs from that of Rhamphochromis in being expanded anteriorly so that it covers the border of the postorbital process of the sphenotic. Other cichlid taxa In all the other cichlid taxa examined, the dilatator operculi is as described for Rhampho- 84 MELANIE L. J. STIASSNY chromis. The anterior expansion of the dilatator in Bathybates species is interpreted as a synapomorphy uniting the species of that genus. The levator arcus palatini. The position and structure of the levator arcus palatini is constant in the taxa examined in this study, although variations were observed in its size and shape. The following brief description of the levator in Rhamphochromis is representative of the modal cichlid arrangement. exs nlf5 nlf4 nlf3 postorb calyx met postorb pop Fig. 12 Superficial postorbital musculature (lateral view) of A. Rhamphochromis longiceps. B. Bathybates ferox. PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 85 Rhamphochromis (Fig. 1 2 A) The levator is approximately conical in shape with its apex situated dorsally. Origin is from the ventral face of the postorbital process of the sphenotic. The fibres fan out from here and pass ventrally to insert musculously onto the hyomandibular transverse zone and flange, and into the calyx (Barel et al., 1 976). No trenchant differences in the form of the levator in Bathybates were recognized. Group two muscles All group two muscles are derived from the constrictor hyoideus dorsalis and are innervated by the ramus hyomandibularis of the Vllth cranial nerve (Edgeworth, 1935). The adductor arcus palatini. The adductor is well developed in the Cichlidae; it extends anteriorly to fill the fissura infraorbitalis and forms the floor of the orbit between the neurocranium and the suspensorium. d-bb rf aap ect vl-bb Fig. 13 Ethmovomerine region of A. Rhamphochromis longiceps. B. Bathybates ferox. The insertion area of the adductor arcus palatini muscle is indicated by broken lines. 86 MELANIE L. J. STIASSNY Rhamphochromis (Fig. 1 3 A) The adductor originates musculously from the parasphenoid ventral crest and wing, and from the prootic anteroventral to the lateral commissure. The posterior border of the adductor is demarcated by the prootic-parasphenoid crest. Fibres originate from the length of the dorsoventral crest of the parasphenoid, and the rostral margin of the muscle lies below the transverse level of the preorbital process. The fibres insert musculously on the medial face of the suspensorium. A sheet of connective tissue originates from the ventromedial face of the lateral ethmoid bones and fans out ventrally to cover the dorsal surface of the adductor. The posterior fibres pass caudoventrally and insert musculously on the medial face of the hyomandibular flange and the dorsomedial face of the metapterygoid. Anteriorly the muscle fibres slope sharply forward and insert on the medial face of the palatine fossa and the ectopterygoid. The adductor is considerably thicker posteriorly although a slight thickening of the section inserting onto the palatine is discernible. nlf 2 nlfl mes intorb-s aap end ect Fig. 14 Ethmovomerine region of Trematocara unimaculatum. The insertion area of the adductor arcus palatini muscle is indicated by broken lines. Bathybates and other cichlid taxa (Figs 1 3B & 14) The adductor of Bathybates is essentially similar to that of Rhamphochromis although as can be seen from the accompanying figure (Fig. 13B) the angle of orientation of the fibres inserting onto the palatine fossa is not as acute and the fibres pass ventrally. This seems to be a result of the position of the lateral ethmoid-palatine articulation facet which is situated more anteriorly in Rhamphochromis. An arrangement similar to that of Rhamphochromis is present in other elongate piscivores of Lake Malawi e.g. 'Haplochromis' caeruleus and 'Haplochromis' spilorhynchus, and those of Lake Victoria e.g. Prognathochromis macrognathus, whilst one similar to that of Bathybates is present in other Lake Tanganyika genera e.g. Haplotaxodon, Callochromis, Hemibates, Ectodus, and Aulonocranus. Unfortunately insufficient outgroup data are available to enable the polarity of this character complex to be determined. PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 87 In Trematocara marginatum the arrangement is as in Bathybates but in the remaining species of Trematocara the adductor has apparently migrated anteriorly along the parasphenoid and onto the ventrolateral face of the vomer from where the fibres pass posteroventrally to insert into the palatine fossa (Fig. 14). In common with Chichoki's (1976) observations, I have found that the adductor in all of the South American, Madagascan, and Asian species examined does not contact the medial face of the palatine and extends only as far rostrally as the medial face of the endopterygoid bone. The levator operculi. The levator passes between the lateral neurocranial wall and the operculum, caudal to the dilatator operculi. The levator is constant in form and position throughout the Cichlidae. It is an approxi- mately triangular muscle with the apex situated dorsally. Origin is from the ventrocaudal region of the pterotic facet of the neurocranium, and the fibres fan out ventrally to insert on the medial face of the operculum, caudal to the site of insertion of the adductor operculi and the adductor hyomandibulae muscles. The anterior portion of the muscle inserts musculously onto the levator ledge on the medial face of the operculum, whilst the posterior fibres merge into a connective tissue sheet situated near the posterior margin of the operculum. The adductor operculi. This cylindrical muscle connects the neurocranium with the medial face of the operculum, at a point adjacent to its articulation with the hyomandibula. Rhamphochromis The adductor originates musculously from the lateral region of the lateral awning (Barel et al., 1976), its fibres passing laterally to insert on the adductor process which lies just behind the suspensoriad articulation socket on the levator ledge of the operculum. No trenchant differences in the form of the adductor in Bathybates were recognized. The adductor hyomandibulae. In the Cichlidae a small adductor hyomandibulae usually is present. It has apparently developed from the anterior fibres of the adductor operculi. The muscle slip originates, with the adductor operculi, from the prootic and inserts onto the medial face of the hyomandibular head adjacent to the opercular condyle of that bone. The pharyngeal jaw apparatus (PJA) It appears that throughout the cichlid radiation the full complement of perciform branchial muscles and bony elements of the PJA is retained and that no major changes occur in their spatial relationships to one another. However, within this configuration a seemingly endless spectrum of minor morphological variation is expressed. This is realized through differ- ences in the relative size and robustness of the pharyngeal bones, the shape and distribution of their teeth, and through proportional changes in the various muscles coupled with slight differences in their sites of origin and insertion. Osteological features of the PJA The lower pharyngeal element. The lower pharyngeal element in cichlid fishes is com- posed of the suturally united ceratobranchials of the fifth branchial arch and their associated tooth plates. The size and shape of the resultant element, as well as the form and distribution of the pharyngeal teeth, shows considerable interspecific variation. Indeed, tooth form varies not only between taxa but also in different fields of the element in one individual. An antero- posterior gradient of increasing tooth specialization is considered as being the standard cichlid arrangement (Liem, 1978). Typically the lower pharyngeal element, which is the 88 MELANIE L. J. STIASSNY largest element of the PJA, is shaped somewhat like an indented triangle with the apex lying rostrally. Posterolaterally the bone terminates, on either side, in a posteriorly directed horn (=muscular process of Liem, 1 973) and a ventral keel is developed. Liem (1973, 1978) and Liem & Osse (1975) have drawn attention to convergence in pharyngeal tooth form in trophically related groups. Amongst piscivores Liem (1978) illustrates the pharyngeal teeth of Lamprologus compressiceps and 'Haplochromis' compressiceps, fishes which belong to distinct phyletic lineages, but which have an almost identical pharyngeal dentition. h-lpe keel Fig. 15 Lower pharyngeal element. Left, Bathybates ferox. A. Dorsal view. B. Ventral view (right half). C. Lateral view. Right, Rhamphochromis longiceps. D. Dorsal view. E. Ventral view (right half)- F. Lateral view. Rhamphochromis The lower pharyngeal element of Rhamphochromis is illustrated in Fig. 1 5D, E, F and as can be seen the bone is relatively elongate and is of a form commonly encountered amongst piscivorous cichlids (Poll, 1956; Greenwood, 1962, 1974; Barel, van Oijen, Witte & Witte- Maas, 1 977; Hoogerhoud & Barel, 1978; Liem, 1978). Seen ventrally, the medial region of the bone is rounded and convex and the lateral fossae are narrow. The dentigerous surface is roughly triangular in outline and the teeth are relatively fine. As in most cichlids the pharyngeal teeth located in the posterior field of the bone are the larger ones. In the anterior field the teeth are slender unicuspids in which the acutely pointed cusp is frequently directed caudally. The teeth become increasingly bevelled and the final two rows of somewhat enlarged teeth bear slightly hooked major cusps below which are a series of two to four accessory cusps. On the lower pharyngeal element these cusps are borne on the anterior edge of the teeth. It is this series of accessory cusps that gives these teeth their serrated appearance. The pharyngeal teeth are 'crowded' caudolaterally but uniformly distributed over the rest of the dentigerous surface. PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 89 Bathybates The lower pharyngeal element is elongate and the tooth form and gradation are similar to those described for Rhamphochromis. A caudolateral 'crowding' of the teeth also occurs in Bathybates although the remaining teeth are sparsely and less uniformly arranged over the rest of the dentigerous surface. As can be seen from the accompanying illustration (Fig. 1 5 A, B, C) the dentigerous area is enlarged and covers most of the pharyngo-buccal face of the lower pharyngeal element. In ventral view the lateral fossae are expanded and the bone is usually flattened but in some individuals it is concave. Of the Bathybates species examined, Bathybates ferox displays the most extreme modifi- cation of the lower pharyngeal element, but all the species share an enlargement of the dentigerous area, a dorsoventral flattening of the caudal region of the bone, and an increased fossa width. Both the enlargement of the dentigerous surface and the dorsoventral flattening of the caudal region of the pharyngeal element in Bathybates species are, on the basis of their limited distribution, interpreted as derived characters. The lower pharyngeal element of Rhamphochromis is rather typical of that found in the majority of piscivorous, elongate cichlids and no trenchant differences are recognized. The pharyngeal dentition is similar in all of the piscivorous cichlids examined. The occurrence of this specialized tooth form amongst the piscivorous lineages in Lake Victoria (Greenwood, 1974), in Lamprologus compressiceps and other distantly related piscivores (Liem & Osse, 1975; Liem, 1978) seems to indicate that an independent evolution of this type of pharyngeal dentition is not uncommon amongst piscivorous cichlids. For this reason the extreme similarity in tooth form and gradation between Bathybates and Rhampho- chromis is interpreted as convergence. The upper pharyngeal jaws. As with the lower pharyngeal element, the size and shape of the constituent elements of the upper pharyngeal jaws as well as the form of their dentition, show considerable interspecific variation. Less consideration has, however, been given to the upper pharyngeal elements as a potential source of taxonomically useful characters. This is partly because of their relative inaccessability but also, since both upper and lower elements operate as a functional unit, adaptational changes in the lower element are reflected in concomitant changes in the upper elements. Rhamphochromis (Fig. 1 6 A) The principal bony elements constituting the upper pharyngeal jaw in Rhamphochromis, in common with all cichlids, are the paired second and third* pharyngobranchials (and their associated tooth plates) and the fourth upper tooth-plate. A pair of first pharyngobranchials (Pbl) serve to suspend the jaws from the neurocranium. Each is a stick-like ossified element connecting the anterior arm of the first epibranchial with the prootic just below the lateral commissure. The first pharyngobranchials lack associated tooth plates. Each second pharyngobranchial (Pb2) is situated anterior to the third and is linked to the dorsomedial face of that element by a tract of connective tissue so that it lies with its caudal border closely apposed to the third pharyngobranchial (Pb3). Fused to the ventral surface of each Pb2 is a narrow, relatively elongate tooth plate with five or six teeth arranged in a single row. Each tooth bears a hooked major cusp with two accessory cusps above it. These cusps are borne on the lateral side of the teeth. The third pharyngobranchials are the largest elements of the upper jaws. On the dorsal *The term pharyngobranchial as used in this study refers to the infrapharyngobranchial. Amongst recent teleosts, ossified suprapharyngobranchials only occur in the first arch of elopids and alepocephalids and are thus not considered. The problem of the correct homology of the element with which the third and fourth epibranchials articulate is discussed by Ismail (1979) and Ismail & Verraes (in prep.). Throughout the present study it is assumed that this element represents the third pharyngobranchial alone. 90 MELANIE L. J. STIASSNY art.fc-pb3 pb2 up-4 pb3 con.tis.tract ep1 Fig. 16 Right upper pharyngeal jaw (dorsal view) in A. Rhamphochromis longiceps B. Tremato- cara marginatum. surface of each bone there is a prominent raised facet which articulates with a corresponding process, the pharyngeal apophysis, on the base of the neurocranium (see Greenwood, 1978). Fused to the ventral surface of each Pb3 is a large tooth plate in which all the teeth have a hooked major cusp with two or three accessory cusps below it. In the caudal region the teeth become progressively smaller and more numerous. The fourth upper tooth-plates (UP4) are well ossified and each is closely apposed to the caudal border of the corresponding Pb3. Each tooth plate bears numerous small hooked teeth also with tiny accessory cusps. Along the caudal margin of these bones there is a 'frayed zone' (the crista pharyngobranchialis of Goedel, \914b) composed of numerous, very small, irregularly scattered unicuspid teeth. A strong connective tissue link exists between the Pb3 and UP4 of either side of the pharynx so that the whole complex apparently functions as a single unit. Bathybates As in Rhamphochromis the upper pharyngeal jaw complex as a whole is not robust and it is relatively elongate. The form and arrangement of the teeth are similar to those described in Rhamphochromis (see Liem, 1978 for other piscivorous cichlids). In the course of this investigation, no trenchant differences were recognized in the form and composition of the upper pharyngeal jaws of Rhamphochromis and Bathybates. Both genera possess the typical piscivorous cichlid configuration (Hoogerhoud & Barel, 1978; Liem, 1978). Other cichlid taxa Of the outgroup taxa examined, all the Trematocara species share a characteristic feature of the third pharyngobranchial bones, viz the anterior margin of Pb3 is deeply indented, giving the margin a 'U'-shaped outline (Fig. 16B). This feature of Pb3 is not found in any other taxa and is assumed to be a synapomorphy uniting the species of the genus. The epibranchial skeleton. Because of the intimacy of their association with the pharyngo- branchials, both in terms of structure and function, the epibranchials are considered as part of the upper pharyngeal jaw complex. Barel et al. (1976) have produced an admirable description of the epibranchial series in PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 9 1 Astatotilapia elegans, and since the form of these bones in Bathybates and Rhamphochromis differ little from those of A. elegans, the following description will only highlight those features which are of presumed phylogenetic significance. Rhamphochromis (Fig. 1 6 A) The first epibranchials (Epl) are 'Y'-shaped elements. In each the rostral arm contacts the medial face of its respective Pb2 via a thin tract of connective tissue. An interarcual cartilage (Rosen & Greenwood, 1976) is completely absent in the majority of specimens examined. One specimen did retain a small nubble of cartilage suspended in this connective tissue tract, but the cartilage did not contact either Epl or Pb2. Each second epibranchial (Ep2) contacts the dorsolateral face of Pb2 and the uncinate process contacts the anterior margin of Pb3 on its dorsolateral face. The main body of each Ep2 is rostrally expanded and the anterior margin projects medial to the uncinate process of Epl . This expanded margin is capped with a flange of cartilage. The third epibranchials (Ep3) are the smallest of the series. The main shaft of each contacts Pb3, and a well developed caudolateral process articulates with a corresponding facet on the fourth epibranchial. Each forth epibranchial (Ep4) has a rostrally expanded head (the quadrangular region of Barel et al., 1976). This enlarged head is slightly incurved and forms a large, posteriorly directed cupped area that overlies the dorsal face of UP4. The dorsal tip of the head of Ep4 articulates with a dorsal eminence formed at the junction of Pb3 and UP4. The rostral margin of each Ep4 bears an articulatory facet for the caudolateral process of Ep3; just below this facet the shank of Ep4 is produced into a well developed shank spine. Bathybates No trenchant differences in the form or arrangement of the epibranchials of Bathybates and Rhamphochromis were recognized Other cichlid taxa In the great majority of taxa examined the epibranchials are of the form described above. However, amongst the Lake Tanganyika genera all species of Trematocara possess a characteristically shaped Ep4 (Fig. 16B). In these fishes the head of Ep4 is not expanded or cupped, and the shank spine is reduced in size. The absence of this expanded head in Trematocara is interpreted as a secondary loss within the family rather than as the retention of the plesiomorphic perciform condition. It has been suggested elsewhere (page 77) that Trematocara is the sistergroup of Hemibates and that together these two form the sistergroup of Bathybates. A typical cichlid expansion of the head of Ep4 is present in both Hemibates and Bathybates. It must be assumed therefore that the ancestor of the whole group possessed an Ep4 with an expanded head. If this was the case then the expansion must have been secondarily lost in Trematocara species. Myological features of the PJA The dorsal branchial muscles develop from the muscle plates which are formed in each of the branchial segments, and may be grouped according to their developmental origin and innervation (Edgeworth, 1935). Group one muscles The levatores externi muscles. The four pairs of levator externus muscles connect the neuro- cranium with the four pairs of epibranchials and the lower pharyngeal element. They originate, together with the levatores interni, as a single muscle mass from the hyoman- dibulad shell of the neurocranium. 92 MELANIE L. J. STIASSNY Iev-int1 ca m.t-p2 m.c-p2 m.t-e2 lev-ext 1 ad5 art.fc-pb3 td-p oes ep4 lev- post Fig. 17 Rhamphochromis longiceps, dorsal view of isolated pharyngeal jaw apparatus. Rhamphochromis (Fig. 1 7) FIRST LEVATOR EXTERNUS. This rostral levator passes caudoventrally from the hyoman- dibulad shell to insert musculously onto the dorsal face of Epl. The insertion site is quite extensive and is situated on the dorsomedial face of Epl at the junction of its rostral and caudal arms. SECOND LEVATOR EXTERNUS. This muscle originates, with the first levator, from the rostral part of the lateral rim of the hyomandibulad shell. It inserts onto the dorsolateral face of Ep2 at a point just above the triangular edge of the caudal border of that bone. The anterolateral fibres are tendinously associated with Ep2; the remaining fibres insert musculously. THIRD LEVATOR EXTERNUS. This is the smallest of the series and is closely apposed to the rostrolateral side of the fourth levator for its entire length. It terminates in a long tendon which inserts onto the tip of the caudolateral process of Ep3. FOURTH LEVATOR EXTERNUS. This is the largest of the levator series. The lateral section originates from the lateral rim of the hyomandibulad shell, the remaining section originating behind the levatores interni from the medial part of the shell. The two sections merge ventrally to form a single muscle mass. A small slip of lateral fibres inserts, via a short tendon, onto the well developed shank spine. The mass of the levator bypasses the shank spine, passing medial to it, and tapers onto a tendon which inserts on the horn of the lower pharyngeal element medial to the insertion of the fifth adductor muscle. Bathybates (Fig. 18) The levatores externi series of Bathybates is similar to that of Rhamphochromis. However, in Bathybates the third levator inserts onto a small aponeurosis and a long tendon is not developed. Other cichlid taxa In other cichlids the mode of insertion of the first three levatores externi is as described for PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 93 Rhamphochromis. The elongation of the tendon of the third levator externus in Rhamphochromis longiceps (and to a lesser extent that of the other Rhamphochromis species) is unusual although the length of this tendon in some other species, for example Prognathochromis prognathus from Lake Victoria, does approach the Rhamphochromis condition. In the majority of taxa examined the third levator externus inserts onto Ep3 via a well developed, but not elongate, tendon. The reduction of this tendon, like that in Bathybates, is also found in other Lake Tanganyika genera, for example Hemibates, Trematocara, Lamprologus, Limnochromis, Aulonocranus, and Ctenochromis. Insufficient data on the distribution of various character states of the third levator insertion are available to permit interpretation of polarity in the character at an intergeneric level. According to Liem & Osse (1975) and Liem (1978) the fourth levator externus in all cichlid fishes is composed of a small strap-like lateral head inserting onto the shank spine of Ep4, and a large medial head which passes medial to the spine to insert tendinously onto the posterior horn of the lower pharyngeal element. It was this shift in the major insertion site of the fourth levators (from the fourth epibranchials to the lower pharyngeal element) that Liem (1973) considered to be part of the 'key innovation' of the Cichlidae. In Trematocara the fourth levator externus does not retain an attachment on Ep4 and the entire muscle passes over the head of Ep4 to insert tendinously on the horn of the lower pharyngeal element. The loss of a lateral section inserting onto Ep4 is interpreted as a synapomorphy uniting the species of the genus. The levator posterior muscle. According to Edgeworth (1935) the levator posterior represents a fifth levator externus that has migrated caudally to originate from the ventrolateral face of the pterotic region of the neurocranium. The levator posterior is separated from the rostral levators by a hiatus within which the adductor operculi muscle is situated. Rhamphochromis (Fig. 1 7) The fibres of this large muscle pass ventrocaudally from their origin on the ventral face of the pterotic and intercalar bones to insert musculously on the dorsolateral margin of Ep4 and the dorsal border of the shank spine. Some ventromedial fibres intercalate with the dorsal fibres of the medial section of the fourth levator externus. Bathybates (Fig. 18) The origin of the levator posterior is situated further medially in Bathybates than in Rhamphochromis, and is restricted to the intercalar bone. Other cichlid taxa In the majority of cichlid taxa examined the levator posterior originates from the ventro- lateral region of the lateral awning, usually from the pterotic and intercalar bones (see Hoogerhoud & Barel, 1978). In Hemibates the origin of the levator has migrated medially onto the exoccipital bone at a site just anterior to the vagus foramen. This shift in origin does not appear to be correlated with the presence of an inflated otic bulla since in Aulonocranus, another genus with an inflated bulla, the origin of the levator is in the usual position. Similarly in Trematocara although the origin of the levator is more medially situated than in Rhamphochromis it is still restricted to the lateral awning. The medial migration of the levator posterior in Hemibates is interpreted as the derived condition and it represents the end point of a trend visible in both Bathybates and Trematocara. The levatores interni muscles. The two pairs of levatores interni muscles originate, together 94 MELANIE L. J. STIASSNY lev-extl Iev-int1 ca m.t-p2 m.c-p2 \ \ \ ep1 m.t-e2 cart.ext-ep2 Iev-int2 Iev-ext4 ep4 od-p td-p ad 5 Fig. 18 Bath ybates ferox, dorsal view of isolated pharyngeal jaw apparatus. with the levatores externi 1-3, from the hyomandibulad shell between the lateral and medial sections of the fourth levator externus. Rhamphochromis (Fig. 1 7) FIRST LEVATOR INTERNUS. This rostral levator originates on the hyomandibulad shell medial to the second levator internus. It is a relatively large muscle that passes ventrally (and slightly caudally) and tapers into a short aponeurosis that inserts on the dorsomedial junction of Pb2 and Pb3 just caudal to the uncinate process of Ep 1 . SECOND LEVATOR INTERNUS. The second levator is slightly pinnate and originates lateral to the first levator internus. Fibres pass ventrocaudally to insert via a short aponeurosis onto the ventrolateral face of Pb3 just anterior to the headofEp3. Bathybates (Fig. 18) The levatores interni have similar sites of origin and insertion to those described above for Rhamphochromis. The first levator inserts musculously and is slightly expanded at its base. Other cichlid taxa The size of the levatores interni varies considerably amongst cichlid fishes (see Liem & Stewart, 1976). Hoogerhoud & Barel (1978) interpret the relatively large levatores interni of piscivorous species to be part of a complex of adaptations associated with the trituration of prey. Insufficient outgroup data render ambiguous any phylogenetic interpretation of the differences between the insertion of the first levator internus in Rhamphochromis (via a short aponeurosis) and that in Bathybates (musculously). PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 95 m.t-e2 cart.ext.ep2 Iev-int1 t.lev-intl Iev-int2 od-a m.t-p2 m.c-|_2 art.fc-pb3 od-p Fig. 19 Trematocara marginatum, dorsal view of anterior region of isolated pharyngeal jaw apparatus (left side). In Hemibates and Trematocara (Fig. 19) the first levator internus muscles are rostrocaudally expanded and insert via well developed strap-like tendons. The rostrocaudal expansion of the first levator occurs in other species with inflated otic bullae eg. Aulonocranus and Aulonocara (cf. Hoogerhoud & Barel, 1978) but the presence of a well developed strap-like tendon is not so correlated. It is therefore assumed that this strap-like tendon is a synapomorphy shared by Trematocara and Hemibates. The adductores muscles. Rhamphochromis With the exception of the fifth muscle the adductores are situated in the angle between the dorsal part of the ceratobranchial and the shank of the epibranchial of each arch. The fifth adductor is the caudal representative of the series whose site of attachment has shifted onto Ep4 with the loss of Ep5. The fifth adductor is spindle shaped and it passes from the shaft of Ep4 at a point below the shank spine to insert onto the horn of the lower pharyngeal element. A few ventromedial fibres intercalate with those of the fourth levator externus. No trenchant differences in the adductor series ofBathybates were recognized Group two muscles The remaining muscles of the dorsal gill arches are all derived from the sphincter oesophagi (Edgeworth, 1935) which is itself a derivative of the upgrowth around the oesophagus of the ventral ends of the muscle plates of the fifth branchial arch (Holstvoogd, 1965). All these muscles are innervated by branches of the vagus nerve (Edgeworth, 1935). The sphincter oesophagi. The sphincter oesophagi forms a continuous muscle sheath around the oesophagus. No separate muscle bundle crosses the dorsal midline anterior to the entrance of the retractor dorsalis into the oesophageal tissues. The transversi dorsales muscles Rhamphochromis (Fig. 1 7) As in all cichlids there are two distinct parts of this muscle, the transversus dorsalis anterior and posterior. 96 MELANIE L. J. STIASSNY The transversus dorsalis anterior is a large tripartite muscle. The anterior part (the musculus transverus pharyngobranchialis 2 of Anker, 1978) is a relatively small, well defined muscle. It arises from the lateral part of the rostral face of Pb2 and its fibres pass medially across the midline to attach to the Pb2 of the opposite side. The second part of this muscle complex (the musculus cranio-pharyngobranchialis 2 of Anker, 1978) connects the second pharyngobranchials with each other as well as with the neurocranium. The anterior fibres pass medially and insert on a complex median aponeurosis which in turn is attached to the parasphenoid bone at the base of the pharyngeal apophysis. The bulk of the muscle passes caudomedially to join the central aponeurosis of the third part of the muscle and the fibres of the opposite side. The third part (the musculus transversus epibranchialis 2 of Anker, 1978) originates from the dorsal face of Ep2; its fibres pass medially to insert on a flat, strip-like aponeurosis which traverses the anterior face of the articulatory facets of the third pharyngobranchials. The transversus dorsalis posterior (the musculus transversus epibranchialis 4 of Anker, 1978) is separated from the anterior muscle complex by a hiatus so that the articulatory facets of the third pharyngobranchials are exposed to form a diarthrosis with the pharyngeal apophysis. The strap-like transversus dorsalis posterior originates from the caudal eminence formed at the junction of Pb3 and UP4. The fibres pass medially and are not interrupted by a central aponeurosis. Bathybates (Fig. 1 8) The arrangement of this muscle complex is similar to that just described. However, the musculus cranio-pharyngobranchialis 2 is considerably larger in Bathybates than is its counterpart in Rhamphochromis. This is also the case in other Lake Tanganyika genera, whilst in other Lake Malawi genera and in the riverine Serranochromis the muscle proportions are as described in Rhamphochromis. Unfortunately too few outgroup data preclude any phylogenetic interpretation of these differences. The obliqui dorsales muscles. Winterbottom (1974) discusses the nomenclatural confusion that has centred around the muscles connecting the posteromedial face of Ep4 and the dorsal tip of the fifth ceratobranchial. It is considered that, with the loss of the fifth pharyngo- branchial, the obliquus posterior represents a part of the obliquus dorsalis whose medial site of attachment has shifted to Ep4. Rhamphochromis (Fig. 17) The obliquus dorsalis anterior is a well developed muscle. It originates from the lateral wall of the articulatory facet of Pb3 and its fibres pass caudomedially to insert along the inner face of the expanded head of Ep4. Some fibres also insert along the caudolateral process of Ep3. The obliquus dorsalis posterior is also well developed. It connects the expanded head of Ep4 with the horn of the lower pharyngeal element. The fibres of the obliquus posterior insert onto the tendon of the fourth levator externus, and together these two muscles insert tendinously on the horn. Bathybates (Fig. 18) The obliquus posterior is markedly reduced and is represented by a small strand of fibres originating from the dorsomedial face of the head of Ep4; as a result most of the head of that bone is exposed. Other cichlid taxa Liem (1978:346) states that '. . . the obliquus posterior is either weakly developed or absent in piscivorous cichlids.' Observations made during the course of this study contradict that assessment. For example in the Rhamphochromis, Boulengerochromis, Serranochromis, Lamprologus, and Hemibates species examined, as well as in "Haplochr amis' caeruleus, "Haplochromis* dimidiatus, 'Haplochromis' woodi and Prognathochromis prognathus, many PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 97 of which species were studied by Liem, an obliquus posterior is present and is not markedly reduced. On the basis of these data, as well as the widespread occurrence of a well developed obliquus posterior in the majority of taxa from other trophic groups, the reduced area of origin and total size of the obliquus posterior in Bathybates is interpreted as a derived character. The retractor dorsalis muscles. The bilaterally paired muscles connect the posterior pharyngobranchial elements with the vertebral column. The size of the retractor varies markedly within the Cichlidae (Hoogerhoud & Barel, 1978) and the site of origin is from the ventral face of the anterior vertebrae and includes the apophysis that is developed on the third, fourth or fifth abdominal vertebrae (Trewavas, 1964; Greenwood, 1979). Rhamphochromis (Fig. 1 7) The retractor originates from the ventrolateral face of the first, second, and third abdominal vertebrae and from the ventral apophysis on the third vertebra. Its fibres pass rostro- ventrally and insert musculously as two distinct bundles on the mediocaudal face of the third pharyngobranchials; a few fibres insert on the rostral region of UP4. No trenchant differences in the retractor of Bathybates were recognized. Bathybates Hemibates Trematocara D-F A-C Fig. 20 Cladogram illustrating the hypothesis of relationship for Bathybates. Apomorphic characters defining the genera Bathybates, Hemibates and Trematocara are discussed in the text (see Stiassny, 1980 for additional data). Synapomorphies: A. Laterally expanded ascending process of the anguloarticular. B. Palatine-lateral ethmoid ligament absent. C. The dorsal bony bridge reduced or absent. D. Tendon of the first levator internus broad and strap-like. E. Palatine-mesethmoid ligament present. F. Elongate rostral cartilage extending below the articular processes of the premaxillae. 