Bulletin of the British Museum (Natural History) Puellina (Bryozoa; Cheilostomata; Cribrilinidae) from British and adjacent waters J. D. D. Bishop & B. C. Househam Zoology series Vol 53 No 1 24 September 1987 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in four dentine 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), 1987 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Mr J. F. Peake Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISBN 0565 050338 ISSN 0007-1498 British Museum (Natural History) Cromwell Road London SW7 5BD Zoology series Vol 53 No. 1 pp 1-63 Issued 24 September 1987 Puellina (Bryozoa; Cheilostomata; Cribri British and adjacent waters J. D. D. Bishop Departments of Palaeontology and Zoology, British Museum (Natural Hi' London SW7 5BD B. C. Househam Department of Oceanography, University of Liverpool, Brownlow Hill, Liverpool L69 3BX JM Contents Synopsis 1 Introduction Material and methods 2 Ovicell development 4 Systematic account 4 Generic and subgeneric definitions 4 P. gattyae (Landsborough, 1852) P. setosa (Waters, 1899) 11 P.modican.sp 16 P. praecox n.sp 21 P. directa n.sp 25 P. venwsta Canu & Bassler, 1925 28 P. innominata (Couch, 1 844) 33 P. bifida (d'Hondl, 1970) .... P. arrecta n.sp 43 P.flabellifera (Kirkpatrick, 1 888) . 48 P. corbula n.sp 52 P. pedunculata Gautier, 1956 55 Discussion 59 Acknowledgements .61 References 61 Synopsis Cribrilaria Canu & Bassler, 1929 is treated as a subgenus of Puellina Jullien, 1 886, and Glabrilaria n.subgen. is established within Puellina to accommodate P. pedunculata Gautier, 1956 and related species. The following nine species of Puellina are recognised from British and adjacent waters: P. arrecta n.sp., P. bifida (d'Hondt, 1970), P. corbula n.sp., P. directa n.sp., P. gattyae (Landsborough, 1852), P. innominata (Couch, 1844), P. modica n.sp., P. praecox n.sp., and P. venusta Canu & Bassler, 1925. A table of key characters for the identification of these species is given. The distribution of Puellina around the British Isles shows a marked southerly and westerly bias. There seems to be no evidence for the occurrence of P. setosa (Waters, 1899) further north than the southern part of the Bay of Biscay. A neotype is designated for P. gattyae. Lectotypes are selected for Lepralia steindachneri Heller, 1867 and L. cribrosa Heller, 1 867, both of which are junior synonyms of P. gattyae, and for P. setosa, P. bifida, P. pedunculata and P.flabellifera (Kirkpatrick, 1888). P. praecox n.sp. also occurs in the Plio-Pleistocene Red Crag of eastern England. P. venusta is found as a fossil in the Badenian (Middle Miocene) of the Vienna Basin. P. bifida is extremely similar to P. kollmanni (David & Pouyet, 1974) from the Badenian of the Vienna Basin. P. rarecostata (Reuss, 1848) from the same rocks is intermediate in appearance between the two Recent species P. setosa and P. modica. Three categories of ovicell development within the genus are noted, and their significance discussed. Bull. Br. Mus. not. Hist. (Zool.) 53(1): 1-63 Issued 24 September 1987 2 J. D. D. BISHOP & B. C. HOUSEHAM Introduction Perhaps because of difficulties of observation caused by their small zooid size, there has been a tendency in the past to assign most Recent Puellina or Cribrilaria specimens to a limited number of nominal species within which extensive morphological variation has been accommodated. The wide limits of variation assigned to Cribrilina radiata by Hincks (1880) is a notable example of this approach. The subsequent task of refining species concepts within the group has been given impetus by the increased morphological detail available through the use of the Scanning Electron Microscope (SEM). This work was initiated by Harmelin (1970) in a revision of Mediterranean species and has been continued by the same author (1978; 1984) for NE. Atlantic and Mediterranean faunas. The present paper is an account of Recent species from British and adjacent waters, in which mention is also made of relevant fossil material from the European Neogene. Following Gordon (1984), Cribrilaria is treated as a subgenus of Puellina. A new subgenus within Puellina is established for certain species formerly placed in Cribrilaria. Material and methods Material studied during the present work is located in the following institutions: BMNH: British Museum (Natural History), London, U.K.; Recent specimens, Zoology Department; fossil specimens, Palaeontology Department. IP: Institut de Paleontologie, Museum National d'Histoire Naturelle, Paris, France; Canu Collection. IZUI: Institut fur Zoologie der Universitat Innsbruck, Innsbruck, Austria; Heller Collection. Table 1 Details of Channel Stations Station Latitude Longitude Depth Deposit 183 50°39-3'N 01°22-6'E 48m Clean coarse sand 230 49°36-3'N 03°26-7'W 75m Coarse shell sand 240 49°30-4'N 01°59-0'W 29m Coarse shell sand 247 49°10-8'N 02°52-1'W 64m Clean fine shell gravel, some stones 248 49°08-1'N 03°02-3'W 73m Clean fine shell gravel, some stones 249 49°05-1'N 03°13-5'W 77m Clean fine shell gravel, some stones 250 49°02-5'N 03°25-0'W 80m Stones, some clean fine shell gravel 251 49°00-0'N 03°36-5'W 77m Coarse shell sand, some stones 252 48°57-0'N 03°48-7'W 82m Clean shell gravel, some stones 253 48°52-2'N 03°54-0'W 79m Clean coarse sand, some gravel 256 48°50-7'N 04°13-1'W 91m Clean coarse shell sand, some stones 257 48°48-7'N 04°22-5'W 96m Clean coarse shell sand, some stones 258 48°46-2'N 04°30-7'W 99m Clean coarse shell sand, some stones 260 48°51-5'N 04°42-0'W 106m Coarse silty sand, some gravel and stones 261 49°00-0'N 04°44-5'W 101m Clean coarse shell sand, some gravel and stones 263 49°15-2'N 04°50-0'W 102m Coarse silty shell sand, some shells 275 49°38-2'N oroo-o'w 38m Clean coarse sand, some gravel and stones 314 49°12-4'N 03°14-8'W 73m Clean shell gravel, some stones 315 48°57-5'N 04°00-7'W 88m Clean fine shell gravel, some stones 316 49°07-2'N 04°02-5'W 86m Stones, some clean fine shell gravel 317 49°18-0'N 04°04-2'W 86m Shell gravel, some stones 318 49°27-6'N 04°07-0'W 86m Clean coarse shell sand, some stones BRITISH PUELLINA Fig. 1 Measurements on autozooids, interzooidal avicularia (bottom left) and erect avicularia (bottom right). Abbreviations: Lav = length of rostrum and frontal non-calcified area plus proximal gymnocyst of avicularium; Lf = length of frontal shield of autozooid; If = width of frontal shield of autozooid; Lfg = length of frontal shield plus gymnocyst of autozooid; Lo = length of orifice of non-ovicellate autozooid; lo = width of orifice of non-ovicellate autozooid; loo = width of orifice of ovicellate auto- zooid; Lov = length of ovicell; lov = width of ovicell; Lr = length of rostrum plus frontal non-calcified area of avicularium. LBIMM: Laboratoire de Biologic des Invertebres Marins et de Malacologie, Museum National d'Histoire Naturelle, Paris, France. MM: Manchester Museum, Manchester, U.K.; microslide preparations from the Waters Collection. NMV: Naturhistorisches Museum, Vienna, Austria. Many of the records presented here are from encrusted shells collected during 1960 to 1962 for a study of the Recent benthos of the English Channel by Holme (1966), subsequently used for an account of bryozoan assemblages by Grant & Hayward (1986), and now deposited in the collec- tions of the BMNH. Station details taken from Holme (1966) are given in Table 1. In the list of material examined for each species, these stations are referred to simply as 'Channel Stn . . .'; they are mostly on the French side of the Channel, with a few in mid-Channel. Locality details of other BMNH material are given as they appear on specimen labels or in the specimen catalogue; any additional information is given in square brackets. SEM observations were made on uncoated material using an International Scientific Instru- ments model 60 A with charge-free anticontamination system (CFAS). Images were generated at a 4 J. D. D. BISHOP & B. C. HOUSEHAM working voltage of 1 5 or 25 KV from back-scattered electrons using a Robinson detector. The magnifications of the resulting micrographs were calculated by light microscope calibration of each specimen, since the readings given by the SEM itself were inaccurate. Stereo pairs were taken at an angular separation of 10°. Morphological measurements were taken from scanning electron micrographs. The measure- ments, and the abbreviations employed for them, are explained in Figure 1. Orifice width was measured to the proximo-lateral corners of the orifice. The papilla pores were taken to mark the edge of the frontal shield. Ovicell length was measured to the distal margin of the ovicell opening. The apertural bar was not included in counts of the number of costae. Values of the coefficient of variation (C.V.) were corrected as described by Simpson, Roe & Lewontin (1960:101, 102). The term pluriserial is used in the descriptions below to indicate encrusting colonies forming lobes of fewer than about 10 autozooids abreast (Figs 29, 74). Multiserial colonies consist of broader lobes or extensive sheet-like growths. Ovicell development In Puellina, and in many other cribrimorphs, the process of colony growth is apparently dis- continuous, with pauses between the budding of successive zooid generations at the edge of the colony (Lidgard, 1985). Based on the timing of production of the ovicell in relation to the budding of the maternal autozooid and of the zooid distal to it, the following three categories of ovicell may be recognised in the species discussed here: A — The ovicell is produced during the budding of an autozooid distal to the maternal autozooid, and is a proximal component of the distal autozooid. Thus the completed ovicell never lies at the extreme edge of the colony. The frontal walls of the ovicell and of the distal autozooid are continuous. See Figures 40, 46, 50, 62 and 79. B — The ovicell is produced during the budding of a kenozooid (which has an area of costate frontal shield) distal to the maternal autozooid, and is a proximal component of the kenozooid. If the area of costate frontal shield is relatively small, the newly completed ovicell may lie very close to the extreme edge of the colony. The frontal walls of the ovicell and of the kenozooid are continuous. See Figures 70, 94, 95 and 96. C — The ovicell appears to be a distal component of the maternal autozooid, and is not associated with an area of kenozooidal costate frontal shield. Thus the newly completed ovicell lies at the extreme edge of the colony. Zooids distal to the maternal autozooid and ovicell are budded subsequently, and their frontal walls are separate from that of the ovicell. See Figures 2, 19, 23, 33 and 83. Ovicells of category C of certain species have a number of uncalcified subcircular spots on the frontal surface (Figs 2, 12), similar in appearance when viewed by SEM to the pelmatidia of the costate frontal shield. These occur in a more or less regular arrangement, generally