98 MELANIE L. J. STIASSNY Discussion A number of apomorphic characters has been identified amongst the cichlid taxa studied and on the basis of these features it is possible to establish the monophyly of Bathybates and to generate an hypothesis of phylogenetic relationships for that genus (Fig. 20). The relatively high degree of morphological differentiation exhibited by the Lake Tanganyika Cichlidae renders that flock a most suitable subject for cladistic analysis at the intergeneric level. This was not found to be the case with the Lake Malawi Cichlidae. Although the range of morphological differentiation throughout the whole flock is greater in Lake Malawi than in Lake Victoria, with respect to the piscivorous species at least, a similar situation exists in the two lakes. The range of morphological variation amongst the various taxa is narrow and in most cases if structures are simply regarded in terms of presence or absence no character differences are distinguishable. In this respect Greenwood's (1974) conclusion that the different characteristics identified in the Lake Victoria 'Haplochromis' flock are but slight variants of a basic 'bauplan' developed and differentiated by ontogenetic reorganization, may be broadened to include the piscivorous grade of Lake Malawi. This is not to imply that a close phylogenetic relationship exists between the Lake Victoria and Lake Malawi haplochromines or that the Lake Malawi piscivores are necessarily a phylogenetic lineage. None of the character complexes investigated in the preceding sections has revealed apomorphic character states in Rhamphochromis. Although all of the species currently included in this genus do have a highly distinctive 'fades' such an overall similarity does not, in itself, constitute evidence of monophyly. Given the extreme appearance of these Rhamphochromis species it is surprising that very few apomorphic characters can be found to define the genus. Regan (1921) was of the opinion that the beak-like expansion of the premaxillae (Fig. 6A) characterized the genus, but similar premaxillary expansion is also present in 'Haplochromis' caeruleus, 'Haplochromis' strigatus, and 'Haplochromis' compressiceps from Lake Malawi as well as in some Lake Victoria piscivores. Regan also noted that the anterior teeth of the second series in the upper jaw are enlarged in Rhamphochromis species. This feature does appear to characterize all the species of Rhamphochromis. In all Rhamphochromis species the urohyal bone lacks an anterodorsal process. In the great majority of other cichlid species the urohyal bears a distinct anterodorsal spine (Barel el al., 1976) and thus its presence is interpreted as a plesiomorphic character within the Cichlidae. Rhamphochromis is unique amongst Lake Malawi Cichlidae in lacking this spine. The loss of the spine occurs mosaically amongst the Lake Tanganyika genera and therefore is assumed to have taken place independently a number of times within that lake. All of the Lake Victoria taxa examined have an anterodorsal spine on the urohyal. All Rhamphochromis are large, elongate, streamlined fishes and similar features also characterize the piscivorous grade of Lake Victoria 'Haplochromis' (Greenwood, 1962, 1974). In Lake Victoria body elongation is not accompanied by a marked increase in the total number of vertebrae. The range of vertebral counts for these species is 30-32, whilst that of the more 'generalized' forms is 27-30 (Greenwood, 1962; Greenwood & Barel, 1978). This slight increase in total number involves an increase in the number of caudal, rather than abdominal vertebrae (Greenwood, 1979). Amongst the piscivores of Lake Malawi an increase in size and elongation of the body is frequently accompanied by an increase in the total number of vertebrae, and many species have a higher total count than their Lake Victoria counterparts. For example; 'Haplo- chromis' strigatus 30-32 (13-14+ 17-18), 'Haplochromis' dimidiatus 32-33 (13-14+ 18- 20), 'Haplochromis' caeruleus 34 (15 + 19), 'Haplochromis' spilorhynchus, 33-34 (15 + 17-18), 'Haplochromis' lepturus 33-35 (15-16+ 18-19), 'Haplochromis' macrostoma 31-32 (13-14+17-18), 'Haplochromis' compressiceps 31-32 (13-14+17-18), Aristo- chromis christyi 32(14-15 + 17-18), Diplotaxodon argenteus 33(16+17). Rhamphochromis stands alone amongst the Lake Malawi Cichlidae in possessing a much PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 99 higher total number of vertebrae (as many as 39 in some specimens of Rhamphochromis leptosomd). Furthermore, Rhamphochromis is easily distinguished from the other Lake Malawi (and Lake Victoria and Lake Tanganyika) cichlids by the fact that the increase in the total number of vertebrae involves an increase in the number of abdominal vertebrae. In all species of Rhamphochromis the number of abdominal vertebrae is 1 7 or more. This increase in the total number of vertebrae, and the increase in the number of abdominal vertebrae, are both interpreted as apomorphic characters which serve to distinguish Rhamphochromis from other Lake Malawi haplochromines. But as can be seen from the figures given above '//.' caeruleus, '//.' spilorhynchus '//.' lepturus, Aristochromis and Diplotaxodon display a slight increase in the number of abdominal vertebrae, and a similar, though more marked increase is found in the members of the Serranochromis lineage (Greenwood, 1979; Trewavas, 1964; Bell-Cross, 1975). Greenwood (1979) has suggested that Serranochromis may have contributed to the Lake Malawi flock; the shared apomorphic character of an increased number of abdominal vertebrae found in Serranochromis and Rhamphochromis, and to a lesser extent, also in '//.' caeruleus, '//.' spilorhynchus '//.' lepturus, Aristochromis and Diplotaxodon may reflect a close relationship amongst these fishes. In many ways Rhamphochromis represents an endpoint in an evolutionary trend towards the production of a highly specialized morphotype. Though often extreme, the characters involved in the production of this large mouthed, streamlined fish are linked through a gradal series to those found in the less modified piscivores of Lakes Malawi and Victoria. From a knowledge of intra- and interspecific variation in meristic characters, neurocranial and dental morphology, and from ecological and biological data for much of the Lake Victoria flock, Greenwood was later able to breakdown the piscivorous grade into two phyletic lineages (1980). Unfortunately, comprehensive data of this nature are not available for the majority of the Malawi Cichlidae and in their absence a similar breakdown of the Malawi piscivores cannot be achieved. The level of discrimination provided by this purely anatomically based cladistic investigation has proven to be adequate when applied to an intergeneric analysis of the Lake Tanganyika cichlids but insufficient to detect salient character differences within the Lake Malawi flock. It seems that in the face of low morphological differentiation combined with a high level of species proliferation, a cladistic approach relying upon purely anatomical data is stretched to its limits. If the phylogenetic relationships of Rhamphochromis and the other Lake Malawi cichlids are to be resolved then ecological, ethological and possibly physio- logical characters must be employed within the same framework (see also Greenwood, 1980). Acknowledgements I am indebted to the Trustees of the British Museum (Natural History) for allowing me access to the facilities and collections of the museum, and to Prof. Dr P. Dullemeijer, Prof. G. Chapman and Dr G. J. Sheals in whose departments I carried out this research. I would also like to thank Drs P. H. Greenwood and C. D. N. Barel for their guidance, encouragement and critical reading of the manuscript. I am grateful to all the members, past and present, of the fish section (BMNH) who have been so generous with their time and skills. In particular I thank Augustine Baddokwaya, Dr Keith Banister, Jim Chambers, Maggie Clarke, Dr Di Kershaw, Dr Rosemary Lowe- McConnell, Alice Parshall, Dr Gordon Reid, and Robert Travers for making my stay in the section so enjoyable. A special word of thanks goes to Dr Ethelwynn Trewavas for her kindness and support, and to Gordon Howes whose help and friendship have been so unstintingly given and invaluable. 1 00 MELANIE L. J. STIASSN Y References Alexander, R. McN. 1967. The functions and mechanisms of the protrusible upper jaws of some acanthopterygian fishes. J. Zoo/., Land. 151: 43-64 Anker, G. Ch. 1978. The morphology of the head-muscles of a generalized Haplochromis species: H. elegans Trewavas 1933 (Pisces, Cichlidae). Neth. J. Zoo/. 28: 234-271. Barel, C. D. N., Witte, F. & van Oijen, M. J. P. 1976. The shape of the skeletal elements in the head of a generalized Haplochromis species: H. elegans Trewavas 1933 (Pisces, Cichlidae). Neth. J. Zoo/. 26: 163-265. Barel, C. D. N., van Oijen, M. J. P., Witte, F. & Witte-Maas, E. L. M. 1977. An introduction to the taxonomy and morphology of the haplochromine Cichlidae from Lake Victoria. Neth. J. Zoo/. 27: 333-389. Bell-Cross, G. A. 1975. A revision of certain Haplochromis species (Pisces: Cichlidae) of Central Africa. Occ. Pap. natn. Mus. Rhod. ser. B. 5 (7): 405^464. Boulenger, G. A. 1898. Report on the collection of fishes made by Mr. J. E. S. Moore in Lake Tanganyika during his expedition, 1895-1896. Trans, zoo/. Soc. Lond. 15: 1-30. 1915. Catalogue of the fresh- water fishes of Africa 3. London. Chichoki, F. 1976. Cladistic history of cichlid fishes and reproductive categories of the American genera. Ph.D. thesis, Univ. Michigan, Ann Arbor. Coulter, G. W. 1966. The deep benthic fishes of the south of Lake Tanganyika with special reference to the distribution and feeding in Bathybates species, Hemibates stenosoma and Chrysichthys species. Fish. Res. Bull. Zambia4: 33-38. 1967. Low apparent oxygen requirements of deep-water fishes in Lake Tanganyika. Nature, Lond. 215:317-318. Dingerkus, G. & Uhler, L. D. 1977. Enzyme clearing of Alcian blue stained whole small vertebrates for demonstration of cartilage. Stain Tech. 52 (4): 229-232. Edgeworth, F. H. 1935. The cranial muscles of vertebrates. 493 pp. London. Fryer, G. 1959. The trophic interrelationships and ecology of some littoral communities of Lake Nyasa with especial reference to the fishes, and a discussion of the evolution of a group of rock frequenting Cichlidae. Proc. zoo/. Soc. Lond. 132: 153-281. 1969. Speciation and adaptive radiation in African lakes. Verh. int. Verein. theor. angew. Limnol. 17: 303-322. Fryer, G. & lies, T. D. 1972. The cichlid fishes of the Great Lakes of Africa. Their biology and evolution. Edinburgh. Galis, F. & Barel, C. D. N. 1980. Comparative functional morphology of the gills of African lacustrine Cichlidae (Pisces, Teleostei). Neth. J. Zoo/. 30: 392^30 Goedel, W. von 1974a. Beitrage zur vergleichenden und funktionellen Anatomic des Kopfes von Tilapia (Cichlidae, Teleostei). Teil I. Zoo/. Jb. (Anat.)92: 220-274. 19740. Beitrage zur vergleichenden und funktionellen Anatomic des Kopfes von Tilapia (Cichlidae, Teleostei). Teil II. Zoo/. Jb. (Anat.) 92: 321-383. Greenwood, P. H. 1962. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae), Part V. Bull. Br. Mus. nat. Hist. (Zool.)9: 139-214. 1974. Cichlid fishes of Lake Victoria, east Africa: the biology and evolution of a species flock. Bull. Br. Mus. nat. Hist. (Zool.) Suppl. 6: 1-134. 1978. A review of the pharyngeal apophysis and its significance in the classification of African cichlid fishes. Bull. Br. Mus. nat. Hist. (Zool.) 33: 297-323. 1979. Towards a phyletic classification of the 'genus' Haplochromis and related taxa. Part I. Bull. Br. Mus. nat. Hist. (Zool.) 35: 265-322. 1980. Towards a phyletic classification of the 'genus' Haplochromis and related taxa. Part II. Bull. Br. Mus. nat. Hist. (Zool.) 39: 1-101. & Barel, C. D. N. 1978. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae). Part VIII. Bull. Br. Mus. nat. Hist. (Zool.) 33: 141-192. Holstvoogd, C. 1965. The pharyngeal bones and muscles in Teleostei, a taxonomic study. Proc. K. ned. Akad. Wet. (Zool.) Ser C. 68 (3): 209-2 1 8. Hoogerhoud, R. J. C. & Barel, C. D. N. 1978. Integrated morphological adaptations in piscivorous and mollusc-crushing Haplochromis species. In Zodiac symposium on adaptation. Wageningen, The Netherlands: 52-56. Ismail, M. H. 1979. The ontogeny of the head parts in Haplochromis elegans Trewavas, 1933 (Teleostei, Cichlidae). Ph.D. thesis, Faculty of Science, State Univ. of Ghent, Belgium. PHYLOGENETIC VS CONVERGENT RELATIONSHIP BETWEEN PISCIVOROUS CICHLID FISHES 1 0 1 Kesteven, H. L. 1943. The evolution of the skull and cephalic muscles. I. The Fishes. Mem. Aust. Mus. 8 (2): 65-1 32. Liem, K. F. 1973. Evolutionary strategies and morphological innovations: cichlid pharyngeal jaws. Syst. Zool. 22:425-441. 1978. Modulatory multiplicity in the functional repertoire of the feeding mechanism in cichlid fishes. Part I. Piscivores. /. Morph. 158: 323-360. 1979. Modulatory multiplicity in the feeding mechanism in cichlid fishes, as exemplified by the invertebrate pickers of Lake Tanganyika. J. Zool., Lond. 189: 93-125. & Osse, J. W. M. 1975. Biological versatility, evolution, and food resource exploitation in African cichlid fishes. Am. Zool. 15: 427-454. & Stewart, D. J. 1976. Evolution of the scale-eating cichlid fishes of Lake Tanganyika: a generic revision with a description of a new species. Bull. Mus. comp. Zool. Harv. 147: 3 19-350. Lowe-McConnell, R. H. 1975. Fish communities in tropical freshwater s. 337 pp. London and New York. Nelson, G. J. 1969. Gill arches and the phytogeny of fishes with notes on the classification of vertebrates. Bull. Am. Mus. nat. Hist. 141: 475-552. Patterson, C. 1975. The braincase of pholidophorid and leptolepid fishes, with a review of the actinopterygian braincase. Phil. Trans. R. Soc. Ser B. 269 (899): 275-574. & Rosen, D. E. 1977. Review of the ichthyodectiform and other Mesozoic teleost fishes and the theory and practice of classifying fossils. Bull. Am. Mus. nat. Hist. 158: 81-1 72. Pellegrin, J. 1903. Contribution a 1'etude anatomique, biologique et taxonomique des poissons de la familledescichlides. Mem. Soc. zool. Fr. 16: 14^402. Poll, M. 1956. Poissons Cichlidae. Result, sclent. Explor. hydrobiol. Lac Tanganika (1946-47) 3, Fasc. 5b: 1-619. 1978. Contribution a la connaissance du genre Lamprologus Schth. description de quatre espece nouvelles rehabitation de Lamprologus mondabu et synopsis remanie des especes du Lac Tanganyika. Bull Acad. r. Belg. 64: 725-758. Regan, C. T. 1920. The classification of the fishes of the family Cichlidae. I. The Tanganyika genera. Ann. Mag. nat. Hist. (9) 5: 33-53. 1921 . The cichlid fishes of Lake Nyasa. Proc. zool. Soc. Lond. 1921: 675-727. 1922. The cichlid fishes of Lake Victoria. Proc. zool. Soc. Lond. 1922: 1 57-191. Ricardo-Bertram, C. K., Borley, H. J. H. & Trewavas, E. 1942. Report on the fish and fisheries of Lake Nyasa. Crown Agents, London. Rosen, D. E. 1973. Interrelationships of higher euteleosteans. In Interrelationships of fishes (Greenwood, P. H., Miles, R. S. & Patterson, C., Eds) Suppl. 1 , Zool. J. Linn. Soc. 53: 397-5 13. & Patterson, C. 1969. The structure and relationships of the paracanthopterygian fishes. Bull. Am. Mus. nat. Hist. 141: 359^174. & Greenwood, P. H. 1976. A fourth neotropical species of synbranchid eel and the phylogeny and systematics of the synbranchiform fishes. Bull. Am. Mus. nat. Hist. 157: 1-70 Schaeffer, B., Hecht, M. K. & Eldredge, N. 1972. Phylogeny and paleontology. Evolut. Biol. 6: 3 1-57. Stiassny, M. L. J. 1980. The anatomy and phylogeny of two genera of African cichlid fishes. Ph.D. thesis, Univ. London, London. Trewavas, E. 1935. A synopsis of the cichlid fishes of Lake Nyasa. Ann. Mag. nat. Hist. (10) 16: 65-118. 1949. The origin and evolution of the cichlid fishes of the great African lakes, with special reference to Lake Nyasa. 13th Int. Cong. Zool. 1949: 365-368. - 1964. A revision of the genus Serranochromis Regan (Pisces, Cichlidae). Annls. Mus. r. Congo Beige Ser. 8vo, Zool. no. 125: 1-58. 1973. On the cichlid fishes of the genus Pelmatochromis with the proposal of a new genus for P. congicus; on the relationship between Pelmatochromis and Tilapia and the recognition of Sarotherodon as a distinct genus. Bull Br. Mus. nat. Hist. (Zool.) 25: 1-26. Vetter, B. 1874. Untersuchungen zur vergleichenden Anatomic der Kiemen- und Kiefermusculatur der Fische. Jena Z. Naturw. 8: 405^58. Winterbottom, R. 1974. A descriptive synonomy of the striated muscles of the Teleostei. Proc. Acad. nat. Sci. Philad. 125 (12): 224-317. Witte, F. in press. Trophic composition and breeding seasonality of the haplochromine Cichlidae (Pisces: Teleostei) from the Mwanza Gulf of Lake Victoria. Manuscript accepted for publication 1 8 September 1980 British Museum (Natural History) 1881-1981 Centenary Publications Chance, change & challenge Two multi-author volumes from one of the foremost scientific institutions in the world. General Editor: P. H. Greenwood The Evolving Earth Editor: L. R. M. Cocks The Evolving Biosphere Editor: P. L. Forey In the first volume, The Evolving Earth, twenty scientists have been asked to review the present state of knowledge in their particular field, ranging from the origin of the Earth, through ocean sediments and soils to continental drift and palaeogeography. In the companion volume, The Evolving Biosphere, museum scientists have chosen an evolutionary concept — speciation, coevolution, biogeography etc. and related this to the group of animals or plants in which they are specialising. Thus beetles and birds exemplify sympatric and allopatric speciation, butterflies mimicry and certain fishes explosive evolution. In both volumes the text is supplemented by over one hundred specially-commissioned pieces of two-colour artwork. These two books will be invaluable to all sixth-form and undergraduate biology and geology students. The Evolving Earth: 276x219 mm, 280pp, 138 line illustrations, 42 halftones The Evolving Biosphere: 276x219 mm, approx. 320pp, 133 line illustrations Publishing: Spring 1981 Co-published by the British Museum (Natural History), London and Cambridge University Press, Cambridge. Titles to be published in Volume 40 Eugene Penard's Slides of Gymnamoebia : re-examination and taxonomic evaluation. By F. C. Page Japanese earthworms: a synopsis of the Megadrile species (Oligochaeta). By E. G. Easton Phylogenetic versus convergent relationship between biscivorous cichlid fishes from Lakes Malawi and Tanganyika. By M. L. J. Stiassny Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester GENE -4 AU Bulletin of the British Museum (Natural History) Miscellanea Zoology series Vol 40 No 4 30 July 1981 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four scientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology, and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself, available separately, and individually priced. Volumes contain about 300 pages and several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the individual parts. Orders and enquiries should be sent to : Publications Sales, British Museum (Natural History), Cromwell Road, London SW7 5BD, England World List abbreviation: Bull. Br. Mus. nat. Hist. (Zool.) © Trustees of the British Museum (Natural History), 1981 ISSN 0007-1498 Zoology series Vol 40 No 4 pp 103-209 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 30 July 1981 *> GENERAI " 4 AUG J LIBRARY Miscellanea Contents Page The calceolus, a sensory structure of gammaridean amphipods (Amphipoda: Gammaridea). By R. J. Lincoln & D. E. Hurley 103 A new species of Lernaea (Copepoda: Cyclopoida) from Papua — New Guinea. By G. A. Boxshall 117 Some type specimens of Isopoda (Flabellifera) in the British Museum (Natural History), and the isopods in the Linnaean Collection. By J. Ellis . . 121 Conchoecia hystrix n. sp. a new halocyprid ostracod for the Porcupine Bight region of the Northeastern Atlantic. By M. V. Angel & C. Ellis ..... 129 The Conchoecia skogsbergi species complex (Ostracoda, Halocyprididae) in the Atlantic Ocean. By A. J. Gooday 137 The calceolus, a sensory structure of gammaridean amphipods (Amphipoda: Gammaridea) Roger J. Lincoln Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Desmond E. Hurley New Zealand Oceanographic Institute, DSIR, P.O. Box 12-346, Wellington North, New Zealand Introduction The calceolus is a microscopic external surface structure, presumed to serve a sensory function, found on the antennae of a select group of amphipods belonging to the suborder Gammaridea. It occurs in only about 10 per cent of the known gammaridean species, and is absent elsewhere in the Crustacea. First noted by Milne-Edwards in 1830 and referred to as the 'cupule membraneuse', it later acquired the name 'calceolus' because of its slipper- shaped profile when viewed under a microscope. A good account of contemporary knowledge was provided by Blanc (1883, 1884). Despite this early recognition, calceoli have since received only limited attention from taxonomists and physiologists, and remain largely enigmatic inconsistently-documented structures. Their occurrence is uncertain in many species, they are poorly understood in terms of morphology and ontogenetic development, and their precise function has yet to be established. The small size of calceoli (20-300 /zm) is the probable explanation for this lack of attention, since good resolution of their intricate surface structure is almost impossible using conventional light microscopy, and they are easily overlooked at the lower magnifi- cations often used in figuring antennae for taxonomic work. A few attempts have been made by taxonomists to draw calceoli at high magnification under a light microscope, and some idea of the general profile and surface pattern has been obtained, but the true three- dimensional complexity of the structure cannot be appreciated. Some of the earliest drawings of calceoli are as good as or better than most illustrations in recent literature. Calceoli have not always been reliably distinguished from aesthetascs. These also occur frequently on amphipod antennae but have a much simpler structure. Confusion has been especially noticeable in taxonomic work on freshwater amphipods which may have aesthetascs of unusually large size. Unlike calceoli, aesthetascs are found widely throughout the Crustacea, and are thought to function as chemoreceptors. Aesthetascs in amphipods generally have a very simple spatulate shape and are restricted to the flagellum of antenna 1 . The structurally more bizarre calceoli are found on antenna 2 or both antenna 1 and 2, but not on antenna 1 alone. In some species, as for example Eusirus antarcticus Thomson, calceoli and aesthetascs occur together on the flagellar articles of the same individual dispelling any thoughts that calceoli and aesthetascs might simply be variants of the same surface structure. In E. antarcticus the calceolus is about one-third the length of the aesthetasc. We have assembled a considerable amount of data on the occurrence and distribution of calceoli amongst amphipods but have found surprisingly little ecological or biological pattern in this information. For certain, calceoli do not occur outside the suborder Gammaridea, but of the 80 or so families of gammarideans presently recognised only 19 Bull. Br. Mus. not. Hist. (Zool.) 40 (4) : 103-1 16 Issued 30 July 198 1 104 R. J. LINCOLN & D. E. HURLEY Table 1 Superfamilies and families of gammaridean amphipods (after Bousfield, 1978). Calceoliferous families are shown in bold capital letters. Phoxocephaloidea Oedicerotoidea Pardaliscoidea PHOXOCEPHALIDAE OEDICEROTIDAE Pardaliscidae UROTHOIDAE Stilipedidae PLATYISCHNOPIDAE Leucothoidea Hyperiopsidae Pleustidae Astyridae Lysianassoidea Laphystiopsidae Vitjazianidae LYSIANASSIDAE Amphilochidae Leucothoidae Liljeborgioidea Pontoporeioidea Anamixidae Liljeborgiidae HAUSTORIIDAE Maxillipiidae Sebidae PONTOPOREIIDAE Colomastigidae Salentinellidae Pagetinidae Gammaroidea GAMMARIDAE Nihotungidae Stenothoidae Dexaminoidea Atylidae ACANTHOGAMMARIDAE Cressidae Anatylidae ANISOGAMMARIDAE MESOGAMMARIDAE Thaumatelsonidae Lepechinellidae Dexaminidae Gammaroporeiidae Macrohectopidae Talitroidea Hyalidae Prophliantidae Typhlogammaridae Pontogammaridae Hyalellidae Talitridae Ampeliscoidea Ampeliscidae Crangonyctoidea CRANGONYCTIDAE NEONIPHARGIDAE PARAMELITIDAE Niphargoidea Niphargidae Bogidielloidea Bogidiellidae Eusiroidea EUSIRIDAE PONTOGENEIIDAE CALLIOPIIDAE GAMMARELLIDAE AMATHILLOPSIDAE Bateidae Paramphithoidae Ceinidae Dogielinotidae Najnidae Eophliantidae Phliantidae Temnophliantidae Kuriidae Stegocephaloidea Stegocephalidae Acanthonotozomatidae Ochlesidae Lafystiidae Melphidippoidea Melphidippidae Melitoidea Hadziidae Melitidae Carangoliopsidae Corophioidea Photidae Isaeidae Ischyroceridae Ampithoidae Biancolinidae Aoridae Cheluridae Corophiidae Podoceridae contain calceoliferous species (Table 1), and these are restricted to just 7 of the 19 super- families (as proposed by Bousfield (1978) in a recent revision of the group). Even within these families the calceoli are far from uniformly distributed; some genera are entirely non- calceoliferous, others have both calceoliferous and non-calceoliferous species. Ecologically, the calceoliferous species show no special pattern — they may occur in marine, brackish water or freshwater (including hypogean) habitats, from shallow to abyssal depths, in polar, temperate or tropical regions, and may be active swimmers, or burrowers, or live in algae. We could find no obvious correlation of the presence or absence of calceoli with behavioural patterns. An additional dimension of variability is suggested by recent ecological work which affirms that calceoli may be present or absent in different populations of the same species, or from different samples of the same population taken at different seasons of the year (Minkley & Cole, 1963; Cole, 1970; Goedmakers, 1972; Croker & Gable, 1977), although we have CALCEOLUS OF GAMM ARIDEAN AMPHIPODS 1 05 some reservations as to the universality of these statements. Jazdzewski (1977) suggests that other considerations such as the appearance of calceoli only in males of a certain age may also account for apparent variability of occurrence within species' populations. A checklist of all known calceoliferous amphipod species is given in a recent paper by Hurley (1980). This tabulation includes only those species that have actually been described or figured in the literature as having calceoli, excluding any that are calceoliferous by inference alone (e.g. generic diagnosis). The list comprises 584 species and subspecies belonging to 134 different genera, from an estimated 5000 species and an estimated 1000 genera in the Gammaridea as a whole. Within the 19 calceoliferous families, of approxi- mately 375 genera and 2000 species, the proportion of species possessing calceoli is a little less than one-third. Hurley's compilation shows up some trends in the location of the calceoli on the antennae in the different families. In haustoriids, phoxocephalids and lysianassids, only the male has calceoli which may occur on antenna 1 and antenna 2, although in lysianassids they are absent from peduncular articles. Gammarids, acanthogammarids, anisogammarids and mesogammarids have a few species with calceoliferous females, but calceoli are typically restricted to the flagellum of antenna 2 in males. The crangonyctids have a similar pattern to the 4 gammaroid families above except that the calceoli occur on the peduncle as well as the flagellum of antenna 2. Eusiroids (Eusiridae, Pontogeneiidae, Calliopiidae, Gammarellidae, Amathillopsidae) are commonly calceoliferous in both sexes and on both antennae. The function of calceoli has received very little direct attention and is far from resolved. They have variously been considered organs for clasping, copulation, and taste, and more recently linked with pheromone reception (Dahl et al, 1970) but only the latter hypothesis is supported by direct experimental evidence. However, from structural and other evidence we would argue against a chemosensory role for calceoli. We believe the structural complexity of the calceoli involves some form of sound, vibration or pressure wave sensitivity. Scanning electron microscopy was used to examine the calceoli of more than 60 different amphipod species in some 40 genera representing most of the calceoliferous families. We looked first at the morphology of a wide range of calceoli and applied this information to the problem of function. An unexpected bonus, following from the recognition of distinct structural designs amongst the calceoli, has been the rewarding prospect of using them as indicators of phylogenetic affinity. Material and methods All the scanning work for this study was carried out in the E.M. Unit of the British Museum (Natural History) using either a Cambridge 2 A or a Stereoscan 600. Satisfactory results were obtained with antennal preparations that were simply oven dried before coating, although this was later replaced by routine critical point desiccation followed by sputter coating with gold. A variety of different methods for fixing preparations to stubs were tried and most proved adequate, but use of a thin film of Araldite was eventually adopted as the simplest and most effective. All source material from which dissections were made came either from the collections of the BM(NH) or from the N.Z. Oceanographic Institute. Some of this material had been in preservative for many years and had rather too much attached debris for high resolution photomicrography, but in all instances the basic configuration of the calceolus was quite clear, and material preserved in spirit for over a century still gave useful, if not spectacular, results. Results The calceoli of the sixty or so species examined showed considerable morphological diversity from the relatively simplistic condition found in Phoxocephalus and Urothoe, to 106 R. J. LINCOLN & D. E. HURLEY the highly complex structures in Eusirus, Amathillopsis, Chosroes, and others. Despite this architectural variety, a certain basic design was evident throughout. The typical calceolus (Figs la, 3a) has two surface components, which we have designated the proximal (p.e.) and distal elements (d.e.), more or less closely attached to the basal receptacle (r.), and a slender stalk (st.). The distal element is characterised by a series of ridges or annulations, or may comprise a number of separate or partially overlapping plates. The proximal element, in contrast, is a single component, either a concave crescent-shaped plate closely applied to the proximal margin of the distal element, or a discrete circular cup that sits freely on the receptacle attached only by a narrow base. In the majority of calceoli examined, except those of phoxocephalids, urothoids and crangonyctids, there is a large bulbous swelling or bulla (b.) at the proximal end of the receptacle close to the attachment of the stalk. In a few of the eusirid calceoli studied the proximal element had been dislodged during the preparation revealing a circular opening in the receptacle through which the base of the proximal dish appeared to connect to the underlying bulla. Within our sample we have been able to recognise just 9 distinct structural types, and have described and illustrated each of these nine different designs, and listed those species allocated to each group. The 9 categories are designated after the significant family com- ponent; gammarid, bathyporeid, lysianassid, pontogeneiid, eusirid, gammarellid, oedicerotid, phoxocephalid and crangonyctid. 1. Gammarid (Fig. la-c) Gammarus duebeni Liljeborg Gammarus locusta (L.) Gammarus pulex (L.) Eulimnogammarusfuscus (Dybowsky) Eulimnogammarus verrucosus (Gerstfeldt) Echinogammarus veneris (Heller) Eogammarus confervicolous (Stimpson) Odontogammarus calcaratus (Dybowsky) Micruropus talitroides (Dybowsky) Micruropus vortex (Dybowsky) Micruropus wahli (Dybowsky) The gammarid calceolus represents one of the simplest configurations. The proximal element forms a weakly concave crescentic plate closely applied along its inner margin to the distal element. The distal element usually has well defined transverse banding ranging from an observed maximum of 25-30 bands in Eulimnogammarus verrucosus and Eogammarus confervicolous, through 10 in Micruropus wahli to as few as 2-3 poorly defined bands in Micruropus vortex and M. talitroides. High magnification of the distal element reveals that the banded markings are not simple ridges but a series of closely overlapping transverse plates, typical of the distal elements of almost all calceoli investigated. Gammarid calceoli are usually confined to males, are typically few in number, and there is only one calceolus on each flagellar article. 2. Bathyporeid (Fig. Id) Bathyporeia guilliamsoniana (Bate) Bat hyporeia pilosa Lindstrom Bathyporeia sarsi Watkin Zaramilla kergueleni Stebbing The bathyporeid calceolus is basically similar to the gammarid-type but is characterised by short tentacle-like projections along the posterior margin of the proximal element. The proximal element is a quite small shallow crescent shaped plate in close contact with the banded distal element. The banding is very much as in the gammarid pattern, and varies Fig. 1 a, Gammarus pule.x, d.e., distal element; p.e., proximal element; r., receptacle; b., bulla; St., stalk: b,, Micruropus wahli: c, Eulimnogammarus verrucosus: d, Bathyporeia sarsi: e, Oediceroides lahillei: f, Parawaldeckia thomsoni. Bar scales = 10 ^m. 108 R.J. LINCOLN &D. E. HURLEY from 7 bands in Bathyporeia sarsi to about 25 in Zaramilla kergueleni. In both Zaramilla and Bathyporeia species the calceoli are present on the flagellum of antenna 1 and antenna 2 in the male only, and there is only one calceolus on any one flagellar article. 3 . Lysianassid (Figs 1 f, 2a-f) Amaryllus macrophthalma Haswell Cheirimedon similis Thurston Hippomedon denticulatus (Bate) Hippomedon holbolli (Kroyer) Lepidepecreum cingulatum Barnard Orchomene plebs (Hurley) Parawaldeckia thomsoni (Stebbing) Pseudorchomene coatsi (Chilton) Socarnes vahli (Kroyer) Tryphosella kergueleni (Miers) Uristes gigas Dana Waldeckia obesa (Chevreux) The two surface elements of the lysianassid calceolus are more or less flattened and partially overlap, the distal element uppermost. The proximal element ranges in shape from a small crescent in Hippomedon holbolli to an almost circular disc in Waldeckia obesa, Parawaldeckia thomsoni and Cheirimedon similis. The distal element has weak surface banding which radiates sublongitudinally from a point close to the proximal margin in Waldeckia, Orchomene, Parawaldeckia, Pseudorchomene, Lepidepecreum and Amaryllus. In Hippomedon, Tryphosa and Socarnes, it appears quite smooth. The proximal element is only weakly concave with a slightly raised outer margin. The distal element is typically flattened, and is rather membraneous at the distal free margin. In lateral view, both surface elements rest rather freely on the receptacle with a small area of attachment near the centre. To support both surface elements the receptacle is elongated and extends almost to the distal margin of the distal element. The bulla is always well developed in the lysianassid calceolus. One surprising and rather anomalous exception to the typical lysianassid design is found in Uristes gigas (Fig. 20 which has the distal element of the calceolus strongly banded in concentric ridges that have their centre of origin close to the distal margin. This configuration has some resemblance to the pontogeneiid-type described below. 4. Pontogeneiid (Fig. 3a-d) Apherusajurinei (Milne-Edwards) Bovallia gigantea Pfeffer Calliopius laeviusculus (Kroyer) Eusiroides monoculoides (Haswell) Eusiropsis riisei Stebbing Eusiroides stenopleura Barnard Haliragesfulvocinctus (Sars) Halirages mixtus Stephensen Paracalliopefluviatilis (Thomson) Paramoera gregaria (Pfeffer) Pontogeneia sp. The pontogeneiid calceolus is constructed along similar lines to the lysianassid-type, but is typically more robust with a distinctly concave proximal element and a large strongly banded distal element. The proximal element has the shape of an almost complete cup in Bovallia gigantea and the 5 species of Eusiroides Eusiropsis and Halirages, and is larger than the distal element in E. monoculoides, subequal in E. stenopleura, and smaller in E. riisei. In contrast, a relatively small crescent-shaped proximal element is found in Calliopius laeviusculus, Apherusa jurinei and Paramoera gregaria, partially overlapped by the larger Fig. 2 a, Orchomene plebs: b, Waldeckia obesa: c, Pseudorchomene coatsi: d, Hippomedon holbolli: e, Socarnes vahli: f, Uristes gigas. Bar scales a-e = 1 0 /^m, f = 20 //m. Fig. 3 a, Calliopius laeviusculus, d.e., distal element; p.e., proximal element; r., receptacle; b., bulla; st., stalk: b, Paramoera gregaria: c, Eusiroides stenopleura: d, Apherusa jurinei: e, Chosroes incisus: f, Crangonyx pseudogracili.s. Bar scales a-d, f = 10 //m, e = 2 /urn. CALCEOLUS OF GAMMARIDEAN AMPHIPODS 1 1 1 distal element. The banding of the pontogeneiid calceolus is usually transverse, sometimes weakly curved around a distal centre. Attachment of the surface elements to the elongate receptacle is like that in the lysianassid calceolus, and the bulla is similarly well developed. 5. Eusirid (Fig. 4a-d) Eusirus antarcticus Thomson Eusirus microps Walker Eusirus perdentatus Chevreux Rhachotropis aculeatus (Lepechin) Rhachotropis helleri (Boeck) Rhachotropis macropus Sars Schraderia gracilis Pfeffer Amathillopsis australis Stebbing The special feature shared by the eusirid, gammarellid and oedicerotid types that immediately distinguishes them from other calceoli is the distinct separation of the proximal and distal elements and the remarkable cup-shaped configuration of the former. The proximal cup is robust, deeply concave, often set well apart from the distal element, and is attached to the receptacle only by a small basal connection. The following approximate measurements were obtained for the diameter of the proximal cup: Eusirus antarcticus 20-25 um, E. microps 23-25 //m, E. perdentatus 25-60 //m, Rhachotropis aculeatus 45-70 um, R. helleri 23-27 um, R. macropus 23-30 um, Schraderia gracilis 10 um, Amathillopsis australis 25-40 um. The distal element is elongated and carries a series of discrete crescentic plates, ranging from as few as 4 in Rhachotropis species and Amathillopsis australis, to 15 in Eusirus antarcticus, 25 in Eusirus microps, and more than 100 in Eusirus perdentatus. The multiplate distal element of the Eusirus species gives rise to an extremely elongate calceolus. The bulla at the base of the receptacle is pronounced in all eusirid calceoli. Of all species studied the greatest development of the 'parabolic' proximal dish belongs to Amathillopsis australis. The largest calceolus was that sported by Eusirus perdentatus. The 'pore' in the apex of the distal element reported by Dahl (1975) for Rhachotropis macropus is not a true feature, but is an artefact produced by the rolling-up of the distal plate, probably the result of prolonged exposure to the electron beam or an excessive current. We have included Schraderia gracilis in this group since it has a calceolus with an essentially eusird-type design, although the structure of the surface elements is unusual. The proximal cup in particular is enormously enlarged and saucer-shaped extending well outside the supporting receptacle, and unlike those in other eusirids appears flexible with a frayed edge to its outer margin (The somewhat collapsed state of the proximal element may be an artefact of the s.e.m. preparation). 6. Gammarellid (Fig. 3e) Gammarellus angulosus (Rathke) Gammarellus homari (Fabricius) Chosroes incisus Stebbing The calceoli of these three species differ from the eusirid-type in the presence of a second cup-shaped element between the basic proximal and distal elements. Apart from this additional cup, the resemblance to the calceolus of Rhachotropis is quite strong. The proximal cup has a diameter of only about 8 um in Chosroes incisus and 7 um in Gammarellus angulosus, the intermediate cup measuring about 4-5 um and 3'OjUm respectively. Gammarellus and Chosroes are further united by the particular arrangement of the calceoli on the articles of the antennae. In both genera, the calceoli are situated in rows that extend all around the distal margins of the articles, unlike all other species examined in which calceoli are restricted to just one surface of the antenna. With the exception of Urothoe, the gammarellid calceoli were the smallest calceoli examined during this study. Fig. 4 a, Eusirus antarcticus: b, Eusirus perdentatus, mid part of distal element: c, Amathillopsis australis: d, Amathillopsis australis, proximal element: e, Phoxocephalus regium: f, Urothoe elegans. Bar scales a-e = 10 wm, f= 2 urn. CALCEOLUS OF GAMMARIDEAN AMPHIPODS 1 1 3 7. Oedicerotid (Fig. le) Oediceroides calmani Walker Oediceroides lahillei Chevreux Oediceropsis brevicornis Liljeborg A discrete proximal cup embraced by a broad lamellar receptacle, a small suboval distal element, and a waisted receptacle characterise the oedicerotid calceolus. The distal element, which has the distal half marked with distinct transverse ridges, is attached to the spatulate extension of the receptacle at the point of the slight surface depression. The bulla at the base of the receptacle is well developed. The proximal cup has a diameter of about 25-30 /zm in Oediceroides lahillei. 