approximating to an outer ring of spots sometimes enclosing a number of others. This condition is referred to as punctate in the relevant descriptions below. In ovicells of category C of other species, and of categories A and B, small uncalcified patches sometimes occur on the frontal surface (Figs 50, 73), particularly on the crests of ridges and the tops of tubercles. However the size, shape and occurrence of these patches are inconstant and their spatial arrangement is relatively haphazard, and in such cases the ovicell is described below as non-punctate. Systematic account Genus PUELLINA Jullien, 1886 Type species: P. gattyae (Landsborough, 1852), by original designation. Jullien (1886), in addition to establishing Puellina, selected Eschara radiata Moll, 1803 as the type BRITISH PUELLINA 5 species of Cribrilina Gray, 1848. Since Lepralia punctata Hassall, 1841 is the type species of Cribrilina by monotypy, Jullien's selection was invalid, but it did serve to indicate clearly that he excluded E. radiata from the genus Puellina. However, E. radiata is the type species of Cribrilaria Canu & Bassler, 1929, which is treated in the present paper as a subgenus of Puellina. The definition of the genus Puellina adopted here is therefore broader than Jullien's (1886), but coincides with that of Levinsen (1909), who referred E. radiata to the genus, and of Gordon (1984), who included Cribrilaria in Puellina as a subgenus. Jullien's original concept of the genus was in fact equivalent to that of the subgenus Puellina employed here. REVISED DIAGNOSIS. Autozooid with pericystal frontal shield of fused but discernible spines ( = costae) in more or less pinnate or radiating pattern; each costa generally with 1 or more small pelmatidia ( = lumen pores); fusion of successive costae interrupted by 1 or more lacunae ( = inter- costal pores) which are short (i.e. do not take a slit-like form elongate parallel to the costae); non-calcified papilla (presumed to be evagination of frontal membrane) protruding from single papilla pore (bordered on one side by gymnocyst) between bases of successive costae around margin of frontal shield. Calcified (secondary) orifice D-shaped, with straight proximal edge, abutting similarly-shaped operculum around its entire margin; orifice without condyles, opercu- lum hinged at extreme proximal edge. Apertural bar (presumed to be modified distalmost pair of costae) just proximal to orifice usually with 1 or more sub-oral ( = sub-orificial) lacunae on midline between left and right components of bar. Oral ( = orificial) spines erect, unfused, jointed at or near base, present around lateral and distal margins of orifice in non-ovicellate autozooids, restricted to lateral margins of orifice in ovicellate zooids. Autozooids with several pore chambers around distolateral and distal margins. Ovicell sub-globular, hyperstomial, cleithral. Avicularium (if present) with discrete gymnocystal cystid. Ancestrula tatiform, without pronounced cryptocyst, frequently regenerated as zooid with cribrimorph frontal shield. The possession of non-calcified papillae protruding through papilla pores around the periphery of the frontal shield is regarded here as a synapomorphy of the genus Puellina. Subgenus PUELLINA Jullien, 1886 Puellina Jullien, 1886; 607. Puellina: Prenant & Bobin, 1966; 597. Hayward & Ryland, 1979; 66. Harmelin, 1984; 81. Type species: P. gattyae (Landsborough, 1852), by original designation. Other NE. Atlantic species included: P. setosa (Waters, 1899), P. modica n. sp. Proximal gymocyst of autozooid relatively broad; pericyst relatively small (approximately circular in outline). Oral spines without apophyses, each with brown chitinous articulation at base; 2 spines in ovicellate zooids (5 in non-ovicellate zooids). Setiform papillae from most distal pair of papilla pores. External uncalcified windows of pore chambers relatively small. Ovicell of category C (p. 4), punctate. No avicularia. Median proximal spine of ancestrula with simple (not bifid) tip; ancestrula regenerated as miniature autozooid (or kenozooid). Subgenus CRIBRILARIA Canu & Bassler, 1929 Cribrilaria Canu & Bassler, 1929; 33. (Cribrilaria Canu & Bassler, 1928; 27 is regarded as a nomen nudum.) Cribrilaria: Hayward & Ryland, 1979; 62. Gordon, 1984; 63. Type species: P. radiata (Moll, 1803), by original designation. NE. Atlantic species included: P. venusta Canu & Bassler, 1925, P. innominata (Couch, 1844), P. bifida (d'Hondt, 1970), P. hexaspinosa (Harmelin, 1978), P. octospinosa (Harmelin, 1978), P. arrecta n.sp. Proximal gymnocyst of autozooid relatively narrow; pericyst relatively large. Oral spines with or without apophyses, bases not brown; 4 in ovicellate zooids. Setiform papillae from most distal pair of papilla pores. External uncalcified windows of pore chambers relatively large. Ovicell of cate- gory A (p. 4) (sometimes B in P. bifida), non-punctate. Avicularia interzooidal. Median proximal spine of ancestrula generally with bifid tip; ancestrula regenerated as kenozooid. J. D. D. BISHOP & B. C. HOUSEHAM _0 o o jy o "I oo X Id C 00 3 x 03 „ ^ c a ed O a-6 o o o ,0 BRITISH PUELLINA 1 Subgenus GLABRILARIA n.subgen. Type species: P. pedunculata Gautier, 1956. Other species included: P. africana (Hayward & Cook, 1983), P. corbula n.sp., Cribrilaria pedunculata in the sense of Harmelin (1968) (not Gautier, 1956), P. cristata (Harmelin, 1978). Proximal gymnocyst of autozooid relatively narrow; pericyst relatively large. Oral spines without apophyses, bases not brown; 4 in ovicellate zooids. Ordinary (short) papillae from most distal pair of papilla pores. External uncalcified windows of pore chambers of moderate size. Ovicell of category B or C (p. 4), non-punctate. Avicularia small, erect or semi-erect, generally budded from single pore chamber. Median proximal spine of ancestrula with simple (not bifid) tip; ancestrula regenerated as kenozooid. ETYMOLOGY. From a modification of Cribrilaria, from which the new subgenus is distinguished, combined with the Latin glaber, 'without hair' or 'smooth skinned', a reference to the absence of setiform papillae in this group. Pucllina (Puellina) gattyae (Landsborough, 1852) (Figs 2-10 and Table 2) Lepralia gattyae Landsborough, 1852; 326, 327, pi. 18 fig. 71. Lepralia gattyae: Busk, 1854; 73, 74, pi. 83 fig. 6. Lepralia steindachneri Heller, 1867; 109, 1 10, pi. 2 fig. 5. Lepralia cribrosa Heller, 1867; 109, pi. 2 fig. 6. Lepralia gattyae: Waters, 1879; 36, pi. 9 fig. 6. Cribrilina gattyae: Hincks, 1880; 198, 199, pi. 25 fig. 10. not Cribrilina gattyae: Norman, 1903; 98 (as 'Gattyae'') 'form encrusting a shell', pi. 9 fig. 5. not Puellina gattyae var. balearica Barroso, 1919; 340, 341 , figs 1-5 ( = Collarina sp.). Puellina gattyae: Prenant & Bobin, 1966; 598-600, fig. 2081. Puellina gattyae: Hayward & Ryland, 1979; 66, 67, fig. 18A-D. Puellina gattyae: Harmelin, 1984; 81-82, 85-87, figs 2-4, 7. MATERIAL EXAMINED BY SEM. Neotype (designated here): BMNH 1854.1 1.15.125 (Sidmouth, [South] Devon). BMNH: 1 897.5. 1 .623 (South Devon, three colonies); 1 899.7. 1 . 1 3 1 7A (Sidmouth); 1911.10.1 .735A (Belfast, on shell). IZUI: 135/9855 (Adriatic Sea, lectotype of Lepralia steindachneri Heller, 1867, selected here); 125/17021 (Adriatic Sea, lectotype of Lepralia cribrosa Heller, 1867, selected here). OTHER MATERIAL. BMNH: 1879.4.25.17 (Naples); 1899.5.1.717 (Birterbuy [ = Bertraghboy] Bay [Galway], labelled Cribrilina punctata var., colony on shell fragment 1); 1899.7.1.1317 and 1899.7.1.1318 (Sidmouth); 1899.7.1.1319 (Weymouth); 1911.10.1.734 (Naples); 1911.10.1.736 (British); 1975.1.12.444, 1975.1.12.518-522, 1975.7.1.5 and 1975.7. 1.10 (Chios). MM: 987 and 988 (Naples); 992 and 1171 (Menton). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material multiserial, with a few hundred autozooids. Autozooid of variable shape in frontal view, generally sub-oval (longer than broad). Lateral gymnocyst of autozooid narrow or of moderate width; proximal gymnocyst broad, with proximal extension(s) between neighbouring zooids; gymnocyst with fine radiating striations. Frontal shield ( = pericyst) generally almost circular in outline, moderately to heavily calcified, slightly to strongly convex, with 5-9 costae (usually 7 or 8) fused extensively in centre of shield; central region of costal fusion may form blunt median umbo (frequently well developed in ovi- cellate zooids; Fig. 2). Each costa generally without clearly differentiated basal and subsequent portions, without distinct ornamentation; minute pelmatidium frequently seen near base, and larger pelmatidium occasionally near centre of shield. Intercostal pores subcircular, of variable size, without denticles, usually only 1 pore between successive costae. Lipped papilla pore between bases of successive costae, larger than adjacent intercostal pore. J. D. D. BISHOP & B. C. HOUSEHAM BRITISH PUELLINA 9 Orifice of non-ovicellate autozooid D-shaped, clearly broader than long, proximal edge straight. Orifice of ovicellate zooid not appearing truncate in frontal view; width 1-1-1-2 times that of orifice of non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid (6 in one disrupted autozooid of neotype), 2 in ovicellate zooid; with brown chitinous joint at base but without subsequent segmen- tation, without apophyses. Apertural bar sloping upwards to variable extent from proximal mar- gin of orifice to first pair of costae; without tubercles; median suture and minute pelma- tidium either side of midline occasionally discernible. Single median sub-oral lacuna, of variable shape and size although generally slightly larger than an intercostal pore, without denticles, flanked on either side by single intercostal pore (occasionally 2) slightly proximal to it. Each autozooid with 6-8 distolateral pore chambers; uncalcified external windows of pore chambers (seen in zooids on edge of colony: Fig. 4) relatively small (about size of sub-oral lacuna). Ovicell of category C (p. 4), length 1-0-1-2 times that of frontal shield of autozooid; frontal surface punctate, without ridges or discernible median suture. Avicularium not present in available material. Ancestrula (Fig. 9) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Erect spines jointed at base; details of subsequent jointing (if any) not seen. Median proximal spine blunt-tipped and of relatively uniform thickness throughout length. Spines without inwardly directed (or other) apophyses. Ancestrula sometimes regenerated as small autozooid with 3 or 4 oral spines (Fig. 10). Table 2 Puellina gattyae (Landsborough): measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 47-51 48-7 3-17 9 Lf 99-150 130 8-87 24 b 45-53 48-1 6-32 8 124-155 140 10-0 10 c 42^9 45-5 12-3 2 88-97 94-0 5-99 3 d 43^7 44-4 4-43 7 100-148 124 11-5 15 a lo 59-66 62-7 4-29 6 Lfg 193-301 235 13-2 24 b 65-70 67-6 3-55 8 237-390 279 19-3 10 c 56-60 58-0 5-49 2 193-222 211 8-18 3 d 57-75 64-4 15-8 5 218-346 273 14-9 14 a loo 70-87 77-8 7-63 4 If 115-169 140 10-5 24 b 85-88 86-5 2-76 2 108-147 132 8-80 10 c — — — 0 84-111 94-3 16-8 3 d 72-87 80-3 7-21 7 93-157 125 18-3 15 a Lov 123-155 138 8-14 10 lov 158-190 176 4-96 10 b 144-155 147 2-64 8 166-187 176 4-48 9 c — — — 0 — — — 0 d 134-171 148 9-68 6 154-179 165 5-76 7 a = neotype (BMNH 1854.11.15.125); b = BMNH 1911.10.1.735A; c = IZUI 9855 (lectotype of Lepralia steindachneri Heller); d = IZUI 17021 (lectotype of Lepralia cribrosa Heller). Figs 5-10 Puellina gattyae (Landsborough): (5) non-ovicellate autozooid, BMNH 1897.5.1.623, South Devon, x 163; (6) ovicellate and non-ovicellate autozooids, BMNH 1911. 10.1. 735A, Belfast (on shell), x 113; (7) non-ovicellate autozooids, IZUI 135/9855, Adriatic (lectotype of Lepralia steindachneri), x 123; (8) ovicellate and non-ovicellate autozooids, IZUI 125/17021, Adriatic (lecto- type of Lepralia cribrosa), x 101; (9) ancestrula, BMNH 1897.5.1.623, South Devon, x 240; (10) ancestrula regenerated as autozooid, BMNH 1899.7. 1.1 3 17A, Sidmouth, x 290. 10 J. D. D. BISHOP & B. C. HOUSEHAM REMARKS. Type material of Puellina gattyae does not seem to have been formally recognised by previous authors. Despite this, there has apparently been little uncertainty concerning the identity of P. gattyae occurring on algae, although the synonyms steindachneri and cribrosa were created by Heller (1867), and Barroso (1919) erroneously assigned a new variety to the taxon. However there has been confusion, outlined below, regarding the occurrence of the species on hard substrata. P. gattyae also has at least two very close relatives, P. setosa (which has been the subject of consider- able taxonomic uncertainty itself) and P. modica n.sp. Therefore an evaluation of possible type material is now considered worthwhile. Landsborough (1852:326) introduced the name Lepralia gattyae, but attributed the name to Busk. Landsborough's description of the species was preceded by quotation marks and was stated ( 1 852:326) to be 'from the pen of Mrs. Gatty', who, furthermore, was credited with the discovery of the species. It had been found (1852:327) on algae 'during two successive winters, at Sidmouth (1851 and 1 852), and is to be met with also in Jersey'. Landsborough's figure (but not his text) was cited in Busk's (1854) account of L. gattyae; again the discovery of the species was credited to Mrs Gatty and material was reported from the same two localities, Sidmouth (Mrs Gatty) and Jersey (J. Alder). Hence it would appear that the original concept of the species employed by both Landsborough and Busk was due to Mrs Gatty, and that the name gattyae was probably suggested to Landsborough by Busk. Despite the fact that Landsborough's use of the name appeared first, Busk has been cited, incorrectly, as the author of Lepralia gattyae by several subsequent workers, including Hincks (1880), Jullien (1886), Waters (1899), Levinsen (1909), Barroso (1919) and Eggleston (1969). No collection of material associated with Landsborough's work is known. The Busk collection at the BMNH does not contain any material of the species from Jersey and only one microslide preparation, BMNH 1854.1 1.15.125, a single colony on alga from Sidmouth pro- vided by Mrs Gatty, is undoubtedly contemporary with Busk's catalogue (1854). The label reads: 'Sidmouth G L. ['coronata' crossed out] Gattyae PI. 83'. The letter 'G' on the label identifies Mrs Gatty as the source of the specimen, as may be confirmed by checking other microslides with this mark against the entries in Busk's catalogue; Flustra coronata Audouin had been likened by Busk (1854) to L. gattyae; plate 83 refers to the illustration of L. gattyae in Busk (1854). However, BMNH 1854.11.15.125 is not recognisable as the colony illustrated by either Landsborough or Busk. It is noted as 'type' in the MS species catalogue of the BMNH, but this supposed type status is presumed to stem from the specimen's connection with Busk's (1854) account rather than Landsborough's (1852). Although it is probable that this colony was part of the material on which Mrs Gatty based her description of the species published by Landsborough, it is not possible to prove this conclusively and it would not be valid to select it as lectotype. However, BMNH 1854.1 1.15.125 is clearly the best available indicator of the true nature of Landsborough's species, and in recognition of its significance it is selected here as neotype. The colony has lost much of its earlier growth, leaving two almost disjunct lobes totalling c. 200 autozooids. Examination of IZUI specimen 135/9855 (Fig. 7) confirmed the conspecificity of Lepralia steindachneri Heller, 1867 and P. gattyae; this was first suggested by Waters (1879) and has been generally accepted by later authors. The arrangement of zooids in IZUI 135/9855 is reminiscent of Heller's plate 2 figure 5, but is not precisely the same; it is not known if this is the illustrated specimen. However, no other material attributable to Heller's account of this species is known (W. Schedl, Institut fur Zoologie der Universitat Innsbruck, pers. comm. to JDDB, 1986) and specimen 135/9855 is hereby selected as lectotype of L. steindachneri, confirming the established synonymy. The identity of L. cribrosa Heller, 1867 has been uncertain in the past, but the name has been commonly cited as a junior synonym of Collarina balzaci (Audouin). However, Heller's figure (1867: pi. 2 fig. 6) shows setiform papillae, which are not present in C. balzaci auct. The type material received on loan (IZUI 1 25/ 1 702 1 ; Fig. 8) is in fact referable to P. gattyae. Once again, it is not known if this is the specimen illustrated by Heller, but no other material referable to Heller's account is available (W. Schedl, Institut fur Zoologie der Universitat Innsbruck, pers. comm. to JDDB, 1986). IZUI 125/17021 is therefore selected as lectotype of L. cribrosa, the name of which passes into the synonymy of P. gattyae. Heller's Adriatic material (of both nominal species) and BRITISH PUELLINA 1 1 the BMNH specimens from the Aegean (Chios) listed above have a relatively smaller, flatter, and more lightly calcified frontal shield than is generally seen in material from further west in the Mediterranean basin and from the NE. Atlantic. In the Chios specimens, the median umbo may be relatively narrow and pointed. Whilst there has been little confusion concerning the identity of P. gattyae material growing on algae, the name has been applied to several taxa encrusting shells etc., apparently in the belief that the appearance of the species is transformed on hard substrata. In fact P. gattyae on shell (BMNH 1899.5. 1 .717 and 191 1 . 10. 1 .735A: Fig. 6) looks much the same as on algae and has similar dimen- sions (Table 2). Norman (1903) contrasted a form he tentatively attributed to P. gattyae found on a shell from Guernsey with the form on algae. However, judging by the shape of the intercostal pores illustrated by Norman (1903: pi. 9 fig. 5), his shell form was certainly not P. gattyae. Grant & Hay ward (1986) recorded P. gattyae on shells from the English Channel, but the species has not been found on re-examination of their material. It would appear that their records of P. gattyae refer largely to material of P. modica n.sp. Other material on shell substrates in the collections of the BMNH originally registered under the name gattyae has been incorrectly identified with one exception, 1911.10.1.735A; the remainder comprises P. arrecta n.sp., P. bifida, P. corbula n.sp. and, most commonly, P. modica n.sp. P. gattyae has been widely reported in the Mediterranean, including the Adriatic and Aegean Seas. It is known from SW. England and the Channel Islands. The most northerly records that could be substantiated were both on shell, from Gal way (BMNH 1899.5.1.717) and Belfast (BMNH 1911. 10. 1.735A). Puellina (Puelttna) setosa (Waters, 1899) (Figs 11-1 9 and Table 3) not Cribrilina radiata var. a Hincks, 1 880; 186, 1 90 ( = Puellina modica n.sp.). not Cribrilina radiata 'with vibraculoid setae': Hincks, 1880; pi. 25 fig. 5 ( = Puellina innominatd). Cribrilina setosa Waters, 1899; 8, 9. not Puellina setosa: Osburn, 1950; 186, pi. 29 fig. 4. Puellina setosa: Harmelin, 1968; 1 196, 1 198, fig. 3(3, 4). not Cribrilaria setosa: Harmelin, 1970; part 88-91, fig. 1m, pi. 2 figs 4, 5 ( = C. cassidainsis Harmelin, 1984). not Cribrilaria setosa: Harmelin, 1978; 189 ( = C. cassidainsis Harmelin, 1984). Puellina setosa: Hayward & Ryland, 1979; 68, fig. 19A-D. Puellina setosa: Harmelin, 1984; 81-82, 85-87, figs 1, 5, 6, 14. MATERIAL EXAMINED BY SEM. Lectotype (selected here): MM 1173 (Madeira, the colony occupying an extensive section of the upper margin of the shell fragment and touching a balanid barnacle near the right (narrower) end of the substrate, viewed with the microslide's labels the correct way up). BMNH: 191 1.10.1. 727a-c (Fosse de Cap Breton, colonies on stone). MM: 1 173 (Madeira, two paralectotype colonies on same shell as lectotype); 1 137 (Naples, three colonies on alga). OTHER MATERIAL. BMNH: 191 1.10.1.727 (Fosse de Cap Breton, several colonies on stone). MM: 1173 (Madeira, one paralectotype colony on same shell as lectotype); 989 and 1006 (Naples, both labelled gattyae). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material multiserial, with several hundred autozooids. Autozooid of variable shape in frontal view, gener- ally sub-oval (longer than broad), occasionally proximally bifid. Gymnocyst of autozooid narrow distolaterally, of moderate width proximolaterally, and broad proximally, with proximal extension(s) between neighbouring zooids; gymnocyst often with fine radial striations. Frontal shield ( = pericyst) generally almost circular in outline, moderately calcified, with little or no differential median thickening, with 7-12 costae (usually 9 or 10). Each costa with broad, very short, moderately ascending basal portion at periphery of frontal shield leading to tapering, nearly horizontal subsequent portion contributing to slightly convex central region of shield. 12 J. D. D. BISHOP & B. C. HOUSEHAM Fig. 11 Puellina setosa (Waters): stereo pair of BMNH 191 1. 10. 1.727c, Fosse de Cap Breton, x73. Costae without distinct ornamentation; first pair slightly prominent (sometimes forming very small median umbo), remainder low. Each costa frequently with pelmatidium near centre of shield and smaller pelmatidium near base. Intercostal pores sub-rectangular or reniform, relatively uniform in size throughout shield, often with denticle (sometimes 2) adjacent to intercostal suture; usually 2 or 3, occasionally 4, pores between successive costae. Lipped papilla pore between successive costae, about same size as adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, slightly broader than long, proximal edge straight. Orifice of ovicellate zooid not appearing truncate in frontal view; width 1 -2-1 -3 times that of orifice of non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 2 in ovicellate zooid; with brown chitinous joint at base but without subsequent segmentation, without apophyses. Apertural bar sloping slightly upwards from proximal margin of orifice to first row of intercostal pores, with pelmatidium on either side of median suture; without tubercles. Median sub-oral lacuna, about same size as an intercostal pore, frequently with denticle (occasionally 2) on proximal margin; one colony, BMNH 191 1.10.1. 