8. Phoxocephalid (Fig. 4e, f) Metaphoxusfultoni (Scott) Metaphoxus pectinatus Walker Paraphoxus rostratus (Dana) Phoxocephalus regium Barnard Urothoe elegans Bate The phoxocephalid and crangonyctid calceoli differ in a number of ways from those already described although either could be derived from preceding types by a reduction in complexity. In the phoxocephalid, the slender stalk and bulbous receptacle are absent and the surface elements are supported on a simple paddle-shaped lobe. No differentiated proximal element is apparent; instead the receptacle carries 3 to 6 oval, weakly concave plates which are probably homologous with the distal element of other calceoli. Pontharpinia rostrata has only 3 such plates of which the basal plate is much the largest and may represent the missing proximal element. There are 4 plates in Phoxocephalus regium, and 6 in Metaphoxus pectinatus, M.fultoni, and Urothoe elegans. 9. Crangonyctid (Fig. 3f) Crangonyx pseudogracilis Bousfield Synurella sp. The crangonyctid calceolus is a greatly extended version of the phoxocephalid design. Once again there is no discrete stalk or bulbous receptacle, but a paddle-shaped lobe supporting a series of narrow plates. The plates are crescent-shaped and separated one from another proximally, but become more closely packed and much narrower distally. There are about 20 plates in Synurella sp. and 35 in Crangonyx pseudogracilis. We interpret the series of plates as the equivalent of the distal element of other calceoli, although the generally simplistic design of both the crangonyctid and phoxocephalid-types could indicate separate evolutionary development or developments. Discussion (a) Calceoli function Although various suggestions have been made as to the function of calceoli the only direct experimental work of any importance is that of Dahl and colleagues in a series of controlled aquarium experiments devised to investigate the occurrence of pheromones in amphipods (Dahl et al, 1970; Dahl, 1970, 1975). Adult females of Gammarus duebeni Liljeborg were fed a radioactive diet of 3H labelled fish liver, and were introduced into an aquarium contain- ing unlabelled male amphipods. The two sexes were kept apart by a fine nylon-mesh partition. After 30 and 60 minutes the amphipods were isolated and specimens selected for scintillation counting and microscope autoradiography. The males had by then become radioactive, and the 3H label was localized on the second antenna, either within or very close to the calceoli. Dahl and colleagues concluded that a labelled pheromone produced by the female was dispersed in the aquarium water and taken up selectively either by the male 114 R. J. LINCOLN & D. E. HURLEY calceoli or by the tissue in the immediate vicinity of the calceoli. (The limited resolution of the microscope autoradiographic technique used did not permit precise localization of the uptake site). In Gammarus duebeni only males have calceoli. It is our belief that the labelled pheromone may have been taken up by accompanying setae seen in Dahl's figures alongside the calceoli, perhaps indicated by the presence of two uptake sites in the same transverse section (Dahl et al, 1970 Fig. 3 A). The solitary nature of the calceoli in duebeni (Fig. la, b) would seem to preclude the presence of two calceoli in a single transverse or obliquely transverse section. Alternatively, the calceoli may secondarily provide an avenue for pheromone uptake that is incidental to their main function. We believe that they are structurally too complex for chemoreception to be their primary role. We note that chemoreceptors in Crustacea are typically simple sac-like structures (e.g. aesthetascs), or hair-like, or funnel-canals or pores (Barber, 1961), pegs or pits. The model of a protein sieve envisaged as the basis of an aquatic chemoreceptor does not demand the range of architectural novelty characteristic of calceoli. Calceoli have a morphological complexity greater than any other aquatic receptor of equivalent size that we have encountered in the literature. The occurrence of calceoli on males and not on females is a common feature which suggests that calceoli have some involvement in the amphipod's reproductive behaviour. Calceoli are not always confined to one sex (Hurley, 1980) but when they are it is always the male that is calceoliferous. This is consistent with Dahl's pheromone theory, but since the vast majority of amphipod species are non-calceoliferous one would have to postulate that in these pheromone receptivity had been taken over by some other receptor, or that pheromones were not part of the behavioural strategy. We had hoped that a survey of the ecological and behavioural features of calceoliferous versus non-calceoliferous species would provide a clue to function but this was not the case. Calceoliferous species are found throughout almost the full spectrum of habitats characteristic of their family groups, and from what little is known about behaviour, calceoli can be present and absent in closely allied species apparently having similar habits and modes of life. Other arguments against chemosensitivity as a primary role — admittedly based on comparative external arrangement only — are the orientation and directionality of calceoli. Calceoli are always arranged in one or more well defined rows along the axis of the antenna, normally the underside of antenna 1 and the upper surface of antenna 2. This arrangement permits a forwardly-directed 'array' of calceoli in an animal with antenna 1 raised and antenna 2 in a lowered posture. In addition, they are clearly organised to point in the same direction relative to the antennal axis. In some species, for example Eusirus perdentatus and Amathillopsis australis, although there is only one calceolus per segment they are ranged in repetitive pairs or triplets each slightly offset from its neighbour. Directionality is a property of the calceolus itself and is most obvious in the 'parabolic' cup reminiscent of a radar reflector found in the most specialised forms. Searching for an explanation that is compatible with complexity, orientation and directionality we are drawn to one satisfying possibility — a sensitivity to water borne pressure waves whether produced by sound waves, animal vibrations or other disturbances in the water. One could envisage the advantages of disturbance sensors in identifying the presence or approach of other animals, whether of the same species or not, in identifying movement- disturbances or behavioural characteristics of prey, or of water disturbances in streams around stones and ripples which would enable them to seek or avoid particular ecological situations. This sonar or phono-receptor theory is not supported by experimental evidence but we are hopeful that the photographs and discussion in this paper will attract the attention of biologists and in particular electrophysiologists with the experimental facilities to probe this possibility. (b) Phylogenetic considerations Despite the embryonic state of knowledge about phylogenetic relationships of higher gammaridean taxa it is noteworthy that the calceoliferous families have generally been 115 recognised as having some evolutionary affinity (Barnard, 1969; Bousfield, 1978). Notwith- standing the passing doubts occasioned by the structure of calceoli in Phoxocephaloidea and Crangonyctoidea it seems probable that calceoli have arisen only once in the Gammaridea and have undergone limited structural radiation during the evolution of the group. In the absence of evidence pointing to convergence, similarity of calceolus design may be taken as an indicator of geneological affinity. The discovery of close structural similarities between the calceoli of many species traditionally placed in the same genus or family and discontinuities between species from different groups, has given us confidence that calceolus architecture has phylogenetic significance. Most of the species examined and allocated to the 9 calceolus-types are in good agreement with established family groupings but there are anomalies that suggest incorrect classification. Some of the species or genera which we considered wrongly designated during the early part of our study have since been relocated in a manner consistent with the calceoli evidence (Bousfield, 1978). The important anomalies are discussed below. Amongst the first amphipods studied were species of Bathyporeia and Urothoe, two genera that for a long time have been placed together in the Haustoriidae. The calceoli are quite different, however, pointing to separate relationships, and it was further discovered that Urothoe shares the calceolus type of Phoxocephalus. This supports fully the recent revision by Bousfield (1978) in which Urothoe is moved from the haustoriids to a new family along- side Phoxocephalus in the superfamily Phoxocephaloidea. The bathyporeid-type calceolus is shared by Zaramilla, a genus placed in the Eusiridae by Barnard (1969), although special reference was made to its apparent 'haustoriid' affinities. There can be little doubt that Zaramilla belongs close to Bathyporeia, and the marked similarity of their calceoli to the gammarid-type (especially Anisogammarus confervicolous) must be further evidence for the proximity of the Pontoporeiidae to the gammaroid families. Calceoli may prove particularly useful in re-assessing the Eusiridae, a family recently made very large and unwieldly by the inclusion of the families Pontogeneiidae and Calliopiidae (Barnard, 1969, 1972). Eusirid amphipods are frequently calceoliferous and have some of the largest and structurally most complex calceoli known. We have recognised 3 types within the eusirid complex — pontogeneiid, eusirid and gammarellid — and our allocation of species to each of these tends to cut across previously accepted family boundaries. Thus, the 'calliopiids5 Calliopius, Apherusa, Halirages and Paracalliope share the same type of calceolus as the 'pontogeneiid' Pontogeneia, and the 'eusirids' Eusiroides, Eusiropsis, Bovallia and Paramoera. Eusirus and Rhachotropis, traditionally confamilial, must be joined by Amathillopsis, a genus having a chequered history being variously allocated to the Gammaridae, Amathillopsidae and the Paramphithoidae. We have placed Schraderia with our eusirids since it shares the same basic calceolus design, although it differs somewhat in detail and relative proportions. The third group mentioned in the eusirid context, the gammarellid-type, brings together Gammarellus and Chosroes, linked by the common possession of an intermediate cup- shaped surface element. Gammarellus was, until its transfer to a new family (Bousfield, 1977) assigned to the Gammaridae, and Chosroes was with the calliopiids. If Bousfield's new family Gammarellidae receives general acceptance by amphipodologists, then Chosroes must be considered for inclusion also. It is particularly satisfying to note that, as well as having similar calceoli, Gammarellus and Chosroes (figured Sars, 1894; Stebbing, 1888) show a surprising similarity in many characters. New perspectives produced by this SEM study should encourage other taxonomists to pay greater attention to this microscopic antennal receptor so often ignored in systematic descriptions, and, we hope, encourage some physiological work on their structure and function. References Barber, S. B. 1961. Chemoreception and thermoreception. In Waterman, T. H. (Ed.) The physiology of Crustacea Vol. 2 : 109-128. Academic Press, New York and London. 116 R.J. LINCOLN &D. E. HURLEY Barnard, J. L. 1969. The families and genera of marine gammaridean Amphipoda. Bull. U.S. natn. Mus.211 : 1-535. 1972. Gammaridean Amphipoda of Australia. Part 1 . Smithson. Contr. Zool. 103 : 1-333. Blanc, H. 1883. Structure des cupules membraneux ou 'calceoli' chez quelques Amphipodes. Zool Anz. 6 (142): 370-372. 1884. Die Amphipoden des Kieler Bucht nebst einiger histologischen Darstellung der 'Calceoli'. Nova Acta Acad. Caesar. Leop. Carol. 47 (2) : 39-96. Bousfield, E. L. 1977. A new look at the systematics of gammaroidean amphipods of the world. Crmtaceana Suppl 4 : 282-3 16. 1978. A revised classification and phylogeny ofamphipod crustaceans. Trans. R. Soc. Can. Ser 4, 16 : 343-390. Cole, G. A. 1970. Gammarus minus: geographic variation and description of new subspecies G. m. pinicollis (Crustacea, Amphipoda). Trans. Am. microsc. Soc. 89 (4) : 5 14-523. Croker, R. A. & Gable, M. F. 1977. Geographic variation in western Atlantic populations of Gammarus oceanicus Segerstrale (Amphipoda). Crustaceana32 : 55-76. Dahl, E. 1970. Pheromone transport and reception in an amphipod. Science 170 : 739-740. 1975. Pheromones in aquatic invertebrates. In Regnell, G. (Ed.) Biological signals : 95-1 14. Kungl. Fysiografiska Sallskapet: Lund, (Lund). Dahl, E., Emanuelsson, H. & Mecklenburg, C. von. 1970. Pheromone reception in the males of the amphipod Gammarus duebeni Lilljeborg. Oikos 21 : 42^7. Goedmakers, A. 1972. Gammarus fossarum Koch, 1835: Redescription based on neotype material and notes on its local variation (Crustacea, Amphipoda). Bijdr. Dierk. 42 (2) : 124-137. Hurley, D. E. 1980. A provisional checklist of Crustacea Amphipoda known to have calceoli. NZOI /tec. 4 (8): 7 1-1 20. Jazdzewski, K. 1977. Remarks on the morphology of Gammarus fossarum Koch, 1835, and Gammarus kischineffensis Schellenberg, 1937. Crustaceana Suppl 4 : 201-21 1. Milne-Edwards, H. 1830. Extrait des recherches pour servir a 1'histoire naturelle des Crustaces Amphipodes. Annls Sci. nat. 20 : 353-399. Minkley, W. L. & Cole, G. A. 1963. Ecological and morphological studies on gammarid amphipods (Gammarus spp.) in spring-fed streams of northern Kentucky. Occ. Pap. C. C. Adams Cent. ecol. Stud.W: 1-35. Sars, G. O. 1 894. An Account of the Crustacea of Norway, with short descriptions and figures of all the species : 473-67 1 . Christiania and Copenhagen (Cammermeyers). Stebbing, T. R. R. 1888. Report on the Amphipoda collected by H.M.S. Challenger during the years 1873-1876. Rep. Sclent. Results Voy. Challenger (Zoology) 29 : 1-1737. Manuscript accepted for publication 12 September 1980 G. A. Boxshall Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction In his recent synopsis of the genus Lernaea Linn., 1758 Kabata (1979) recognized 37 species. All of these parasitise freshwater fishes although some are also known to occur on amphibian tadpoles. Eleven species have been reported from India, S.E. Asia and the Far East (Kabata, 1979) but no records of Lernaea from the Australasian zoogeographic region have as yet been published. In October 1979 some Lernaea were collected by Dr I. L. Owen of the Central Veterinary Laboratory, Port Moresby, Papua-New Guinea. The specimens were found to represent a new species which is described in detail below. Specimens were examined and dissected in lactophenol. Drawings were made using a camera lucida and terminology is adapted from Kabata (1979). Description of new species Lernaea papuensis n. sp. Postmetamorphosis adult female (Fig. 1A); cephalothorax small, hemispherical, bearing antennae and mouthparts ventrally and with nauplius eye visible through integument dorsally (Fig. 1 B). Holdfast apparently comprising 6 subequal arms, probably representing an unbranched dorsal pair and a ventral pair, with each member being divided at its origin into 2 equal branches. Holdfast usually arranged in anteroposterior plane, sometimes dorsal pair pass perpendicularly into dorsoventral plane (Fig. 2D). Holdfast arms of largest paratype (Fig. 2E) distorted and overlapping due to site on host. Neck, comprising second to fourth leg-bearing somites, passing imperceptibly into genital somite. Neck expanding in girth posteriorly but marked with conspicuous swellings, each delimited by constrictions, anterior to legs 2 and 3. Genital somite with simple hemispherical genital prominence posteriorly. Abdomen conical, unsegmented and bearing uropods distally. Total body length from anteriormost tip of holdfast to posterior tip of uropod ranging from 5'4 to 10 mm. First antenna (Fig. 1C) indistinctly 4-segmented; segments 1 to 4 bearing 10, 3, 4, and 10 armature elements respectively. Second antenna (Fig. ID) indistinctly 3-segmented, segments 1 and 2 unarmed and comprising half total length of appendage, segment 3 with 3 setae on posterior margin, and a claw-like spine, 5 slender setae and 1 setule distally. Labrum (Fig. IB) a flattened triangular plate, overlying mandibles and first maxillae. Second maxilla (Fig. IE) 2-segmented with 2 curved claws apically. Maxilliped (Fig. IF) basal segment with small papilla, armed with an apical setule, on distal part of medial margin; terminal segment with 5 strong curved spines apically. Thoracic legs 1 to 4 more or less regularly spaced along body; leg 1 situated on ventral surface of cephalothorax at posterior border of cephalothorax (Fig. 1 A 1 ), legs 2 to 4 (Fig. 1 A 2-4) on ventral surface of neck. Legs biramous with 3-segmented rami (Figs 1G, 2A-C), armature formula as follows: Bull. Br. Mus. nal. Hist. (Zool.) 40 (4) : 1 1 7-1 20 Issued 30 July 1 98 1 Fig. 1 Lernaea papuensis n. sp. female. A, Holotype, dorsal; B, paratype head and cephalic appendages, ventral; C, first antenna, ventral; D, second antenna, ventral; E, second maxilla, posteroventral; F, maxilliped, posteroventral; G, first leg, anterior; H, uropods, dorsal. Scales 50 //m unless otherwise stated. NEW SPECIES OF LERNAEA 119 Legl Leg 2 Leg 3 Leg 4 coxa 0-1 0-1 0-1 0-1 basis 1-1 1-0 1-0 1-0 endopod exopod Several long pinnules present near mediodistal angle of basis of legs 2 to 4. Lateral margins of all endopod segments armed with pinnules. Leg 5 not observed. Uropods (Fig. 1H) subcylindrical, about 1'5 times longer than wide and bearing a very long plumose seta on distal margin, a short lateral seta, dorsal seta and posterolateral seta. Fig. 2 Lernaea papuensis n. sp. paratype female. A, second leg, posterior; B, third leg, posterior; C, fourth leg, posterior; D, smallest ovigerous paratype, dorsal; E, largest paratype, dorsal. Scales 50 //m unless otherwise stated. 120 G. A. BOXSHALL MATERIAL EXAMINED. Holotype 9, 9 paratype 99 from gill arches, pectoral fins and body surface of Jardine's Barramundi, Scleropages jardini Kent, caught in the Bensbach river, Western Province, Papua-New Guinea. BM(NH) Reg. Nos holotype 1980. 122, paratypes 1980.123-131. REMARKS. Although the size and shape of the holdfast ofLernaea species is variable with age and especially with position on the host (Fryer, 1961) it is the gross morphology of the holdfast and trunk which provides virtually all the characters used to distinguish between the species (Harding, 1950). This situation arises because of the high degree of uniformity of appendage structure and armature throughout the genus (Harding, 1950; Kabata, 1979). The new species differs from all others in the possession of a holdfast comprising 6 slender, elongate arms which are more or less equal in length. The holdfast of L. papuensis could be derived from the condition exhibited by L. senegali Zimmermann, 1923 which possesses a pair of simple dorsal arms and a pair of ventral arms which are branched near their tips. The 4 ventral arms of the holdfast of L. papuensis are probably homologous with the branched ventral pair of L. senegali but they are branched at their bases. The genital prominence is hemispherical in both the new species and L. senegali. Another important taxonomic character of L. papuensis is the shape of the neck. The conspicuous expansions of the neck, delimited by constrictions, anterior to leg 2 and particularly to leg 3, are more marked in this species than in the majority ofLernaea species. Acknowledgements I would like to thank Dr I. L. Owen for donating the material to the collections of the BM(NH) and for providing relevant data. References Fryer, G. 1961. Variation and systematic problems in a group of lernaeid copepods. Crustaceana 2 : 275-285. Harding, J. P. 1950. On some species ofLernaea (Crustacea, Copepoda: Parasites of fresh-water fish). BullBr. Mus. nat. Hist. (Zool.) 1 : 1-27. Kabata, Z. 1979. Parasitic Copepoda of British Fishes. Ray Society, London. 468 pp, 199 pi. Manuscript accepted for publication 10 September 1980 Some type specimens of Isopoda (Flabelliferd) in the British Museum (Natural History), and the isopods in the Linnaean Collection Joan Ellis Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction British Museum (Natural History) collection The first catalogue of the crustacean collection was compiled by Adam White and published in 1847. The foundation of that collection was W. E. Leach's material, acquired by the Museum in 1826. This contained the type material of approximately 140 crustacean species described by Leach, 37 of which belong to the suborder Flabellifera. Between 1826 and the publication of White's catalogue the collection was augmented from various sources. The most notable additions to the marine Isopoda were from the collections of George Montagu, Thomas Say, and those made during the voyages of H. M. Ships Erebus and Terror to the Antarctic. In 1863 the Museum acquired from the Linnean Society Sir Joseph Banks' collection. Unlike the insects, the crustaceans from Banks' collection were not registered, and no record exists of the species it contained. Five species of 'Oniscus' described by J. C. Fabricius (1775) were from Banks' collection, and although the types of all five were considered by Zimsen (1964) to be lost, recent investigations have established that two of these, Ceratothoa imbricata and Serolis paradoxa are still in the BM(NH) collections*. The other three species are amphipods (see Stebbing, 1 888). In the latter half of the nineteenth century the dry collection of isopods was augmented from various sources, especially from the voyages of H. M. Ships Herald and Rattlesnake (see Miers, 1884 p. 179). It also included the type material of several species described by E. J. Miers, the Assistant in charge of Crustacea from 1872 to 1885 (see Gordon, 1971). Although parts of the Crustacea collection have been catalogued since 1847 (see Bell, 1855; Bate, 1862; Thurston & Allen, 1969; Lincoln & Ellis, 1974; Lincoln & Hurley, 1974; Ellis & Lincoln, 1975), this is the first attempt since White's to catalogue the marine isopods, and this paper is intended as a precursor to a catalogue of the entire flabelliferan type collection which contains an estimated 200 species from some 50 genera. The Linnaean collection Linnaeus' collection was bought from his widow by J. E. Smith in 1774. It remained in Smith's possession until his death in 1828, and a year later it was purchased by the Linnean Society of London. The isopod collection, preserved dry, now consists of 13 specimens (see p. 127) labelled in Linnaeus' handwriting and which must be considered holotypes. (A specimen labelled 'cet? included among these is the parasitic amphipod Cyamus ceti (L., 1758).) There are also 28 The holotypes of Serolis paradoxa and Ceratothoa imbricata were collected during Capt. Cook's first voyage (1 768-1 77 1 ). Much of the zoological material collected during Cook's voyages has become widely dispersed and/or lost, and these specimens were overlooked when Whitehead (19b9) compiled his account of the traceable specimens. Bull.Br.Mus.nat.Hist.(Zool.)40(4): 121-128 Issued 30 July 1981 122 J.ELLIS unlabelled specimens that have been provisionally determined. These comprise: 4 Exosphaeroma gigas (Leach, 1818); 2 Exosphaeroma sp.; 4 Parisocladus perforates (Edwards, 1840); 1 Idotea carinata, Miers 1881 (=Synisoma sp.); 1 Paridotea ungulata (Pallas, 1772); 1 Paridotea rubra Barnard, 1914; 1 Aega tridens Leach, 1815; 2 Ceratothoa imbricata (Fabricius, 1775); 1 amphipod (family Stegocephalidae); 1 Ligidium sp.\ 1 Asellus aquaticus (L., 1758) (anterior half only); 1 Nerocila serra ScHiodte & Meinert, 1881; 1 Idotea granulosa Rathke, 1843; 1 Idotea chelipes (Pallas, 1766); 1 Ligia dilatata Brandt, 1833; 1 Ligia oceanica (L., 1 767); 4 Oniscoidea. Smith was known to have augmented Linnaeus' zoological collection with material of his own (Jackson, 1 890) and these unlabelled specimens may have been added by him. Methods Until the latter half of the nineteenth century, small crustaceans were curated at the BM(NH) in a similar manner to entomological specimens i.e. impaled with entomological pins and stored dry in insect cabinets. Throughout the last 1 50 years the condition of many of these crustaceans has deteriorated, chiefly resulting from repeated movement of the dry collection. Appendages have become detached, specimens broken in half, and a few have disintegrated completely. To prevent further deterioration of valuable type material and to make it more easily accessible for study, a decision was taken to transfer specimens to alcohol for incorporation in the Museum's spirit collection. The techniques used for rendering the dried specimens suitable for alcohol preservation are given below. In order to allow restorative chemicals to penetrate the body tissues air must be expelled from the specimen. Sometimes this can be achieved following direct immersion into 80% IMS (industrial methylated spirit). If the specimen fails to sink gentle heat should be applied until the solution reaches boiling point, immediately after which the container should be allowed to cool. Following either procedure the air-free specimen is placed into distilled water for one hour and then relaxed using one of the chemical methods described below, each of which has particular advantages or disadvantages as noted. After this chemical treatment the specimen is again immersed into distilled water for one hour and then transferred to 80% IMS for storage. Chemical treatments All solutions were made with distilled water. (a) 1-2% sodium chloride. Length of time: dependent on size of specimen, 4-5 hours for small specimens; up to 24 hours for large specimens. This method takes longer than others, but this can be an advantage with small or fragile specimens. (b) 2% tri-sodium orthophosphate. Length of time: 4-6 hours. A close watch must be kept on the specimens from 4 hours onwards, as they may become too flaccid. A further disadvantage of this technique is that a cloying, flocculent precipitate may develop, mostly at the bases of the appendages. As this precipitate has to be removed by mechanical means, damage to the specimen can occur unless extreme care is taken. (c) 0-5% formaldehyde OR 5% sodium sulphate. Length of time: 3£-6 hours, but may safely be left overnight as prolonged immersion does not seem to be deleterious. (d) 2% citric acid and 20% sodium citrate in equal parts. Length of time: up to 4 hours. In the limited trials of this technique, the specimens have a tendency to become too flaccid. (e) Sandison's technique: 90% ethyl alcohol— 30 volumes; 0*5% formaldehyde— 50 volumes; 5% sodium citrate — 20 volumes. Length of time: 1-24 hours. The results produced by this technique were very variable. Reasonable results were obtained when fragile specimens were immersed for 1-2 hours, but larger, more robust specimens sometimes remained largely unaffected. TYPE SPECI MENS OF ISOPODA 1 23 British Museum (Natural History) Collection AEGIDAE Aega bicarinata Leach (1818 : 349) HOLOTYPE: 1979 : 306 : 1 . Mediterranean? Presented by W. E. Leach. Aega emarginata Leach ( 1 8 1 5 : 370) [transferred to Aega psora (L., 1 758)] HOLOTYPE: 1979 : 309 : 1 . Locality unknown. Presented by W. E. Leach. Aega meinerti Miers (1 884 : 305) HOLOTYPE: 1858: 172. King George Sound, Western Australia. Presented by F. M. Rayner, H.M.S. Herald. Aega monophthalma Johnston (1834 : 233) HOLOTYPE: 1979 : 299 : 1. Berwick Bay, Northumberland. On large codfish. Presented by G. Johnston. Rocinela danmoniensis Leach (1818 : 349) HOLOTYPE: 1979 : 328 : 1. Plymouth Sound. Presented by W. E. Leach. CIROLANIDAE Cirolana cranchi Leach (1818: 347) HOLOTYPE: 1979 : 279 : 1 . Falmouth, Cornwall. Presented by Mr Cranch. Cirolana harfordi (Lockington) (1 877 : 46, as Aega harfordi) PARATYPES: 1878:9 (4 specimens). Santa Rosa Island. Under stones, MT, in muddy places. Collected by W. G. W. Harford. Presented by W. N. Lockington & W. G. W. Harford. Cirolana rossi Miers (1 876a : 228) SYNTYPES. 1856:33 (2 specimens). New Zealand. H.M.S. Herald. Purchased of Mr Erebus — The Antarctic Voyage'. Presented by Lieut. A. Smith. SYNTYPES: 1844 : 3 (7 specimens). Auckland Is. Antarctic Expedition. Presented by Capt. James Clarke Ross. SYNTYPE: 1848 : 80. New Zealand. Collected by Rev. W. Colenso. Presented by Dr J. Hooker. SYNTYPE: 1856:33 (2 specimens). New Zealand. H.M.S. Herald. Purchased of Mr Cuming. Cirolana tenuistylis Miers (1884 : 303) SYNTYPES: 1862 : 53 (2 specimens). Locality unknown. Collected by F. M. Rayner, H.M.S. Herald. Presented by Mr Warwick. Conilera montagui Leach (1818 : 348) [transferred to Conilera cylindracea (Montagu, 1804)] HOLOTYPE: 1979 : 289 : 1. Salcombe, Devonshire. Presented by G. Montagu. Eurydice pulchra Leach (1818 : 370) SYNTYPES: 1979 : 290 (2 specimens). Bantham, Devonshire. Presented by W. E. Leach. Nelocira swainsoni Leach (1818 : 347) [transferred to Cirolana cranchi Leach, 1818] SYNTYPES: 1979 : 288 (3 specimens). Sicily. Collected and presented by W. Swainson. CYMOTHOIDAE Anilocra capensis Leach (1818 : 350) SYNTYPES: 1979 : 329 (3 specimens). Cape of Good Hope. Presented by W. E. Leach. Anilocra cuvieri Leach (1818 : 350) [transferred to Anilocra physodes (L., 1 758)] SYNTYPES: 1979 : 332 (3 specimens). Island of Ivica (Ibiza?), Mediterranean. Presented by G. Cuvier. Anilocra mediterranea Leach (1818 : 350) [transferred to Anilocra physodes (L., 1 758)] HOLOTYPE: 1979 : 330 : 1. Sicily. Presented by W. E. Leach. Canolira rissoniana Leach (1818: 350) [= Anilocra sp.] HOLOTYPE: 1979 : 33 1 : 1 . Locality unknown. Presented by W. E. Leach. Ceratothoa imbricata (Fabricius) (1 775 : 296, as Oniscus) 124 J.ELLIS HOLOTYPE: 1979 : 403 : 1. New Zealand, Capt. Cook's first voyage. Sir Joseph Banks' collection. Presented by the Linnean Society of London. Cymothoa banksii Leach (1818:353) [transferred to Ceratothoa imbricata (Fabricius, 1775)] HOLOTYPE: 1979 : 402 : 1. New Zealand. Presented by W. E. Leach. Cymothoa dufresni Leach (1818 : 352) [transferred to Cymothoa oestrum (L., 1758)] SYNTYPES: 1979 : 405 (2 specimens). Locality unknown. Presented by W. E. Leach. Cymothoa leschenaultii Leach (1818:352) [transferred to Cymothoa eremita (Briinnich, 1783)] HOLOTYPE: 1979 : 406 : 1. Pondicherry, India. Presented by Leschenault. Cymothoa mathieui Leach (1818: 353) [transferred to Cymothoa eremita (Briinnich, 1 783)] SYNTYPES: 1979 : 407 (2 specimens). 'Ile-de-France'. Presented by M. Mathieu. SYNTYPE?: 1979:408. 'Ile-de-France'. Sir Joseph Banks' collection. Presented by the Linnean Society of London. Cymothoa trigonocephala Leach (1818:353) [transferred to Ceratothoa imbricata (Fabricius, 1775)] SYNTYPES: 1979 : 404 (2 specimens). Locality unknown. Presented by W. E. Leach. Lironeca contracta Miers (1 880 : 466, footnote) SYNTYPE: 1844 : 105. Australia. Presented by the Earl of Derby. SYNTYPE: 1846 : 89. Australia. Presented by J. B. Jukes. Lironeca desmarestii Leach (1818: 352) [transferred to Lironeca redmani Leach, 1818] SYNTYPES: 1979 : 336 (3 specimens). Locality unknown. Presented by W. E. Leach. Lironeca laticauda Miers (1877 : 677) SYNTYPES: 1862 : 96 (2 specimens). Manchuria. Purchased of Stevens. Lironeca micronyx Miers (1880 : 466, footnote) SYNTYPES: 1846 : 104 (3 specimens). Mauritius. Presented by R. Templeton. Lironeca novaezealandiae Miers (1876a:228) [transferred to Lironeca raynaudi Milne Edwards, 1840] SYNTYPES: 1845 : 30 (2 specimens). New Zealand. Presented by Mr Earl. Lironeca ovalis (Say) ( 1 8 1 8a : 394, as Cymothoa) SYNTYPE: 1979 : 413. N. America. Presented by T. Say. Lironeca rafmeskii Leach (1818 : 352) SYNTYPES: 1979 : 338 (2 specimens). Cape of Good Hope? (but see original description — 'localite inconnu'). Presented by W. E. Leach. Lironeca redmani Leach (1818: 352) HOLOTYPE: 1979 : 401 : 1. Jamaica. Presented by Lieut. Redman. Lironeca vulgaris Stimpson (1857 : 508). See also Stimpson 1859 : 88 SYNTYPES: 1878 : 9 (3 specimens). San Francisco Market. Presented by W. N. Lockington & W. G. W. Harford. Nerocila blainvillei Leach (1818:351) SYNTYPES: 1979:400 (2 specimens). Sicily? (but see original description — 'localite inconnu'). Presented by W. E. Leach. Nerocila congener Miers (1 880 : 468, footnote) HOLOTYPE: 1979 : 337. Philippine Islands. Presented by Mr Cuming. Nerocila longispina Miers (1880 : 468) HOLOTYPE: 1849 : 86. Malabar. Presented by I. Ward. Nerocila macleayi Miers (1884 : 301) [nom. nov. for Nerocila imbricata (Fabricius): Miers \816avide Miers 1884] SYNTYPE: 1845:30 (as Nerocila imbricata (Fabricius): White 1847) New Zealand. Presented by Mr Earl. SYNTYPE: 1979:414: 1 (as Nerocila imbricata (Fabricius): White 1847). Locality and donor unknown. SYNTYPE: 1979:417: 1 (as Cilonera macleayii White & Doubleday 1843 nom. nud.}. 'Africa'. Presented by W. E. Leach. TYPE SPECIMENS OF ISOPODA 125 SYNTYPE: 1842 : 44 (as Cilonera macleayii White & Doubleday 1843 nom. nud.). New Zealand. Presented by Andrew Sinclair. SYNTYPE: 1847 : 104. New Zealand. Presented by Andrew Sinclair. SYNTYPE: 1979 : 415 : 1. New Zealand 'taken out of the mouth of a fish'. Presented by Andrew Sinclair. SYNTYPE: 1850 : 12. West coast, America. Presented by Capt. Kellett & Lieut. Wood. SYNTYPE: 1 979 : 4 1 6 : 1 . Locality and donor unknown. Nerocila trichiura (Miers) (1877: 677, as Anilocra) HOLOTYPE: 1846 : 104. Mauritius. Presented by Robert Templeton. Olencira lamarckii Leach (1818:351) [transferred to Olencira praegustator (Latrobe, 1 802)] HOLOTYPE: 1979 : 333 : 1. Locality unknown. Presented by W. E. Leach. LIMNORIIDAE Limnoria terebrans Leach (1814: 433) [transferred to Limnoria lignorum (Rathke, 1 799)] SYNTYPES: 1979:445 (3 specimens). Bell Rock, Scotland. Collected by R. Stevenson. Presented by W. E. Leach. SEROLIDAE Scrolls fabricii Leach (1818: 340) [transferred to Serolis paradoxa (Fabricius, 1 775)] HOLOTYPE: 1979 : 452 : 1 . Senegal. Presented by W. E. Leach. Serolis latifrons Miers (187 5a : 74) SYNTYPE: 1843:70. Rendezvous Cove, Auckland Is. 'Collected in the Erebus — The Antarctic Voyage'. Presented by Lieut A. Smith. Serolis paradoxa (Fabricius) (1 775 : 296, as Oniscus) HOLOTYPE: 1979 : 451 : 1. Tierra del Fuego, Capt. Cook's first voyage. Sir Joseph Banks' collection. Presented by the Linnean Society of London. Serolis septemcarinata Miers (1 8756 : 1 16) SYNTYPES: 1843 : 70 (2 specimens). Crozet Islands. 'Collected in the Erebus during the Antarctic voyage'. Presented by Lieut. A. Smith. SPHAEROMATIDAE Ancinus depressus (Say) (1 8 1 86 : 483, as Naesd) SYNTYPE: 1979 : 444 : 1 . North America. Presented by Thomas Say. Campecopea cranchi Leach (1818:341) [transferred to Campecopea hirsuta (Montagu, 1804)] SYNTYPES: 1979:441 (2 specimens). Falmouth, Cornwall. Collected by Mr Cranch. Presented by W. E. Leach. Campecopea hirsuta (Montagu) (1 804 : 7 1 , as Oniscus) SYNTYPES: 1979 : 440 (3 specimens). Devon. Presented by G. Montagu. Cassidinidea ovalis (Say) (1 8 1 86 : 484, as Naesa) SYNTYPE: 1979 : 422 : 1 . North America. Presented by Thomas Say. Cilicaea antennalis Miers (1884:310) HOLOTYPE: 1979 : 442 : 1 . Swan River, W. Australia. Presented by Sir J. Richardson. Cilicaea latreillei Leach (1818: 342) SYNTYPES: 1979 : 443 (2 specimens). Locality unknown. Presented by W. E. Leach. Cilicaea latreillei var. longispina Miers (1884 : 3 10) HOLOTYPE: 1851 : 33. Bass Strait. HMS Rattlesnake. Presented by J. Macgillivray. Cymodoce bifida Leach (1818 : 343) HOLOTYPE: 1979 : 43 1 . Locality unknown. Presented by W. E. Leach. Cymodoce con vexa Miers (1876a : 229) SYNTYPES: 1852 : 43 (3 specimens). New Zealand. Purchased of Mr Cuming. 