727c, frequently with second, smaller median lacuna proximal to first lacuna and immediately distal to first row of intercostal pores. Each autozooid with c. 8 distolateral pore chambers; uncalcified external windows of pore BRITISH PUELLINA 13 Figs 12-15 Puellina setosa (Waters): (12) ovicellate autozooids, BMNH 191 1 . 10. 1 .727b, Fosse de Caj Breton, x 127; (13) edge of colony viewed at angle, BMNH 1911.10.1.727c, Fosse de Cap Breton x70; (14) non-ovicellate autozooid, same colony, x 144; (15) ovicellate and non-ovicellat autozooids, lectotype (MM 1173), Madeira, x 101. 14 J. D. D. BISHOP & B. C. HOUSEHAM Figs 16-19 Puellina setosa (Waters): (16) kenozooid and non-ovicellate autozooids, BMNH 191 1.10.1. 727b, Fosse de Cap Breton, x 125; (17) ancestrula (distal end faces bottom left), paralecto- type (MM 1173), Madeira, x 190; (18) ancestrula regenerated as autozooid, lectotype (MM 1173), Madeira, x 290; (19) ovicellate and non-ovicellate autozooids, MM 1 137, Naples, x 60. chambers (seen in zooids on edge of colony: Fig. 13) relatively small (only slightly larger than an intercostal pore). Ovicell of category C (p. 4), length 0-8-1-0 times that of frontal shield of autozooid (sometimes 1-1 times in MM 1 137, Naples); frontal surface punctate, frequently with very low median ridge, without discernible median suture. Avicularium not present in available material. Kenozooid (Fig. 16) infrequent, of variable size and shape, with cribrimorph frontal shield of costae in radiating pattern. Ancestrula (Fig. 1 7) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Details of jointing (if any) of BRITISH PUELLINA 15 spines not seen. Median proximal spine broad near tip. Spines without inwardly directed (or other) apophyses. Ancestrula sometimes regenerated as small autozooid (BMNH 1911.10.1.727a and lectotype: Fig. 18) which may lack oral spines, or sometimes as kenozooid (MM 1137, largest colony). REMARKS. When Waters (1899) introduced the name Cribrilina setosa, he included only Cribrilina radiata var. a Hincks (1880) in the synonymy of the new species. It is therefore apparent that the specific name setosa was intended to replace Hincks' appelation var. a. However, there is no suggestion in Waters' unillustrated account that he actually examined material considered by Hincks to belong to var. a, and no specimen labelled C. radiata var. a is to be found amongst Hincks' material in the BMNH. The only indication concerning material available to Waters is as follows (1 899:8): 'From an examination of specimens from Madeira and Naples, I cannot agree in calling this form with vibraculae a variety of radiata.' No type material of Cribrilina setosa seems to have been formally recognised by later authors. The Madeiran collection which formed the main subject matter for Waters' paper was stated (1 899:7) to have been sent by J. Y. Johnson. Microslide MM 1 173 is labelled in Waters' hand as follows: 'Cribrilina setosa. Waters. Madeira. J.Y.J.'s coll. ovicell 0-24 & Crib. Balzaci no suboral pore'. There is thus little doubt that this was part of the material described by Waters, even though there is no reference to the publication and no desig- nation as type material on the labels. (A section on Cribrilina balzaci followed that on C. setosa in Table 3 Puellina setosa (Waters): measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 48-58 53-3 6-28 10 Lf 150-237 196 11-1 21 b 52-58 56-3 4-17 6 160-235 197 11-9 14 c 56-58 56-6 1-66 5 157-173 167 4-01 6 d 54-59 56-6 3-48 7 143-191 170 7-71 19 e 58-64 61-1 4-31 7 180-238 202 9-30 10 f 54-60 57-4 3-78 10 177-206 191 4-54 16 a lo 56-70 62-5 5-67 13 Lfg 201-367 280 15-4 21 b 58-66 63-3 5-83 8 242-397 301 14-2 12 c 62-73 67-4 6-67 5 294-324 308 4-65 5 d 71-78 74-6 4-63 5 266-347 305 7-66 18 e 62-78 71-6 7-34 8 242-374 318 14-8 8 f 69-75 71-8 3-61 10 272^18 328 11-3 16 a loo _ 0 If 140-207 181 11-3 21 b 77 — — 1 170-216 192 9-36 8 c — — 0 146-167 154 4-71 8 d 93-99 97-0 3-10 4 154-269 190 15-3 18 e 94-96 95-0 1-68 2 162-220 188 9-12 10 f — — 0 177-199 190 3-29 17 a Lov _ 0 lov — — — 0 b 125-158 150 9-93 5 177 — — 1 c — — 0 — — — 0 d 157-183 171 5-57 7 190-200 196 2-46 7 e 174-181 178 3-14 2 172-176 174 1-83 2 f — — 0 0 a = lectotype (MM 1173); b = paralectotype (MM 1173); c and d = MM 1137 (2 colonies); e = BMNH 1911.10.1.727b; f= BMNH 191 1.10.1.727c. 16 J. D. D. BISHOP & B. C. HOUSEHAM Waters' paper.) Microslide MM 1 1 37 is labelled in Waters' hand: 'Cribrilina ['setosa Wat.' crossed out] ['Gattyae' crossed out] Off Nisida Naples. 30 metres 30/10/['87' crossed out]79 5 spines without lateral avicularia or vibracula radiata'. There is therefore less evidence to connect this microslide directly with Waters' account, and Waters was apparently unsure of the identity of his Mediterranean material (which is nevertheless considered here to be conspecific with MM 1 173 from Madeira). No other material labelled setosa has been located in the Waters Collection. Accordingly, the lectotype (Figs 15, 18) has been selected from the colonies on MM 1 173. Waters' material (MM 1 173 and 1 137, including the lectotype) is not in fact considered to be the species named Cribrilina radiata var. a by Hincks (1880:186). The number of costae on the frontal shield of the material from Madeira and Naples clearly exceeds that specified in Hincks' account, the arrangement of costae does not recall that described by the earlier author, and the ovicell of Puellina setosa does not bear nodular projections. Hincks' variety is considered to be the form named Puellina modica n.sp. below (see the Remarks on that species). Osburn (1950) referred Cribrilina setosa to Puellina. Soule (1959) noted that at least part of Osburn's material had vicarious avicularia. The identity of the Pacific form with the specimens described by Waters is at best extremely doubtful, but Osburn's generic assignation is nevertheless correct with respect to Waters' material. The illustrations of P. setosa by Hay ward & Ryland (1979: fig. 19) are of BMNH 191 1 . 10. 1 .727. The ancestrula they figured (1979: fig. 19D) has in fact regenerated as an autozooid, as suggested by Harmelin (1984) and Aristegui (1984). A similar regeneration in the lectotype is illustrated in the present paper (Fig. 18). As noted by Waters (1899), P. setosa is very similar to P. gattyae. P. setosa may be distinguished from P. gattyae by the presence of more than 1 intercostal pore between successive costae; by the sub-rectangular or reniform shape of the intercostal pores, which have denticles (cf. rounded, without denticles); by the absence in all zooids of a well developed median umbo; and by the relatively less extensive proximal gymnocyst. Furthermore, the most distal pair of costae are slightly more prominent than the rest in P. setosa, a pattern not seen in P. gattyae. In the Mediterranean, well documented records of P. setosa exist from the French coast, Naples, the Sicilian Channel and Greece (Waters, 1899; Harmelin, 1968; 1984). The species was recorded from Madeira by Waters (1899) and from the Canaries by Aristegui (1984). Having discounted Cribrilina radiata var. a Hincks, 1880 (see above), the most northerly Atlantic record for P. setosa that could be substantiated was from Fosse ( = Gouf) de Cap Breton in the southern part of the Bay of Biscay (BMNH 191 1.10.1.727). In view of the uncertainty that has surrounded the identity of the species, the record from the Irish Sea near the Isle of Man by Eggleston ( 1 969) must be regarded as extremely doubtful. There seems to be no real evidence that the species occurs in British or adjacent waters. Puellina (Puellina) modica n.sp. (Figs 20-28 and Table 4) Lepralia innominata: Busk, 1854; pi. 86 fig. 3 ( = BMNH 1847.9.16.79). Cribrilina radiata var. a Hincks, 1880; 186, 190. not Cribrilina radiata 'with vibraculoid setae': Hincks, 1880; pi. 25 fig. 5 ( = P. innominata). Holotype: BMNH 1985.1 1.20.58 (Channel Stn 249, 49°05-1'N 03°13-5'W, 77 m). Paratypes (examined by SEM): BMNH: 1847.9.16.79 (British); 1936. 12.30.33 1C (no locality [probably Liverpool Bay]); 1936.12.30.335A-C (Liverpool Bay); 1985.1 1.20.46 (Channel Stn 230); 1985.1 1.20.69i and 1985.11.20.71b,c (Channel Stn 252); 1985.11.20.106a,b (Channel Stn 275); 1985.11.20.112b (Channel Stn 315). OTHER MATERIAL. BMNH: 1847.9.16.139 and 1847.9.18.66 (British); 1899.5.1.1028 (Antrim, colony on shell A); 1899.5.1.1089 (Antrim, colony on shell A); 1899.5.1.1387 (Antrim, deep water); 1911.10.1.737C,D (Guernsey); 1936.1 2.30.329 A-C (Isle of Man); 1963.2. 12.349 (Hastings); 1985. 11.20.50b,c and 1985.11.20.51 (Channel Stn 247); 1985. 11.20.52f and 1 985. 11.20.54h (Channel Stn 248); 1985.11.20.55 and 1985.11.20.57c (Channel Stn 249); 1985.11.20.60b, 1985.1 1.20.61b, 1985. 11.20.63c and 1 985. 11.20.64i-m (Channel Stn 250); BRITISH PUELLINA 17 18 J. D. D. BISHOP & B. C. HOUSEHAM 1985.1 1.20.65b, 1985.11.20.66 and 1985. 11.20.67c (Channel Stn 251); 1985.11.20.69j-m, 1985.1 1.20.70b-d, 1985.11.20.71d and 1985.11.20.721 (Channel Stn 252); 1985.11.20.85b, 1985.11.20.87c,d and 1985.11.20.88c-h (Channel Stn 257); 1985.11.20.91c-h, 1985.11.20.92b, 1985.11.20.93b, 1985.11.20.94e, 1985.11.20.95a-e, 1985.11.20.96b,c and 1985.11.20.98c (Channel Stn 258); 1985.11.20.100d (Channel Stn 260); 1985.11.20.104i (Channel Stn 263); 1985.11.20.105h,i and 1985.11.20.106c,d (Channel Stn 275); 1985.1 1.20.108e,f (Channel Stn 314); 1985.1 1.20.1 15 (Channel Stn 317); 1985.1 1.20.1 161 (Channel Stn 318). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material pluriserial to multiserial, with over 100 autozooids. Autozooid of variable shape in frontal view, generally round or sub-oval (longer than broad). Lateral gymnocyst of autozooid of moderate width or broad; proximal gymnocyst broad, frequently with short proximal extension(s) between neighbouring zooids; gymnocyst often with fine radial striations. Frontal shield ( = pericyst) generally almost circular in outline, moderately to heavily calcified, with little or no differential median thickening and no umbo, slightly convex, with 4-6 (most commonly 5) short tapering costae converging towards single point on midline. Each costa without clearly differentiated basal and subsequent portions; costal ridge, with longitudinal crest, occupy- ing almost entire length of costa and generally decreasing gradually in height from periphery to centre of shield; without tubercles. Most distal pair of costae forming very shallow V in frontal view, with their costal ridges sometimes continuous from side to side. Intercostal pores sub- rectangular or reniform, small, relatively uniform in size throughout shield, often with denticles adjacent to intercostal suture; 1 or 2 pores between successive costae. Lipped papilla pore between bases of successive costae, larger than adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, slightly broader than long, proximal edge straight. Orifice of ovicellate zooid not appearing truncate in frontal view; width 1 • 1-1 -2 times that of orifice of non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 2 in ovicellate zooid; with brown chitinous joint at base but without subsequent segmentation, slightly curved or straight, without apophyses. Apertural bar without marked proximal-distal slope; frequently with minute pelmatidium on each side of median suture; without tubercles. (Pelmatidia not observed elsewhere on shield.) Single, large, median sub-oral lacuna (Fig. 24), considerably larger than an intercostal pore, frequently with pair of denticles on proximal margin. Each autozooid with 5-8 lateral and distolateral pore chambers; uncalcified external windows of pore chambers, seen in zooids on edge of colony, relatively small (slightly smaller than sub-oral lacuna). Ovicell of category C (p. 4), length 1-2-1-5 times that of frontal shield of autozooid; frontal surface punctate, uncalcified spots being borne on c. 5 indistinct ridges in more or less radiating pattern or on tubercles; median suture occasionally discernible. Avicularium not present in available material. Kenozooid (Fig. 25) infrequent; single example seen, smaller than an autozooid, with rudimen- tary cribrimorph frontal shield of costae in more or less radiating pattern. Ancestrula (Figs 26, 28) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Erect spines jointed at base; subsequent segmentation not observed. Median proximal spine tapering towards tip (not bifid). Spines without apophyses. Ancestrula sometimes regenerated as autozooid (Fig. 27) or kenozooid. ETYMOLOGY. L. modicus — not very large, humble; referring to the modest size of both zooids and colonies of this species. REMARKS. In addition to the radiata and innominata forms, Hincks (1880) included three varieties, a, P and y, in his concept ofCribrilina radiata. Cribrilina radiata var. a was originally characterised (Hincks, 1880: 186) by its very small zooids, each with 5 very prominent costae on the frontal shield; the distal pair of costae formed a 'wall' proximal to the orifice; the ovicell bore nodular projections. Hincks also mentioned the presence of two slender vibraculoid setae (i.e. setiform papillae) in var. a, and this observation alone was taken to characterise the taxon by subsequent workers. Hincks made no reference to any particular figure in his account. Nevertheless a number of later authors seem to have taken Hincks' plate 25 figure 5 to depict var. a, although only Figs 23-28 Puellina modica n.sp.: (23) ovicellate and non-ovicellate autozooids at edge of colony, BMNH 1985.1 1.20.71c, Channel Stn 252, x97; (24) distal part of non-ovicellate autozooid, same specimen, x 310; (25) kenozooid, ovicellate and non-ovicellate autozooids, BMNH 1985.1 1.20.1 06a, Channel Stn 275, x 148; (26) ancestrula, holotype (BMNH 1985.1 1.20.58), Channel Stn 249, x 250; (27) ancestrula regenerated as autozooid, BMNH 1985.11.20.46, Channel Stn 230, x250; (28) ancestrula, BMNH 1985.11.20.69i, Channel Stn 252, x 181. 20 J. D. D. BISHOP & B. C. HOUSEHAM Table 4 Puellina modica n.sp.: measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 42-55 47-6 8-17 9 Lf 111-148 125 9-70 12 b 48-57 52-9 5-41 15 108-149 127 8-92 21 c 46-52 48-0 6-19 5 111-136 124 8-76 5 d 47-49 48-0 1-54 5 104-137 120 6-93 16 a lo 53-60 56-3 4-13 8 Lfg 178-272 229 13-6 12 b 56-65 59-9 4-92 17 192-319 234 12-7 20 c 57-64 60-6 4-68 5 171-226 202 12-5 5 d 59-64 61-7 4-42 3 177-247 203 9-82 16 a loo 61-69 65-0 9-79 2 If 89-118 105 8-44 12 b 69-72 69-6 2-03 5 93-124 105 8-17 20 c — — — 0 89-106 100 7-23 5 d 70 70-0 0-00 2 84-113 101 8-82 13 a Lov 137-141 139 2-30 2 lov 141-151 146 5-45 2 b 140-158 147 5-57 4 173-200 183 6-73 5 c — — — 0 — — — 0 d 137-151 143 4-81 4 135-167 153 8-66 5 a = holotype(BMNH 1985.11.20.58); b = BMNH 1847.9.16.79; c= 1936.12.30.335B;d = BMNH 1985.1 1.20.71c. Hayward & Ryland (1979) categorically stated this interpretation of the figure, while Prenant & Bobin (1966) seemed to attribute a similar opinion to Osburn (1950). The figure was apparently assumed to portray var. a because the legend stated [Cribrilina radiata] 'with vibraculoid setae'. Waters (1899) introduced the name Cribrilina setosa with the apparent intention of elevating Cribrilina radiata var. a Hincks, 1880 to specific rank. The name setosa has accordingly been used subsequent to its introduction for material considered to resemble Hincks' plate 25 figure 5 and with conspicuous setiform papillae. The interpretation of Hincks' plate 25 figure 5 as a depiction of Cribrilina radiata var. a in the sense of Hincks (1880) is not accepted here. As was noted by Hayward & Ryland (1979), the greater number and different disposition of costae in the figure are clearly at variance with Hincks' written account of the variety. The figure is in fact probably of Puellina innominata, an opinion shared by Harmelin (1970:85). Nor is var. a considered to be Puellina setosa as defined by the lectotype of P. setosa chosen above from Waters' material. Rather, Hincks' account is thought to have referred to Puellina modica n.sp., with which it agrees closely. It is not entirely clear whether Hincks regarded the 'vibraculoid setae' as unique to var. a. These structures are referred to elsewhere in his (Hincks, 1880) general account of C. radiata: p. 186 ('orifice semicircular, ... on each side of it a very delicate setiform appendage, often wanting') and p. 189 ('The small vibraculoid appendages are often wanting'). They are indeed shown by him in plate 25 figure 2, in the 'innominata form'. However, Hincks (1883:443) later referred to 'the form which bears vibraculoid setae' in a way which implied that he did consider that these structures characterised a single form or variety within his concept of C. radiata. In fact setiform papillae occur throughout the subgenera Puellina and Cribrilaria as defined above. Hincks (1880:190) indicated the provenance of his var. a material as 'Antrim, deep water (Hyndman)'. No specimen labelled C. radiata var. a has been found amongst Hincks' material in the BMNH. However, P. modica is to be found on a number of shells and stones (bearing other encrusting species and labelled with their names) from Antrim in the Hincks collection (e.g. BRITISH PUELLINA 21 BMNH 1899.5.1.1028 and 1899.5.1.1089). In particular, a colony of the species occurred on (but has since been accidentally detached from) a stone encrusted by the cyclostome Stomatoporina incurvata (Hincks, 1859) which forms part of the type series (BMNH 1899.5.1.1387) of the cyclos- tome and is clearly documented as material provided by Hyndman from deep water off Antrim. There seems little doubt that P. modica is the form intended by Hincks as C. radiata var. a. Puellina modica may be distinguished from its close relatives P. gattyae and P. setosa by its prominent costal ridges. Furthermore, each autozooid has relatively few (typically 5) costae, and the ovicell of P. modica has distinct ridges or tubercles whereas those of P. gattyae and P. setosa are almost smooth. P. modica is also very similar to P. rarecostata (Reuss, 1848) from the Badenian (Middle Miocene) of the Vienna Basin (material of P. rarecostata examined by SEM: two syntypes, both NMV 1848/38/76; BMNH D55591; all from Eisenstadt). However, P. rarecostata has less promi- nent costal ridges than P. modica and a smooth ovicell. The fossil species is in fact intermediate in appearance between P. modica and P. setosa. It is regarded as a member of the subgenus Puellina as defined above. The occurrence in P. modica, P. bifida and P. praecox n.sp. of ridges or tubercles on the ovicell and of costal ridges that are higher towards the edge of the frontal shield, and the absence of a median umbo in each, impart a similar appearance to these three species. In P. bifida and P. praecox, however, the exposed gymnocyst is less extensive than in P. modica, and the frontal shield more elongate and with more numerous costae. In addition, the oral spine bases are not brown. Unlike P. modica and P. praecox, P. bifida has category A or B ovicells (p. 4). P. modica is known from the English Channel off Sussex, Normandy, Brittany and Guernsey at depths (where known) of 38-106 m; from the Irish Sea; and 'off Antrim' (including material from 'deep water'). Puellina praecox n.sp. (Figs 29-35 and Table 5) Holotype: BMNH 1985. 1 1 .20.52e (Channel Stn 248, 49°08- 1'N 03°02-3'W, 73 m). Paratypes (examined by SEM): BMNH: 1911.10.1.1182A (off Antrim); 1936.12.30.335A-C (Liverpool Bay); 1985.1 1.20.61a (Channel Stn 250); 1985.1 1.20.72h (Channel Stn 252); D55583 (Red Crag, Plio- Pleistocene, Walton-on-the-Naze, Essex); D55584, D55585, D55586a,b and D55587 (Red Crag, Plio- Pleistocene, Shottisham, Suffolk) . OTHER MATERIAL. BMNH: 1936.12.30.332D (Liverpool Bay); 1936.12.30.333A (no locality [probably Liverpool Bay]); 1 985. 1 1 .20.9 1 i (Channel Stn 258). DESCRIPTION. Colony encrusting, unilaminar, minute, zooids distinct; largest colonies in available material pluriserial, with only c. 20 autozooids. Autozooid generally sub-oval in frontal view (longer than broad). Exposed lateral gymnocyst of autozooid narrow or absent (except at edge of colony); proximal gymnocyst narrow but occasionally with proximal extension(s) between neighbouring zooids. Frontal shield ( = pericyst) lightly calcified, without umbo or any other median thickening; with 8-10 costae each with short, broad, steeply ascending basal portion leading to tapering, nearly horizontal subsequent portion contributing to slightly convex central region of shield. Each costa generally with ridge, decreasing in height gradually from periphery to centre of shield, plus tubercle (occasionally with pelmatidium) near base; degree of development of ridge and tubercle variable. Intercostal pores of variable shape (circular, sub-oval, reniform or sub-rectangular), of variable size throughout shield but often larger in first intercostal row (i.e. just proximal to orifice) than elsewhere, sometimes with denticles adjacent to intercostal suture; 4 (occasionally 3) pores between adjacent costae. Lipped papilla pore between bases of successive costae, about size of adjacent intercostal pore. Orifice of autozooid D-shaped, slightly broader than long, proximal edge straight. Orifice of ovicellate zooid not appearing truncate in frontal view; width 1-1-1-2 times that of orifice of 22 J. D. D. BISHOP & B. C. HOUSEHAM BRITISH PUELLINA 23 Figs 32-35 Puellina praecox n.sp.: (32) whole colony, BMNH 1936.12.30.335C, Liverpool Bay, x 87; (33) ovicellate autozooid, same specimen, x210; (34) ancestrula and periancestrular autozooids, BMNH 1936.12.30.335A, Liverpool Bay, x 183; (35) whole colony, BMNH D55586b, Red Crag (Plio-Pleistocene), Shottisham, Suffolk, x 100. 24 J. D. D. BISHOP & B. C. HOUSEHAM Table 5 Puellina praecox n.sp.: measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 44-51 47-3 5-12 11 Lf 169-210 198 6-23 10 b 49-54 51-5 7-73 2 173-219 193 946 6 c — — — 0 180 — — 1 d 40-44 41-7 542 3 113-139 128 7-59 7 e 41-45 43-0 740 2 168-195 182 11-8 2 f — — — 0 164-189 179 7-27 4 g 38 — — 1 141-179 163 12-0 4 h — — — 0 137-163 150 13-8 2 a lo 53-59 56-6 3-83 11 Lfg 203-300 254 12-9 10 b 54-59 56-5 7-04 2 203-301 241 18-8 6 c — — — 0 210 — — 1 d 47-48 47-5 1-68 2 135-173 162 8-29 7 e 54-57 55-5 4-30 2 217-260 239 144 2 f 49-53 51-0 6-24 2 — — — 0 g 47 — — 1 181-231 212 11-8 4 h — — — 0 191-199 195 3-26 2 a loo 63-72 67-5 10-6 2 If 143-172 157 6-08 9 b 62-67 64-5 6-17 2 141-166 154 5-81 6 c 65-68 66-5 3-59 2 139-149 144 5-52 2 d 47-55 51-0 12-5 2 97-114 105 7-44 4 e 64 — — 1 139-148 144 4-98 2 f — — — 0 138-188 156 154 4 g 55 — — 1 145-162 155 5-10 4 h 55 — — 1 119-134 127 9.44 2 a Lov 128-148 138 11-6 2 lov 154-167 161 645 2 b 130-173 149 12-9 4 160-193 180 9-00 4 c 135-138 137 1-74 2 187-193 190 2-51 2 d 119 — — 1 124-147 134 941 3 e 118 — — 1 145 — — 1 f 155 — — 1 230 — — 1 g 158 — — 1 180 — — 1 h — — — 0 — — — 0 a = holotype (BMNH 1985.1 1.20.52e); b = BMNH 1936.12.30.335A; c = BMNH 1936.12.30.335C; d = BMNH 1985.11.20.61a;e = BMNH 1911. 10.1. 1182A; f=BMNH D55586a; g = BMNH D55586b; h = BMNH D55584. non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 4 in ovicellate zooid; jointed just above base but without subsequent segmentation, without apophyses. Apertural bar with low tubercle on each side of median suture, and pelmatidium on each tubercle. Median sub-oral lacuna larger than an intercostal pore, of variable shape, occasionally with denticles on proximal margin; occasion- ally a second lacuna proximal to main lacuna (Fig. 33), smaller than most intercostal pores, denticles not observed. Each autozooid with c. 5 distolateral pore chambers; uncalcified external windows of pore chambers, seen in zooids on edge of colony, relatively small (about the size of an intercostal pore). Ovicell of category C (p. 4), length 0-7-0-9 times that of frontal shield; frontal surface usually with 2-5 prominent tubercles, sometimes with uncalcified tops. Kenozooid not present in available material. BRITISH PUELLINA 25 Avicularium not present in available material. Ancestrula (Fig. 34) tatiform, with median proximal spine plus 5 pairs of spines around edge of gymnocyst (i.e. total of 11 spines). Spine morphology unknown. Ancestrula sometimes regenerated as kenozooid (Fig. 30). ETYMOLOGY. L. praecox — ripe before its time, premature, precocious; referring to the development of ovicells early in astogeny in this species. REMARKS. The dimensions of BMNH 1985.11.20.61a were noticeably smaller than those of the other material measured (Table 5). However, no further differences were noted between this specimen and the others. P. praecox n.sp. closely resembles P. bifida, especially in details of the frontal shield (including the apertural bar). However, the frontal shield of P. praecox is less heavily calcified than in P. bifida, with the costal ridges less well developed. The oral spines of P. praecox lack apophyses. The ovicell is of category C (p. 4) (cf. category A or B in P. bifida) and typically bears tubercles rather than ridges. The two species may occur at the same locality (Channel Stns 248, 250 and 252) and even on the same shell substrate (BMNH 1985. 1 1 .20.52 and 1985. 1 1 .20.72). Puellina praecox would be a typical member of the subgenus Cribrilaria as defined above but for its category C ovicell. The absence of avicularia may be regarded as a correlate of the small colony size reached by this species. P. praecox is known from the English Channel off Brittany at depths of 73-99 m; from the Irish Sea; from off Antrim; and as a fossil on shells from the Red Crag (Plio-Pleistocene) of Essex and Suffolk (North Sea basin). Puellina directa n.sp. (Figs 36-42 and Table 6). Holotype: BMNH 1985.1 1.20.76b (Channel Stn 253, 48°52-2'N 03°54-0'W, 79 m). Paratypes (examined by SEM): BMNH: 1985. 11.20.64e (Channel Stn 250); 1 985. 11.20.76c,d (Channel Stn 253); 1985. 11.20.94d (Channel Stn 258); 1985.11.20.109cand 1985. 11. 20. HOa (Channel Stn 3 14). OTHER MATERIAL. BMNH: 1985.1 1.20.54f,g (Channel Stn 248); 1985.11.20.56a,b and 1985.11.20.59h (Channel Stn 249); 1985. 11. 20.62 and 1985. 11.20.64f,g (Channel Stn 250); 1985.1 1.20.65a and 1985.1 1.20.67b (Channel Stn 251); 1985.11.20.69e,f and 1985.1 1.20.721 j (Channel Stn 252); 1985.11.20.83d-f (Channel Stn 256); 1985.11.20.90c,d and 1985.11.20.97b (Channel Stn 258); 1985.11.20.105e-g (Channel Stn 275); 1985.1 1.20. 108g and 1985.1 1.20.1 lOb (Channel Stn 314). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material pluriserial to multiserial, with c. 100 autozooids. Autozooid of variable shape in frontal view, generally sub-oval (longer than broad). Lateral and proximal gymnocyst of autozooid broad, frequently with short proximal and proximo-lateral extensions between neighbouring zooids. Frontal shield ( = pericyst) moderately to heavily calcified, with little or no differential median thickening, strongly convex, with 7-10 costae (usually 7 or 8). Each costa without clearly differ- entiated basal and subsequent portions; costal ridge (with narrow longitudinal crest) occupying almost entire length of costa. Most distal two (occasionally three) pairs of costae usually almost parallel, with left and right members of each pair abutting on midline and forming straight line across shield; ridges of these distal costae especially prominent (with crest often irregularly serrate), sometimes continuous from one side of shield to other but more commonly interrupted by narrow V-shaped notch at midline (the latter arrangement especially noticeable when viewed obliquely: Fig. 38). Intercostal pores sub-rectangular or reniform, relatively uniform in size throughout shield, often with denticles adjacent to intercostal suture; usually 3 or 4 pores between successive costae, but frequently fewer between short proximal costae. Lipped papilla pore between bases of successive costae, about same size as, or slightly larger than, an intercostal pore. Orifice of non-ovicellate autozooid D-shaped, clearly broader than long, proximal edge straight. Orifice of ovicellate zooid not appearing truncate in frontal view; width c. 1 -3 times that of orifice of 26 J. D. D. BISHOP & B. C. HOUSEHAM Fig. 36 Puellina directa n.sp.: stereo pair of BMNH 1985.1 1.20.1 lOa, Channel Stn 314, x 103. non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 2 in ovicellate zooid; jointed some distance above base (Fig. 38), without apophyses. Apertural bar sloping upwards from proximal margin of orifice to first row of intercostal pores; with tubercle each side of median lacuna; minute pelmatidium on each tubercle, and median suture, occasionally discernible. (Pelmatidia not seen elsewhere on shield.) Single median sub-oral lacuna larger than an intercostal pore, frequently with 1 or 2 denticles (which may subdivide lacuna) on proximal margin. Each autozooid with 5-8 distolateral pore chambers; uncalcified external windows of pore chambers (seen in zooids on edge of colony) of variable size, generally rather small but con- siderably larger than an intercostal pore. Ovicell of category A (p. 4), length 0-7-1-1 (usually c. 0-9) times that of frontal shield of autozooid; frontal surface non-punctate, generally with median ridge (occasionally with concurrent suture) plus 1 or 2 shorter ridges on each side radiating from median ridge. Avicularium not present in available material. Kenozooid (Fig. 39) infrequent, up to size of autozooid, with cribrimorph frontal shield of costae in radiating pattern; gymnocyst broad. BRITISH PUELLINA 27 Figs 37^12 Puellina directa n.sp.: (37) whole colony, holotype (BMNH 1985. 1 1 .20.76b), Channel Stn 253 x 38- (38) edge of colony viewed at angle, holotype, x 95; (39) kenozooid with ovicellate and non-ovicellate autozooids, BMNH 1985.1 1.20.76c, Channel Stn 253, x97; (40) ovicellate and non- ovicellate autozooids at edge of colony, holotype, x 129; (41) non-ovicellate autozooid, holotype, x 182; (42) ancestrula, BMNH 1985.1 1.20.76d, Channel Stn 253, x 240. 28 J. D. D. BISHOP & B. C. HOUSEHAM Table 6 Puellina directa n.sp.: measurements in microns Specimen Range Mean C.V. n Range Mean C.V. a Lo 40-48 43-2 7-70 6 Lf 128-162 149 7-19 8 b 43^9 46-0 4-69 7 154-206 175 8-88 9 c 37^8 43-3 6-02 20 131-212 164 11-8 24 a lo 51-60 55-6 5-88 7 Lfg 173-257 211 12-4 8 b 55-66 60-7 6-21 6 210-294 241 11-2 8 c 58-66 61-0 3-72 20 209-326 241 14-4 23 a loo 0 If 117-137 124 5-70 8 b 77-83 79-0 4-76 3 122-148 132 8-40 8 c 75-80 77-0 2-21 9 110-157 135 8-68 29 a Lov 0 lov 0 b 166 166 0-00 2 225 — — 1 c 128-157 140 7-37 9 151-198 175 9-47 9 a = holotype(BMNH 1985.11.20.76b); b = BMNH 1985.11.20.76c;c = BMNH 1985.11.20.110a. Ancestrula (Fig. 42) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of slender erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Median proximal spine with strongly bifid tip. Spines without apophyses. Regeneration of ancestrula not seen. ETYMOLOGY. L. directus — set straight, arranged in a straight line; referring to the disposition of the more distal pairs of costae in this species. REMARKS. P. directa n.sp. resembles members of the subgenus Puellina as defined above (P. gattyae, P. setosa and P. modica n.sp.), except that the ovicell is non-punctate and of category A (p. 4). The species is best characterised by the arrangement of the more distal pairs of costae in parallel or sub-parallel lines right across the strongly convex frontal shield, by the strong costal ridges which tend to be higher towards the centre of the shield, and by the broad gymnocyst which is about as wide laterally as proximally. P. directa is known only from the English Channel off Normandy and Brittany, at depths of 38-99 m. Puellina (Cribrilaria) venusta Canu & Bassler, 1925 (Figs 43^9, 99 and Table 7) part Cribrilina radiata, radiata form: Hincks, 1880; 185-190, pi. 25 figs 3, ?6. Puellina venusta Canu & Bassler, 1925; 22, pi. 2 fig. 5. Cribrilaria crenulata Harmelin, 1970; 91-93, fig. li-k, pi. 2 figs 1-3. Cribrilaria venusta: Harmelin, 1978; 180, 182, 184, pi. 2 figs 3-5. Cribrilaria venusta: Hay ward & Ryland, 1978; 146. part or whole Cribrilaria radiata: Hayward & Ryland, 1979; 62, 63, fig. 16A,B. Cribrilaria venusta: Hayward & Cook, 1983; 43, 45, fig. 1 1 A. MATERIAL EXAMINED BY SEM. BMNH: 1911.10.1.711 (Fosse de Cap Breton); 1985. 11. 20.68 (Channel Stn 252); 1 985. 