126 J.ELLIS Cymodoce emarginata Leach (1818 : 342) SYNTYPE: 1979 : 434 : 1. Mount Edgecombe, near Plymouth, Devon. Variety 'a'. Collected and presented by W. E. Leach. SYNTYPE: 1979:435: 1. Falmouth, Cornwall. Variety 'b\ Collected by Mr Cranch. Presented by W. E. Leach. Cymodoce lamarcki Leach (1818: 343) [transferred to Cymodoce truncata Leach, 1814] SYNTYPES: 1979 : 436 (2 specimens). Sicily. Collected by W. Swainson. Presented by W. E. Leach. Cymodoce truncata Leach ( 1 8 1 4 : 433) HOLOTYPE: 1979 : 439 : 1 . Devon. Presented by W. E. Leach. Dynamene montagui Leach (1818:344) [transferred to Dynamene bidentata (Adams, 1800)] SYNTYPES: 1979 : 430 (3 specimens). Devon. Presented by G. Montagu. Dynamene rubra Leach (1818 : 344) [transferred to Dynamene bidentata (Adams, 1 800)] SYNTYPES: 1979 : 428 (4 specimens). Devon. Presented by W. E. Leach. Dynamene viridis Leach (1818 : 344) [transferred to Dynamene bidentata (Adams, 1 800)] SYNTYPES: 1979 : 429 (3 specimens). Devon. Presented by W. E. Leach. Exosphaeroma coats i Tattersall (1913:885) SYNTYPE: 1979:425 (fragment). St. Paul Island, Indian Ocean. Presented by J. Macgillivray. Exosphaeroma gigas (Leach) (1818 : 346, as Sphaeromd) SYNTYPE: 1941 : 6 : 27 : 5. Locality unknown. Presented by W. E. Leach. SYNTYPE: 1979 : 420 : 1. Locality unknown. Sir Joseph Banks' collection. Presented by the Linnean Society of London. Exosphaeroma kraussi Tattersall (1913 : 884) SYNTYPES: 1881 : 19(14 specimens). Sea Point, Cape Town, South Africa. Presented by A. E. Craven. Exosphaeroma lanceolatum (White) (1843 : 345, as Sphaeroma gigas var. lanceolatum) SYNTYPES: 1842 : 4 (White (1847 : 102) refers to 4 specimens, but there are 7 with this registration number). Falkland Islands. Presented by W. E. Wright. Isodadus spiniger var. recurvatus Miers (1 876/7 : 113) [transferred to Isocladus armatus (Milne Edwards, 1840)] HOLOTYPE: 1856 : 164. New Zealand. Presented by Dr Sinclair. Isocladus tristensis (Leach) (1818 : 345, as Sphaeromd) SYNTYPES: 1979 : 453 (4 specimens). Tristan d'Acunha. Collected by Capt. Carmichael. Presented by W. E. Leach. Sphaeroma curtum Leach (1818 : 345) [transferred to Cymodoce truncata Leach, 1814] HOLOTYPE: 1 979 : 4 1 8 : 1 . Britain. Presented by W. E. Leach. Sphaeroma dumerilli Leach (1818 : 345) [transferred to Cymodoce truncata Leach, 1814] SYNTYPES: 1979 : 432 (2 specimens). Genoa. Presented by W. E. Leach. SYNTYPE: 1979 : 433. Sicily. Presented by W. E. Leach. Sphaeroma hookeri Leach (1 8 14 : 433) SYNTYPES: 1979 : 421 (4 specimens). Suffolk [Leach (1814) gives locality as Norfolk, but see Leach (1815: 369; 1818: 345)]. Collected by W. J. Hooker. Presented by W. E. Leach. Sphaeroma olivacea Lockington (1877 : 45) [transferred to Gnorimosphaeroma oregonensis (Dana, 1853)] SYNTYPES: 1878 : 9 (3 specimens). San Francisco Bay. Presented by W. N. Lockington & W. G. W. Harford. Sphaeroma prideauxianum Leach (1818 : 345) [transferred to Cymodoce truncata Leach, 1814] HOLOTYPE: 1979 : 423 : 1 . Devon. Collected by C. Prideaux. Presented by W. E. Leach. Sphaeroma quadridentatum Say ( 1 8 1 8# : 400) SYNTYPES: 1979 : 419 (4 specimens). St Catherine Island, Georgia, U.S.A. Presented by T. Say. TYPE SPECIMENS OF ISOPODA 127 Sphaeroma rugicauda Leach (1 8 14 : 405 & 433) SYNTYPES: 1979 : 424 (5 specimens). Ulva I., Argyll, Scotland. Presented by W. E. Leach. Zuzara diadema Leach (1818: 344) HOLOTYPE: 1979 : 426 : 1 (fragment). Australia. Collected by R. Brown. Presented by W. E. Leach. Zuzara semipunctata Leach (1818: 344) HOLOTYPE: 1979 : 427 : 1. Locality unknown. Presented by W. E. Leach. Linnaean Collection Oniscus asilus L. (1758 : 636) [redet. Nerocila sp.] Cymothoa oestrum (L.) (1758 : 636) Aega psora (L.)( 1758: 636) Anilocraphysodes(L.)(\158 : 636) Saduria entomon (L.) (1758 : 636) Oniscus marinus L. (1758 : 637) [=Idotea neglecta Sars, 1897. See Heegard & Holthuis (I960)] Idotealinearis(L.)(\161 : 1060) Asellus aquaticus(L.)(\158 : 637) Ligia oceanica (L.) ( 1 767) : 1 06 1 ) Oniscus assimilis L. (1 767 : 106 1 ) [redet. Sphaeroma serratum (Fabricius, 1 787)] Oniscus asellus L. (1758 : 637) Oniscus armadillo L. (1758 : 637) [redet. Armadillidium vulgare (Latreille, 1804)] Acknowledgements Especial thanks are due to Reg Harris for generously sharing with me his wide experience of relaxing dried specimens. I am indebted to Elizabeth Young for permission to examine the Linnaean collection, and to Alwyne Wheeler for guidance through the collection. Finally, my thanks go to Drs Ray Ingle and Tony Fincham for their invaluable advice and criticism in the preparation of this paper. References Bate, C. Spence 1862. Catalogue of the specimens of amphipodous Crustacea in the collection of the British Museum. London (British Museum), 392 pp. Bell, T. 1855. Catalogue of Crustacea in the collection of the British Museum. Part L Leucosidae. London (British Museum), 24 pp. Ellis, J. P. & Lincoln, R. J. 1975. Catalogue of the types of terrestrial isopods (Oniscoidea) in the collections of the British Museum (Natural History) II. Oniscoidea, excluding Pseudotracheata. Bull. Br. Mus. nat. Hist. (Zool.) 28 : 65-1 OOe. Fabricius, J. C. 1775. Systema Entomologiae, sistens Insectorum classes, ordines, genera, species, adjectis synonymis, locis, descriptionibus, observationibus. Flensburgi et Lipsiae, 832 pp. Gordon, I. 1971. Biographical note on Edward John Miers, F.Z.S., F.L.S. Researches Crust. 4: 123-128. Heegaard, P. E. & Holthuis, L. B. 1960. Proposed use of the plenary powers to validate the generic name Idotea J. C. Fabricius, 1 798 (Class Crustacea, Order Isopoda) and matters connected therewith. Z.N.(S.)412. Bull. zool. Norn. 17 : 178-184. Jackson, B. D. 1890. History of the Linnean collections, prepared for the centenary anniversary of the Linnean Society by the senior secretary. Proc. Linn. Soc. Lond. 1887-88 : 1 8-34. Johnston, G. 1834. Illustrations in British zoology. Mag. nat. Hist. 1 : 230-235. Leach, W. E. 1814. Crustaceology. In: Brewster's Edinburgh Encyclopaedia Vol. VII Pt. II: 385-437. 1815. A tabular view of the external characters of four classes of animals, which Linne arranged under Insecta; with the distribution of the genera composing three of these classes into orders, etc. and descriptions of several new genera and species. Trans. Linn. Soc. Lond. 11 : 306^400. 1818. Cymothoadees. In: Cuvier, F. (Ed.) Dictionnaire des sciences naturelles. Paris & Strasbourg Vol. 12:338-354. 128 J.ELLIS Lincoln, R. J. & Ellis, J. P. 1974. Catalogue of the types of terrestrial isopods (Oniscoidea) in the collections of the British Museum (Natural History) I. Superfamily Pseudotracheata. Bull Br Mus nat. Hist. (Zool.)27 : 189-246 h. & Hurley, D. E. 1974. Catalogue of the whale-lice (Crustacea: Amphipoda: Cyamidae) in the collections of the British Museum (Natural History). Bull. Br. Mus. nat. Hist. (Zool.) 27 : 65-72. Linnaeus, C. 1 758. Systema Naturae per regna tria naturae, secundum classes, ordines, genera, species, cumcharacteribus, differentiis, synonymis, locis. (10th edition) Vol. 1 : 1-824 Holmiae. - 1767. Systema Naturae . . . (12th edition) Vol. 1 Pt. 2 : 533-1327 Holmiae. Lockington, W. N. 1877. Description of seventeen new species of Crustacea. Proc. Calif. Acad. Sci. (1)7:41-48. Miers, E. J. 1875#. Descriptions of new species of Crustacea collected at Kerguelen's Island by the Rev. A. E. Eaton. Ann. Mag. nat. Hist. (4) 16 : 73-76. 18756. Descriptions of three additional species of Crustacea from Kerguelen's Land and Crozet Island, with remarks upon the genus Paramoera. Ann. Mag. nat. Hist. (4) 16 : 1 1 5-1 18. 1876fl. Descriptions of some new species of Crustacea, chiefly from New Zealand. Ann. Mag. nat. Hist. (4)17:218-229. 1 8766. Catalogue of the stalk- and sessile-eyed Crustacea of New Zealand. London (New Zealand Colonial Museum and Geological Survey Department), 136 pp. 1877. On a collection of Crustacea, Decapoda and Isopoda, chiefly from South America, with descriptions of new genera and species. Proc. zool. Soc. Lond. 1877 : 653-679. 1880. On a collection of Crustacea from the Malaysian Region. Part IV. Penaeidea, Stomatopoda, Isopoda, Suctoria and Xiphosura. Ann. Mag. nat. Hist. (5)5 : 457-472. 1884. Crustacea. In: Report on the zoological collections made in the Indo-Pacific Ocean during the voyage of H. M.S. 'Alert' 1881-82. London [British Museum (Natural History)]: 178-326 & 513-575. Montagu, G. 1804. Description of several marine animals found on the south coast of Devonshire. Trans. Linn. Soc. Lond. 7 : 61-85. Say, T. 1 8 1 8a. An account of the Crustacea of the United States. /. Acad. nat. Sci. Philad. 1 : 374^0 1 . 1 8 1 86. Description of three new species of the genus Naesa. J. Acad. nat. Sci. Philad. 1 : 482^485. Stebbing, T. R. R. 1888. Report on the Amphipoda collected by H.M.S. Challenger during the years 1873-1876. Rep. scient. Results Voy. Challenger (Zoology) 29 : 1-1737. Stimpson, W. 1857. The Crustacea and Echinodermata of the Pacific shores of North America. Boston J. nat. Hist. 6: 503-513. 1859. Notices of new species of Crustacea of western North America; being an abstract from a paper to be published in the journal of the society. Proc. Boston Soc. nat. Hist. 6 : 84-89. Tattersall, W. M. 1913. The Schizopoda, Stomatopoda, and non-Antarctic Isopoda of the Scottish National Antarctic Expedition. Trans. R. Soc. Edinb. 49 : 865-894. Thurston, M. H. & Allen, E. 1969. Type material of the families Lysianassidae, Stegocephalidae, Ampeliscidae and Haustoriidae (Crustacea : Amphipoda) in the collections of the British Museum (Natural History). Bull. Br. Mus. nat. Hist. (Zool) 17 : 347-388. White, A. 1843. Descriptions of apparently new species and varieties of insects and other Annulosa, principally from the collection in the British Museum. Ann. Mag. nat. Hist. 12 : 342-346. 1847. List of specimens of Crustacea in the collection of the British Museum. London (British Museum), 143 pp. & Doubleday, E. 1843. List of the annulose animals hitherto recorded as found in New Zealand, with the descriptions of some new species. In: Dieffenbach, E. Travels in New Zealand: with contributions to the geography, geology, botany, and natural history of that country. Volume 2 : 265-29 1 London (John Murray). Whitehead, P. J. P. 1969. Zoological specimens from Captain Cook's voyages. /. Soc. Biblphv nat. Hist. 5: 161-201. Zimsen, E. 1964. The type material of I. C. Fabricius. Copenhagen (Munksgaard), 656 pp. Manuscript accepted for publication 12 September 1980 Conchoecia hystrix n. sp. a new halocyprid ostracod for the Porcupine Bight region of the Northeastern Atlantic Martin V. Angel & Celia Ellis Institute of Oceanographic Sciences, Wormley, Godalming, Surrey Introduction During Discovery cruise 105 investigations were made into the influence of the proximity of the sea-bed on the midwater fauna on the continental slope just south of the Porcupine Sea Bight off south-west Ireland in the Northeastern Atlantic. Samples were collected to within 15 m of the seabed at four stations over soundings ranging from about 1000-1650 m using a multiple RMT 1 + 8 system (Roe & Shale, 1979). Height above the bottom was measured by the reflection of the sound signal from the net monitor used to control the opening and closing the nets and to telemeter back to the ship the depth of fishing, net speed and the in situ temperature. Each set of samples included a series of three successive samples of a plankton net with a nominal 1 m2 mouth area and a mesh size of 0'33 mm, and a micronekton net with a nominal mouth area of 8 m2 and a mesh size of 5 mm. Each successive sample was fished closer and closer to the sea bed. Since the soundings varied along the ship's track, the nets were kept within a constant range of the sea bed. Initial analysis of the planktonic ostracods has revealed the presence of a new species of halocyprid ostracod that fits the old concept of the genus Conchoecia (sensu Miiller, 1906), but does not fit any of Muller's groupings. Poulsen (1973) has recently subdivided the genus Conchoecia, but as many of his new genera are heterogenous, and since no type species were designated (Martens, 1979), this new species is described here as Conchoecia hystrix. Conchoecia hystrix n. sp. The specific name is derived from the latin name for the porcupine after the name of the type locality. The holotype, a male was taken at Discovery station 10108 (station details in Table 1). It is mounted in 'Euparal' on slides in the British Museum (Natural History) No. 1980. 132. MALE. The holotype has a carapace length of 1'23 mm and the only other male specimen taken, also at the type locality, was 1'26 mm long. The breadth of the carapace is equal to its height, and a little less than half its length (Table 2). Viewed laterally the carapace tapers anteriorly (Fig. 1 A) with the ventral edge curving smoothly into the posterior edge. Viewed ventrally (Fig. 1 B) the sides of the carapace curve smoothly. There are no spines at the posterior dorsal corner and there is an absence of sculpturing. The asymmetric gland on the left valve opens just posterior of the posterior hinge, but just dorsal to the male dorsal glands. Three groups of small edge glands with granular contents open on the right valve between the opening of the asymmetric gland and the posterior ventral corner, and there are corresponding groups of similar but slightly larger glands on the left valve (Fig. 1C). Frontal organ The stalk of the frontal organ or Organ of Bellonci (Andersson, 1977) ends level with the end of the limb of the first antenna (Fig. ID). The capitulum is bare of armature. It is downturned with a broad base which initially tapers rapidly but the distal two thirds are parallel-sided. The tip is rounded. Bull. Br. Mm. nat. Hist. (Zool.) 40 (4) : 1 29-1 35 Issued 30 July 1 98 1 130 M. V. ANGEL &C. ELLIS Table 1 Station data and length data for all the specimens of Conchoecia hvstrix collected on Discovery cruise 105. Station 10108 haul 8 RMT 1 (Net 3) start position 49°23'6'N 12'47'6'W Depth of fishing 1410-1425 m (1 5-30 m above sea bed) Date 6 September 1979 1347-1420 hours. Distance travelled by net 1'39 km 99 1-34 (mounted specimen), 1-36,1-32, 1-32, 1-38, 1'38, 1*42, 1'32, 1'32, 1'38, l'34mm cW 1 -23 (holotype), 1 -26 mm Stage VI 1-02, 1'02, 1'02, 1'04, 1-06, 1'08, MO mm Stage V 0-78, 0-84 mm Stage IV 0-66 mm Station 10108 haul 7 RMT 1 (Net 2) start position 49°25'4'N 12'49-1'W Depth of fishing 1350-1410 m (30-90 m above sea bed) Date 6 September 1979 1247-1347 hours. Distance travelled by net 3" 73 km 99 1-34, 1-38 mm Station 101 10 haul 5 RMT 1 (Net 3) start position 49° 17-0'N 11'50-8'W Depth of fishing 935-1000 m (1 5^0 m above sea bed) Date 7 September 1 979 2224-2324 hours. Distance travelled by net 2-37 km Stage VI 1 -02 mm Table 2 Meristic characters of the carapaces, frontal organs and first and second antennae of the male holotype and female paratype expressed as percentages of the total carapace length. d 9 Carapace length mm 1-23 1-34 Carapace breadth 47-8 41-1 Carapace height 47-8 52-1 Frontal organ stalk 40-7) 35-0 capitulum 19-0* J J \J Ant. 1 Segt 1 18-2 - Segt2 18-2 - Total 36-5 17-3 dorsal seta - 2-0 a 23-9 x b c 49-5 } 6-5 J 17-7 d 45-1 Ant. 2 Protoprodite 48-6 40-6 Exopodite segt 1 20-8 19-0 Exopodite segt 2-8 9-6 8-7 Longest swimming seta 49.4 45-3 Endopodite seta f 51-2 35-9 Endopodite seta g 38-1 30-3 Endopodite setae h, i, j 16-4 24-6-30-3 First antenna (Fig. 1 D) The first two segments are subequal and bare. The dorsal seta on the second segment that holds the stalk of the frontal organ in place is inserted just over a third of the way from the segment's proximal end. The a seta is long and reflexed back parallel to the limb and almost extends to its base. The c seta is short. The e seta is only slightly longer than the d and b setae, and carries two rows of alternating spines with 20-21 NEW HALOCYPRID OSTRACOD 131 Fig. 1 Conchoecia hystrix male holotype. A. Outline of carapace lateral view. B. Outline of carapace ventral view. C. Detail of the carapace glands viewed from the outside showing the asymmetric gland on the left valve opening posterior of the hinge and its position relative to the openings of the male dorsal corner glands, and also the distribution of the edge glands with granular contents. D. Frontal organ and first antenna. E. Endopodite of the left second antenna. F. Hook appendage of the right second antenna. G. Hook appendage of the left second antenna. Scales in millimetres. spines in each row. There is some suggestion of a weakly developed pad on the b seta, other- wise both the b and d setae are bare. Second antenna The prptopodite is just less than 50% of the carapace length and similar in length to the longest swimming seta. The f seta on the endopodite is slightly longer and is nearly 4/3's the length of the g seta and three times the lengths of the h, i and j setae. All these 132 M. V. ANGEL &C. ELLIS Fig. 2 Conchoecia hystrix male holotype. A. Toothed edge of the coxale and tooth lists of the mandible. B. Basale and endopodite of the mandible. C. Fifth limb. D. Caudal furca and copulatory appendage. E. Sixth limb. Scales in millimetres. setae are bare. The a and b setae are also bare (Fig. 1 E). The hook appendage on the right endopodite (Fig. IF) has a short straight basal part, followed by a right angle bend and a slightly tapering curved arm which has a ridged rounded end. The left hook appendage (Fig. 1G) is a much weaker structure lacking the marked right angle bend and the ridging on its tip. NEW HALOCYPRID OSTRACOD 133 Mandible The first segment of the exopodite (Fig. 2B) carries two long setae and two very short setae on its inner face. The longest terminal claw seta is nearly as long as the total length of the exopodite. The toothed edge of the pars incisa is normal for the genus. There are rows of hairs in the region below the two spine teeth. The coxale toothed edge consists of ten teeth (Fig. 2 A). The distal list has two large teeth, the second of which has secondary serrations, followed by 17-18 small less well defined teeth. The proximal list has a large tooth followed by two smaller teeth, another large tooth and a further fifteen teeth that become progressively smaller. The inner toothed surface, used by Poulsen (1973) as one of the main criteria for separating his genera, appears to be undivided. Labrum It is slightly notched, although the type specimen appears to have an aberrant structure. Maxilla The basal segment (Fig. 3E) carries a seta. The first endopodite segment has six anterior, one lateral and three posterior setae. Fifth limb The first segment of the exopodite (Fig. 2C) has a group of three setae ventrally near its base, and a further two setae inserted laterally on its outer face. A further two setae occur ventrally towards its distal end together with a lateral seta on both the inner and outer faces. There is also a long dorsal seta which extends to the tip of the limb. The second segment has one dorsal and two ventral setae. The most ventral of the terminal claw setae is relatively short and thin. Sixth limb The basal segment (Fig. 2E) carries two very small setae. The second segment has a single ventral seta and the third segment one seta on both dorsal and ventral surfaces. Caudal Furca The furca (Fig. 2D) carries the normal eight claw setae for the genus with a dorsal unpaired seta. The first claw setae are more curved than usual. Copulatory organ The organ (Fig. 2D) tapers towards its rounded end. There are about five bands of oblique muscles. FEMALE. The female paratype has a carapace length of 1 '34 mm. It, too, is mounted on slides in 'Euparal' and deposited in the British Museum (Natural History) No. 1980. 133. The range in length of the other thirteen specimens is 1'32-1'42 mm, averaging 1*35 mm. The carapace shape (Fig. 3 A, B) is similar to that of the male's, although the height is relatively greater and the breadth relatively slimmer. The positions of the asymmetrical glands are similar to the male, and there is the same distribution of edge glands with granular contents. There is a faint longitudinal sculpturing on the carapace of some specimens. Frontal organ The capitulum is not well differentiated from the stalk (Fig. 3C), but it is bent down very slightly. The ventral and lateral surfaces of the capitulum are covered with short spines. The tip of the capitulum is produced into a spine. The whole organ is about twice the length of the limb of the first antennae. First antenna The limb is not well differentiated into segments (Fig. 3C) and it is bare of supplementary armature. There is a very short dorsal seta which does not even reach the end of the limb. The a-d setae are half the length of the e seta, which is about a third the carapace length. The e seta carries a few relatively long distally pointing spinules on its trailing edge halfway along its length, and a scatter of similar spinules more distally on the leading edge. Second antenna The protopodite (Fig. 3D) is around 40% of the total carapace length. It carries a patch of hairs close to the insertion of the endopodite. The first exopodite segment is just less than half the length of the protopodite and just over twice the length of the other exopodite segments. The longest swimming seta is a little longer than the protopodite. On the endopodite the a and b seta are bare. There are no c, d or e setae. The f seta which is equal in length to the protopodite, is bare but is slightly flattened distally. The g seta is similar in structure and length to the h, i and j setae. Gut contents The gut contents included a multilayered block of folded membranous material of unidentifiable origin. The contents were rich in densely staining granules that appeared to be coccoid bacteria about 1 //m in diameter. It also included a few rounded mineral particles 2-4 //m in size, suggesting the species may feed on the sea bed. 134 M. V. ANGEL &C. ELLIS Fig. 3 Conchoecia hystrix female paratype. A. Carapace lateral view. B. Carapace ventral view. C. Frontal organ and first antenna. D. Second antenna. E. Endopodite of maxilla. F. Sixth limb. Scales in millimetres. NEW HALOCYPRID OSTRACOD 1 35 DISCUSSION. The depth range of this species appears to be restricted. It was absent from samples collected at depths of 1600-1700 m and only a single specimen was taken at a shallower depth than at the type locality. The occurrence of novel species close to the sea bed even in relatively shallow depths suggests that there are distinct environmental conditions in this poorly explored habitat (e.g. Wishner, 1980). References Andersson, A. 1977. The organ of Belloni in ostracodes: on ultrastructual study of the rod-shaped organ or frontal organ. Acta zool, Stockh. 58 : 197-204. Martens, J. M. 1979. Die pelagischen Ostracoden der Expedition Marchile I (Siidost-Pazifik), II: Systematik und Vorkommen (Crustacea: Ostracoda: Myodocopida). Mitt. hamb. zool. Mus. Inst. 76 : 303-366. Miiller, G. W. 1906. Ostracoda. Wiss. Ergebn. dt. Tie/see- Exped. Valdivia 8 (2) : 29-1 54. Poulsen, E. M. 1973. Ostracoda-Myodocopa Pt. Illb Halocypriformes-HalocyprididaeConchoecinae. Dana Rep. 84 : 1-244. Roe, H. S. J. & Shale, D. M. 1979. A new multiple rectangular midwater trawl (RMT 1 +8 M) and some modifications to the Institute of Oceanographic Sciences' RMT 1+8. Mar. Biol. 50 (3): 283-288. Wishner, K. F. 1980. The biomass of the deep-sea benthopelagic plankton. Deep-Sea Res. 27 (3): 203-2 16. Manuscript accepted for publication 10 September 1980 The Conchoecia skogsbergi species complex (Ostracoda, Halocyprididae) in the Atlantic Ocean A. J. Gooday Institute of Oceanographic Sciences, Wormley, Godalming, Surrey, U.K. Introduction The planktonic ostracod genus Conchoecia includes a heterogeneous array of 105 described species with at least another 15-20 recognized, but as yet undescribed (Angel, personal communication). The named species have been separated into a number of groupings (Miiller, 1 906# ; Skogsberg, 1 920). Some of these were given generic status by Granata & di Caporiacco (1949) and Poulsen (1973) but this nomenclature is not universally accepted because several of the genera appear to be unnatural units (Angel and Fasham, 1975 : 71 1). One of the more important, and probably natural, assemblages is the rotundata group ( = Metaconchoecia Granata & di Caporiacco, 1949), which comprises 11 or 12 previously described species, most obviously united by the location of the left asymmetric gland in an anterior position, just behind the rostrum. Although most of the rotundata group species are fairly well understood, two names, C. rotundata Miiller, 1890 and C. skogsbergi lies, 1953, have not been applied consistently. C. rotundata has been the subject of particular confusion. These long standing taxonomic problems have been compounded in recent years by Angel's recognition that 'C. rotundata' in Atlantic Discovery material comprises an array of very similar 'forms' (Angel, 1972; 1979 : 68-73; Angel & Fasham, 1975 : 71 1). In the present paper, which is largely based on the same Discovery material, C. rotundata and C. skogsbergi are redescribed and eight new species are established. All but one of these ten species are embraced by Angel's 'C. rotundata forms 1-15'. They together make up a closely related assemblage within the rotundata group which is referred to here as the skogsbergi complex. It is obvious from material in other collections that many, perhaps all, of the skogsbergi complex species were seen by earlier workers. Understandably, they were usually identified as C. rotundata or, in more recent literature, as C. skogsbergi. Materials and methods Most of the material was collected by the RRS Discovery between 1968 and 1974 in the N. Atlantic at a series of stations situated approximately along the 20°W meridian between 60°N and the equator (Angel & Fasham, 1975) and also along a 32°N transect from Africa to Bermuda (Angel, 1979). With a few exceptions, these samples were taken with the RMT 1 component of the RMT 1 + 8 opening and closing net system which is able to sample discrete horizontal horizons within the water column (Baker, Clarke & Harris, 1973). The gear was usually fished at four depth horizons down to 100 m, then 100 m horizons down to 1000 m and broader bands between 1000 m and 2000 m, thus allowing reasonably precise data on the depth distribution of planktonic organisms to be obtained. Further details of sampling procedures with the RMT 1 +8 are given elsewhere (Angel & Fasham, 1976; Badcock & Merrett, 1976; Angel, 1979). At Discovery Station 6665, a modified Indian Ocean Standard Net (Nl 13) fitted with a catch dividing bucket (CDB, Foxton, 1963) was used. A smaller number of specimens from the S. Atlantic and the Atlantic sector of the Southern Ocean Bull. Br. Mus. nat. Hist. (Zool.) 40 (4) : 1 37-209 Issued 30 July 1 98 1 138 A. J. GOODAY were collected by Discovery II between 1936 and 1938 with vertically hauled 70cm nets (N70V). The station data are deposited in the library of the British Museum (Natural History). The Discovery material of lies (1953), the remaining specimens of which have been reexamined, was collected from the RRS William Scoresby with N70V nets. Some important museum material from the Atlantic, Indian, Pacific and Southern Oceans, including specimens studied by G. W. Miiller, Skogsberg, Fowler and Poulsen, was reexamined. However, a comprehensive examination of all relevant material in other collections was not attempted. Details of this material are given below and in Tables 1 & 2 and Appendix 1. The following abbreviations are used when referring to examined specimens: BM(NH) -British Museum (Natural History), London. DC, Wormley -Discovery Collections, Institute of Oceanographic Sciences, Wormley, U.K. NR, Stockholm -Naturhistoriska riksmuseet, Stockholm, Sweden. SI, Washington -Smithsonian Institution, Washington, D.C., U.S.A. ZIZM, Hamburg -Universitat Hamburg, Zoologisches Institut und Zoologisches Museum, Hamburg, Federal Republic of Germany. ZM, Copenhagen -Zoologisk Museum, Copenhagen, Denmark. ZM, Berlin -Zoologisches Museum, Berlin, German Democratic Republic. S. A.E. -Swedish 'Antarctic' Expedition 1 90 1-1 903. In the laboratory, the ostracods were examined, measured and dissected under a Wild M5 Stereomicroscope. Mounted animals were examined under a Wild Ml 5 microscope. Carapace outlines were executed with the aid of an M5 'Zeichentubus' and the line drawings of appendages with an Ml 5 'Zeichentubus'. Carapace lengths and breadths were measured with the animal lying on its back. In the text, mean carapace lengths are given with their standard deviations. Potentially ambiguous measurements are defined by Gooday (1976:59). The following abbreviations for morphological characters are used: Al First antenna A2 Second antenna Exl, 2 etc. First, second etc. segment of exopodite Enl, 2 etc. First, second etc. segment of endopodite LSS Longest swimming seta F.O. Frontal organ L Carapace length H Carapace height B Carapace breadth LAG Left asymmetric gland RAG Right asymmetric gland Historical Review (i) 1890-1920. Conchoecia rotundata was established by Miiller (1890), in the earliest of his halocyprid papers, on the basis of a few specimens taken at a depth of 1000^000 m of wire at two stations in the tropical Pacific. These specimens were up to 1-15 mm long. They were inadequately described and the species cannot now be recognized with any confidence, although its identity is speculated on below. A few years later, Miiller (1894) gave a fuller CONCHOECIA SKOGSBERGI SPECIES COMPLEX 139 description of a smaller form (L = 0'80mm) from the Mediterranean which had a more rounded lateral outline. The situation becomes further confused with Miiller's (1906#) Valdivia report in which specimens of C. rotundata, collected over a wide area (40°N to 62°S in the Atlantic, Indian and Southern Oceans), were said to vary considerably in size and outline with height : length ratios of 4/7 (57'1%) to 8/19 (42-1%) and lengths of 1-40 mm to 1-75 mm for Antarctic specimens and 0*80 mm to 1*40 mm for those from warmer water. Miiller (1906a : 83, pi. XVII, figs 23-26) distinguished two distinct carapace types in the Valdivia material, one long and elongate and the other relatively short and more rounded, the latter corresponding closely to the Mediterranean form (Miiller, 1 894) of C. rotundata. The same author gives additional records of C. rotundata in his Siboga and Gauss reports (Miiller, 19066, 1908) but in neither of these papers is the material described. The account of this species in Miillers (1912) comprehensive treatise on the Ostracoda is drawn from the description in the Valdivia report. It is shown below that Miiller's Gauss and Valdivia specimens, at least, belong to a number of species and his published descriptions (Miillers, 1906#, 1912) of 'C. rotundata' greatly oversimplify the nature of this material. In an unorthodox but stimulating contribution to halocyprid taxonomy, Fowler (1909) recognized the taxonomic difficulties created by Miiller's inclusion of two widely different forms in one species. Fowler believed both forms were present in his material from the Bay of Biscay and resolved the problem by regarding the elongate carapaces as adults (Stage I) and the short carapaces as penultimate instars (Stage II) of the same species. The Stage I instars had mean lengths of 1 -0 mm (d1) and 1 • 1 mm (9), the Stage II instars had mean lengths of 0*75 mm (rf) and 0*79 mm (9) (Fowler, 1909 : 273). It is significant that these temperate specimens of the elongate form were markedly smaller than Miiller's (\906a) elongate form from the Southern Ocean, a point returned to below. This initial period of research ended when Skogsberg (1920) described 24 specimens of C. rotundata from the SW Atlantic, ranging in length from 1-45 mm to 1'60 mm (both sexes). These correspond well in size and lateral shape to Miiller's (\906a) long form. With characteristic thoroughness, Skogsberg (1920) reviewed the C. rotundata problem and concluded that the long and short forms were distinct taxonomic entities, the short form, of which he had no material, perhaps being the same as C. nasotuberculata Miiller, 1906, and the elongate form being closer to Miiller's original concept of C. rotundata. (ii) lies' contribution. One of the key contributions was that of lies (1953) who studied Discovery samples from the Benguela Current in the SE Atlantic. In many of these samples, lies identified both adults and juveniles of the long and short forms of C. rotundata as well as C. nasotuberculata. lies noted that the long and short forms in his material were morpho- logically quite distinct and also had different depth distributions. He therefore concluded that Skogsberg (1920) had been correct in suspecting that the long and short forms of Miiller (1906#) and Fowler (1909) were separate species. lies believed that the short form was conspecific with the Mediterranean C. rotundata of Miiller (1894) because of its similar size (L = 0'80-0'90 mm) and lateral carapace outline. In addition, he pointed out that despite Miiller's (1890) inadequate description, there were sufficient differences between the short form and Miiller's original concept of C. rotundata to justify describing it as a distinct species, C. teretivalvata lies, 1953. A second new species, C. skogsbergi, was erected for the long form, which clearly differed in size and lateral outline from the original C. rotundata. However, lies (1953) did not describe the Benguela Current material of C. skogsbergi but referred to Skogsberg's (1920) account of C. rotundata from the SW Atlantic as the type description. Following lies (1953) report, C. rotundata was then left as the name of a valid but unrecognizable taxon which included only Miiller's (1890) specimens. (iii) 7967 onwards. Despite the unresolved nature of C. rotundata, the taxonomic legacy left by Skogsberg and lies appeared to be fairly satisfactory in that two species, C. skogsbergi and C. teretivalvata, both previously identified as C. rotundata, were now clearly recognized as distinct and were adequately described. This was the situation when the lull in planktonic 140 A.J. GOODAY ostracod studies, which followed the publication of lies (1953) paper, was terminated in the late 1960s. During this most recent and continuing period of research, both of lies (1953) species have been widely recognized. There are records of C. skogsbergi from the following high latitude areas in the northern hemisphere: the Norwegian Sea (Angel, 19680), near the North Pole (Leung, 1972; 1973), the Bering Sea (Chavtur & Shornikov, 1974), the Okhotsk and Bering Seas and the Kurile- Kamchatka trench (Chavtur,. 