11.20.82a (Channel Stn 256); 1985.1 1.20.1 17 (Channel Stn 318); D55588 (Badenian, Hartl-lucke, Eisenstadt). LBIMM: BRY-6527 (Marseille, holotype of Cribrilaria crenulata Harmelin, 1970). OTHER MATERIAL. Holotype: IP, Canu Collection (no registration number) (34°12'N, 07°07'W, off Rabat, Morocco, 140-1 50m). BRITISH PUELLINA 29 30 J. D. D. BISHOP & B. C. HOUSEHAM BMNH: 1911.10.1.706 (Guernsey, two colonies on separate substrates); 1911.10.1.712 (Birturbuy [ = Bertraghboy] Bay [Galway]); 191 1.10.1.732 (Shetland, 320m, colony and substrate, a pebble, broken in two); 1919.6.24.39 (Guernsey); 1985.1.2.54, 1985.1.2.83 and 1985.1.2.88 (northern Bay of Biscay); 1985.1 1.20.83b (Channel Stn 256); 1985.1 1.20.87b (Channel Stn 257); 1985.1 1.20. 103b,c (Channel Stn 261); 1985.1 1.20. 104c, d (Channel Stn 263); 1985.1 1.20.1 13 (Channel Stn 316). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material multiserial, with well over 1000 autozooids. Autozooid of very variable shape in frontal view, generally sub-oval but commonly broadly bifid proximally (when passing to both sides of proximal zooid) or elongate. Exposed lateral and proximal gymnocyst of autozooid very narrow or absent (except at edge of colony), but frequently with short proximal extension(s) between neighbouring zooids. Frontal shield ( = pericyst) lightly to moderately calcified, without median thickening (except on apertural bar — see below), with 12-23 costae (most commonly 15-18; high numbers generally occurring in proximally bifid zooids, in which the pattern of costae may be partially duplicated, or in elongate zooids). Each costa with short, steeply ascending basal portion leading to tapering, nearly horizontal subsequent portion contributing to slightly convex central region of shield; without ornamentation; up to 3 minute pelmatidia along length of each costa. Intercostal pores sub-oval, of variable size throughout shield but generally slightly larger in row immediately proxi- mal to apertural bar than in rest of shield, without denticles; 3 or 4 pores between successive costae. Lipped papilla pore between bases of successive costae, about size of adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, very clearly broader than long, proximal edge straight or indistinctly crenulate. Orifice of ovicellate zooid frequently appearing distally truncate to varying extent in frontal view because of overhanging calcification of ovicell; width 1-0-1-1 times that of orifice of non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 4 in ovicellate zooid; jointed just above base, with adjacent constriction (Fig. 47); apophyses not seen. Apertural bar forming short triangular area, sloping steeply upwards from proximal margin of orifice to form median sub-oral mucro; mucro of variable development, sometimes spiniform or bifid; apertural bar without tubercles; pelmatidia and median suture not seen. Without median sub-oral lacuna. Each autozooid with about 9 distolateral pore chambers; uncalcified external windows of pore chambers, seen in zooids on edge of colony (Fig. 48), often sub-rectangular, very large (much larger than an intercostal pore). Ovicell of category A (p. 4), with rather variable outline in frontal view, length 0-5-0-6 times that of frontal shield of autozooid; frontal surface non-punctate, with variably developed median or diagonal ridge, sometimes with concurrent suture. Avicularium frequent, interzooidal, smaller than autozooid; palate elongate-triangular, inclined at low angle to substrate; rostrum generally directed between autozooids. Mandible equal in length to palate and of same shape, i.e. fitting into palate. Proximal gymnocyst short (much shorter than palate plus frontal non-calcified area). Kenozooid (Fig. 48) infrequent, up to size of autozooid, with cribrimorph frontal shield of costae in radiating pattern. Ancestrula (Fig. 49) tatiform, with median proximal spine plus 5 pairs of spines around edge of gymnocyst (i.e. total of 1 1 spines). Details of segmentation (if any) of spines not seen. Morphology of median proximal spine not seen. Other spines each with upwardly directed spiniform apophysis on inner surface near base. Regeneration of ancestrula not seen. Figs 46-49 Puellina venusta Canu & Bassler: (46) ovicellate and non-ovicellate autozooids and avicu- larium, BMNH 1985.1 1.20.82a, Channel Stn 256, x 1 16; (47) non-ovicellate autozooids (some with basal portions of oral spines intact) and avicularia (with mandibles intact), BMNH 1985.11.20.68, Channel Stn 252, x 145; (48) edge of colony, maternal autozooid (top left) arrowed, BMNH 1911.10.1.711, Fosse de Cap Breton, x 49; (49) ancestrula, BMNH 1 985. 1 1 .20. 1 1 7, Channel Stn 3 1 8, x230. BRITISH PUELLINA 31 32 J. D. D. BISHOP & B. C. HOUSEHAM Table 7 Puellina venusta Canu & Bassler: measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 48-57 53-5 6-83 6 Lf 277-345 302 6-72 9 b 51-61 57-9 5-91 8 256-350 306 9-20 9 c 49-65 57-0 6-59 26 283-372 326 742 29 d 53-71 60-9 10-7 7 274-378 317 124 6 a lo 80-95 85-1 6-64 7 Lfg 342-445 371 8-85 9 b 77-100 86-7 8-27 10 281^33 343 13-0 10 c 81-100 91-5 4-68 25 353^86 413 942 27 d 82-94 87-3 6-19 7 352-444 386 11-7 5 a loo 87-94 90-3 4-21 3 If 244-322 267 8-79 10 b 87 — — 1 169-336 253 214 11 c 97 — — 1 254-406 304 13-0 18 d 82-94 86-8 6-43 4 219-279 247 9-23 9 a Lov 141-167 153 7-68 4 Lr 187-209 198 8-84 2 b 147 — — 1 156 — — 1 c 158-170 164 5-82 2 160-230 180 16-5 5 d 167-197 179 8-29 4 195-203 199 2-18 3 a lov 152-168 163 6-15 3 Lav 220-241 231 7-25 2 b 155 — — 1 233 — — 1 c 184-228 206 17-0 2 183-284 221 184 5 d 195-213 207 4-37 4 220-245 231 6-06 3 a = BMNH 1911. 10.1.711; b = BMNH 1985.11.20.68; c = BMNH 1985.11.20.82a; d = LBIMM-BRY-6527 (holotype of Cribrilaria crenulata Harmelin). REMARKS. In a taxonomic account of Mediterranean species referred to Cribrilaria, Harmelin (1970) redescribed C. radiata (Moll, 1803), of which a neotype was designated, and described as new a species to which he gave the name C. crenulata. C. crenulata was later realised (Harmelin, 1978) to be a geographical variant of the formerly overlooked species Puellina venusta Canu & Bassler, 1925, which had been described from the Atlantic off Morocco. The two forms P. radiata and P. venusta seem to be very closely related. P. radiata, as defined by the neotype, is thought to be restricted to the Mediterranean (Harmelin, 1978:174 and pers. comm. to JDDB, 1985), although the name has been widely applied to Atlantic material referable to P. venusta. Traces of the crenulate proximal margin of the orifice shown by Mediterranean specimens described as Cribrilaria crenulata by Harmelin (1970) may be seen, albeit indistinctly, in some Atlantic material (BMNH 1911.10.1.711). As noted by Harmelin (1970), avicularia in Recent Mediterranean material frequently show multiple regenerations stacked one on top of the other; this was also reported in Atlantic material studied by Harmelin (1978) and Aristegui (1983). Autozooids of the holotype of C. crenulata have up to 12 distolateral pore chambers. Harmelin (1978:180) stated that oral spines of P. venusta sometimes have 'une couronne de spinules legere- ment au-dessus de leur base'. The median proximal spine of the ancestrula of material studied by Harmelin (1970:92 and fig. 3b) has a bifid tip. The specimen figured by Hayward & Ryland (1979: fig. 16A,B, as Cribrilaria radiata) is BMNH 1911.10.1.711 (Fig. 48). Part of the material studied by Hayward & Ryland ( 1 978) is in the BMNH collections (1985. 1.2.54, 1985.1.2.83 and 1985.1.2.88). BRITISH PUELLINA 33 P. venusta is a common fossil species encrusting shells from the Badenian (Middle Miocene) of Eisenstadt in the Vienna Basin. The fossil (Fig. 99) has slightly more extensive gymnocyst than has most Recent material (Figs 43-49), but agreement is otherwise close. The Badenian of Eisenstadt is the type locality of a very similar species, Puellina scripta (Reuss, 1848). The holotype of P. scripta, NM V 1 848/38/75, is illustrated in Figure 98, and is the only colony of the species examined. In this specimen the avicularium is consistently larger than in P. venusta, the sub-oral mucro of the autozooid is poorly developed or absent, and the frontal shield almost flat. P. scripta is therefore very close to the Recent Mediterranean species P. radiata. It differs only in the presence of 5 oral spines in all non-ovicellate autozooids whereas in P. radiata the usual complement is 4 spines, with 20-60% of zooids in rare colonies having 5 (Harmelin, 1970). The name P. scripta has been placed in synonymy with P. radiata by several authors, although it has been applied at least once to Recent material (Manzoni, 1871). It is interesting to note that the co-occurrence of the apparently closely related entities P. venusta and P. scripta/ P. radiata in the Paratethys during Middle Miocene times persists to the present day in the Mediterranean. The taxonomic status of these forms requires clarification. In the English Channel material studied in the present paper, P. venusta occurred off Brittany at depths of 82-102 m. It was reported at greater depths (180-182 m and 250-270 m) in the northern Bay of Biscay by Hayward & Ryland (1978). P. venusta appears to have a wide present-day latitudinal range. The most northerly material available is from deep water (320 m) off the Shetlands (BMNH 191 1 . 10. 1 .732). The species was described originally from material collected off Rabat (Morocco), and has been recorded from the Canaries by Aristegui (1983; 1984) and off the eastern coast of South Africa at a depth of 90 m by Hayward & Cook (1983). Puellina (Cribrilaria) innominata (Couch, 1844) (Figs 50-58 and Table 8) Lepralia innominata Couch, 1844; 1 14, 1 15, pi. 22 fig. 4. Lepralia innominata: Johnston, 1847; 319, 320, pi. 55 fig. 12. not Lepralia innominata: Busk, 1859; 40, pi. 4 fig. 2. part or whole Cribrilina radiata, innominata form: Hincks, 1880; 185-190, pi. 25 figs 1, 2, 5. Cribrilaria innominata: Hayward & Ryland, 1979; 64, fig. 17A-D. Cribrilaria innominata: Bishop, 1986; 96, 98, figs 1-8. MATERIAL EXAMINED BY SEM. Neotype: 1847.9.16.32 (Goran [Cornwall]; selected by Bishop, 1986). BMNH: 1897.5. 1.632 (Hastings); 1 899.5. 1.720A (British); 1899.5.1.723 (Antrim); 1911.10.1.723 (British); 1936.12.30.335B (Liverpool Bay); 1985.11.20.41b (off Craggan Rocks, S. Cornwall); 1985.1 1.20.73a and 1985.11.20.76a (Channel Stn 253); 1985.11.20.102a (Channel Stn 260); 1985.11.20.103d (Channel Stn 261); 1985.1 1.20.1 16b (Channel Stn 318). OTHER MATERIAL. BMNH: 1847.9.16.53 (British); 1847.9.16.122 (Sana Island [W. Scotland]); 1847.9.16.161 (British, colony numbered 2); 1 847.9. 1 6. 162 and 1 899.5. 1 .7208, D (British); 1899.5. 1 .722 (Guernsey, colonies on three shells); 1911.10.1.720 and 1911.10.1.721 (British, numerous colonies on shell fragments); 1911.10.1.722 (Shetland, several colonies on pebbles); 1911.10.1.725 (Guernsey, several colonies on shell fragments); 1911.10.1.726 (Birturbuy [ = Bertraghboy] Bay [Galway], numerous colonies on shell fragments); 1919.6.24.35 (Birterbuy [ = Bertraghboy] Bay [Galway]); 1936. 