1976; 19770; 19776). In the Southern Ocean it has been recorded by Hillman (1967; 1968; 1969; as "C. rotundata') and Deevey (1974; 19786). These recent, cold water, high latitude reports of C. skogsbergi are fairly convincing since where descriptions, or at least length data, are given, the specimens conform more or less closely to Skogsberg's (1920) definitive description. C. skogsbergi has also been reported from warmer, temperate, and tropical waters but here the situation is more complex and probably none of these identifications is correct. Angel's recent studies of N. Atlantic Discovery material have revealed an array of undescribed ostracod species, closely related to C. skogsbergi but smaller, although generally larger than C. teretivalvata. Angel & Fasharn (1975; see also Angel, 1972; 19770; 19776; 1979) distinguished 1 5 forms within this complex. In the present study, which is largely based on the same material, Angel's forms are placed in nine species, seven of them new. Some of these skogsbergi complex species are the same as those assigned by other authors to C. rotundata or low latitude forms of C. skogsbergi. The male C. rotundata long form of Fowler (1909: pi. 24, fig. 205) belongs to Angel & Fasham's (1975) form 3 while the smaller, warmer water variety of C. skogsbergi reported by Deevey (1968; 1974; 19780) from the Sargasso Sea, the Atlantic between 30°S and 35°S and from off Venezuela, is at least in part equivalent to forms 1, 5 and 13. From the same area Deevey (1968) described, as C. rotundata, a small form which also belongs to this complex (form 1 5). The rather different species from the W. Atlantic and tropical Pacific assigned to C. rotundata by Poulsen (1973), is close to form 4. Another species, said to be distinct from both C. skogsbergi and C. teretivalvata, was reported by George (1969) and George, Purushan & Madhupratap (1975) from the NW Indian Ocean but was not described and so cannot be assessed; the same comment applies to C. rotundata of Chavtur (19770) from the subtropical Pacific. The C. rotundata material of G. W. Muller Several fairly large collections of G. W. Miiller's material still exist (Athersuch, 1976). During this study, animals collected by the Valdivia (Deutsche Tiefsee-Expedition 1898-1899; 642 specimens: see Muller, 19060) and the Gauss (Deutsche Sud-Polar Expedition, 1901-1903; 617 specimens: see Muller, 1908), and identified by Muller as C. rotundata, were examined. These specimens are from the Atlantic, Southern and Indian Oceans. None are from the Pacific. It was not possible to examine the Bay of Naples samples (Muller, 1894) mentioned by Athersuch (1976). The Valdivia and Gauss material is all preserved in alcohol. In general, the Valdivia specimens are poorly preserved with the valves widely splayed, often making identification impossible, or at best tentative. The Gauss material is generally in better, often surprisingly good condition and most specimens can be identified. The 20 or so species found in these samples are listed in Tables 1 and 2 and Appendix 1 . About half of them (C. brachyaskos Muller, 1906, C. elegans Sars, 1865, C. hyalophyllum Claus, 1890, C. kyrtophora Muller, 1906, C. macrochiera Muller, 1906, C. macromma Muller, 1906, C. nasotuberculata, C. procera Muller, 1894, C. pseudoparthenoda Angel, 1971 (very close to C. parthenoda Muller, 1906) and C. spinirostris Claus, 1974) are each represented by only a few specimens and were probably simply misidentified by Muller. The remainder comprise C. arcuata Deevey 1978, C. rotundata sensu Deevey, 1968, C. skogsbergi, C. teretivalvata and five of the new species described herein, all of which were presumably embraced by Miiller's understanding of C. rotundata. These nine species, although closely related, differ from each other to varying degrees and in some cases are rather strikingly different (for example C. skogsbergi and C. teretivalvata). It is therefore CONCHOECIA SKOGSBERGI SPECIES COMPLEX 141 somewhat surprising that Miiller was content to leave them unseparated. However, when examining his samples, for example the one from Gauss Station 12.11.01 which contains 196 specimens belonging to nine different species (Table 2), one senses that he was exasperated by the diversity of closely related species within 'C rotundata' and chose (Miiller, 1906a, 1912) not to proceed beyond dividing this 'species' into long and short forms. The present status of C. rotundata and C. skogsbergi Following the historical review and the discussion of Miiller's material, we are now in a position to consider how the names C. rotundata and C. skogsbergi are applied in this paper. Miiller's (1890) original description of C. rotundata includes the following taxonomically useful information. The carapace is moderately elongate and strongly tapered with the greatest height somewhat > half the length; the posterior end is strongly curved with the extremity at about half the height; the maximum breadth is somewhat < half the length; the length is up to 1 • 1 5 mm; the frontal organ capitulum is of somewhat variable shape; there are ten pairs of spines on the male first antenna e seta. Miiller also illustrates a side view of the female carapace, the female second antenna endopodite, the male first antenna e seta armature and two male and two female frontal organ capitulums. As pointed out above, this information, together with the figures, is not sufficient to allow the species to be identified. However, it is possible to speculate. The lateral outline (Miiller, 1890: pi. XXVIII, fig. 42) is reminiscent of several skogsbergi complex species, although less elongate than any of them. On the other hand, the outline is more elongate than in C. kyrtophora, C. nasotuberculata and C. teretivalvata and these three species are also smaller (< 1-0 mm) than the largest of Miiller's (1890) specimens. The male e seta spines (pi. XXVIII, fig. 43) resemble these spines in the skogsbergi complex but are unlike those of C. kyrtophora. On the other hand, one of the female and one of the male frontal organ capitulums (pi. XXIX, figs 14a, d) are closely similar to the capitulum in C. kyrtophora and some related rotundata group species, while the other two capitulums (pi. XXIX, figs 14b, c) are more reminiscent of this structure in the skogsbergi complex species. It is therefore likely that Miiller ( 1 890) had at least two species in his material, perhaps C. kyrtophora and a species of the skogsbergi complex. Miiller's Pacific types of C. rotundata are presumed lost but, as discussed above, the Gauss and Valdivia material of this 'species' includes a variety of species, many of them belonging to the skogsbergi complex. None of this material is from the Pacific and it can therefore cast no direct light on the original identity of C. rotundata. However, it does demonstrate that Miiller (1906<2, 1908) had a very broad concept of C. rotundata when writing the Valdivia and Gauss reports and adds weight to the above suggestion that this species was originally polytypic. Since there is no clear idea of what Miiller (1890) meant by C. rotundata, the name is applied here in the sense of Deevey's (1968). There are, admittedly, some clear discrepancies between Deevey's species and Miillers (1890) original account of C. rotundata. Miiller's specimens were considerably larger and the lateral outline, although similar, was relatively higher; also the male first antenna e seta (Miiller, 1890: pi. XXVIII, fig. 43) had more spines (20). Several of the new skogsbergi complex species are closer to Miiller's (1890) specimens in size and lateral outline and may be better candidates for this name. However, less confusion will be caused if the name is retained for Deevey's species which is now adequately described and well understood. This pragmatic approach is further justified by the presence of C. rotundata sensu Deevey in Miiller's Gauss material. C. skogsbergi presents rather different problems. As outlined above, lies established the species after studying specimens from the Benguela Current which convinced him that the long form (C. skogsbergi) and the short form (C. teretivalvata) of C. rotundata were distinct species. However, lies did not actually describe the Benguela specimens that he referred to C. skogsbergi but directed the reader to Skogsberg's (1920) description and figures which 'may 142 A. J.GOODAY be taken as typical for this species'. This statement is followed by the sentence 'Type material will be deposited at the British Museum' (lies, 1953 : 265), but in fact there are no specimens of C. skogsbergi in the BM(NH) collections. In July 1976, Dr lies kindly sent me his remaining Benguela Current material of C. skogsbergi. It included a dissected male, mounted on two slides, each labelled 'type'. However, the slides did not have a catalogue number and were clearly not part of a museum collection. Since the Benguela Current material was not described, since the type specimens selected from it were not deposited in a museum and since the type description was said to be Skogsberg's (1920) account of the long form of C. rotundata, the species should clearly be based on Skogsberg's material. I therefore obtained, through the courtesy of Dr R. Olerod, Skogsberg's remaining material of C. rotundata in the Naturhistoriska riksmuseet, Stockholm, and selected a type specimen from it. This procedure, apart from being nomenclaturally correct, has the advantage of basing C. skogsbergi firmly on a typical specimen from the SW Atlantic, rather than on the undescribed material of lies (1953). In fact, lies specimens have proved, on reexamination, to belong to another of the skogsbergi complex species and hence additional nomenclatural confusion has been avoided. Taxonomic characters For the purpose of routinely identifying planktonic ostracods, it is obviously an advantage if the important taxonomic characters are external features of the carapace, rather than appendage details which require dissection to be seen. It is therefore fortunate that in the skogsbergi complex, and probably in the genus Conchoecia as a whole, speciation seems always to be expressed morphologically in the size and shape of the carapace. The disposition and armature of the limb segments, and other internal features are, on the other hand, relatively conservative. With experience, and reasonable preservation, the species described here can therefore be recognized by their external morphology, although examination of the limbs is occasionally essential to confirm an identification. In this section, the taxonomically important carapace characters, and some less important internal characters, are discussed. (i) Carapace outline. In material which is reasonably undistorted, species can usually be separated by consistent, although sometimes subtle, differences in their lateral and ventral outlines. In lateral view, particular attention should be paid to the relative height of the carapace, the way it tapers, the curvature of the posterior end and the position, above or below the mid-point, of the posterior extremity. In ventral view, the relative breadth of the carapace, the curvature of the sides and the shape of the an tero ventral region, which may be sharply or bluntly pointed, are important. The ventral appearance of the carapace should always be illustrated. There is often some intraspecific variability in the carapace outline, for example in the degree to which specimens taper in lateral view. To show the range of variation, a series of carapaces are illustrated for each of the species described in the paper. (ii) Carapace length. Certain closely related species of Conchoecia can be readily dis- tinguished because their size ranges do not overlap (Angel, 1973; Gooday, 1976). This is not usually the case in the rotundata group, most species of which have overlapping size ranges. However, considering the skogsbergi complex by itself, two species do have characteristic carapace lengths: C. rotundata is consistently smaller, and C. skogsbergi is usually larger than other species. As shown below in the key, carapace length data may also be of value in separating similar species. Measurements of material spanning the known range of each species often reveal slight, geographically related size variations, lengths usually tending to decrease southwards. These complete data on the geographical variation in carapace length are available in tables stored in the BM(NH) library. CONCHOECIA SKOGSBERGI SPECIES COMPLEX 143 (iii) Position of the left asymmetric gland. In the rotundata group, the position of this gland varies from <9% (C. nasotuberculata) to >25% (C. glandulosa Miiller, 1906, C. macromma} of the carapace length behind the tip of the rostrum, but does not uniquely characterize any of the species. Gland positions for the skogsbergi complex species are summarized in Table 9. In one of these species the gland is situated more posteriorly than in another member of the complex. The gland position may also be of value in discriminating between closely related species. (iv) Other morphometric characters. The relative lengths of the frontal organ capitulum and the antennal segments and setae, expressed as percentages of the carapace length, have proved to be of taxonomic value in the genus Conchoecia. Within the skogsbergi complex, two species can be recognized using this character alone and pairs of species may also be separated in this way. However, the measurements are tedious and time consuming to obtain and although they may be used to confirm doubtful identifications, they are rarely of primary taxonomic importance. (v) Frontal organ capitulum. This prominent structure is almost always described and illustrated in species descriptions and hence its usefulness as a taxonomic character needs to be considered. In the rotundata group it is of variable shape. The male capitulum of C. glandulosa, C. kyrtophora, C. macromma, C. pusilla Miiller, 1906 and C. nasotuberculata is usually elongate, often slightly curved with a rounded, sometimes rather bulbous end. In the skogsbergi complex, and in C. teretivalvata, the male capitulum is relatively shorter and distally tapered, the ventral margin is clearly convex proximally and there is a corre- sponding, although less pronounced, concavity of the dorsal margin. In C. isocheira Miiller, 1906 and C. arcuata, the male capitulum is rather similar to that of skogsbergi complex species. The female capitulum of C. arcuata, C. isochiera, C. kyrtophora, C. macromma, C. nasotuberculata and C. pusilla is not clearly delimited from the shaft, elongate, and rather bulbous distally, with a terminal spine in C. macromma. C. glandulosa has a similar, but more bulbous female capitulum. In the skogsbergi complex, and in C. teretivalvata, the female capitulum is more clearly delimited from the shaft, more or less tapered and the end pointed and downturned, or narrowly rounded and less clearly downturned. Hence, within the rotundata group, the shape of the capitulum may be characteristic of certain species, or species groupings. Among members of the skogsbergi complex, the shape of the capitulum displays rather little interspecific variation. In males, the relative height may vary somewhat while the female capitulum is rounded in three species, pointed and clearly downturned in the remainder. Thus none of the species described in this paper has a characteristically shaped capitulum. Moreover, the shape sometimes varies within a species; this is particularly so in the case of C. skogsbergi (Figs 18, 19). For these reasons, the capitulum shape has very limited taxonomic value in the skogsbergi complex. The number and the distribution of capitulum spines have been used as characters to separate C. skogsbergi and C. rotundata (Poulsen, 1973 : 73). However, the value of these spines as taxonomic characters was not confirmed during the present study. (vi) Armature of the male first antenna e seta. This is another of the characters usually described by taxonomists. Among rotundata group species, C. kyrtophora is unique in having square ended spines lying at right angles to the seta (Angel, in press) and in C. isochiera the spines bear a single, distal row of 'moderately large, oval, hyaline appendage(s)' (Skogsberg, 1920:658). Miiller (1912:62) used differences in the e seta armature to characterize these species in his key. All other members of the rotundata group have pointed, paired or staggered spines, lying almost flat against the seta. The number of these spines varies from 14 to 31 in the skogsbergi complex (Tables 5 & 13). However, the numbers show considerable overlap between species and are therefore of rather limited usefulness in taxonomy, although C. rotundata usually has fewer spines than the other members of the complex. 144 A. J. GOOD AY Key to the Conchoecia rotund at a group (Figs 1-9, pp. 1 63, 1 7 1-1 74) 1. LAG > 20% of length behind tip of rostrum 2 LAG < 20% of length behind tip of rostrum 4 2. Length usually < 1-25 mm (0-90-1 -29 mm)* macromtna Length usually > 1*25 mm 3 3. Anterior end curved and produced well forwards below rostrum; posterior end bluntly pointed where RAG opens above mid-point. L = 1 -40-2- 1 0 mm . . glandulosa Anterior end not produced forwards; posterior end smoothly curved. L= 1-60-1-85 mm (c?); 1-38-1-55 mm (9) abyssalis 4. Length > 1-30 mm skogsbergi Length < 1-30 mm 5 5. Carapace short and round in lateral view; H > 60% of length 6 Carapace more cylindrical; H < 50% of length 8 6. Sides of carapace evenly curved in ventral view. L = 0-80-1- 15 mm . . .teretivalvata Sides of carapace constricted behind insertions of A2 when viewed ventrally ... 7 7. Carapace with lateral tubercles close to posterior dorsal corner, cf Al e seta spines lie pointing dorsally, almost parallel to seta. LAG opens on rostrum, in front of anterior end of hinge nasotuberculata Carapace without tubercles.cf Al e seta with spines set at right angles to seta. LAG opens just behind rostrum posterior to anterior end of hinge L = 0-72-1-0 mm . . kyrtophora 8. RAG situated below posterior dorsal corner, opening at end of triangular process which makes distinct angle in upper part of posterior margin 9 RAG opens near posterior dorsal corner, riot on a process so that posterior margin of carapace appears smoothly curved 10 9. cf Al e seta bears single row of 7-9 spines, each bearing a distal hyaline appendage. L = 0-80-1 -00 mm (d), 0-95-1- 11 mm (9) isochiera cf Al e seta bears about 1 6 simple paired spines. L = O'70-l • 1 3 mm . . . pusilla 10. Carapace with maximum height near middle, ventral margin arcuate; incisure tends to be deep and curved, cf A2 i and j setae with side branches. LAG > 1 5% of length behind tip of rostrum. L = 0-93-1- 12 mm arcuata Carapace with maximum height nearer posterior end, lateral outline tapered with gently curved ventral margin; incisure tends to be fairly shallow, cf A2 i and j setae simple. LAG > 15% of length behind tip of rostrum in one species only .... 11 1 1 . LAG > 1 5% of length behind tip of rostrum; in ventral view carapace narrow with B usually <40% of length. On Al, b and d setae only slightly shorter than e seta. L = 0'95-1'18 mm discoveryi sp.n. LAG < 15% of length behind tip of rostrum; except in one species, carapace wider in ventral view, with B consistently >40%of length; d A 1 b and d setae markedly shorter than eseta 12 12. LAG usually 12%-15% of length behind tip of rostrum. Posterior end strongly and symmetrically rounded or slightly downturned; in cf 8 = 38-42% of length. L> 1-06 mm (1-10-1-40 mm) /ow/m'sp. n. LAG perched above and just behind rostrum, < 12% of length behind tip; in cf, B is always > 40% of length. Length < 1-20 mm ' 13 13. Carapace less strongly tapered, anteroventral region, below rostrum, is not sharply pointed in ventral view. Ventral outline not strongly biconvex. cfAl e seta with 23-28 spines . 14 Carapace more strongly tapered, anteroventral region, below rostrum, is pointed in ventral view. Ventral outline more strongly biconvex, cf A 1 e seta with 1 4-26 spines ... 15 14. L= 1-06-1-26 mm. Posterior end asymmetrically curved in lateral view with extremity below the mid-point wolferisp.n. L = 0'9 1-1 -06 mm. Posterior end more symmetrically curved in lateral view with extremity around the mid-point obtusa sp.n. 15. Carapace unusually broad in ventral view, particularly in 9; B >50% (cf), >46'0% (9) of length. Length = 0-97-1 -16 mm inflatasp.n. Carapace less broad in ventral view; except in one species, B < 50% (cf), < 46% (9) of length 1 6 *Although Deevey ( 1 974) gives lengths of 1 -50 mm (cf), 1 -27-1 -40 mm (9) for C. macromma. CONCHOECIA SKOGSBERGI SPECIES COMPLEX 145 16. Length < 0-8 7 mm (9), < 0-8 3 mm (c?);rfAl e seta with 14- 18 spines. . . rotundata Length > 0'87 mm (9), > O83 mm (rf);rfAl e seta with 18-24 spines 17 17. Carapace height usually >50% of length; rf Al LSS >60% of length, 9 A2 LSS >47% of length. Length = 0-98-1 '08 mm australis sp.n. Carapace height usually <50% of length; c?Al LSS <60% of length, 9 A2 LSS <47% of length 18 18. Length = 0-95-1 -20 mm. d A2 f seta >40%, g seta >43% of length; 9 A2 LSS >41% f-j setae > 20% of length subinflata sp.n. Length = 0-85-1 -01 mm. cfA2 f seta < 38%, g seta < 42% of length; 9 A2 LSS < 41%, f-j setae < 20% of length acuta sp.n. Systematic descriptions Species of the skogsbergi complex show virtually no interspecific variation in the structure and setation of the mandible, maxilla, 5th, 6th and 7th limbs, labrum and caudal furca, or in the dentition of the mandibular tooth lists and cutting edge. The basic morphology of the first and second antenna in both sexes is also constant with only the proportional lengths of the main setae and segments (summarized in Tables 3, 4, 7, 8, 11, 12) and details of the setal armature (Tables 5, 13) varying between species. In the systematic section that follows, a complete account is therefore given only for C. fowleri sp. nov. and descriptions of other species are limited to those characters in which they differ morphologically from C. fowleri sp. nov. The description is not based on C. skogsbergi itself because the relatively large size and polytypic character of this species make it rather atypical of the skogsbergi group as a whole. Most of the type material and other figured Discovery specimens are deposited in the BM(NH), under registration numbers 1979.690-827, 1980.141-145. There is also a collection of Discovery specimens of most of the described species deposited in the SI, Washington under registration numbers USNM 158124-158132. Museum specimens are undissected, except where otherwise stated. Dissected specimens are stained with lignin pink and mounted on slides in Euparal. Undissected specimens are preserved in 80% alcohol. Conchoecia fowleri sp. nov. (Figs 10-17, 18A-J, 19A-I) Conchoecia rotundata Miiller, 1890.— Miiller, 1908:69-70 (in part).— Fowler, 1909:249-251 (in part),? not pi. 23, figs 217 (? = C.0>rwata Deevey, 1978), 206, 208-210,212, 214,216,218,220,222, 224 ( = ?), pi. 24, figs 205 ( = C. teretivalvata lies, 1953), 207 ( = C. subinflata sp. nov.), 2 1 1 , 2 1 3, 22 1 , 223 (? = C. pusilla Miiller, 1 906), 2 1 5 ( = ?), 2 1 9 ( = C. teretivalvata lies, 1953). Conchoecia rotundata Miiller, 1890, forms 1, 5, 9. Angel & Fasham, 1975 : 737 (distribution). — Angel, 1977a : 246 (vertical distribution). Not Conchoecia rotundata Miiller, 1890. Conchoecia skogsbergi lies, 1953. Angel, 19686:308, Fig. 8 (vertical distribution). — Deevey, 1968 : 54-55, Figs 20a-d, 21a, d, f-h, 22a.— Angel, 1969 : 518, 539 (vertical distribution).— Deevey, 1974 : 364 (not Fig. 5b, = C. skogsbergi lies, 1953).— Deevey, 19780 : 70. Not Conchoecia skogsbergi lies, 1953. ETYMOLOGY. Named after Dr G. H. Fowler, one of the first authors to work on planktonic ostracods from the NW Atlantic. DIAGNOSIS. Lateral carapace outline elongate, rather gently tapered in anterior 2/3 to 3/4; posterior end symmetrically rounded, or slightly upturned. Length = 1 -10-1 -28 mm. Ventral carapace outline relatively narrow and weakly biconvex, in c? B = 38%-43% of length; anteroventral part of each valve not sharply pointed in ventral view. LAG usually lies 12%- 1 5% of length behind tip of rostrum. TYPE MATERIAL. Holotype: dissected d (BMNH 1979.695). Paratypes: 299, dissected (BM(NH) 1979.696-697); 18cW,2l99(BM(NH) 1979.746-755). 146 A. J. GOODAY TYPE LOCALITY. Discovery Station 7711, haul 32; 52054-7'-52056'5'N; 20°12-6'-20°7-7'W; depth 605-700 m; date 22 May 1971; time 2249-0049 hr; gear RMT 1 . OTHER MATERIAL EXAMINED, (i) Approximately 2500cW, 350099 and 5000JJ (DC Wormley). (ii) 21cfcf, 2099 Discovery Station 7406, haul 44 (SI Washington, USNM 158126). (iii) 49$, 2dtf in Fowler's (1908) material of C. rotundata (BM(NH) 1910.72.114). (iv) 19, Valdivia Station 182, in Muller's (1906fl) material of C. rotundata (ZIZM, Hamburg K- 18937; tentative identification), (v) 499, 4dtf, Gauss Stations 19. 10.0 Id, 12.1 1.01, 4.9.03, 30.9.03b, in Miiller's (1908) material of C. rotundata (ZM, Berlin, 26476). (vi). Ip, U, Dana Stations 3613-8, 3624-4, in Poulsen's (1973) material of Metaconchoecia rotundata (ZM, Copenhagen). DESCRIPTION OF THE MALE. Carapace (Figs 10, 1 1 A-S). In lateral view the ventral margin is almost straight or gently curved and joins the posterior end evenly. The dorsal margin is either straight or in the shape of a very broad V with the apex just behind the second antenna protopodite insertion; it joins the posterior end at a rounded angle. The rostral incisure is usually fairly shallow. In ventral view, the sides of the carapace are only gently curved. The right asymmetric gland opens near the posterodorsal corner. The left gland opens in an anterior position somewhat behind the incisure. There is no surface ornamentation. Frontal organ (Figs 14F, 18A-J). The shaft does not extend beyond the end of the first antenna. In general, the proximal half of the capitulum is expanded, with a strongly convex ventral margin, and the distal part is slightly tapered or parallel sided. However, it is some- what variable and may be narrower, parallel sided, with a strong proximal downflexure (Figs 1 8F, G, I). Stout ventral spines are developed, particularly on the proximal part. First antenna (Figs. 14D, E). The segmentation is fairly distinct. Segment 2 bears fine lateral spines. The a seta extends back parallel to the limb, except proximally where it loops down and is rather expanded. The b seta is slightly shorter than the d seta and bears 5-1 1 closely spaced anterior spines, followed by 5-10 more widely spaced spines with 6-12 spines on the posterior side. The d seta has 8-15 anterior spines, followed by 2-6 more widely spaced spines. The e seta armature comprises 24-30 (mean 26*3, 56 observations) spines which lie at an acute angle to the seta and are paired or less commonly staggered. Second antenna (Figs 14A-C). The protopodite bears a patch of short hairs behind En 1. Ex 1 has a short distal ventral seta and an area of proximal outer hairs. Posteriorly, En 1 bears 3 triangular ridges covered by fine hairs. The processes mamillaris is bulbous with a beak-like extension pointing slightly forwards. The right hook appendage is strongly developed with a long curved distal section and a number of subterminal ridges; the left hook appendage is smaller, with a short straight distal section and no subterminal ridges. The b seta is > twice the length of the a seta, the d seta is slightly shorter than the c seta and the e seta is a short spine. The f seta is rather longer than the g seta and on its anterior side bears 3-1 1 small spines; 2-3 spines are sometimes visible on the g seta. Mandible (Figs 16B-E). The coxale cutting edge has a straight anterior section followed by 1 1-20 (usually 1 1-16) teeth. The distal list has a large pointed posterior tooth, followed by 18-26 small teeth. The proximal list has a large pointed, posterior tooth, 1-5 very small teeth, a second large tooth followed by 18-26 very small teeth, one of which, near the middle, is larger than the remainder; the inner surface of this list is covered with papillae. The cutting edge of the basale has two spine teeth, the posterior one pointed, the anterior one more rounded and both devoid of spines or hairs; these are followed by six serrated teeth of which the most posterior lacks secondary cusps. The anterior inner tooth is triangular with small serrations. Near the cutting edge of the basale there are two short setae inside the posterior margin, a longer seta on the anterior margin and a long median seta. The basale also has a long distal median seta below the endopodite. Ex 1 bears an outer distal plumose seta and two setae on the inner edge, one longer than the other. On Ex 2 there are three outer distal setae, one long, one of medium length and the other short, and two setae on the inner edge which are similar to those on Ex 1 . Ex 3 has three outer setae, two long and claw like and the third short, and four short inner setae. CONCHOECIA SKOGSBERGI SPECIES COMPLEX 147 Maxilla (Fig. 16G). The anterior margin bears three long setae, of which the most proximal is the longest, and a rather shorter seta arising from just inside the margin. The posterior margin has three fairly long setae. The basal seta extends just beyond the end of the limb. There are 5-7 short spines on the bottom of the main segment. Labrum (Fig. 16 A). The hyaline membrane is interrupted by a deep V-shaped notch. On each side of the notch are 1 1-12 flaccid, inward facing teeth. Fifth limb (Fig. 16F). Ex 1 bears seven ventral setae, two of these are posterior (distal), two are median and there are three smaller ones at the anterior end; there is also a long distal dorsal seta and a short lateral distal seta. Ex 2 has a medium sized ventral seta and a slightly longer dorsal seta. Ex 3 bears two fairly long claw-like setae and a shorter ventral seta. Sixth limb (Fig. 14G). Ventrally, segment 1 bears three short posterior (distal) setae, two median setae, one of them short and the other longer and plumose, and two anterior (proximal) setae, one of them plumose. Segment 2 has a minute ventral seta which points out from the limb. Segment 3 has a minute ventral seta and a similar dorsal seta lying parallel to the limb. The terminal setae are fairly short and are armed with hairs only distally. Seventh limb (Fig. 161). The terminal segment bears two setae, one about three times the length of the other. Caudal furca (Fig. 16H). There is an unpaired seta above the smallest claw seta. The claw setae are unusually straight. Penis (Fig. 14H). The end of the penis is obliquely truncated. The terminal part of the vas deferens, which is narrow and tubular, lies free in a depression at the end of the penis; this depression is bounded posteriorly by a distinct lobe. There are 3-5 transverse muscles. DESCRIPTION OF THE FEMALE. Carapace (Figs 12, 13A-O). The lateral outline has a more strongly rounded posterior end, but is otherwise like that of the male. In ventral view, the sides of the carapace are almost straight or only slightly curved. Frontal organ (Figs 15E, 19A-I). The shaft extends well beyond the end of the first antenna and is about twice the combined length of segments 1 and 2. The capitulum is not differentiated from the shaft. It is rather expanded proximally and tapers to a narrowly rounded end. The dorsal surface is sometimes slightly concave near the middle; the ventral surface is usually slightly concave distally. In occasional specimens the end is more pointed and downturned. First antenna (Figs 15C, D). Segment 2 is about twice the length of segment 1 and bears minute scattered lateral spines. The segmentation is fairly distinct. There is no dorsal seta. The a-d setae are almost twice the length of the e seta and are somewhat expanded beyond their basal stalk. The e seta bears 30-37 posterior spines, extending from near the distal end to just above the middle of the seta; the anterior side has 34-45 spines situated rather more proximally. The e seta tapers to a point and is not flattened. Second antenna (Figs 15 A, B). The armature of the protopodite, Ex 1 and En 2 is similar to that of the male, although the protopodite hairs and Ex 1 spines are not always visible. The a seta is about half the length of the b seta and both carry fine hairs. The c (or ?d) seta is minute and often not visible or absent. The f-g setae are rather > half the length of the protopodite and are devoid of armature. Sixth limb (Fig. 15F). Segment 1 bears a plumose distal dorsal seta, four distal ventral setae, the longer two of which are plumose, two median ventral setae, the longer one plumose, and a plumose proximal ventral seta. Segment 2 has a single ventral seta. Segment 3 has a ventral median seta and a rather longer dorsal median seta. The terminal segment bears three claw setae, the median one is the longest and the dorsal seta is rather longer than the ventral seta. DIMENSIONS, rf Carapace length: 1-10-1-28 mm, mean 1-21 ± 0*02 mm (n = 978). 9: !• 12-1 -28 mm, mean 1-21 ± 0'02 mm (n = 560). See Tables 3 & 4 for other morphometric data and Table 9 for left asymmetric gland positions. 148 A. J. GOODAY REMARKS. Judging from the carapace outlines illustrated by Deevey (1968: Fig. 20a, b, 2 1 a, d), her small form of C. skogsbergi from the W. Atlantic (Deevey, 1968; 1974; 1978a)is, at least in part, C.fowleri. However, the length range of Deevey's specimens extends down to I'OO mm. This is well below the lower size limit of C.fowleri and suggests that she included at least one other species of the skogsbergi complex with C.fowleri. One of the specimens figured by Fowler (1909 : 124, fig. 215) was identified by Angel, (1977a: Table 5) as C. rotundata form 1 ( = C.fowleri}. However, it cannot be C.fowleri, and in fact is impossible to assign to any rotundata group species, because it combines an elongate lateral outline with a left asymmetric gland which opens very near the tip of the rostrum. GEOGRAPHICAL DISTRIBUTION. Atlantic Ocean: abundant in the E. Atlantic between 18°N, 25°W and 60°N, 20°W, less common at the equator and 1 1°N, 20°W (Angel & Fasham, 1975; Angel, 1977#; 1979); occasional specimens in Gauss samples from between 6°N and 35°S in the E. Atlantic (Table 2); in the W. Atlantic, fairly common around 32°N, 64°W (Deevey, 1968; Angel, 1979); probably present in the SW Atlantic and off Venezuela (Deevey, 1974; 1978#; as C. skogsbergi). Indian Ocean: single tentatively identified specimen in one Valdivia sample from E. Indian Ocean (Table 1). Pacific Ocean: two specimens in Dana material from the SW Pacific. VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 400 m-1250 m; at 60°N, 20°W and 53°N, 20°W it is most common between 400 m and 700 m, further south around 1 8°N, 25°W and 1 1 °N, 20°W it is most abundant at rather greater depths. Conchoeciafowleri form A (FigsllT-Z, 13P-X, 17) Conchoeciafowleri form 13. — Angel & Fasham, 1975 : 737. MATERIAL. 2699, 19cfcf, Discovery GATE Stations; 9299, 38 15% of length behind tip of rostrum, cf Al e seta only slightly longer than b and d setae and relatively shorter than in other species of skogsbergi complex. TYPE MATERIAL. Holotype: dissected d (BM(NH) 1979.693). Paratypes: 1 dissected 9 (BM(NH) 1979.694); 149$, 1 3dtf (BM(NH) 1979.785-794). TYPE LOCALITY. Discovery Station 7711, haul 32; 52°54-7'-52°56-5'N, 20°12-6'-2007-7'W; depth 605-700 m; date 22 May 1971, time 2249-0049 hr; gear RMT 1 . OTHER MATERIAL EXAMINED, (i) Approximately 380099, 1500cfcf and 2900JJ (DC, Wormley). (ii) 1099, lOcfcf from Discovery Station 771 1, haul 13 (SI, Washington, USNM 158132). (iii) U Dana Station 3624-7, in Poulsen's (1973) material of Metaconchoecia rotundata (ZM, Copenhagen, tentative identification), (iv) 299, Id1 in Fowler's (1909) material of C. rotundata (BM(NH) 19 10.72. 116). SUPPLEMENTARY DESCRIPTION. Male. Carapace (Fig. 20). The carapace is only slightly tapered in lateral view. Frontal organ (Fig. 22B). The capitulum is similar in shape to that of C. fowleri but the ventral spines are rather smaller and more numerous. Female Carapace (Fig. 21). The lateral outline is like that of the male but rather less elongate; in ventral view the carapace is almost parallel sided. Frontal organ (Fig. 22H). The capitulum is less strongly tapered than in C. fowleri and has a more bluntly rounded end. DIMENSIONS, cf Carapace length: G'97-1'14 mm, mean 1'05 ±0*02 mm (n = 664). 9: Q'95-1'18 mm, mean 1-07 ± 0'02 mm (n = 919). See Tables 3 & 4 for other morphometric data and Table 9 for left asymmetric gland positions. REMARKS. The lateral outline of C. discoveryi tends to be slightly less tapered than that of C fowleri and the posterior end is rather more clearly upturned. The carapace is also somewhat shorter. C. discoveryi is distinguished from all other species of the skogsbergi complex, including C. fowleri, by the relatively posterior position of the left asymmetric gland and the relative shortness of the male first antenna e seta. GEOGRAPHICAL DISTRIBUTION. Atlantic Ocean: abundant in the E. Atlantic between 40°N, 20°W and 60°N, 20°W, less common between 40°N, 20°W and the equator (Angel & Fasham, 1975; Angel, 1979); common in the W. Atlantic around 32°N, 64°W (Angel, 1979). Pacific Ocean: single tentatively identified juvenile in Dana material from the SW Pacific. VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 600 m-1500 m; it is most abundant between 900m and 1500m and particularly between 1000m and 1250m. At 53°N 20°W the females are most abundant at 1000-1500 m and the males and juveniles at 600-800 m. A similar separation occurs around 32°N 64°W (Angel, 1979). Conchoecia obtusa sp. nov. (Figs 23-25) ?Metaconchoecia rotundata (Miiller, 1890).— James, 1975 : 1 14-118, p. XXI, figs 1, m, pi. XXII, figs a-1. Conchoecia rotundata form 2, Angel & Fasham, 1975:737 (distribution).— Angel, 19770:246 (vertical distribution). — Angel, 1979 : 69-70 (vertical distribution). ETYMOLOGY. L. obtusus, blunt: referring to the fact that the lateral carapace outline is not strongly tapered and the anteroventral part of each valve is rounded or bluntly pointed in ventral view. DIAGNOSIS. Lateral carapace outline usually only slightly tapered in anterior 1/2-2/3. Ventral outline not strong biconvex, in cf B >41% of length; anteroventral part of each valve not sharply pointed in ventral view. LAG lies < 12% of length behind tip of rostrum, cf Al e seta with 23-28 spines. Posterior end gently, usually symmetrically rounded. Length = 0-9 1-1 -06 mm. 150 A.J.GOODAY TYPE MATERIAL. Holotype: dissected rf (BM(NH) 1979.700). Paratypes: 1 dissected 9 (BM(NH) 1979.701); 41 99, 42cTcT(BM(NH) 1979.726-735). TYPE LOCALITY. Discovery Station 7856, haul 2; 29°58-l'-29°53-6'N, 23°09'-23'T8'W; depth 405-505 m; date 3 1 March 1972; time 09 10-1 1 10 hr; gear RMT 1. OTHER MATERIAL EXAMINED, (i) Approximately 190099, 1400cfcf and 1650JJ (DC, Wormley). (ii) 2099, 22cW, from Discovery Station 7856, haul 8 (SI, Washington USNM 158128). (iii) 19 Dana Station 3583-1 , (ZM, Copenhagen). SUPPLEMENTARY DESCRIPTION. Male. Second antenna (Fig. 25D). The left hook appendage is rather more strongly developed than in C. fowleri whereas the right hook is less well developed. Female. Frontal organ (Fig. 25H). The end of the capitulum is more pointed and downturned than in C. fowleri and the distal part of the ventral margin is clearly concave. DIMENSIONS, d Carapace length: 0'9 1-1 -00 mm, mean 0*96 ± 0*02 mm (n = 626). 9: 0-91-1-06 mm, mean 0'97 ± 0*04 mm (n = 672). See Tables 3 and 4 for other morphometric data and Table 9 for left asymmetric gland positions. REMARKS. C. obtusa is consistently smaller than C. fowleri and has a rather less tapered, more rectangular outline. In the male, the ventral outline is relatively broader. It is distinguished from C. discoveryi by the more symmetrically rounded posterior end, the relatively broader ventral outline of the male and the clearly more anterior position of the left asymmetric gland. Judging from the carapace length and outline, C. rotundata of James (1975) may be this species. GEOGRAPHICAL DISTRIBUTION. Atlantic Ocean: in the E. Atlantic, abundant at 30°N, 23°W, common at 40°N, 20°W, uncommon at 1 1°N, 20°W and 18°N, 25°W, rare at the equator, 53°N, 20°W and 60°N, 20°W (Angel & Fasham, 1975; Angel, 1979) and uncommon around 44'N, 13°W (Angel, 1977a); in the W. Atlantic, abundant at 33°N, 64°W (Angel, 1979). Indian Ocean: possibly occurs in the NE Indian Ocean (as C. rotundata of James, 1975). Pacific Ocean: single female in Dana material from the SW Pacific. VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 100 m-800 m; it is most abundant between 400 m and 700 m. Conchoecia skogsbergi lies, 1953 (Figs 18K-P, 19J-0, 26-29) Conchoecia rotundata Miiller, 1890. — Miiller, 1906fl : 83-84 (in part, long form only), pi. XVII, figs 25-33 (not figs 23, 24 = C teretivalvata lies, 1953).— Miiller, 1908 : 69, 70 (in part). Miiller 1912 : 77 (long form only). — Skogsberg, 1920:649-658, Fig. CXXII (designated type description by lies, 1953).— Hillman, 1967:200 (mentioned only).— Hillman, 1968:158 (listed).— Hillman, 1969: Map 9 (distribution). Conchoecia rotundata Miiller, 1890, form 10, Angel & Fasham, 1975 : 737 (distribution). Not Conchoecia rotundata Miiller, 1 890. Conchoecia skogsbergi lies, 1953.— Leung, 1972 : 31-32.— Leung, 1973 : 10-1 1.— Deevey, 1974 : 364 (in part, 9 >l'40mm, cT >l-35 mm only). — Chavtur & Shornikov, 1974:286 (mentioned). — Deevey 19786: 54, 55, Fig. 10. ?Conchoecia skogsbergi lies, 1953. — Angel, 1968a: 1-6, figs 1-10. Not Conchoecia skogsbergi lies, 1953 ( = C. subinflata sp. nov.). — Angel 19686 ( = C. fowleri sp. nov.).— Deevey 1968 : 54-55, Figs 20a-d, 2 la, d, f-h, 22a ( = C. fowleri). ?Metaconchoecia skogsbergi (lies, 1953).— Poulsen, 1973: 73-74, Fig. 35.— Chavtur, 1976: 105-106.— Chavtur 1977a : 30 (listed).— Chavtur. 19776 : 145-146. Conchoecia (Metaconchoecia) skogsbergi lies, 1953. — Deevey, 19786 : 54-55, Fig. 10. CONCHOECIA SKOGSBERGI SPECIES COMPLEX 151 DIAGNOSIS. Lateral carapace outline tapered in anterior 2/3-3/4 and relatively higher than in C. fowleri, posterior end approximately symmetrically rounded. Ventral outline rather weakly biconvex, in cf >42'5% of length. LAG lies 1 1%-15% of length behind tip of rostrum. Length > 1'30 mm. TYPE MATERIAL. Holotype: cf (NR Stockholm reg. no. 3101). Paratypes: 2rfcT, 499, 2JJ (NR Stockholm reg. no. 3101). The type material is part of the collection on which Skogsberg (1920:657) based the description of C. rotundata that was later designated the type description of C. skogsbergi by lies (1953). Neither the holotype nor the paratypes correspond obviously to Skogsberg's (1920 : Figs CXXII, 1 and 2) figured carapaces, which were, however, also from the type locality. TYPE LOCALITY. Station 64b of the Swedish 'Antarctic' Expedition of 1901-1903: 48°27'S, 42°36'W; depth 2500-0 m, date 23 June 1902. OTHER MATERIAL EXAMINED, (i) 299, U, S.A.E. Station 70b in Skogsberg's ( 1 920) material of C. rotundata (NR, Stockholm, 238). (ii) 3699, 21cTcT 12JJ in Muller's (1908) Gauss material of C. rotundata (Table 2 for station details; ZM, Berlin 26467, 26479, BM(NH) 1924.7.19.184-187). (iii) Specimens collected by Discovery II: 299, 2JJ (Discovery Station 1773), 19 (1775), 3c7c? (1776), 2$9, 4dtf, 5JJ (1777), 599, \<*, 7JJ (1778), 499, !42% of length; anteroventral parts of each valve not sharply pointed in ventral view. LAG lies < 12-0% of length behind tip of rostrum, cf Al e seta with 24-28 spines. Posterior end symmetrically rounded or somewhat downturned. Length = 1*06-1 '26 mm. TYPE MATERIAL. Holotype: dissected d (BM(NH) 1979.704). Paratypes: 1 dissected 9 (BM(NH) 1979.705), 2l9$, 20dtf(BM(NH) 1979.805-814). TYPE LOCALITY. Discovery Station 6665, haul 4; 10°32-7'N, 19°57-4'W; depth 400-295 m; date 22 January 1968; time 1559-1 731 hr;gearNl 13 CDB. OTHER MATERIAL EXAMINED, (i) Approximately 32099, 200^ and 37JJ (DC, Wormley). (ii) 1999, 17rfcT from Discovery Station 6665, haul 8 (SI, Washington, USNM 158127). (iii) 19, Dana Station 3583-1 in Poulsen's (1973) material of Metaconchoecia rotundata (ZM, Copenhagen, tentative identification). CONCHOECIA SKOGSBERGI SPECIES COMPLEX 153 DIMENSIONS, rf Carapace length: 1 -06-1 -20 mm, mean 1-13 ± O02 mm (n = 405). 9: 1-10-1-26 mm, mean 1-17 ± 0-03 mm (n = 521). See Tables 3 and 4 for other morphometric data and Table 9 for left asymmetric gland positions. REMARKS. This species is similar in shape to C. fowleri and C. discoveryi but is broader in ventral view with a somewhat downturned, rather than upturned, posterior end, and a more anteriorly situated left asymmetric gland. C. wolferi is consistently larger than C. obtusa, the lateral outline is relatively more elongate and the posterior end tends to be more downturned. GEOGRAPHICAL DISTRIBUTION. Atlantic Ocean: fairly common around 1 1°N, 20°W and 18°N, 25°W in E. Atlantic and on the equator. Pacific Ocean, single tentatively identified female in Dana material from SW Pacific. VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 300-800 m. At 1 1°N, 20°W it is most abundant betweeen 300 m and 400 m, at 18°N, 25°W it is most abundant at depths of 400 m to 600 m. Conchoecia acuta sp. nov. (Figs 32-34) Conchoecia rotundata Miiller, 1890.— Muller, 1908 : 69, 70 (in part). Conchoecia rotundata forms 4, 12 Angel & Fasham 1975 : 737 (distribution).— Angel, 1979 : 71, fig. 59 (vertical distribution). ETYMOLOGY. L. acutus, pointed: referring to the rather strongly tapered lateral outline and the sharply pointed shape of the anteroventral region when viewed ventrally. DIAGNOSIS. Lateral outline relatively higher than in C. fowleri and clearly tapered in anterior 2/3 to 3/4. Posterior end approximately symmetrically rounded. Ventral outline weakly biconvex in 9 (B = 35-46% of length), more strongly biconvex in d (B = 42-5-50% of length); anteroventral part of each valve pointed in ventral view. LAG lies < 1 1-5% of length behind tip of rostrum. rfAl e seta with 20-24 spines. Length = 0-85-1-00 mm. rfA2 f seta <38%, g seta <42%, 9 A2 LSS <41%, f-j seta <20% of length. TYPE MATERIAL. Holotype: dissected rf (BM(NH) 1979.690). Paratypes: 1 dissected 9 (BM(NH) 1979.691); 1 dissected cf (BM(NH) 1979.692); 20dy, 24$9 (BM(NH) 1979.736-745). TYPE LOCALITY. Discovery Station 7089, haul 19; 17°48'N, 25°22'W; depth 197-1 12 m; date 1 5 November 1 969; time 1 307-1 537 hr; gear RMT 1 . OTHER MATERIAL EXAMINED, (i) Approximately 95099, 550dtf, 350JJ (DC, Wormley). (ii) 1699, 16dy from Discovery Station 7856, haul 22 (SI, Washington USNM 158131). (Hi) Specimens in Miiller's (1908) material of C rotundata: 8199, 29dtf, 10JJ (Gauss Station 19.10.01d); 2699, 13dtf (19.10.01c); 1099, 7dtf (26.10.01); 19, 2dd (5.1 I.Ola); 399, 2rfrf (12.11.01); ?l9 (18.5.03) (ZM, Berlin, 26465, 26474). (iv) 599, 3dtf in Muller's (1906a) material of C. rotundata from Valdivia Station 26 (ZM, Berlin, 16483, all tentative identifications). SUPPLEMENTARY DESCRIPTION. Male. Second antenna (Figs 34D, E). The hook appendages are similar to those of C. fowleri but more nearly equal in size. Female. Frontal organ (Fig. 34H). The capitulum is similar in shape to that of C. fowleri but has a rather more pointed and downturned tip. DIMENSIONS. cT Carapace length: Q-85-0'99 mm, mean 0'9 1± 0*04 mm (n = 457). 9: 0-85-1-01 mm, mean 0'93 ± O'Ol mm (n = 723). See Tables 11 and 12 for other morphometric data and Table 9 for left asymmetric gland positions. 154 A.J.GOODAY REMARKS. C. acuta resembles Miiller's (1890, pi. XXVIII, fig. 42) original C. rotundata in lateral view but is more elongate and also considerably smaller. Compared with C. obtusa, the lateral outline is more tapered and the ventral outline is more convex with the antero- ventral part of each valve being pointed rather than rounded and the rostrum more produced. C. acuta is consistently smaller than C. fowleri and C. wolferi and the left asymmetric gland has a clearly more anterior position than in C. discovery!. GEOGRAPHICAL DISTRIBUTION. North Atlantic Ocean: common in the E. Atlantic between the equator and 30°N, 23°W (Angel & Fasham, 1975; this paper), but its reported occurrence at 40°N, 20°W has not been confirmed; common in the W. Atlantic around 32°N, 64°W (Angel, 1979; this paper). South Atlantic Ocean: fairly common in Gauss material from 19°S to 35°S in the E. Atlantic. Indian Ocean: 19 from the SE Indian Ocean in Gauss material (tentative identification). VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 50-500 m; it occurs mainly in the top 300 m, particularly between 100 m and 200 m. Conchoecia aff. acuta (Figs 35, 36) Conchoecia rotundata Miiller, 1890.— Miiller, 1908 : 69, 70 (in part). Metaconchoecia rotundata (Miiller, 1890). — Poulsen, 1973 : 71-72; Figs 34a-j. MATERIAL EXAMINED. Broad form. Dana Stations 3587-6 (Id1), 3611-2 (Id, 1J), 3613-3 (2599), 3613-8 (399, 4dd), 3613-9 (399, 2dd), 3613-10 (1799, 14dd, 5JJ), 3623-2 (19, Id1), 3623-5 (2999, lOdd, 4J), 3624-2 (16$9, 7dd), 3624-3 (299, Id), 3624-4 (899, 2dd), 3624-5 (2799, 1 1 50% of length, d A2 LSS >60% and 9 A2 LSS >47'5% oflength. 155 TYPE MATERIAL. Holotype: dissected d (BM(NH) 1980.141). Paratypes: 1 dissected, 9, 3 dissected cfc? (BM(NH) !980.142-145);499,646% (9) >50% (d) of length, biconvex with strongly curved sides in d1, curved or almost straight sides in 9. TYPE MATERIAL. Holotype: dissected d (BM(NH) 1979.698). Paratypes: 1 dissected 9 (BM(NH) 1979.699); 899, 10d'd1(BM(NH) 1979.716-725). TYPE LOCALITY. Discovery Station 6665, haul 24; 10'31'3'N, 19°58'W; depth 1490-1260 m; date 25 February 1 968; time 0404-0835 hr; gear Nl 1 3 CDB. OTHER MATERIAL EXAMINED, (i) Approximately 100099, 850rfcf, 1250JJ (DC, Wormley). (ii) 1099, 12d-d- from Discovery Station 7856, haul 15, (SI, Washington, USNM 158129). (iii) 399, 156 A. J. GOODAY U, Dana Station 3587-6, in Poulsen's (1973) material of Metaconchoecia rotundata (ZM, Copenhagen), (iv) Specimens in Miiller's (1908) material of C. rotundata (ZM, Berlin 26480): 2d<5 (Gauss Station 19.10.01d), 299, 4rf(12.1 1.01), 599, H (4.9.03), 2$9 (30.9.03b).(v) 299, Id" Valdivia Station 26, in Miiller's (1906a) material (ZM, Berlin 26481, tentative identifications). SUPPLEMENTARY DESCRIPTION. Male. Frontal organ (Fig. 4 1 B). The capitulum is similar in shape to that ofC.fowleri but is relatively higher. Second antenna (Figs 4 ID, E). The right hook appendage is only slightly larger than the left hook. Penis (Fig. 42). The number of transverse muscles is unusually variable, ranging from two to six. Female. Frontal organ (Fig. 41H). The capitulum is relatively higher than in C. fowleri, the end is pointed and downturned and the ventral margin distally concave. DIMENSIONS, rf Carapace length: O'99-l- 16 mm, mean l-08±0'02mm (n = 493). 9: 0-97-1-16 mm, mean 1*07 ± 0*03 mm (n = 699). See Tables 1 1 and 12 for other morpho- metric data and Table 9 for left asymmetric gland positions. REMARKS. C. inflata is comparable in size to Miiller's (1890) original C. rotundata. The lateral outline is also similar although relatively lower and more elongate. However, as discussed above, the initial description of C. rotundata was inadequate and a proper comparison with other species such as C. inflata is now impossible. Martens (1979 : 351) records a form similar to C. inflata, possibly a subspecies, from the SE Pacific. The broad ventral outline, which is particularly obvious in the female, distinguishes C. inflata from other skogsbergi complex species. The very gently curved, almost vertical posterior end is also characteristic. GEOGRAPHICAL DISTRIBUTION. North Atlantic Ocean: in the E. Atlantic, abundant at 30°N, 23°W, fairly uncommon at 11°N, 20°W, 18°N, 25°W and 40°N, 20°W, rare at 53°N, 20°W (Angel & Fasham, 1975; Angel, 1979; this paper) and uncommon at 44°N, 13°W (Angel, 1977a); in the W. Atlantic, uncommon at 32°N, 64°W (Angel, 1979; this paper). South Atlantic Ocean: occurs in Gauss material from 12°S to 35°S in the E. Atlantic. Pacific Ocean: a few specimens in Dana material from the SW Pacific; a related form, possibly a subspecies, occurs off the Chilean coast (Martens, 1979). VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 200-1 500m; it occurs mainly between 200 m and 500 m, but is usually rather deeper (500-600 m) at 32°N, 64°W. Conchoecia rotundata Miiller, 1890 (Figs 43, 44) Conchoecia rotundata Miiller, 1890:275, pi. XXVIII, figs 41^13, pi. XXIX, fig. 44.— Miiller, 1908 : 69-70 (in part).— Deevey, 1968 : 51-54 (in part: $ = 0-77-1-00 mm; rf = 0'75-0'97 mm), Figs 20e-j, 2 1 b, c, i, j, k (not 2 1 e, too large: cf L = 0'97 mm), 22c-e (not b, too large: c? L = 0-97 mm). Not Conchoecia rotundata Miiller, 1890.— Hillman, 1967:200 ( = C. skogsbergi lies, 1953).— Hillman, 1968:158 ( = C. skogsbergi).— Hillman, 1969: Map 9 ( = C. skogsbergi).— Deevey, 1970 : 810 (too large: 9 = 0-85-1-1 5 mm, rf = 0-85-l'10 mm).— Poulsen, 1973 : 71-72, text-figs 34a-j ( = C. aft", acuta).— Deevey, 1974:364, Fig. 5h (in part) (too large: 9 = 0'87-0'95 mm, rf = 0-80-0-85 mm).— Williams, 1975 : 225, 227, text-fig. 8a( = C. teretivalvata lies, 1953).— Deevey, 1 978a: 70 (too large: 40%, g seta > 43%, 9 A2 LSS > 4 1 %, f-j setae > 20% of length. TYPE MATERIAL. Holotype: dissected d (BM(NH) 1979.702). Paratypes: 1 dissected 9 (BM(NH) 1979.703); 4499, 47dd,(BM(NH) 1979.815-824). TYPE LOCALITY. Discovery Station 7709, haul 27; 60°8-7'-60°l 1'0'N, 19°59-9'-19°51'5'W; depth 500-600 m; date 28 April 1971;time 1751-1951 hr;gearRMT 1. OTHER MATERIAL EXAMINED, (i) Approximately 380099, 2400dd, 4600JJ (DC, Wormley). (ii) 4499, 47dd, Discovery Station 7709, haul 23 (SI, Washington, USNM 158130). (iii) 699, 2dd, 5JJ, William Scoresby Station 977, in lies (1953) material of C. skogsbergi (DC, Wormley). (iv) 2dd in Fowler's (1909) material of C. rotundata, redetermined by lies as C. skogsbergi (BM(NH) 1910.72.117). (v) Specimens in Miiller's (1908) material of C. rotundata (ZM, Berlin, 26478, 26466): Id (Gauss Station 19.10.01c), 1699, 3dd (19.10.01d), 19 (5.1 I.Ola), 399, 3dd (12.1 1.01), 19 (2.5.03, tentative identification), (vi) 1699, 6dd, Valdivia Station 121d in Miiller's (1906a) material of C. rotundata (ZM, Berlin 26482; tentative identifications), (vii) 19, Dana Station 3587-6, in Poulsen's (1973) material of M. rotundata (ZM, Copenhagen), (viii) Id, Valdivia Station 182, in Miiller's (1906a) material (ZIZM Hamburg K-18937, tentative identification). SUPPLEMENTARY DESCRIPTION. Male. Carapace (Fig. 45). In ventral view the carapace may be slightly constricted behind the second antenna insertion. Female. Carapace. In an occasional variant (Figs. 46U, CC, DD) the posterior half of the carapace is variably inflated, the inflation being visible only with the animal on its back. Frontal organ (Fig. 47H). The end of the capitulum is pointed and downturned and the distal part of the ventral surface is concave. DIMENSIONS, d Carapace length: 0-95-1-20 mm, mean 1-04 ±0-02 mm (n = 880). 9: 0-95-1-18 mm, mean 1-05 ± 0-03 mm (n= 1508). See Tables 1 1 and 12 for other morpho- metric data and Table 9 for left asymmetric gland positions. REMARKS. This species has been identified in some of Fowler's (1909) C. rotundata material from the Bay of Biscay and Fowler's (1909: fig. 205) figured 'Stage I' male has a closely similar lateral outline. On the other hand, the female carapace outline (Fowler, 1909: fig. 2 1 5) is quite different and cannot be assigned to any known skogsbergi complex species. The specimens from the Benguela Current which lies (1953) placed in C. skogsbergi have been reexamined and identified as C. subinjlata. A single, rather damaged male in Poulsen's (1973) material of Metaconchoecia rotundata may belong to C. subinjlata although the left asymmetric gland is much less prominent and situated further back than in typical specimens. C. subinjlata is usually longer than C. acuta but the size ranges overlap and specimens of similar size are difficult to separate. C. subinjlata has a rather less strongly curved posterior margin but can only be reliably distinguished from C. acuta by differences in the relative lengths of various seta on the second antenna. Compared with C. obtusa, this species usually CONCHOECIA SKOGSBERGI SPECIES COMPLEX 159 has a more clearly tapered lateral outline, the posterior end of the male is less strongly curved and the anteroventral region of each valve is pointed in ventral view rather than being bluntly pointed or rounded. In addition, C. subinflata always has more male first antenna e seta spines than C, obtusa and there are various meristic differences. C. subinflata has a relatively lower lateral outline than C. australis and differs in a number of meristic characters. It is similar in lateral outline to C. inflata but narrower in ventral view, particularly in the female. C. subinflata is consistently larger than C. rotundata and the lateral outline is less elongate. The left asymmetric gland is situated more anteriorly than in C.fowleri and C. discoveryi. The lateral outline is less elongate and more tapered than that of C. wolferi. GEOGRAPHICAL DISTRIBUTION. North Atlantic Ocean: In the E. Atlantic abundant at 60°N, 20°W and 53°N, 20°W, fairly common at 40°N, 20°W, 11°N, 20°W and on the equator, uncommon at 1 8°N, 25°W and 30°N, 1 3°W (Angel & Fasham, 1975; Angel, 1 979; this paper) and common at 44"N, 13°W (Angel, 1977a); rare at 32°N, 64°W in the W. Atlantic (Angel 1979; this paper). South Atlantic: occurs between 19°S and 35°S in the SE Atlantic in Valdivia and Gauss material (Tables 1 & 2); off coast of Namibia in William Scoresby material. Indian Ocean: Irf from NE Indian Ocean in Valdivia material (Table 1; tentative identification); 19 from S. Indian Ocean in Gauss material (Table 2; tentative identification). Pacific Ocean: 19 in Dana material from SW Pacific (tentative identification). VERTICAL DISTRIBUTION. The overall range in the N. Atlantic is 200-900 m. It is most abundant between 200 m and 400 m. Species relationships within the rotundata group The 19 or 20 species of the rotundata group now described fall into three more or less coherent assemblages. (i) C. kyrtophora and C. nasotuberculata: the carapace is short and tapered, relatively high and rounded in lateral view with the left asymmetric gland near the tip of the rostrum or just behind the rostrum; viewed ventrally, the carapace is constricted behind the second antenna insertion. These two species are closely related and have frequently been confused but are now known to be quite distinct (Angel, in press). A third species, C. teretivalvata, has a similar lateral outline but it lacks the constriction and the shape of the frontal organ capitulum suggests that it may be more closely related to the skogsbergi complex. (ii) C. arcuata, C. bathyrotundata Chavtur, 1977 (a possible synonym of C. arcuata), C. isochiera, C. macromma, C. pusilla and C. glandulosa: the ventral margin is arcuate (strongly in C. isochiera and C. arcuata); except in C. arcuata, the right asymmetric gland is displaced some distance down the posterior margin and opens at the end of a triangular process, the left asymmetric gland usually opens in a relatively posterior position, some distance behind the rostrum (particularly in C. glandulosa, C. macromma and, to a lesser extent, C. arcuata). Within this assemblage, C. arcuata and C. isochiera are probably closely related while C. glandulosa, a large species with a distinctive outline, is probably rather distantly related to the remaining species. (iii) The ten species of the skogsbergi complex: viewed laterally, the carapace is variably tapered, more elongate than in assemblage (i) species and differing from assemblage (ii) species in having an almost straight or slightly curved ventral margin; the right asymmetric gland always opens near the posterior dorsal corner; the left asymmetric gland position is rather variable but it generally opens near the tip of the rostrum. The rare deep water species C. abyssalis, the only described member of the rotundata group not represented in the Discovery collections, may also belong to the skogsbergi complex and C. teretivalvata should possibly be placed here rather than in assemblage (i). Within the complex there are two groupings. The first comprises C. fowleri, C. discoveryi, C. obtusa, C. skogsbergi and C. 160 A. J.GOODAY wolferi in which the carapace is more elongate in lateral view, less strongly biconvex in ventral view with the anteroventral part of each valve being bluntly pointed. The second group comprises C. acuta, C. australis, C. injlata, C. rotundata and C. subinflata all of which have relatively shorter, more clearly tapered lateral outlines, more strongly biconvex ventral outlines with the anteroventral part of each valve clearly pointed. Acknowledgements I am grateful to Mr P. M. David and Dr M. V. Angel for critically reading drafts of this paper. Special thanks go to Dr Angel who introduced me to the project, did much of the initial identification work and patiently awaited its much postponed completion. I am also grateful to Dr E. J. lies for his help during the earlier stages of this study, Dr A. L. Rice for his forebearance while ostracods diverted me from other duties and to Mrs P. Talbot who patiently typed the long and difficult manuscript. Finally, I acknowledge with thanks the following people who have kindly made available museum material: Drs G. A. Boxshall, BM(NH); I. Groth, E.-M.-Arndt-Universitat, Greifswald; H.-E. Gruner, ZM, Berlin; J. Martens, ZIZM, Hamburg; R. Olerod, NR. Stockholm; T. Wolff, ZM, Copenhagen. This study was supported by a NERC Research Fellowship. References Angel, M. V. 1968a. Conchoecia skogsbergi (lies), a halocyprid ostracod new to the Norwegian Sea. Sarsia33 : 1-6. 1 9686. The thermocline as an ecological boundary. Sarsia 34 : 299-3 12. 1969. Planktonic ostracods from the Canary Island region; their depth distributions, diurnal migrations and community organisation. J. mar. biol. Ass. U.K. 49 : 5 1 5-553. 1972. Planktonic oceanic ostracods — historical, present and future. Proc. R. Soc. Edinb. (B) 73 (22): 2 13-228. 1973. Conchoecia from the North Atlantic the 'procera' group. Bull Br. Mus. nat. Hist. (Zool.) 21 (7): 259-283. 1977 a. Studies on Atlantic halocyprid ostracods: vertical distributions of the species in the top 1000 m in the vicinity of 44°N, 1 3°W. J. mar. biol. Ass. U.K. 57 (1) : 239-252. 1977/7. Some speculation on the significance of carapace length in planktonic halocyprid ostracods. In Loftier, H. and Danielpol, D. (eds). Aspects of ecology and zoogeography of Recent and fossil Ostracoda. Junk, The Hague: 45-54. — 1979. Studies on Atlantic halocyprid ostracods: their vertical distributions and community structure in the central gyre region along latitude 30°N from off Africa to Bermuda. Prog. Oceanogr. 8:3-124. In press. Conchoecia nasotuberculata Miiller and C. kyrtophora Miiller two planktonic halocyprid ostracods that have frequently been confused. Angel, M. V. & Fasham, M. J. R. 1975. Analysis of the vertical and geographical distribution of the abundant species of planktonic ostracods in the north-east Atlantic. J. mar. biol. Ass. U.K. 55 : 709-737. Athersuch, J. 1976. TheG. W. M tiller ostracod collection. Pubbl. staz. Zool. Napoli4Q : 344-348. Badcock, J. & Merrett, N. R. 1976. Midwater fishes in the eastern North Atlantic — 1. Vertical distribution and associated biology in 30°N, 23°W, with developmental notes on certain myctophids. Prog. Oceanogr. 7 : 3-58. Baker, A. de C., Clarke, M. R. & Harris, M. 1973. The N.I.O. combination net RMT (1+8) and further developments of rectangular midwater trawls. J. mar. biol. Ass. U.K. 53 (1) : 167-184. Chavtur, V. G. 1976. The fauna of pelagic ostracods of the far eastern seas. Sb. Rab. Inst. Biol. morya, Vladivostok 6: 99-109. - \917a. Fauna of the pelagic ostracods of Halocypridae family from the northwestern Pacific. Biologiya morya 1 : 28-38. 19776. The composition and vertical distribution of pelagic ostracods in the Kurilo- Kamchatka trench region. Trudy Inst. Okeanol. 108 : 136-164. 161 Chavtur, V. G. & Shornikov, E. I. 1974. Planktonic ostracoda of the Bering Sea. Zool Zh. 53:285-288. Deevey, G. B. 1968. Pelagic ostracods of the Sargasso Sea off Bermuda. Bull. Peabody Mus. nat. Hist. 26: 1-125. 1970. Pelagic ostracods (Myodocopa, Halocyprididae) from the North Atlantic off Barbados. Proc. biol. Soc. Wash. 82 (62) : 799-824. 1974. Pelagic ostracods collected on Hudson 70 between the equator and 55°S in the Atlantic. Proc. biol. Soc. Wash. 87 (2) : 35 1-380. 1978a. The planktonic ostracods of the Cariaco Trench and adjacent waters. Proc. biol. Soc. Wash. 91(1): 52-73. 1978&. A taxonomic and distributional study of the planktonic ostracods collected on three cruises of the Eltanin in the South Pacific and the Atlantic regions of the South Pacific. In Pawson, D. L. & Kornicker, L. S. (eds). Biology of the Antarctic Seas. VIII. Antarct. Res. Ser. 28 : 43-70. Fowler, G. H. 1909. Biscayan Plankton. XII. The Ostracoda. Trans. Linn. Soc. Lond. (Zool.), Ser. 2, 10: 2 19-358, pis 16-27. Foxton, P. 1963. An automatic opening-closing device for large midwater plankton nets and midwater trawls. J. mar. biol. Ass. U.K. 49 : 603-620. George, J. 1969. A preliminary report on the distribution and abundance of planktonic ostracods in the Indian Ocean. Bull. natn. Inst. Sci. India 38 : 641-648. George, J., Purushan, K. S. & Madhupratap, M. 1975. Distribution of planktonic ostracods along the south-west coast of India. Indian J. mar. Sci. 4 (2) : 201-202. Gooday, A. J. 1976. The taxonomy of Conchoecia (Ostracoda, Halocyprididae) of the gaussi and edentata groups from the north east Atlantic with a note on their ecology. Bull. Br. Mus. nat. Hist. (Zool.), 30 (3): 57-1 00. Granata, L. & Caporiacco, L. di 1949. Ostracodes marins recuillis pendant les croisieres du Prince Albert lerde Monaco. Result. Camp, scient. Prince Albert I, 109 : 1-51, pis 1-4. Hillman, N. S. 1967. Ecology of Antarctic pelagic Ostracoda. Antarct. J. U.S. 2 (5) : 199-200. 1 968. Studies of Antarctic pelagic Ostracoda. A ntarct.J. (7.5.3(5): 157-158. 1969. Halocyprididae. In Hedgpeth, J. W. (ed.). Distribution of selected groups of marine invertebrates in waters south of 35°S latitude. Antarct. Map Folio Ser. 11 : 29-30, pis 1 5-16. lies, E. J. 1953. A preliminary report on the Ostracoda of the Benguela Current. Discoverv Rep. 57 : 259-280. James, C. M. 1975. Studies on the ostracoda of the Indian seas. Unpublished dissertation. University of Kerala, India. 281 pp. Leung, Y. M. 1972. Ostracods of the central Arctic. In Kobayashi, H. (ed.) Taxonomic Guides to Arctic Zooplankton (VI). University of Southern California Department of Biological Sciences Technical Report No 2: 29^0. 1973. On the ostracod fauna of the Arctic Basin. In Fernandes, H. R. (ed.) Final Report on Drift Station Biology: Zooplankton Taxonomy and Sorting Programs. University of Southern California Department of Biological Sciences: 10-1 i . Martens, J. M. 1978. Die pelagischen Ostracoden der Expedition Marchile I (Siidost-Pazifik) als Indikatoren fur Wasserkorper: Systematik, Verbretung und Zoogeographie. Unpublished disserta- tion. University of Hamburg, 198 pp. 1979. Die pelagischen Ostracoden der Expedition Marchile I (Siidost-Pazifik), II. Systematik und Vorkommen (Crustacea: Ostracoda: Myodocopida). Mitt. hamb. zool. Mus. Inst. 76 : 303-366. Miiller, G. W. 1890. Ueber Halocypriden. Zool. Jb. 5 (2) : 253-280, pis. XXVIII-XIX. 1894. Die Ostracoden des Golfes von Neapel und der angrenzenden Meeres-Abschnitte. Fauna Flora Golf. Neapel Monographic 21 : 1-104. \906a. Ostracoda. Wiss. Ergebn. dt. Tiefsee-Exped. 'Valdivia', 8 : 29-154, pis 5-35. 19066. Die Ostracoden der Siboga-Expedition. Siboga-Exped. 30 : 1^40, pis I-IX. 1908. Die Ostracoden der Deutschen Siidpolar Expedition 1901-1903. Sudpolar Exped. 10: 51-182, pis 4-19. 1912. Ostracoda. Das Tierreich, 31 : 1-434. Poulsen, E. M. 1973. Ostracoda-Myodocopa, 3b. Halocypriformes — Halocypridae, Conchoecinae. Dana Rep. 84: 1-223. Rudyakov, Yu.A. 1962. Ostracoda Myodocopida (family Halocypridae) from the N.W. Pacific Ocean. Trudy Inst. Okeanol. 12 : 97-1 12. Skogsberg, T. 1920. Studies on marine ostracodes, 1. Cypridinids, halocyprids and polycopids. Zool. Bidr. Upps. Suppl. Bd. 1 : 1-784. 162 A. J. GOODAY Williams, R. 1975. Continuous plankton records: a plankton atlas of the North Atlantic and North Sea: supplement 4— the Ostracoda in 1963. Bull. mar. Ecol. 8:21 5-228. Manuscript accepted for publication 26 September 1980 Appendix 1 Just before this paper was submitted, I received from the E.M.-Arndt-Universitat, Griefswald a collection of 520 Valdivia specimens (reg. no. II 25095), determined by G. W. Miiller as C. rotundata. These specimens have been re-identified and the new identifications are listed below. The station positions are from Miiller ( 1 906a). Eastern Atlantic Ocean Station Position Species present 32 nr. 43 24°43'N, 1 7° 1 'W C. brachyaskos 1 9; C. inflata 599; C. subinflata ^99; C. teretivalvata 3599, 7dtf, 5JJ; C. rotundata group species indeterminate, 65 specimens; Conchoecia species indeterminate, 7 specimens. 41 8°58'N, 16'27'W C. inflata, 799, 1<*. U; C. macrochiera 1 J; C. nasotuberculata 299; C. procera 299; C. spinirostris 499; C. cf. subinflata 699, 3c?c?, 4JJ; C. rotundata group species indeterminate, 28 specimens. 49 0°20'N, 6°45'W C. acuta 399; C. discover yi Id1; C. elegans 19; C. fowler 7 1299, Id1, U; C. inflata 1499, 3cW, 2JJ; C. procera 299; C. spinirotris 299; C. subinflata /^99, 3dtf, U; C. teretivalvata 499, 3cW; C. rotundata group species indeterminate 799, 3rfcf, 4JJ; Conchoecia species indeterminate 799. 66 3°55'S, 7°48'E C. cf. inflata 499, 1 rf; C. o#' teretivalvata 1 9. Southern Ocean 132 55°20'S, 5°15'E C. skogsbergi 299 (length = 1-645 mm) 2rfd- (length =1-748, 1-773 mm). 1 35 56°30'S, 1 4°29'E C. skogsbergi 1 299, 5dtf, 1 J. There are three size forms. 99: 1-388,1-439-1-491, 1-645-1-722 mm. dW: 1-413, 1-491-1-542, 1-645-1-670 mm. 136 55°58'S, 16'14'E C. skogsbergi 19 (length = 1-456 mm) 2dd (length = 1-4 13 mm). 139 55° 1'S, 21°34'E C. skogsbergi 799, lOdtf. There are two size forms. 99: 1-413, 1-6 19-1 -696 mm. rfd1: 1-413-1-439 mm, 1-619-1-748 mm. 142 55"27'S, 28°58'E C. s kogsbergi 499, 6rfcf, U. There are two size forms. 99: 1-362-1-439, 1-568 mm. dd: 1-413-1-439, 1-568 mm. Indian Ocean 2 1 7 4°56'N, 78° 1 5'E C. hyalophyllum 1 J; C. lophura 2JJ; C. aff. subinflata 2799, 23rfd\ 4JJ; C. teretivalvata 2799, 2* "I o- S co 7°?-- 4 oo TT — 1 >/^in — fS C_? 'I t& • • oo O — — — W °T 7! °_^ II ^ (N VO 00 ON fs) \o in ro in — '- • - on - VO«N ° =<£ 0 7 o |.S S.S -3 1 c D. w co Co • • • "* «r> oo (N — • m Tt O ||| • _i 'o & 3 c -5 a O co ON vc 1 1 1 T O ro r- m — 1 nm ^ c w «C £T 0 */^ r — f^ ^^ \^ "o <*-t_! "o "1 C- — n I | III -- *^- 5> oo IP O E^-S '5b 1> O C — /. ~ "o U OJ Ififl -r a _"O 'oo "O 'Lo •ol « £.SS o 'oo ^J 'oo ^J 'Lo ' £ ° «" W5 _> • oo • ^ oo $$ o SS^^i s.s 3 «a-Bp~=: « w ~ i/) a> ~*— —«-> a c- S- S "-*— -^ 2>-S 2 r f C •— OD _— i r- s --- _ , C • • Mt CQ C/5 CO (U ^ C -^ -^ — O..C 3 - — ' a> c u c -5 •< ^ 'Q. «« 'CL "> "0 _O oo -O to (U ^ § « c f; S .S g "5 S § ^.oo0^ l> C in r-~ rsi OO '^ f*"l d5 q o in — - c J5 O 13 a >*-« M.M 1) c *o T3 •o^ fa' U CO O CO ^;- — * 8 O (N CN OO (N (N MO •QOs c C- CM 0 g 2 <« on ",_ oo II " e Q (N — • Q3 <£ on *? — co" CO +1 ® +| 0 0 0 c •- 11 CO - CO O 4> o 8-g £i enarr || 00 OO — O ^f \,O ^N 4J ex, ^ 5 P 3 S a C^ j= 3 °^ 6s o °" — r- 11^ i_ o ^5 ? ^T 'T (I) 1 I* 1 ZH cS CO ^ 00 0 0 rn op in Tf •^- OO SO VO •*=' s CO £ co o "~> f*^ ^ ,S 6 TT ^f ob5 C^3 "O jO C& rj- m cS ^ T^" _u s: ffl £ ^ u 5 o o .C £ O S.«g c 2 C f^ OO ^ 53 CO Q o p p 1> ^J DC on O (N rn of^ v~"' +1 +1 +1 +1 go op c CO OO ^l" OS g^ '5 u O t~~- do O. a> ^ c ^ « c 00 UH j5 c H"^ £ CO 4J CO CO U £ c 00 T IT" ^^ rO fN || 0 *r = ^ — o oo *o */^ icg ffl CO (U H ZZ JIOQ DC CTJ 170 A. J. GOODAY Table 1 1 Male morphometric characters. C. acuta sp. n. C. australis sp. n. C. inflata sp. n. C. rotundata C. subinflata sp. n. N 41 9 30 20 42 H 46-2 ± 1-3 50-6 ± 1-1 48-0 ± 1-0 47-7 ± 1-4 46-5 ± 0-9 B 48-0 ± 2-6 48-9 53-0 ± 2-1 44-5 ± 2-0 46-4 ± 2-2 P.O. : shaft 30-5 ± 0-6 33-1 ± 0-6 29-3 ± 0-4 29-8 ± 0-74 30-1 ±0-6 capitulum 12-5 ±0-5 12-4 ± 0-4 13-4 ±0-2 12-1 ± 0-6 12-2 ± 0-5 total 42-9 ± 0-8 45-5 ± 0-6 42-7 ± 0-7 41-9 ±0-2 42-3 ± 1-0 Al: seg 1 15-9 ±0-4 16-1 ± 1-0 15-5 ±0-4 15-1 ± 0-2 14-8 ± 0-3 seg2 18-8 ± 0-4 20-3 ± 0-4 18-5 ±0-4 18-6 ±0-6 19-0 ± 0-8 total 34-8 ± 0-5 36-6 ± 1-0 34-0 ± 0-7 33-8 ± 0-8 34-0 ± 0-9 a seta 25-1 ± 2-5 22-4 ± 3-8 26-8 ± 2-1 27-6 ± 2-2 25-2 ± 1-6 bseta 43-9 ± 1-1 49-5 ± 1-0 46-1 ± 0-9 43-8 ± 1-0 44-1 ±0-8 cseta 3-3 ± 0-4 3-1 ± 0-4 3-5 ± 0-7 4-0 ± 0-4 4-0 ± 0-4 dseta 45-6 ± 0-9 50-9 ± 1-3 48-2 ± 1-0 45-4 ± 0-9 46-1 ±0-8 eseta 56-5 ± 0-9 58-7 ± 1-1 60-0 ± 1-2 52-4 ± 1-4 54-3 ± 1-0 A2: protop. 51-7 ±0-8 53-1 ± 1-6 50-8 ± 0-8 49-4 ± 0-8 49-7 ± 0-9 Ex. 1 21-1 ±0-5 24-8 ± 0-5 23-3 ± 0-6 21-8 ±0-5 21-8 ±0-5 Ex. 2-8 9-0 ± 0-3 9-2 ± 0-3 9-1 ±0-3 8-8 ± 0-4 8-8 ± 0-3 LSS 51-4 ± 0-9 62-1 ± 1-1 56-6 ± 1-2 51-7 ± 0-9 55-0 ± 1-2 fseta 36-9 ± 1-0 43-7 ± 1-4 39-2 ± 1-2 39-2 ± 0-8 40-8 ± 1-2 gseta 40-2 ± 1-2 47-9 ± 1-9 43-2 ± 1-2 42-5 ± 0-8 44-4 ± 1-2 h-j seta 13-6 ± 1-1 13-2 ± 1-2 14-5 ± 1-1 15-3 ± 1-4 13-8 ± 1-0 All measurements are expressed as percentages of the carapace length. The measurements are mean values, followed by standard deviations. N = number of observations. Table 12 Female morphometric characters. C. acuta sp. n. C australis sp. n. C. inflata sp. n. C. rotundata C. subinflata sp. n. N 42 4-7 30 20 31 H 46-9 ± 1-0 53-4 ± 1-9 49-9 ± 1-1 48-5 ± 1-0 48-7 ± 1-0 B 41-9 ± 1-8 45-9 49-5 ± 2-1 40-1 ± 2-4 38-0 ± 1-6 P.O. : shaft 26-8 ± 0-9 23-3 26-3 ±0-7 25-0 ±0-5 25-7 ±0-6 capitulum 12-0 ± 0-6 10-8 ± 0-6 12-0 ±0-5 11-8 ±0-4 12-5 ±0-5 total 36-8 ± 1-0 34-0 38-3 ±0-6 36-8 ±0-6 36-7 ±0-7 Al: segs 1 + 2 15-0± 0-4 11-9 15-0 ±0-7 13-3 ± 0-4 12-8 ±0-5 a-d setae 16-6 ± 1-2 14-9 ± 0-9 17-7 ± 1-2 17-7 ± 1-8 17-4 ± 1-4 eseta 31-1 ± 0-8 36-7 ±0-8 35-4 ±0-8 32-8 ± 1-1 35-3 ±0-7 A2: protop. 46-4 ± 0-7 47-2 ±0-8 46-5 ±0-6 43-4 ± 0-8 45-4 ± 0-8 Ex. 1 20-2 ± 0-4 23-1 ± 0-4 22-7 ± 0-4 20-3 ±0-9 20-8 ± 0-5 Ex. 2-8 8-4 ± 0-3 8-5 ±0-5 8-9 ±0-3 8-4 ±0-3 8-4 ±0-2 LSS 38-7 ± 0-9 48-7 ±0-8 43-6 ± 0-6 40-8 ± 0-8 43-2 ± 1-1 f-j setae 18-6 ± 0-9 23-6 ± 1-9 23-1 ± 1-2 22-5 ± 1-6 22-5 ± 1-1 All measurements are expressed as percentages of the carapace length. The measurements are means, followed by standard deviations. N = number of observations. CONCHOECIA SKOGSBERGI SPECIES COMPLEX Table 13 Morphological details of five skogsbergi complex species. 171 C. acuta C. australis C. inflata C. rotundata C. subinflata sp. n. sp. n. sp. n. sp. n. cfAl: b seta spines: ant. side post, side d seta spines e seta spines f seta spines g seta spines Penis muscles gAl: e seta spines: ant. side post, side Mandibular teeth: cutting edge distal list proximal list 7-10:4-7 6-13 9-15:3-7 20-24 1-2 0-5 2-4 20-23 25-35 12-14 18-20 15-20 9-12:3-8 12-18 11-17:3-4 20-23 3-5 1-4 36^0 35^2 14-16 20-21 15-18 6-11 :5-12 3-11 8-13:2-5 19-26 2-3 4-7 2-6 34-14 28-38 11-15 17-22 20-25 6-8 : 4-7 6-1 1:4-11 1-2 7-8 9-12:3-4 8-17:3-5 14-18 18-22 2-5 0-3 4-5 20-33 25-35 12-14 18-20 15-20 3-5 2-3 31-45 25-35 13-15 17-25 20-21 The first range of numbers given for dAl b and d setae refers to the more proximal, closely spaced spines, the second range is for the more distal, widely spaced spines. Note that the cfAl f and g seta spines and some of the spines on the 9A1 e seta and the dAl b and d setae can be seen only be carefully examining these setae under high magnification. B Fig. 2 Lateral and ventral carapace outlines, Conchoecia glandulosa, Station 9756 haul 7: A, 9; B,^. Scale 1-Omm. 172 A. J. GOODAY Fig. 3 Lateral and ventral carapace outlines: A, Conchoecia skogsbergi rf, Station 2501; B, C. abyssalis d (after Rudyakov, 1 962); C, C. skogsbergi 9, 250 1 . Scale 1 -0 mm. Fig. 4 Lateral and ventral carapace outlines: A, Conchoecia nasotuberculata 9, Station 6665 haul 6; B, C. kyrtophora 9, 6665 haul 33; C, C. teretivalvata 9, 7709 haul 2; D, C. naso- tuberculata 1-30 mm) specimens occurring below 1000 m are assigned to C. fowleri form A. The broad length ranges between 1000 m and 1 500 m suggest hybridization. Males black, females white. 182 A. J. GOODAY Fig. 18 Variability in the outline of the male frontal organ capitulum: A-J, Conchoecia fowleri sp. nov.; A, Station 7856 haul 10; B-E, 77 1 1 haul 32; F, 7089 haul 14; G, 6665 haul 23; H, 6665 haul 26; I, 6665 haul 27; J, 6665 haul 20: K-P Conchoecia skogsbergi\ K, small form, 2393; L, small form, 2498; M, small form, 2501; N, intermediate form, 1776; O, intermediate form, 1 778; P, large form, 1781. Scale 0'05 mm. 183 Fig. 19 Variability in the outline of the female frontal organ capitulum: A-I, Conchoecia fowleri, sp. nov.; A-D, F, Station 771 1 haul 32; E, 7089 haul 4; G, 6665 haul 32; H, I, 6665 haul 23: J-O, Conchoecia skogsbergi; J small form, 2018; K, intermediate form, 1779; L, large form, 1 782; M, large form, 1781; N, large form, 2020; O, large form, 2498. Scale 0'05 mm. 184 A. J. GOODAY Fig. 20 Conchoecia discovery! sp. nov. dtf, lateral and ventral carapace outlines: A, B, Station 77 1 1 haul 25; C, 77 1 1 haul "l 5; D-G, 77 1 1 haul 9; H, J, K, N, 7709 haul 63; I, 7709 haul 55; L, M, 7709 haul 35; O, 7406 haul 1 ; P, S, 7406 haul 22; Q, R, 7406 haul 6; T, 7856 haul 48; U, Z, 7089 haul 12; V, X, 7089, haul 5; W, 7089 haul 1 1 ; Y, AA, 7856 haul 48; BB, CC, 6665 haul 13. Scale 1-0 mm. CONCHOECIA SKOGSBERGI SPECIES COMPLEX 185 Fig. 21 Conchoecia discoveryi sp. nov. 99, lateral and ventral carapace outlines: A-F, Station 7709 haul 76; G-J, 77 1 1 haul 32; K, M, 7406 haul 22: L, N, O, 7406 haul 6; P-R, 7856 haul 48; S, T, 7856 haul 50; U, X, Y, 7089 haul 12; V, W, 7089 haul 4; Z, 7089 haul 15; AA, BB 6665 haul 19. Scale I'O mm. 186 A. J. GOOD AY Fig. 22 Conchoecia discoveryi sp. nov., male and female dimorphic parts, A-C, F, holotype BM(NH) reg. no. 1979.693; G-I, paratype BM(NH) reg. no. 1979.694: A, <5 first antenna; B, 2000m 1-4 N.Atlantic 1-7 Fig. 29 Conchoecia skogsbergi. Carapace lengths in mm, plotted against water depth in the Southern Ocean and NE Atlantic. Males are black, females are white. 194 A. J. GOODAY Fig. 30 Conchoecia wolferi sp. nov. lateral and ventral carapace outlines, A-P, 99; Q-DD, dd; A, B, G, Station 7089 haul 17; C, D, F, H, 7089 haul 22; E, 7089 haul 24; I, K-P, 6665 haul 4; J, 6665 haul 8; Q, R, T, V, W, 7089 haul 24; S, U, 7089 haul 10; X-Z, 6665 haul 8; AA-DD, 6665 haul 4. Scale 1 -0 mm. CONCHOECIA SKOGSBERGI SPECIES COMPLEX 195 Fig. 31 Conchoecia wolferi sp. nov. male and female dimorphic parts, A-D, F, holotype BM(NH) reg. no. 1979.704; G-I, paratype BM(NH) reg. no. 1979.705: A, d first antenna; B, cT frontal organ capitulum; C, cf first antenna, b, d, e setae armatures; D, rf second antenna; right endopodite; E, left hook appendage; F, penis; G, 9 first antenna; H, 9 frontal organ capitulum; 1, 9 first antenna e seta. Scales 0'05 mm except where scale otherwise. 196 A. J. GOODAY Fig. 32 Conchoecia acuta sp. nov. rfrf, lateral and ventral carapace outlines: A-D, G, Station 7856 haul 22; E,F, 7856 haul 17; H, I, 8270; J-L, 8264; M, 8272; N-V, 7089 haul 19- W Y BB 6665 haul 5; X, Z, CC, 6665 haul 1 ; AA, DD, 6665 haul 3 1 . Scale 1 -0 mm CONCHOECIA SKOGSBERG1 SPECIES COMPLEX 197 Fig. 33 Conchoecia acuta sp. nov. 99, lateral and ventral carapace outlines: A, B, D, Station 7856, haul 19; C, E-J, 7856 haul 22; K, 7856 haul 17; L, P, Q, 7856 haul 18; M, S-X, 7089 haul 19; N, 8271; O, 8270; R, 8272; Y, Z, BB, 6665 haul 1; AA, 6665 haul 3; CC, 6665 haul 5; DD, 6665 haul 36. Scale 1-0 mm. 198 A. J. GOODAY Fig. 34 Conchoecia acuta sp. nov., male and female dimorphic parts, A, B, D, E, holotype BM(NH) reg. no. 1979.690; G-I paratype BM(NH) reg. no. 1979.691: A, Contents The larval and post-larval development of the Edible Crab, Cancer pagurus Linnaeus (Decapoda: Brachyura). By R. W. Ingle A taxonomic study of the larvae of four thalassinid species (Decapoda, Thalassinidea) from the Gulf of Mexico. By N. Ngoc-Ho .... The status of Glyphocrangon rimapes Bate 1888 (Crustacea, Decapoda, Glyphocrangonidae). By A. L. Rice Crab zoeae and brachyuran classification: a re-appraisal. By A. L. Rice Page 211 237 275 287 The larval and post-larval development of the Edible Crab, Cancer pagurus Linnaeus (Decapoda: Brachyura) R. W. Ingle Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction The Edible Crab, Cancer pagurus Linnaeus occurs from northern Norway (Christiansen, 1969:43) to Portugal (Nobre, 1936:50); its presence in the Mediterranean requires confirmation (see Zariquiey Alvarez, 1968 : 345-7). In British coastal waters C. pagurus is the object of local but important fisheries and in recent years has been subjected to special studies resulting in a greater understanding of its bionomics (see Edwards, 1978). By comparison, the larval development of C. pagurus is not well documented. There are several accounts of the first zoeal stage (see below) of C. pagurus but the complete larval develop- ment has been described superficially by only Lebour (1928), the early stages from laboratory reared material and the later ones from plankton caught specimens. In 1979 C. pagurus was successfully reared to third crab stage in the BM(NH) and from this material the first account of the complete laboratory larval development of this species is now given. Materials and methods The female crab from which the larvae were reared was collected off Shoalstone Point, Devon (SX937568) from a depth of 1 5 m at a bottom temperature of 10°C in June 1979. The specimen was presented to this Museum by Alan Howard, Fisheries Laboratory MAFF, Burnham-on-Crouch, Essex. The larvae were reared using methods described by Rice & Ingle (1975) and Ingle & Clark (1977). All material was fixed and stored in the preservative formulated by Steedman (1976 : 148) and later transferred to 70% ethanol alcohol. Drawings and meaurements were made with the aid of a camera lucida. Measurements are as follows: T.T. = total lengths of zoeae measured between tips of dorsal and rostral spines; C.L. = carapace lengths measured from between eyes to posteriolateral carapace margin for zoeae, from rostral tip (for megalopa) and frontal margin (for crab stages) to median posterior carapace margin; the C.W. ( = carapace width) of crab stages was taken at the widest part of the carapace. The female and reared material are deposited in the Collections of the Zoology Department, British Museum (Natural History) registration number 1980 : 121-122. Descriptions Cancer pagurus Linnaeus, 1758 Cancer pagurus Thompson, 1828 : PI. VIII, fig. 1 (1st zoea); Cunningham, 1898 : PI. 21, figs 1-2 (1st, 2nd crab); Williamson, 1900 : PI. 1, fig. 4 (1st zoea); 1904 : PI. 4, figs 71-81 (1 1 crab stages); Pearson, 1908:460, PI. 13, figs 83-87 (prezoea, 1st zoea); Nordgaard, 1911:39, figs 1,2 (1st zoea); Williamson, 1911 : 17, PI. 4, figs 50-67 (prezoea, 1st zoea); 1915 : 485, figs 307-310 (1st zoea, 1st Bull. Br. Mm. not. Hist. (Zool.) 40 (5) : 2 1 1-236 Issued 30 July 1 98 1 212 R.W. INGLE crab); Labour, 1928 : 522, figs 2 (1 1-1 5), 4 (22-23), PI. 1, fig. 10, PI. 5, fig. 5, PI. 10, figs 3-5 (lst-5th zoeae, megal., IsMth crab); Gurney, 1942 : fig. 38 A (5th zoea), fig. 42A (2nd or 3rd zoea); Rice, 1975: 237, fig. 1 (1st zoea). FIRST ZOEA Dimensions: T.T. 2'5 mm, C.L. 0'6 mm. Carapace (Fig. la): Dorsal spine long, narrowing distally and slightly curved backwards; rostral spine thin, slightly sinuous, slightly shorter than dorsal spine and minutely spinulate; lateral spines long about ^ carapace length; dorso-median elevation prominent; a pair of anterio-median and posterio-dorsal setae; posterior margin with 6 short setae. Eyes: Partly fused to carapace. Antennule Fig. 20: Unsegmented, with 3 terminal aesthetascs and one short seta. Antenna (Fig. 2a): Spinous process about 3^ x length of exopod, distal { spinulate; exopod with one terminal spine and 2 setae. Mandible (Fig. 3a): Incisor and molar processes well developed, palp absent. Maxillule (Fig. 3d): Endopod 2-segmented with 1 ,6 setae; basal endite with 5 setae-spines, coxal with 7 setae-spines. Maxilla (Fig. 4b): Endopod with large outer and smaller inner lobe with 5 + 3 setae; basal endite with large outer and smaller inner lobe, with 4 + 5 setae; coxal endite with large outer lobe bearing 2 long and one short setae and smaller inner lobe with 3 setae one of which is very long and slightly stouter than others; scaphognathite with 4 marginal setae and a very stout posterior plumose projection. First maxilliped (Fig. 6a): Basis with 10 setae arranged 2,2,3,3; endopod 5-segmented with 3,2, 1 ,2,4+1 setae; exopod incipiently segmented with 4 terminal plumose setae. Second maxilliped (Fig. 7a): Basis with 4 setae; endopod 3-segmented with 1 ,1 ,4 + 1 setae; exopod with 4 terminal plumose setae. Third maxilliped: Not developed. Pereiopods: Not developed Abdomen (Figs 8a, 0: 5-segmented + telson; 2nd segment with pair of dorso-lateral processes; posterio-lateral margins of all segments rounded 3rd-4th with minute spinules. A pair of minute setae near posterio-dorsal margin of segments 2-5. Telson forks long, surfaces minutely spinulate, diverging posteriorly, each with one well developed lateral and one smaller dorsal spine; inner medio-lateral margin of telson with 6 setae, outermost pair with inner margins strongly serrate, median margin of telson strongly convex. SECOND ZOEA Dimensions: T.T. 2*8 mm, C.L. 0'8 mm. Carapace (Fig. Ib): Lateral spines slightly shorter than in first zoea, posterior margin with 9-10 longer setae. Eyes: Now stalked. Antennule (Fig. 2g): With 4 aesthetascs and one seta. Antenna (Fig. 2b): Unchanged. Mandible: Unchanged. Maxillule (Fig. 3e): Endopod setation unchanged; outer margin of basal endite with a prominent plumose seta, distal margin with 7 setae-spines; coxal setation unchanged. Maxilla (Fig. 4c): Endopod, basal and coxal setation unchanged; scaphognathite with 10 plumose setae of equal length. First maxilliped (Fig. 6b): Basal and endopod setation unchanged; exopod with 6 terminal plumose setae. Second maxilliped (Fig. 7b): Basal and endopod setation unchanged; exopod with 6 terminal plumose setae. Third maxilliped: Not developed. Pereiopods: 1-4 present as undifferentiated buds. Abdomen (Fig. 8b): Posterio-lateral margins of segments 3-4 with incipient acute processes; telson forks longer than in first stage. DEVELOPMENT OF CANCER PAGURUS 2 1 3 THIRD ZOEA Dimensions: T.T. 3'7 mm, C.L. 1 '0 mm. Carapace (Fig. Ic): With minute setules on dorsal spine; lateral spines slightly smaller than in previous stage. Eyes: Unchanged. Antennule (Fig. 2h): With 4 aesthetascs and 2 setae. Antenna (Fig. 2c): Spinous process slightly less than 3^x length of exopod; endopod now developed as a broad bud. Mandible: Unchanged. Maxillule (Fig. 30: Endopod setation unchanged; basal endite with 8-9 setae-spines in some specimens, additional setae very small; coxal setation unchanged. Maxilla (Fig. 4d): Endopod and basal endite setation unchanged; outer lobe of coxal endite with 4 setae inner lobe unchanged; scaphognathite with 1 7 setae. First maxilliped (Fig. 6c): Basal and endopod setation unchanged; exopod with 8 terminal plumose setae. Second maxilliped (Fig. 7c): Basal and endopod setation unchanged; exopod with 8 terminal plumose setae. Third maxilliped: Represented as a small bud. Pereiopods: Buds longer than those of previous stage, 5th pair now present as small buds. Abdomen (Fig. 8c): 6th segment now present and almost differentiated from telson; spinous process on posterio-lateral margins of segments 3-4 now conspicuous; inner medio- lateral margin of telson with 8 setae. Pleopods represented as small buds on segments 2-5. FOURTH ZOEA Dimensions: T.T. 4-3 mm, C.L. 1'2 mm. Carapace (Fig. Id): Dorsal and rostral spines slightly stouter than in previous stage and lateral spines smaller; posterior margin of carapace with 1 1-12 setae. Eyes: Unchanged. Antennule (Fig. 2i): Setal formula unchanged but aesthetascs slightly stouter than in previous stage. Antenna (Fig. 2d): Spinous process less than 3^ x length of exopod; endopod bud subequal to exopod. Mandible (Fig. 3b): Outer lip of incisor formed by a prominent row of broad tubercles merging into continuous margin. Maxillule (Fig. 3g): Endopod setation unchanged; basal endite with 10-11 setae-spines; coxal endite with 8 setae-spines. Maxilla (Fig. 5a): Endopod setation unchanged; outer lobe of basal endite unchanged, inner lobe with 6 setae; coxal endite unchanged; scaphognathite with 22 setae. First maxilliped (Fig. 6d): Basal setation unchanged; endopod terminal segment with 6 setae; exopod with 10 distal plumose setae. Second maxilliped (Fig. 7d): Basal and endopod setation unchanged; exopod with 10 distal plumose setae. Third maxilliped: Now with a small exopod. Pereiopods: Longer than in previous stage and incipiently segmented; dactylus differ- entiated on cheliped. Abdomen (Figs 8d, g): 6th segment clearly demarcated from telson; posterio-lateral spinous processes on segments 3-5 longer than in previous stage; telson inner medio-lateral margin with 10 setae. Pleopod buds longer than in previous stage. FIFTH ZOEA Dimensions: T.T. 4*7 mm., C.L. 1'4 mm. Carapace (Fig. le): Setules on dorsal spine and setae on posterior margin of carapace longer than in previous stage, lateral spine smaller. Eyes: Unchanged. Antennule (Fig. 2j): Endopod present as a small bud; exopod with 6 aesthetascs and 4 setae. 214 R. W. INGLE Antenna (Fig. 2e): Spinous process x 3 length of exopod; endopod varying from \ to over | length of spinous process and, in some specimens, incipiently segmented. Mandible (Fig. 3c): Inner margin of incisor process clearly defined as a ridge; mandibular palp present as a small bud. Maxillule (Fig. 4a): Endopod setation unchanged; basal endite with 13-14 setae-spines; coxal endite with 10 setae-spines. Maxilla (Fig. 5b): Endopod setation unchanged; outer lobe of basal endite with 6-7 setae, inner with 7 setae; outer lobe of coxal endite unchanged, inner lobe with 3-4 setae; scaphognathite with 3 1 setae. First maxilliped (Fig. 6e): Basal and endopod setation unchanged; exopod with 1 2 distal plumose setae. Second maxilliped (Fig. 7e): Basal and endopod setation unchanged; exopod with 1 2 distal plumose setae. Third maxilliped: Endopod clearly segmented. Pereiopods: Cheliped well formed, those of 2-5 clearly segmented. Abdomen (Figs 8e, h): Telson forks less divergent than in previous stages; lateral processes on 2nd segment smaller. Pleopod buds long and now present on 6th segment. MEGALOPA Dimensions: C.L. 2*4 mm. Carapace (Figs 9a, b): Longer than broad, narrowing anteriorly; frontal region (f) with slight median furrow, orbital margin expanded; rostrum long, terminally acute and horizontally directed, hepatic regions (h) inflated, protogastric (p) each with a raised carina, epibranchial (e) and mesobranchial (m) regions defined by a carina, cardiac region with a long posteriorly directed horizontal spine. Eyes: Large, elongated. Antennule (Fig. lOa): Peduncle indistinctly 3-segmented, 2nd segment with 1-2 short setae; exopod 4-segmented, 2nd with 8 and 3rd with 4 aesthetascs and 1 seta, 4th with 1 terminal and 1 subterminal seta; endopod with 4 terminal and 1 subterminal setae. Antenna (Fig. lOb): Peduncle with 3 segments; flagellum 6-segmented, setal formula (from distal to proximal) 4,3,0,4,3,0,2, 1,1,1. Mandible (Fig. lOc): Molar and incisor parts not distinguishable one from the other, disto- internal angle acutely produced; mandibular palp 3-segmented, terminal segment longest with 7 distal setae. Maxillule (Fig. lOd): Endopod now unsegmented and with 4 setae; basal endite with 21 setae-spines; coxal endite with 1 3-14 setae. Maxilla (Fig. lOe): Endopod now reduced to acute lobe, unarmed; basal endite with 8 + 9 setae, with 3-4, 5 setae; scaphognathite with 48-49 plumose setae shorter than in last zoeal stage. First maxilliped (Fig. 1 la): Coxal segment partly differentiated from basis and with 11-12 setae on inner margin, basis with 19-22 setae; endopod unsegmented with 4 setae on outer margin and 4 apical setae; exopod 2-segmented, proximal with 2 disto-external setae and terminal segment with 4 apical setae; epipod well developed and with 4-5 setae. Second maxilliped (Fig. 1 Ib): Coxal and basal segments undifTerentiated with 0-2 setae on inner margin; endopod carpus (antipenultimate segment) with 1 disto-internal seta, propodus with 5 disto-external setae, dactylus with 5 spines and 5 setae; exopod 2-segmented, terminal segment with 4 long setae; epipod short and broad. Third maxilliped (Fig. 1 Ic): Basis with 2 setae on internal margin and differentiated from coxa; endopod, ischium with 3-4 setae on outer surface and 13-14 setae placed on or near inner margin that also bears acute tubercles, merus with 2-3 outer disto-external setae and 3 setae on internal margin, carpus with 1 disto-external and 3 disto-internal setae, propodus with 2 disto-external setae and 4 setae on internal margin, dactylus with 7 setae; exopod 2-segmented, distal segment with 4 setae; epipod bifurcate, longest branch with 13-14 setae that extend onto coxal surface. DEVELOPMENT OF CANCER PAGURUS 215 Pereiopods (Figs 12a-e): Cheliped stout, with a prominent ischial spine, inner distal propodal margin with 4 blunt teeth, inner dactylar margin with at least 2 indistinct teeth. Pereiopods 2-5 relatively stout, coxa of 2nd (Fig. 12b) with acute process, dactylus of 5th pereiopod with 3 long terminal setae. Abdomen (Figs 9a,c,d, 12h): With 6 segments + telson; posterio-lateral margins of 2nd-5th broadly truncate and 3rd-5th minutely spinose; a small pair of dorso-median setae present near posterior margins of segments 2-6 along with other setae as shown. Telson (Fig. 12h) broader than long, with three pairs of dorso-median setae and a pair of setae on posterior margin. Five pairs of pleopods, distal segment of pleopod exopods with long plumose marginal setae, 1st (Fig. 120 with 16, 2nd 15-16, 3rd 15-17, 4th (Fig. 12g) and 5th (uropods, Fig. 12h) with 8 setae respectively, endopods of pleopods \-4 with 3 distally placed coupling hooks on internal margins. FIRST CRAB Dimensions: C.L. 2'3 mm., C.W. 2'4 mm. Carapace (Fig. 13a): Maximum width at about 4th pair of anterio-lateral teeth. Dorsal surface minutely denticulate, protogastric and meso-metabranchial regions slightly inflated and with long setae; frontal and orbital margins irregularly denticulate; anterio-lateral margin setose, with 4 large bi- or tridentate teeth with additional spines between them; posterio-lateral and posterior margin of carapace setose. Eyestalks with 2-3 spines. SECOND CRAB Dimensions: C.L. 2-9mm., C.W. 3'9 mm. Carapace (Fig. 13b): Maximum width now at about 7th anterio-lateral tooth. Denticles smaller than in 1st stage and marginal setae almost wholly absent; anterio-lateral margins with 8-9 spinose teeth. THIRD CRAB Dimensions: C.L. 3*7 mm., C.W. 4*9 mm. Carapace (Fig. 1 3c): Denticles now very small; anterio-lateral margins now with 9 defined, obtusely serrate teeth and one posterio-lateral tooth. Spines on eyestalks reduced. Variation The material reared for this present study agrees with previously published accounts of C. pagurus larvae except for the minor details listed in Table 1 in which the various available descriptions of the 1st zoeal stage are compared. In addition, Lebour (1928 : 523) described the fourth zoea as having 'two pairs of extra internal spines to telson' compared with one extra pair acquired at this stage by the present specimens. Gurney (1942, figs 38 A & 42A) figured the maxillule and lst-3rd maxillipeds of C. pagurus. His figure of the maxillule can be attributed to the fifth stage zoea but lacks a seta on the first segment of the endopod whilst the basis of the first maxilliped (perhaps of a 2nd or 3rd stage) has only three setae and an endopod setal formula of 0,2,1 ,2,2, on the first and 1 + 4 on the second and with the exopods of both pairs with 7 terminal setae. Samples of C. pagurus zoeae collected in the southern N. Sea, at 53°50'N: TOO'E from 6-17 July 1976, were compared with the present laboratory reared material. Stages IV and V of the N. Sea samples were found to be considerably larger (i.e. T.T. ZIV 4*7 mm; ZV 5'6 mm) than the reared specimens (i.e. ZIV 4*3 mm; ZV 4'7 mm) and a small percentage of the first zoeae of the plankton specimens was found to have one or both telson forks bifurcated and, in some, an extra medio-lateral telson spine was present as shown in Fig. 13e. In this figured specimen the lateral spine on the right fork of the telson and the small lateral and dorsal spine are absent; an extra small spinule is developed on the left outer bifurcation 216 R.W. INGLE and the seta on the basal segment of the maxillule endopod is also bifurcated. It is not known if these abnormalities are of genotypic or phenotypic origin. Distinguishing features of C. pagurus larvae From the present larval account of C. pagurus it may now be possible to distinguish the zoeae of this species from the early zoeal stages described of many other brachyrhynch crabs that occur in British coastal waters using the following combined features. (1) In C. pagurus only the second segment of the abdomen is armed with a pair of dorso-lateral processes. Early zoeal stages of Polybius henslowii Leach, Bathynectes longipes (Risso), Liocarcinus ( = Macropipus) spp., Goneplax rhomboides (Linnaeus), Geryon tridens Kroyer, Pilumnus hirtellus (Linnaeus), Xantho incisus Leach, Monodaeus couchi (Couch), Pinnotheres pisum (Linnaeus) and P. pinnotheres (Linnaeus) have dorso-lateral processes on more than one segment. (2) C. pagurus has lateral spines on the carapace; these are absent in zoeae of C. maenas (Linnaeus) and Portumnus latipes (Pennant). (3) C. pagurus zoeae have two dorso- lateral spines on each telson fork; the zoeae of Corystes cassivelaunus (Pennant) and Thia scutellata (Fabricius) have only one spine. (4) C. pagurus zoeae have smooth posterio-lateral margins to the abdominal segments except in the first zoeal stage when these margins may have a few very minute spinules. The zoeae of Pirimela denticulata has conspicuously denticulate posterio-lateral margins. (5) In C. pagurus the outer pair of telson mesio-lateral spines have strongly serrate outer margins. By comparison these serrations are far less developed in zoeae examined belonging to Liocarcinus spp., C. maenas, X. incisus, M. couchii, P. hirtellus and G. tridens. Zoeae of C. pagurus are separated less satisfactorily from those of Atelecydus rotundatus (Olivi) by having a relatively straight dorsal spine on the carapace, three distal 'setae' on the antennal exopod and 1 ,6 setae respectively on the endopod segments of the maxillule. Lebour (1928 : 524, fig. 4, 1-5, 26) described the zoeae of A. rotundatus as having a curved dorsal carapace spine in the early stages, two distal setae on the antennal exopod and 1 ,4 setae on the respective segments of the maxillule endopod. Rice (1980 : 336) has included the following two characters in his larval diagnosis of the subfamily Atelecyclinae (to which A. rotundatus belongs), (a) A maxillule endopod with 1 ,6 setae; (b) the diminutive size of the middle seta of the distal two groups of three on the first maxilliped basis. These two subfamilial features are based entirely upon the zoeal descriptions of Erimacrus and Telmessus and their confirmation in larvae belonging to Atelecydus will have to await further laboratory rearing and descriptions of larvae belonging to this genus, particularly of A. rotundatus, to establish the features for separating Atelecydus larvae from those of C. pagurus. An additional feature that may prove of value for separating the early zoeal stages of C. pagurus from corresponding stages of other species is the exceedingly long spine on the inner lobe of the coxa of the maxilla (see Fig. 1 3e). A spine of this proportion is depicted for C. magister Dana by Poole (1966, fig. Ig). This corresponding spine on the coxa of the maxilla of zoeae I, II of C. maenas, Liocarcinus spp., G. tridens, X. incisus and M. couchi never overreaches the other coxal spines as seen in C. pagurus (cf. figs 1 3 d & e). The megalopa of C. pagurus has a prominent cardiac spine on the carapace that separates it from megalopae of Liocarcinus spp., P. henslowii, C. maenas, Xaiva biguttata (Risso), T. scutellata, P. denticulata, G. tridens, G. rhomboides, P. hirtellus, X. incisus, M. couchi and Pinnotheres spp. The narrow styliform dactylus of the 5th pereiopod and the presence of 8 uropodal setae distinguishes the megalopa of C. pagurus from that of P. latipes in which the 5th pereiopod dactylus is lanceolate, the uropods have only 7 setae and the dorsal spine on the carapace arises from the meta- or urogastric regions and is thus further forward than in C pagurus (see Lebour, 1944: fig. 3 d). The absence of spines on the submedian and hepatic regions of the carapace and the presence of 3 setae on the dactylus of the 5th pereiopod separate the megalopa of C. pagurus from that of C. cassivelaunus; the latter species has DEVELOPMENT OF CANCER PAGURUS 217 submedian and hepatic spines (as well as a pair of minute protogastric spinules) and the 5th pereiopod dactylus has only 2 setae (see Ingle & Rice, 1971: figs7F& 8). Further studies are required to establish satisfactory features for separating the megalopa of C. pagurus from that of A. rotundatus. Lebour (1928 : 525, PI. IX, fig. 5) states that the megalopa of A. rotundatus has 10-1 1 uropodal setae (C. pagurus has 8 setae), and depicts 4-5 setae on the 5th pereiopod dactylus (C. pagurus has only 3 setae). Comparisons of larvae of the species of Cancer The first zoeal stage of eleven and the complete larval development of eight species of Cancer have been described. The majority of these accounts omit details of features such as carapace setation of the zoeae, lateral spines of the telson, the armature of the megalopal carapace, proximal segments of the pereiopods and setation of the telson. Distinctive larval features described in these various accounts are summarized in Tables 2-4. No single character (except perhaps the exceedingly large size and exceptionally large number of setae on the antennule and scaphognathite in later stages of C. magister and the presumably 5-segmented abdomen described for all stages of this species (see Poole, 1966)) can be used to separate the known species of Cancer larvae, but they appear to be identifiable on combined features. A comparison of the first zoeal stages of Pacific and Atlantic species of Cancer suggest that those from the former region have more pronounced posterio-lateral spines on the abdominal segments and a tendency to fewer antennular aesthetascs-setae. Zoo- geographic differences are more obvious when the fifth zoeal stages are compared. Pacific species have a maximum of 6 setae on the maxilla endopod; this number is always 8 in Atlantic species except for C. ? bellianus that has 7 setae. There are never more than 9 setae on the 1st maxilliped basis in Pacific zoeae compared with 10 or more in Atlantic ones. Rice (1975 & 1980 : 331) recognized trends towards the existence of two possible groups of Cancer zoeae with respect to the armature of appendages: (a) species with 2 setae on the basal segment of the endopod of the first maxilliped, the maxillule endopod segments armed with 1 ,5 setae respectively and the endopod of the maxilla with 6 setae; (b) species with 3 setae on the basal segment of the endopod, 1 ,6 setae on the maxillule endopod and the endopod of the maxilla armed with 7-8 setae. This present study also suggests that zoeae attributable to group (a) have only 2 and those in group (b) 3 terminal 'spines' on the antennal exopod. Zoeae in which the above mentioned features are not combined are C. amphioetus, C. gibbosulus, C. porteri and zoea V of C. ? bellianus. Acknowledgements I wish to thank Alan Howard, Fisheries Laboratory, MAAF, Burnham-on-Crouch, Essex for kindly providing ovigerous specimens of C. pagurus and Dr John Nichols, MAFF, Fisheries Laboratory, Lowestoft, Suffolk for drawing my attention to the abnormal specimens of C. pagurus zoeae. Dr A. L. Rice kindly read the manuscript. References Aikawa, H. 1937. Further notes on brachyuran larvae. Rec. oceanogr. IV ks Japan 9 : 87-162. Ally, J. R. R. 1975. A description of the laboratory-reared larvae of Cancer gracilis Dana, 1852 (Decapoda, Brachyura). Crustaceana 28 : 23 1-246. Anderson, W. R. 1978. A description of laboratory-reared larvae of the Yellow Crab, Cancer anthonyi Rathbun (Decapoda, Brachyura), and comparisons with larvae of Cancer magister Dana and Cancer productus Randall. Crustaceana 34 : 55-68. 218 R. W. INGLE Bourdillon-Casanova, L. 1960. Le meroplancton du Golfe de Marseille: les larves de crustaces decapodes. Reel. Trav. Sta. mar. Endoume30 : 1-286. Christiansen, M. E. 1969. Marine invertebrates of Scandinavia, No. 2 Crustacea, Decapoda Brachyura. 143 pp. Universitetforlaget, Oslo. Connolly, C. J. 1923. The larval stages and megalops of Cancer amoenus (Herbst). Contr. Can. Biol. Fish (NS) 1 : 337-352. Cunningham, J. T. 1 898. On the early post-larval stages of the common crab (Cancer pagurus), and on the affinity of that species with Atelecyclus heterodon. Proc. zool. Soc. Lond. pp. 204-209. Edwards, E. 1978. The Edible Crab and its fishery in British waters. Fishing News Books 142 pp. Farnham. Fagetti Guaita, E. 1960. Primer estadio larval de cuatro crustaceos Braquiuros de la Bahai de Valparaiso. Revta Biol. mar. 10 : 143-154. Frost, N. 1936. Decapod larvae from Newfoundland waters. Res. Bull. Div. Fish. Res. Newfounld 3 : 1 1-24. Gurney, R. 1942. Larvae of Decapod Crustacea. Ray Society, London i-vi + 306 pp. Ingle, R. W. & Clark, P. F. 1977. A laboratory module for rearing crab larvae. Crustaceana 32 : 220-222. Ingle, R. W. & Rice, A. L. 1971. The larval development of the masked crab, Corystes cassivelaunus (Pennant) (Brachyura, Corystidae), reared in the laboratory. Crustaceana 20 : 27 1-284. Iwata, F. 1973. On the first zoea of the crab, Cancer amphioetus Rathbun from Hokkaido, Japan. Proc. Jap. Soc. Syst. Zool. 9 : 2 1-28. Lebour, M. V. 1928. The larval stages of the Plymouth Brachyura. Proc. zool. Soc. Lond. 2 : 473-560. Mir, R. D. 1961 . The external morphology of the first zoeal stages of the crabs, Cancer magister Dana, Cancer antennarius Stimpson, and Cancer anthonyi Rathbun. Calif Fish. Game 47 : 103-1 1 1 . Nobre, A. 1936. Crustaceos Decapodes e Stomatopodes Marinhos de Portugal. Fauna Marinha de Portugal. IV : i-viii + 2 1 3 pp. Porto. Nordgaard, O. 1912. Faunistiske og biologiske lakttagelser ved den biologiske station i Bergen. K. norske Vidensk. Selsk. Skr. 6 : 1-58. Pearson, J. 1908. Cancer (The Edible Crab). L.M.B.C. Memoirs No. XVI. Proc. Trans. Lpool biol. Soc. 22 :291-499. Poole, R. I. 1966. A description of laboratory-reared zoeae of Cancer magister Dana, and megalopae taken under natural conditions (Decapoda Brachyura). Crustaceana 11 : 83-97. Rice, A. L. 1975. The first zoeal stages of Cancer pagurus L., Pinnotheres pisum (Pennant) and Macrophthalmus depressus Riippell (Crustacea, Decapoda, Brachyura). Bull. Br. Mus. nat. Hist. (Zool.) 28: 237-247. 1980. Crab zoeal morphology and its bearing on the classification of the Brachyura. Trans, zool. Soc. Lond. 35 : 27 1-424. & Ingle, R. W. 1975. The larval development of Carcinus maenas (L.) and C. mediterraneus Czerniavsky (Crustacea, Brachyura, Portunidae) reared in the laboratory. Bull. Br. Mus. nat. Hist. (Zool.)28: 101-119. & Williamson, D. I. 1977. Plankton stages of Crustacea Malacostraca from Atlantic Seamounts. Meteor ForschErgebn. D 26 : 28-64. Roesijadi, G. 1976. Descriptions of the prezoeae of Cancer magister Dana and Cancer productus Randall and the larval stages of Cancer antennarius Stimpson (Decapoda, Brachyura). Crustaceana 31 : 275-295. Sastry, A. N. 1977a. The larval development of the Rock Crab, Cancer irroratus Say, 1817, under laboratory conditions (Decapoda Brachyura). Crustaceana 32 : 155-168. 19776. The larval development of the Jonah Crab, Cancer borealis Stimpson, 1859, under laboratory conditions (Decapoda Brachyura). Crustaceana 32 : 290-303. Steedman, H. F. (ed.) 1976. Zooplankton fixation and preservation. In: Monographs on oceanographic methodology, 350 pp. Paris. Thompson, J. V. 1828. Zoological Researches, and illustrations; or Natural History of nondescript or imperfectly known animals, in a series of memoirs: . . . Memoir I ... On the metamorphoses of the Crustacea, and on zoea, exposing their singular structure, and demonstrating that they are not, as has been supposed, a peculiar Genus, but the larva of Crustacea!! 1 1 pp. Cork. Trask, T. 1970. A description of laboratory-reared larvae of Cancer productus Randall (Decapoda, Brachyura) and a comparison to larvae of Cancer magister Dana. Crustaceana 18 : 133-146. 1974. Laboratory-reared larvae of Cancer anthonyi (Decapoda: Brachyura) with a brief description of the internal anatomy of the megalopa. Mar. Biol. Berlin 27 : 63-74. DEVELOPMENT OF CANCER PAGURUS 2 1 9 Williamson, H. C. 1900. Contributions to the life-history of the Edible Crab (Cancer pagurus). Rep. Fishery Bd Scot I. 19 : 77-142. 1904. Contributions to the life-histories of the Edible Crab (Cancer pagurus) and of other decapod crustacea;-lmpregnation: spawning: casting: distribution: rate of growth. Rep. Fishery Bd Scotl. 22: 100-140. 1911. Report on the larval and later stages of Portunus holsatus, Fabr.; Portunus puber, L.; Portunus depurator, Leach; Hyas araneus (L); Eupagurus bernhardus, L.; Galathea dispersa, Spence Bate; Crangon trispinosus (Hailstone); Cancer pagurus, L. Sclent. Invest. Fishery Bd Scotl. (1909): 1-20. 1915. Crustacea Decapoda Larven. Nord. Plankt. 18 -.315-588. Zariquiey Alvarez, R. 1968. Crustaceos Decapodos Ibericos. Investigacion pesq. 32 : i-xv + 5 10 pp. Manuscript accepted for publication 10 September 1980 .2 s 1 *i OO 1— *"* CO c o ^ i _ c ^ «-• 22 ~ .c -^ IS 4- c ^ u *^ C 3 *^ *^ ^^ __ _A. (U 1 M 2 OO c/5 .^ OO c3 OO ^ r<^ ^ 35 en + - 2 c £ ^/5c^GTr"^5^"^5o t^m ro Tt — CX O .2 •0 C 'C u CO ON c ^c"H hri*^ J3 T^-C-C ^t T_o E U u U B 9 *3 ti^j^ t_ °°1>Cir4-<*'aj O r«^ «0 W5 + _T -4— » r O c £ ^•coEr^^jC-^^0 |«n <**") ^ — ^ C c JS ^^ ^ ^ C */"> ^ _C u. 3 2 o o o *S ° O cO to c/5 3 "O — C/2 £> fi 3 <_• *- o >- CO , . -o J cr o o xj '.£ (UCCc^-e^- cO<-> vO^ 2 . . § | D. — 6 o ON o G ^ _^ •o T3 ^ — in c ^ c CO k. /*? o Ji%_2 ® o CO CO (1 t/3 2? QH CO °° , «~] c 3 "O CO O o Z o •— O "~ _ff .O _? J2-occ<. ^Sea'S i i O 1 IIC 3 CO 1/5 3 oo 1— _O 0 u C CO ON g c c ^ „ c > . u JU £ >^"cO ID "7 <-• c/5 .2 TH "O "O <~; "^ "^ ^ g "i* "cO •—^ -C ^— ^— o -C C O ^ + ;, to r^l (U o N S — ^ ^> C '•— -C 1> £ -> 5 S. o t/5 '— t<— i O CO e s 1 «1 ^ ir' '% « ® .2 " §3 2 . . • • r^ O ^ fj "O _ "O "3 OS <£ o — ~z* ^ ^ ^ 2 ^ Comparat t/5 '— '-^- -*-^ — ^j CX &A .5 (L> ^cojy.Soooc: ac^- T3§gto^^ — - t/3 £ 3 08 U ^^^Jclcg |^"5 ^cc^iio^S 'S'5 s 5 'S.'S. *"" c ^ c ^ ^ o ^"•S-OTj'i « rn x c c o £ x o o o -o 2 co o o a _Q Q. H U- -Jc/3C/)C/5 cC ^ ^ -n ON ON *~ -5 •* & — i*-* i fvj .. E to 5 ~ m •* ~ rs — 3 SP o -^ Q 0- */m ^f-^mm ^oo rf^^l £ 8 5 -2 00 (N 3 •is >/-> m ~™ c ^ fN* ^ S E ^ 00 .C 'I ' ~ 1^ fN— -v^mr^ ^"^ .2?.5 E Tf 1/5 2 •£ U VO /%c. rj_^mm Ttc.. t-. Tt^v, — 'O JU on ,,- D. 2 " Si — •* ^' g E o E • ^^" r— 5. x ^ * 7 *+ ^'T^-S J^ CX) O ^J 3 U>0 in Tt + 'H + 1 D. '»T m — u-^Tt-m rj-ov mTfmvi 1 PJ X Nj ON ^ 2 Tf rs u E s ~ ° •c ~ mm — — . — 3 E t; >, 4J 5 *-• c a> * •— vi aj "7 1 <— > Jo •— - - • I/) O. ' co O Ij- obeoCw*^ oj2n4> li i||E u| -g i a « S » si 1.2 & vf 8. a i i s f. sssis vertex to 1 ON ON .2 U-_c "OO."5 R,"1^ W C ^+ i °y ^ ^ r-" + + r~ °i + ~ "? •? u o 5 —"——"— vq ^r 2 : E t/3 t— "1 vo m" •*" + + *'^»s2::rt+*s!+ ~ ~ T ^ V 2 ** U, .2 .3 a c 00* 71 u. -r: S; 5 C3 X 00 s 2- il 1 •4- "H •*• ~: " "? ° o CO H W (N* o Bj O + "3? c u 00 '6 a 5 U X — TT o -:_ + 1 | JD 0 (U 00 * 3 •P * -H .-S m «r^ 4> •g to ^ (N + E '5 & £ X - -ii _^ TJ O ^ T3 SIS*. ~ TD wl -5- - O c « O Pi — — -le 3 Comparati O al spine/carapace length: :nnule, aesthetascs-setae: :nna, exopod distal spines illule, endopod setae: illule, basis; coxa, ae-spines: ilia, endopod; basis; coxa ae-spines: y • • 1 2 "- 1 i | -rf jj.1 IMS a * E. -a -8 S-ffd 5 -a -8 5SH^«c S5 v '2'S 55±> g> -o -o D.D.EE28 2 Ji Ji = = vw^g. 1 = it '3 '3 1 II 1 M B 'I i 1 1 S g.2 g - lie n Ijlij *? c oo oo oo oo O O O 0 0 0 O & DU u- oo oo oo oo oo oo oo oo oo oo oo ON ON ON ON ON (U a ££& 03 ^ >^ >"» i>> >^ i o .2? C^ C^ C^ QT ob ob 6 o X ^ t-t U* UN . _ Ord cn TJ ^- s^a. cu CL o '> '> 0 O 00 ^ s^s^^eg £££§§> . 6 0 O n — r~- — •/-> — «n C ^ TT 1 | exam a u Q oo rn ^ — fTj Q^ *o ^ o r^- fN r*^ '^ ro CM O (N ON o •* <^^ •^J -fc ^^. ^^. ^^ ^^. Cv ^ v r"^ I^1 1^" J1 [^- 5 .1 O b 22 b r-^ ^- 9 £! o^ 0 *n S; o oo • O T. r~ 7*. 7". — T. — c o b b b ?N v 1 ? 1 ~ jt CN) r~ b C4 — sb 1 o a. rn rn TT ON b b b 2 _ 0 r- ^ Tf o .Is 'C Ik m fN 2 so s I 03 -Si s5 c5 ^ ^ 1, 1 ^^ -~^ "*** ^^* ^ ^ ft ^* Q « <3 5J ;j "•Si ^> -s: ""J2 «^J — c O xj o 2 5 "^ ^ s CB o 00 g i^S ^ 5 > > H &> X ac xotot Q« OH STATUS OF GL YPHOCRANGON RIM APES 283 bosses on the pleura of somites three and four are quite pronounced but, as Bate pointed out, they are not armed with teeth. The pleura of somites 2-5 each carry only two ventro-lateral spines. The outer margin of the scaphocerite has a small notch, as in the other Pacific specimens, but this seems to be more distally placed. Several of the more posterior appendages are damaged but the second pereiopods have the carpus divided into about 20 segments on the right hand side and about 15 on the left. Finally, the dactyls of pereiopods four and five are not bifid, but have a stepped outline and terminate in a single acute tip; they do not have a basal boss. NEW MATERIAL. During a series of cruises undertaken since 1977 to investigate the benthic fauna of the Porcupine Sea-Bight to the south-west of Ireland, a further four female and three male specimens of G. rimapes have been collected at depths ranging from about 3000 m to a little over 4000 m (see Table 1 ). The females all resemble the holotype but are even more similar to the large specimen from station 300 since, like it, they have only three pairs of spines of the rostral margin, the pleura of somite four carry only two spines, while there are consistently three spines on the pleura on each side of somite five. The only significant difference between these specimens and those of the type series is that the spines on the bosses of the pleura of somites three and four are rather less well-developed. The two largest male specimens resemble the recently collected females very closely, although the pleural spines are rather shorter and there is a suggestion of a third spine on the fourth abdominal somite (see Fig. 4C). The main distinction from the females, however, is that while the dactyls of the fourth and fifth pereiopods are more flattened than those of the styliform third pair, they end in a single point and at most have a step near the tip (see Fig. 2E). Thus, as in Glyphocrangon sculpta (see Holthuis, 1971; Barnard, 1950) it is only the females which have the characteristic bifid dactyls. The smallest male (from station 50604) has a well-developed appendix masculina and non-bifid dactyls on the last two pereiopods. However, it differs from the other males, and from most of the females, in having a relatively much longer rostrum which over- reaches the scaphocerite by almost half its own length, in having the posterior rostral spines very much smaller than the more anterior ones, and in having a quite prominent notch in the outer margin of the scaphocerite slightly anterior to its mid-point. DISCUSSION. The suggestion by Holthuis (1971, p. 287) that Bate's specimens might include more than one species was presumably prompted by the wide separation of the localities from which they were obtained, together with the fact that no other species of the genus had been recorded from both the Atlantic and from the Indo-Pacific. From an examination of the type series alone, I would have had little hesitation in concluding that three of the four specimens, that is the holotype from the southern Atlantic, the specimen from off Japan and the larger of the two specimens from Juan Fernandez, belong to the same species despite the slight differences, particularly in the pleural armature, between them. The fourth specimen, that is the smaller female from Juan Fernandez, is rather different since, not only does it have a distinct pleural armature with only two spines on the fifth somite, but it lacks the bifid dactyls on the fourth and fifth pereiopods, it has a very reduced posterior rostral spine, and a relatively longer rostrum. With the additional north-Atlantic material available, however, these variations seem to be acceptable. For although the pleura of the fourth and fifth abdominal somites usually have two and three teeth respectively, these are quite variable. Similarly, while the rostrum always apparently has at least three pairs of lateral spines, the posterior pair may be very small, particularly in juvenile specimens, while additional spines may be present. Finally, the rostrum becomes relatively shorter in older specimens, being almost twice the post-orbital carapace length in the smallest specimen available and less than two thirds as long in the largest. The new material extends the known depth range of G. rimapes from 2500 m to 4100 m 284 A. L. RICE Fig. 4 Glyphocrangon rimapes Bate, 1888. Right hand side and left hand side views of the abdomens of three specimens from the Porcupine Sea-Bight. A, female (C.L. 44-8 mm), Discovery station 9640; B, male, R.V. Challenger station 50604; C, male, Discovery station 9640. The bar scales represent lO'O mm. and the horizontal distribution to include the northeastern Atlantic. In the Porcupine Sea- Bight region rimapes overlaps at the shallower end of its range with G. sculpta (Smith) and is replaced at depths a little over 4000 m by G. atlantica Chace. In order that Holthuis's key should cope adequately with rimapes, his third couplet should be removed and rimapes should be distinguished from all other Atlantic species at the beginning of the key by the presence of at least three pairs of rostral spines in this species. Acknowledgements I am grateful to Dr R. W. Ingle (British Museum (Natural History)) for allowing me to STATUS OF GL YPHOCRANGON RIM APES 285 examine the type material from the Challenger collections. My thanks are also due to Mrs C. E. Darter for making the illustrations and to Mr A. F. Gray for photographing the specimens. Finally, I must thank Jonathan Rees for recognizing the rimapes problem in the first place. References Barnard, K. H. 1950. Descriptive catalogue of South African decapod Crustacea. Ann. S. Afr. Mus. 38: 1-837, figs 1-154. Bate, C. S. 1888. Report on the Crustacea Macrura collected by H.M.S. Challenger during the years 1873-76. Rep. Voy. Challenger, Zool. 24 : i-xc, 1-942, textfigs 1-76, pis 1-150. Holthuis, L. 1971. The Atlantic shrimps of the deep-sea genus Glyphocrangon A. Milne-Edwards, 1881. Bull. mar. Sci. 21 : 267-373, figs 1-15. Manuscript accepted for publication 1 5 September 1980 Crab zoeae and brachyuran classification: a re-appraisal A. L. Rice Institute of Oceanographic Sciences, Wormley, Godalming, Surrey, U.K. Introduction In a recent paper (Rice, 1980a) I reviewed the available information on crab zoeal morphology and attempted to assess its bearing on the classification of the Brachyura. Although zoeal evidence had already been employed several times to try to elucidate specific problems of crab relationships, there had been no previous attempt to relate a general classification based on the larval stages to that based on the adults. The reasoning behind the study was founded on the hope that since the zoeal stages are all adapted for a mid-water existence any classification based on them would be largely free from the problems associ- ated with the convergent and divergent adaptations of the adults to their varied life styles. Using a variety of zoeal features, including details of the appendage setation, I was able to demonstrate, at least to my own satisfaction, a high degree of agreement with the traditional adult classification at the family level, but much less congruence at higher levels. For example, the zoeal stages of the constituent families of the Oxystomata are so distinct from one another that they confirm the doubts about the validity of this grouping which have been expressed by several students of adult crabs. Not only are there no zoeal grounds for grouping the dorippids, leucosiids and calappids together, but there is no support for their separation, either collectively or individually, from the Brachyrhyncha. Similarly, the clear differences between the zoeae of the parthenopids, hymenosomatids and majids argue strongly against their inclusion in a separate oxyrhynchous group, for the former two families, like the oxystomes, clearly seem to belong to the Brachyrhyncha. The majids, on the other hand, did seem to warrant separation from the majority of crabs, for their zoeae exhibit a number of distinctive features which indicate an early divergence and the adoption of a different developmental strategy from that of the rest of the Brachyura. In dividing up the Brachyrhyncha the zoeal stages seemed to be much less helpful. Only two major groups were recognized; first a collection of families with relatively primitive larvae corresponding roughly with Milne Edwards' Cyclometopa or Guinot's (1978) Heterotremata, and a second group with more advanced larvae which corresponds fairly closely with Milne Edwards' Catometopa or Guinot's Thoracotremata. Such a division, however, seemed to be very artificial since it grouped the families according to their general evolutionary level rather than into phylogenetic lineages. Moreover, there were some important discrepancies between the larval divisions and those suggested by Guinot based on the morphology of the sexual organs in the adults. In particular, the Leucosiidae, which Guinot placed in her Heterotremata because at least some members of this family have coxal male sexual openings, have very advanced zoeal stages which seemed to ally them to the catometopous families which Guinot placed in her Thoracotremata. During recent months the debate on brachyuran relationships has progressed somewhat, for in two most interesting notes Saint-Laurent (19800 & b) has generally supported Guinot's division of the Brachyura into the Podotremata, Heterotremata and Thoracotremata, but has disagreed fundamentally in her interpretation of the relationships between them. The main diagnostic differences between Guinot's suggested groups are the positions of the male and female sexual openings: in the Podotremata both the male and female openings are coxal; in the Heterotremata the female openings are all sternal, but at least some species in Bull. Br. Mus. nat. Hist. (Zool.) 40 (5) : 287-296 Issued 30 July 1 98 1 288 A. L. RICE each family have the male openings on the coxa; in the Thoracotremata both the male and female openings are consistently sternal. Guinot suggested that during brachyuran evolution there has been a strong tendency for the sexual openings to move from the primitive decapodan coxal position onto the sternum and that her three sections therefore represent successive stages in this migration. Saint-Laurent, on the other hand, sees great difficulty in deriving these groups from one another. She points out (1980a, p. 1266) that the female genital apparatus in the Podotremata is comparable with that in many other decapodan groups in which the spermatophores are deposited with the aid of the male sexual pleopods into a receptacle, the thelycum, formed from an intersegmentary fold of the integument and without any connection with the oviducts. Fertilization in these forms is external and takes place after egg-laying. Within the Sternitremen crabs (that is the Heterotremata and Thoracotremata together, or the Eubrachyura in Saint-Laurent's terminology) the spermatic mass is deposited, again via the sexual pleopods, in an internal seminal chamber formed by a dilation of the oviduct, within which fertilization takes place. Saint-Laurent finds it difficult to envisage the intermediate stages between one situation and the other involving, as it would, not only the loss of the thelycum but also a change in the orientation of the male sexual pleopods from the thelycum towards the oviducts. Instead, she suggests that the separation between the Podotremata, on the one hand, and the Heterotremata and Thoracotremata, on the other, is cladistic and not simply a difference of evolutionary level. Whether the Eubrachyura were derived from ancestors with or without a thelycum, Saint- Laurent concludes that they represent an independent branch which became separated at an early stage from the primitive brachyuran line. Similarly, Saint-Laurent believes that the distinction between the coxal and sternal position of the male orifice is a fundamental one, indicating that the Heterotremata and Thoracotremata diverged at a very early stage in brachyuran evolution (see Fig. 1). She redefines the Heterotremata as eubrachyurans in which the male genital ducts always pass via the coxae of the fifth legs before opening to the exterior, either on the coxa itself or on the sternum. In the Thoracotremata the ducts always open to the exterior directly through the sternum without passing via the coxae. Guinot had suggested that some heterotrematous groups, such as the Goneplacidae and Leucosiidae, in which the male orifices are, in her terms, sometimes coxo-sternal, represent an intermediate stage towards the thoraco- trematous condition. Saint-Laurent, on the other hand, sees these groups as being truly heterotrematous since the male orifice only appears to be sternal because the tubular pro- longation of the male duct, that is the penis, becomes encapsulated after it leaves the coxa within an integumentary canal at the boundary between sternites seven and eight, to emerge finally in a sternal position. The tendency of the male orifices to move towards the mid-line in both the Heterotremata and the Thoracotremata is seen by Saint-Laurent as a response to the relative narrowing of the anterior abdominal somites compared with the posterior cephalothorax, and the need to bring the orifices close to the bases of the sexual pleopods to ensure successful sperm transer. But she considers the mechanisms by which this has been achieved in the two groups as totally distinct. Consequently, Saint-Laurent's view of brachyuran phylogeny (see Fig. 1) is rather different from that put forward by most authors in the past and implied by Guinot. For while most authors have derived the higher Brachyura from within the Podotremata, Saint-Laurent does not identify ancestors for either the podotrematous groups or the Eubrachyura, but she suggests that they diverged at a very early stage. Similarly, although she indicates that the Heterotremata and the Thoracotremata had a common ancestor, she maintains that the Heterotremata diverged very early on, possibly via more than one line, and that the ancestors of the Thoracotremata are not to be found amongst the extant Heterotremata or their immediate predecessors. Finally, she suggests that the assumption of the thoracotrematous condition, in which the posterior thoracic appendages are freed from any involvement in reproduction, may have allowed the development of highly perfected locomotory mechanisms and enabled this group to colonize a variety of terrestrial habitats. HOMOLOIDEA DROMIOIDEA RANINOIDEA BRACHYURAN CLASSIFICATION HETEROTREMATA 289 PODOTREMATA EUBRACHYURA BRACHYURA Fig. 1 Summary of the phylogenetic relationships of the main brachyuran groups according to Saint-Laurent (modified from Saint-Laurent 19806). The publication of this new approach to brachyuran phylogenetics has prompted me to re-examine the zoeal evidence for crab inter-relationships and, in a later publication, to consider also the megalopa stage. PODOTREMATA Guinot's Podotremata contains the dromiids, homolids, raninids and tymolids, the last three groups being placed together in the subsection Archaeobrachyura. After discussing the zoeal evidence at some length (Rice, 1980#, p. 289 el seq.) I supported Williamson's (1965, 1976) view that the Dromioidea are more closely related to some of the anomuran groups than to the brachyurans and should accordingly not be included in the latter. This conclusion was based on a number of generally anomuran features of dromioid zoeae, but particularly on the presence of the hair-like second telson seta in all known dromiid zoeae, and in the anomurans and thalassinids, but its absence from all higher brachyurans. Knowledge of dromioid larvae was at that time limited to those of the Dromiidae, and Williamson (1976, p. 407) had suggested that larvae of the families Homolodromiidae and Dynomenidae might be much more homolid (that is brachyuran). No homolodromiid larvae have yet been described, but an examination of the late embryos of the dynomenid Acanthodromia erinacea H. Milne Edwards (Rice, in press) has demon- strated that the zoea is very similar to known dromiids, including the presence of a hair-like second seta. Thus, there is still no larval evidence of a more brachyuran branch of the Dromioidea and I therefore remain convinced that they should not be included within the Brachyura. In establishing her Archaeobrachyura, Guinot (1978) recognized that it was not a natural group since, she maintained (p. 232), 'ils comportent a la fois des especes primitives, qui sont sans doute a 1'origine des autres sections (les "vrais Crabes"), et des especes apomorphes avec des caracteristiques speciales aux trois super-families.' From the zoeal evidence I also concluded that this grouping is not natural in the strictly cladistic sense since I believed that the raninids became separated from the primitive brachyuran line at a later stage than that which gave rise to the ancestral lineage of the extant homolids. Thus, I suggested (Rice, 1980a, Fig. 9) that the raninids and the higher brachyurans share a more recent common ancestor than either group does with the homolids. I nevertheless felt that the archae- obrachyuran concept is a useful one since it indicates that although the higher crabs originated from an ancestor within it, they have attained a significantly higher evolutionary 290 A. L. RICE HOMOLOIDEA RANINOIDEA EUBRACHYURA HOMOLOIDEA RANINOIDEA EUBRACHYURA B Fig. 2 Alternative phylograms showing the relationships of the homolids, raninids and eubrachyurans (see text). level so that their separation from the raninids in terms of evolutionary change is much greater than that between the homolids and raninids. Saint-Laurent did not give details of her opinion of the origin of the brachyuran groups. However, since she considers the divergence between the Podotremata and the Eubrachyura to be cladistic, she would presumably favour a phylogram for the homolids, raninids and eubrachyurans like Fig. 2B rather than 2A, that is with the homolids and raninids having a more recent common ancestor than the raninids and the eubrachyurans. It seems to me that the apomorphic characters shared by the zoeae of the raninids and the Eubrachyura, but not by the homolids, argue strongly against this and I would therefore still contend that the homolids became separated from the primitive brachyuran line at an earlier stage than the raninids. EUBRACHYURA (HETEROTREMATA & THORACOTREMATA) Guinot's Heterotremata consists of the superfamilies Dorippoidea, Calappoidea, Cory- stoidea, Portunoidea, Xanthoidea, Majoidea, Parthenopoidea, Bellioidea and Leucosioidea. It therefore corresponds to Milne Edwards' Cyclometopa with the addition of the dorippids (excluding the tymolids), the calappids and the leucosiids from the Oxystomata, and the majids and parthenopids from the Oxyrhyncha. The Thoracotremata contains all the remaining higher crabs and therefore corresponds to Milne Edwards' Catometopa with the addition of the hymenosomatids. BRACHYURAN CLASSIFICATION 291 In attempting to categorize the zoeal stages, and having, like Guinot, dismembered the oxystomes and oxyrhynchs, I thought that I could recognize three main groups. The first of these, the majids, seemed to be a well-defined one in which the zoeal phase is abbreviated to only two stages and well-developed pleopods are present in the second stage. The remaining two groups were much less easily distinguished, but each consisted of a series of families in which the zoeal stages were respectively relatively primitive or relatively advanced. The only feature which seemed consistently to separate these two groups was the number of setae on the endopod of the maxilla, the primitive zoeae having six or more setae and the advanced ones five or fewer. Distinguished in this way, the primitive group corresponded to Guinot's Heterotremata except for the Majidae, which were separated as mentioned above, and the Leucosiidae, Dorippidae and perhaps part of the Calappidae which seemed to be allied with the more advanced families. With the addition of these families, the advanced group there- fore corresponded to Guinot's Thoracotremata. I realized that evolution within the higher brachyurans has been far from simple and has probably involved many separate lines. Nevertheless, the apparent existence of these two large groups of crabs with seemingly primitive and advanced zoeae respectively, together with Guinot's recently published thesis, encouraged me to hope that phylogenetic lines might be discernible from one group to the other. In fact, this hope was not realized, for although I was able tentatively to identify some possible phylogenetic lines within the primitive zoeal groups, I was unable to extend these into the more advanced families because many of them seemed to have a confusing mixture of advanced and primitive features which precluded their derivation from any of the extant groups with generally more primitive zoeae. Adopting the view of eubrachyuran phylogeny suggested by Saint-Laurent, many of these difficulties disappear. For according to this view the Heterotremata and Thoracotremata should be considered as quite distinct taxa with no phylogenetic links between them. On the other hand, this approach poses new problems, for although the zoeae of the Heterotremata are certainly generally more primitive than those of the Thoracotremata, there is much more overlap than I had thought. The distinction between the two groups based on the setation of the endopod of the zoeal maxilla is clearly invalid, for the leucosiids and dorippids are simply highly evolved Heterotremata in which the zoeal morphology has advanced beyond the general level for this group and in a number of features, including the maxilla setation, has approached the thoracotrematous condition. The same is true of the most advanced majids, but in this case a single family, if indeed it is to be regarded as such, contains a whole range of zoeal forms from the relatively primitive oregoninids to the very advanced inachinids. Having eliminated the maxillary endopod setation as a distinction between the hetero- trematous and thoracotrematous zoeae, I can find no other single feature or combination of features which will consistently separate the two groups. This seems rather surprising if, as Saint-Laurent has suggested, the Heterotremata and Thoracotremata have had separate evolutionary histories from a very early stage. However, the key difference between the two groups, that is whether or not the coxae of the fifth legs are involved with the male reproductive apparatus, would not directly affect the larval stages at all. For although this difference may have resulted in the adults evolving along very diverse adaptive lines, the Heterotremata retaining the benthic habit or becoming at least partly pelagic while several of the thoracotrematous groups have invaded the terrestrial environment, the zoeal stages of both groups have retained the primitive planktonic dispersive role. Under these circum- stances, while the selective pressures operating on the adults might be expected to cause the two groups to diverge in features not directly related to the primary distinction between them, adaptation by their larval stages to the same life-style would presumably result in convergence. Assuming that my interpretation of primitive and advanced zoeal characters is correct (Rice, 19800, p. 299 et seq.), such convergence indeed seems to have occurred. Thus, although the most primitive zoeae of the Thoracotremata have a much simpler appendage 292 A. L. RICE setation than most zoeae of the Heterotremata, the same trends are apparent in both groups. In both cases the armature of the carapace and abdomen tends to become reduced, the separation of the sixth abdominal somite from the telson tends to become delayed, the setation of the cephalic appendages tends to become simplified, and there is some fusion of segments, particularly of the endopods of the maxillule and the third maxilliped. As a result of these trends, both sections of the Eubrachyura contain families with at least some representatives having zoeae in which some or all of the carapace spines are absent, the antennae are greatly reduced, the setation of the endopods of the maxillule and maxilla are greatly simplified, the segments of the endopod of the third maxilliped are partly fused and the sixth abdominal somite is fused with the telson throughout the zoeal phase. These conditions are found in the Leucosiidae amongst the Heterotremata and the Pinnotheridae amongst the Thoracotremata, producing a general resemblance between the two which led me to believe that they are closely related (Rice, 1980&). The same trends are apparent in a rather less extreme form in the advanced spider crabs (Inachinae) and the Dorippidae amongst the Heterotremata, and in the Hymenosomatidae amongst the Thoracotremata. However, apart from the advanced features which they share, these groups are all very distinct, four of them, for instance, having the most characteristic telson structures of any of the brachyurans, quite different both from each other and from those found in any other families. They each appear to represent the end point of a different phylogenetic line and suggest that the evolution of each of the sections of the Eubrachyura has involved several of these lines, though not all of them, perhaps, have produced such advanced zoeae. Using only the zoeal characters, in my earlier paper I attempted to trace these lineages amongst the forms which I then considered to be the cyclometopous families, that is the Heterotremata excluding the Leucosiidae and the Dorippidae. Although the philosophy behind this attempt was incorrect in that I hoped to be able to extend the lines into the more advanced zoeal groups, the general conclusions, summarized here in somewhat simplified form in Fig. 3, are still probably valid. From an ancestral form with a zoea similar to the most primitive of the extant xanthids, two major lines are envisaged, one leading to the Portunidae and Geryonidae and thence to the Parthenopidae, and the other to the Corystidae, Cancridae and part of the Atelecyclidae (that is the Corystoidea of Guinot) possibly via the Calappidae. A third major line, comprising the Majidae, is suggested as having separated from the ancestral stock at a level preceding that represented by the most primitive extant xanthids. Apart from the Leucosiidae and the Dorippidae, these three major lines together account for most of the Heterotremata. However, several groups of known zoeae do not fit readily into this simple pattern and indicate the existence of a number of subsidiary evolutionary lines. First, anagenesis in several families has resulted in sub-families with zoeal features which suggest that they are offshoots from the main lines. This seems to be true of the Carcininae and Portuninae within the Portunidae, and of the Pilumninae and Xanthinae within the Xanthidae. Within the Cancridae, two distinctly different types of larvae are found in the single genus Cancer and, according to the criteria I applied, the Corystidae could have been derived only from the more primitive of these. Secondly, some whole families appear to represent short side-branches; the Geryonidae seem to be an off-shoot from a polybiinid ancestor, while the Goneplacidae seem to be very closely allied to the Pilumninae. Thirdly, some heterotrematous zoeae are so unusual that I can only assume that they represent separate, independent lines. One such group is the Bellidae, regarded as a subfamily of the Atelecyclidae by Balss (1957), but completely separated from the Corystoidea by several authors and given superfamily status by Guinot (1978). In my review I discussed only the larvae of Corystoides and Heterozius, for at that time I was unaware of the excellent descriptions of the larvae of Acanthocyclus gayi Milne Edwards and Lucas by Fagetti & Campodonico (1970) and Acanthocyclus albatrossis Rathbun by Campodonico & Guzman (1973). These zoeae resemble those of Corystoides in all essential details, including the very unusual setation of the endopod of the second maxilliped, and confirm the necessity of separating the group totally from the Atelecyclidae. Two other genera which BRACHYURAN CLASSIFICATION Cancridae Parthenopidae Corystidae Portunidae Majidae Atelecyclidae Geryonidae ?Calappidae 293 Fig. 3 Main suggested phylogenetic lines within the Heterotremata. have usually been placed in the Atelecyclidae, Telmessus and Erimacrus, also have some unusual zoeal features which seem to separate them quite clearly from the remainder of the Corystoidea, but do not obviously ally them with any other group. These features include the appearance of the 'exopod' seta on the maxillule in the first stage, the unusually large number of scaphognathite setae in this stage and the presence of lateral setae on the endopod of the first maxilliped. I was, and am, unable to say where these genera belong, but can only suggest that since some of their zoeal characters indicate that they may have abbreviated a longer ancestral series of zoeal stages, they may have evolved from close to the stock which gave rise to the majids by a similar change in developmental strategy. Finally, the zoeae of the monotypic genus Orithyia possess a combination of features quite unlike that of any other known crab. The genus was placed with some doubt in the Calappidae by Ihle (1918) while Guinot (1978) gave it separate family status in her Calappoidea. There is certainly no feature of the zoeae of Orithyia which would rule out the possibility of it being derived from the more primitive calappid zoeae such as those of Calappa or Hepatus. On the other hand, the two groups have little in common which would positively indicate such a relationship. Instead, as I pointed out in the earlier paper, Orithyia zoeae have a superficial resemblance to the dorippids in having very long spinulose dorsal and rostral carapace spines and extended telson forks. Orithyia and dorippid zoeae also share with the higher majids the rather unusual feature of only three medial setae on the basis of the second maxilliped, while the first zoeal stage in both Orithyia and in the spider crabs has rather more marginal setae on the scaphognathite than is usual in the Brachyura. I have linked this last character with the abbreviated development of the majids, and since Orithyia passes through only three zoeal stages the same may be true here. However, none of these features indicate any clear relationship for Orithyia and I am therefore quite unable to suggest where it belongs. I am similarly unable to place the Leucosiidae and Dorippidae into this scheme. Their zoeae are generally much more advanced than those of most other heterotrematous families and their specialized features, particularly their telsons, indicate that they occupy rather isolated positions at the ends of heterotrematous evolutionary lines. Since dorippid zoeae consistently have three setae on the basal segment of the endopod of the first maxilliped they presumably could not have been derived from the portunid-parthenopid branch which have only two setae in this position. Otherwise, however, both dorippid and leucosiid zoeae could have evolved from those of any of the heterotrematous groups. 294 A. L. RICE As noted above, in my earlier review I was unable to identify any possible phylogenetic lines within those brachyuran groups with relatively advanced zoeae. This was partly because I had expected to be able to extend the suggested heterotrematous lineages into the Thoracotremata. Treating the Thoracotremata as a distinct group, as I am here, it is still difficult to identify possible evolutionary lines within it, but some general points can be made. First, the Hymenosomatidae have a number of very advanced zoeal characters which in the past led me to believe that they are fairly closely related to the Pinnotheridae and the Leucosiidae (Rice, 19800, p. 3 1 5). However, since the Leucosiidae are here considered to be advanced Heterotremata, a close relationship between them and the hymenosomatids is precluded. Similarly, although the zoeae of the Hymenosomatidae share with those of the Pinnotheridae a number of advanced features, including reduced carapace spines with the laterals, where present, close to the ventro-lateral margin, reduced antennal exopod, reduced setation of the maxillule and maxilla, and the failure of the sixth abdominal somite to become separated from the telson in all known hymenosomatids and several pinnotherids, the Hymenosomatidae have a number of much less advanced features which argue against a close relationship between the two families. Thus, the proximal segment of the endopod of the maxillule always carries a seta in hymenosomatids but is unarmed in pinnotherids, the endopod of the maxillule carries five setae in the hymenosomatids but only three in the pinnotherids, the endopod of the second maxilliped consists of three segments in the hymenosomatids but only two in the pinnotherids, and the basal segment of the endopod of the first maxilliped carries three setae in the hymenosomatids compared with two in the pinnotherids. In this last feature the Hymenosomatidae are unique amongst the Thoracotremata, suggesting that they could not have evolved from any of the extant groups. On the other hand, hymenosomatids have several very specialized features, particularly the reduced coxal endite on the maxilla, the failure to develop pleopods during the zoeal phase and a total absence of a megalopa stage, which indicate that they could not have been ancestral to any of the other extant groups either. I assume, therefore, that the hymeno- somatids are the sole extant representatives of a thoracotrematous evolutionary line which separated from the remainder at a very early stage. Like the Hymenosomatidae, the advanced Pinnotheridae have a number of specialized zoeal features, particularly the very characteristic telson, which suggest that no other extant group could have evolved from them. Moreover, all pinnotherid zoeae have the antennal exopod vestigial or absent, the basal segment of the endopod of the maxillule unarmed, only three setae on the endopod of the maxilla, and the endopod of the second maxilliped consisting of only two segments, the proximal being unarmed*. This combination of characters is more advanced than that of any thoracotrematous group and is approached only by the Ocypodinae in which, however, the antennal exopod is rarely rudimentary and the endopod of the second maxilliped always consists of three distinct segments. This resemblance does not, of course, necessarily indicate a close relationship, for there are considerable differences between ocypodinid and pinnotherid zoeae. Nevertheless, the appendage setation is so similar in the two groups that it seems likely that they both evolved from the same, or closely related, ancestors. The zoeae of the other ocypodid sub-families, that is the Macrophthalminae and the Scopimerinae, are both somewhat less advanced than those of the Ocypodinae. While either of these more primitive sub-families could have given rise to the Ocypodinae, neither of them could apparently have evolved from the other (see Rice, 19800, p. 344). Either or both of them must therefore be off the postulated line which led to the Ocypodinae and thence to the Pinnotheridae. A similar situation exists in the Grapsidae from which the Ocypodidae were possibly derived. Here, the subfamily Grapsinae contains the most advanced zoeae, derivable from any of the other sub-families. But the Sesarminae, Plagusinae and Varuninae each have *When preparing the general review (Rice, 1980#) I was unaware of the description of the larvae of Pinnixa rathbuni Sakai by Sekiguchi ( 1 978). BRACHYURAN CLASSIFICATION 295 different combinations of advanced and primitive characters which suggest that they each represent a different evolutionary line within the family (see Rice, 1980a, p. 340). Finally, the zoeae of the Gecarcinidae seem to be at an evolutionary level comparable with, or slightly below that of the less advanced sub-families of the Grapsidae and are there- fore the most primitive of the known thoracotrematous forms. This does not mean that the gecarcinids are ancestral to the remainder of the Thoracotremata, but it certainly suggests that they separated from the other evolutionary lines before the zoeae of the latter had attained their present forms. These suggested relationships between the zoeae of the Thoracotremata are far too vague to be formalized into any kind of phylogenetic diagram, even one as tentative as that produced above for the Heterotremata. Nevertheless, they do indicate relative evolutionary levels which may be useful in support of evidence from adult morphology and palaeontology. They also provide a framework to be strengthened or changed as each new piece of larval evidence is obtained, for like all larval studies, they are based on data from only a small proportion of the species known as adults. Conclusion Brachyuran zoeae can provide valuable insights into crab relationships at a variety of taxonomic levels. Since they are all adapted for a relatively similar pelagic existence rather than the very varied environments occupied by the adults stages, they may help to separate groups which have been classified together because of a superficial resemblance between the adults caused by convergence. At the highest level this is the case, for instance, of the Oxystomata and the Oxyrhyncha, while at a lower level the example of the Atelecyclidae sensu Balss (1957) might be cited. However, assuming that Saint-Laurent is correct in her interpretation of the Hetero- tremata and Thoracotremata as having had quite distinct evolutionary histories, this re-examination of the zoeal data has convinced me of a potential danger in the uncritical use of larval information. For groupings based on the larvae, such as those which I thought were recognizable within the Eubrachyura, may be just as misleading as those based on the adults. A major divergence amongst the adult forms, such as the suggested one between the Heterotremata and the Thoracotremata, may not be reflected in the larval stages since parallel adaptation to the same pelagic life-style may cause the zoeae of advanced members of both branches to share apomorphous characters which have apparently been acquired independently, as in the Leucosiidae and the Pinnotheridae. There remains the problem of explaining why the thoracotrematous zoeae are generally so much more advanced than those of most of the heterotrematous groups. One explanation might be that the early Thoracotremata, which presumably had zoeae at more or less the same evolutionary level as those of the bulk of the Heterotremata, have left no extant representatives. But this rather begs the question since it does not explain why these forms should have disappeared while the primitive Heterotremata survived. References Balss, H. 1957. Decapoda. VIII. Systematik. Bronns' Kl. Ordn. Tierreichs Bd. 5, Abt. 1, Buch 7, Leif. 19: 1505-1672. Campodonico, I. & Guzman, L. 1973. Contribucion a la biologia de Acanthocyclus albatrossis Rathbun 1898. An. Inst. Pat. 4 : 373-416. Fagetti, E. & Campodonico, I. 1970. Desarollo larval en el laboratorio de Acanthocyclus gayi Milne-Edwards et Lucas (Crustacea Brachyura: Atelecyclidae, Acanthocyclinae). Revta Biol. mar. 14:63-78. Guinot, D. 1978. Principes d'une classification evolutive des Crustaces Decapodes Brachyoures. Bull. biol. Fr.Belg. 112: 2\ \-292. Ihle, J. E. W. 1918. Die Decapoda Brachyura des Siboga-Expedition. III. Oxystomata, Calappidae, Leucosiidae, Raninidae. Siboga Exped. 39 : b2, 1-322. 296 A- L- Rice, A. L. 1980a. Crab zoeal morphology and its bearing on the classification of the Brachyura. Trans zool. Soc. Lond. 35 : 27 1^24. 1980/). The first zoeal stage of Ebalia nux A. Milne Edwards 1883, with a discussion of the zoeal characters of the Leucosiidae (Crustacea, Decapoda, Brachyura). J. nat. Hist. 14 : 331-337. (in press). The late embryo ofAcanthodromia erinacea A. Milne Edwards (Crustacea, Decapoda, Dynomenidae). J. Crustacean Soc. Saint-Laurent, M. de 1980a. Sur la classification et la phylogenie des Crustaces Decapodes Brachyoures. I. Podotremata Guinot, 1977, et Eubrachyura sect. nov. C.r. hebd. Seanc. Acad. Sci. Pans 290 (D) : 1265-1268. 19806. Sur la classification et la phylogenie des Crustaces Decapodes Brachyoures II. Heterotremata et Thoracotremata Guinot, 1977. C.r. hebd. Seanc. Acad. Sci. Paris 290 (D): 1317-1320. Sekiguchi, H. 1978. Larvae of a pinnotherid crab, Pinnixa rathbuni Sakai. Proc. jap. Soc. svst. zool. 15 : 36-46. Williamson, D. I. 1965. Some larval stages of three Australian crabs belonging to the families Homolidae and Raninidae, and observations on the affinities of these families (Crustacea: Decapoda). Aust. J. mar.freshw. Res. 16 : 369-398. 1976. Larval characters and the origin of crabs (Crustacea, Decapoda, Brachyura). Thalassia jugosl. 10:401-414. Manuscript accepted for publication 10 October 1980 British Museum (Natural History) 1881-1981 Centenary Publications Chance, change & challenge Two multi-author volumes from one of the foremost scientific institutions in the world. General Editor: P. H. Greenwood The Evolving Earth Editor: L. R. M. Cocks The Evolving Biosphere Editor: P. L. Forey In the first volume, The Evolving Earth, twenty scientists have been asked to review the present state of knowledge in their particular field, ranging from the origin of the Earth, through ocean sediments and soils to continental drift and palaeogeography. In the companion volume, The Evolving Biosphere, museum scientists have chosen an evolutionary concept — speciation, coevolution, biogeography etc. and related this to the group of animals or plants in which they are specialising. Thus beetles and birds exemplify sympatric and allopatric speciation, butterflies mimicry and certain fishes explosive evolution. In both volumes the text is supplemented by over one hundred specially-commissioned pieces of two-colour artwork. These two books will be invaluable to all sixth-form and undergraduate biology and geology students. The Evolving Earth: 276x219 mm, 280pp, 138 line illustrations, 42 halftones The Evolving Biosphere: 276x219 mm, approx. 320pp, 133 line illustrations Publishing: May 1981 Co-published by the British Museum (Natural History), London and Cambridge University Press, Cambridge. Titles to be published in Volume 40 Eugene Penard's Slides of Gymnamoebia: re-examination and taxonomic evaluation. By F. C. Page Japanese earthworms: a synopsis of the Megadrile species (Oligochaeta). By E. G. Easton Phylogenetic versus convergent relationship between biscivorous cichlid fishes from Lakes Malawi and Tanganyika. By M. L. J. Stiassny Miscellanea Miscellanea Printed by Henry Ling Ltd, Dorchester