12.30.332A-D (Liverpool Bay); 1936. 12.30.333A-J and 1936.12.30.334 (no locality [probably Liverpool Bay]); 1936.12.30.335C (Liverpool Bay); 1963.3.30.142 (Plymouth, several colonies); 1985.11.20.42a-c, 1985.11.20.43 and 1985.11.20.44a-i (off Craggan Rocks, S. Cornwall); 1985.1 1.20.67a (Channel Stn 251); 1985.11.20.69d, 1985.11.20.70a and 1985. 11.20.72e,f (Channel Stn 252); 1985 11 20 74d and 1985.1 1.20.81a (Channel Stn 253); 1985.11.20.83c (Channel Stn 256); 1985.11.20.88a,b (Channel Stn 257); 1985.1 1.20.91b and 1985.11.20.98b (Channel Stn 258); 1985.11.20.102b (Channel Stn 260)- 1985.ll.20.103e (Channel Stn 261); 1985.11.20.104e-h (Channel Stn 263); 1985.1 1.20.105d and 1985.1 1.20.107c (Channel Stn 275); 1985.ll.20.llla and 1985.1 1.20.1 12a (Channel Stn 315); 1985.1 1.20.1 16c-k (Channel Stn 318). 34 J. D. D. BISHOP & B. C. HOUSEHAM BRITISH PUELLINA Figs 54-58 Puellina innominata (Couch): (54) non-ovicellate autozooid, BMNH 1899.5.1.720A, British, x 143; (55) edge of colony with abraded autozooids, arrows indicate maternal autozooids, same specimen, x 58; (56) ancestrula (partially obscured), BMNH 1985. 1 1 .20. 102a, Channel Stn 260, x 164; (57) ancestrula regenerated as kenozooid (note bifid top of median proximal spine of ances- trula), BMNH 1985.1 1.20.1 16b, Channel Stn 318, x 280; (58) ancestrula regenerated as avicularium, BMNH 1985. 1 1 .20.73a, Channel Stn 253, x 200. 36 J. D. D. BISHOP & B. C. HOUSEHAM DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material multiserial, with well over 1000 autozooids. Autozooid of variable shape in frontal view, generally sub-oval (longer than broad) but occasionally broadly bifid proximally when passing to both sides of proximal zooid. Exposed lateral gymnocyst of autozooid narrow or absent (except at edge of colony); proximal gymnocyst narrow or of moderate width, frequently with proximal extension(s) between neighbouring zooids. Frontal shield ( = pericyst) moderately to heavily calcified, with 9-14 costae (most commonly 1 1 or 12; sometimes more than 14 in proximally bifid zooids in which pattern of costae is partially duplicated), with prominent distal median umbo and variable median ridge proximal to umbo. Each costa with short, broad, steeply ascending basal portion (sometimes seen to bear minute pelmatidium) leading to tapering, nearly horizontal subsequent portion contributing to slightly Table 8 Puellina innominata (Couch): measurements in microns Specimen Range Mean C.V. n Range Mean C.V. n a Lo 47-55 50-1 4-87 11 Lf 216-294 254 7-67 13 b 57-63 58-9 3-08 12 238-291 268 6-09 17 c — — — 0 219-284 247 8-32 12 d 55-60 57-0 3-41 11 284-369 323 8-77 11 e 56 — — 1 293-338 308 7-01 4 f 51-62 56-5 6-01 10 250-308 278 6-33 11 a lo 79-89 83-7 3-78 11 Lfg 293^39 363 12-8 13 b 81-89 85-3 2-72 12 325-454 369 10-7 13 c 74-92 81-8 6-17 11 311^48 363 13-2 11 d 88-103 96-0 4-58 10 404-508 455 8-53 10 e 88-95 92-3 4.44 3 384-449 406 10-0 3 f 81-90 86-3 3-27 10 325-433 384 10-7 9 a loo 89-94 91-0 3-15 3 If 188-266 223 8-85 13 b 91-100 97-1 4-69 8 203-265 229 7-35 18 c 74-86 80-0 11-9 2 184-272 220 13-0 15 d 96-112 105 7-00 4 231-300 266 8-26 15 e 95-115 101 10-3 4 221-268 234 6-86 8 f 91-97 94-0 3-46 3 216-298 251 10-5 11 a Lov 172-177 175 2-28 2 Lr 174-178 176 1-81 2 b 170-188 181 3-12 8 164 — — 1 c 145-174 162 10-2 3 — — — 0 d 191-238 208 9-28 6 230-234 232 1-37 2 e 163-233 195 17-6 4 — — — 0 f 181-195 188 3-23 4 — — — 0 a lov 213 1 Lav 260-267 264 2-12 2 b 195-206 200 2-00 7 180 — — 1 c 184-229 203 12-5 3 — — — 0 d 192-254 229 10-4 6 318-326 322 1-98 2 e 166-214 192 13-7 3 — — — 0 f 201-231 215 6-29 4 — — — 0 a = neotype (BMNH 1847.9.16.32); b = BMNH 1 899.5. 1.720A; c = BMNH 1899.5.1.723; d = BMNH 1911.10.1.723; e = BMNH 1 985. 11. 20. 103d; f= BMNH 1985.11.20.76a. BRITISH PUELLINA 37 convex central region of shield; tubercle or short ridge at junction of the two portions of each costa, near periphery of shield. Intercostal pores of variable shape (round, sub-oval, reniform or sub- rectangular), of relatively uniform size throughout shield, sometimes with denticles adjacent to intercostal suture; 1-5 (most commonly 3 or 4) pores between successive costae. Lipped papilla pore between bases of successive costae, considerably larger than adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, very clearly broader than long, proximal edge straight. Orifice of ovicellate zooid frequently appearing distally truncate to variable extent in frontal view because of overhanging calcification of ovicell; width 1 -0-1-1 times that of orifice of non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 4 in ovicellate zooid; jointed just above base; with inwardly directed, inconspicuous half-circle of denticles near base. Apertural bar form- ing triangular area, sloping upwards from proximal margin of orifice to umbo; with minute pelmatidium each side of indistinct median suture; without tubercles. Single, very large median sub-oral lacuna (much larger than an intercostal pore), variable in shape but frequently sub- triangular, with base of triangle towards orifice; without denticles. Each autozooid with 7-9 distolateral pore chambers; uncalcified external windows of pore chambers (seen in zooids on edge of colony: Fig. 55) relatively large (as large as or larger than sub-oral lacuna). Ovicell of category A (p. 4), length generally 0-6-0-7 times that of frontal shield of autozooid; frontal surface non-punctate, with up to 4 indistinct tubercles or short ridges in more or less radiating pattern, occasionally approximating to central umbo; frequently with median suture. Avicularium (Figs 52, 53) infrequent to fairly frequent, interzooidal, considerably smaller than autozooid. Palate elongate-triangular, inclined at moderate angle to substrate (highest distally); rostrum generally directed between autozooids. Mandible equal in length to palate and of same shape, fitting into it. Length of proximal gymnocyst of avicularium clearly shorter than that of palate plus frontal non-calcified area. Frontal non-calcified area sometimes occupied by cribrimorph frontal shield after regeneration as kenozooid. Kenozooid infrequent (BMNH 1985.11.20.103d only), slightly smaller than autozooid, with cribrimorph frontal shield of costae in radiating pattern; gymnocyst narrow. Ancestrula (Fig. 56) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Details of jointing (if any) of spines not seen. Median proximal spine with strongly bifid tip (Fig. 57). Most distal pair of spines each with short, minutely denticulate, inwardly directed apophysis near base. Remaining eight erect spines each with inwardly directed spiniform apophysis (sometimes with lateral denticles and minutely bifid tip) near base. Ancestrula sometimes regenerated as kenozooid (Fig. 57), or rarely as avicularium (BMNH 1985.1 1.20.73a: Fig. 58). REMARKS. The identity of Puellina innominata has been discussed, and a neotype selected, by Bishop (1986). The species is generally well characterised around the British Isles, where variation is largely limited to the degree of development of the costal ridges and umbo. However, many other forms apparently related to P. innominata as defined by the neotype are known from other parts of the world, and the elucidation of the status of these would involve an immense amount of study. Material from the British Isles and adjacent waters studied in the present paper was collected from the English Channel off Sussex, Normandy, Brittany, Guernsey, Devon and Cornwall (depths, where known, 38-106 m); from Galway; from the Irish Sea; off Antrim; off the W. coast of Scotland; and from the Shetlands. Puellina (Cribrilaria) bifida (d'Hondt, 1970) (Figs 59-72 and Table 9) Colletosia innominata subsp. bifida d'Hondt, 1970; 246-248, fig. 12a-c. MATERIAL EXAMINED BY SEM. Lectotype (selected here); LBIMM-BRY-1 1603 (48°01'N, 05°44'W, 130 m). BMNH: 1911.10.1.735B (Belfast); 1911. 10.1. 1182B (off Antrim); 1936.12.30.331A (two colonies) and B (no locality [probably Liverpool Bay]); 1985.1 1.20.41a (off Craggan Rocks, S. Cornwall); 1985.1 1.20.71a and 1985.1 1.20.72d (Channel Stn 252); 1985.1 1.20.94b, c (Channel Stn 258); 1985.1 1.20. 103a (Channel Stn 261). 38 J. D. D. BISHOP & B. C. HOUSEHAM BRITISH PUELLINA 39 OTHER MATERIAL. BMNH: 1847.9.16.161 (British, colony numbered 1); 1899.5. 1.720C (British); 1899.5.1.953 (Antrim); 1911.10.1.731 (British); 1911.10.1.737B (Guernsey); 1936.12.30.331C,D (no locality [probably Liverpool Bay]); 1936.12.30.332B and 1936.12.30.335C (Liverpool Bay); 1985.11.20.45 (Channel Stn 183); 1985.11.20.47 (Channel Stn 240); 1985.1 1.20.52b-d, 1985.11.20.53 and 1985.11.20.54a-e (Channel Stn 248); 1985.1 1.20.57a and 1985.1 1.20.59b-f (Channel Stn 249); 1985.11.20.60a, 1985.11.20.63a and 1985. 1 1 .20.64a-d (Channel Stn 250); 1 985. 1 1 .20.69b,c (Channel Stn 252); 1 985. 1 1 .20.74c and 1985. 1 1 .20.77a, b (Channel Stn 253); 1 985. 1 1 .20.83a (Channel Stn 256); 1985. 1 1 .20.87a and 1 985. 1 1 .20.89 (Channel Stn 257); 1985.11.20.90a,b, 1985.1 1.20.91a, 1985.1 1.20.92a, 1985.11.20.93a, 1985.11.20.94a, 1985.11.20.96a, 1985.11.20.97a and 1985.11.20.98a (Channel Stn 258); 1985.1 1.20.104a,b (Channel Stn 263); 1985.11.20.105a-c and 1985.1 1.20.107a,b (Channel Stn 275); 1985.1 1.20.108b-d and 1985.1 1.20.109b (Channel Stn 3 14); 1985. 11.20.1 lib (Channel Stn 315); 1985.11.20.114 (Channel Stn 317); 1985.1 1.20.1 16a (Channel Stn 3 18). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material pluriserial to multiserial, with over 200 autozooids. Autozooid of variable shape in frontal view, generally sub-oval (longer than broad). Lateral and proximal gymnocyst of autozooid narrow or of moderate width, but frequently with proximal and proximo-lateral extensions between neighbouring zooids. Frontal shield ( = pericyst) of autozooid moderately calcified, with little or no differential median thickening and no umbo, with 6-1 1 costae (usually 7-9). Each costa with short, broad, steeply ascending basal portion at periphery of frontal shield leading to tapering, nearly horizontal subsequent portion contributing to almost flat central region of shield. Costal ridges prominent, raised near edge of shield but lower centrally, frequently showing abrupt decrease in height between periphery and centre of shield giving characteristic appearance (especially noticeable when viewed obliquely: Fig. 61). Intercostal pores sub-rectangular or reniform, relatively uniform in size throughout shield, often with denticles adjacent to intercostal suture; usually 4, occasionally 5, pores between successive costae, but frequently fewer between short proximal costae. Lipped papilla pore between bases of successive costae, about size of adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, clearly broader than long, proximal edge straight. Orifice of ovicellate zooid frequently appearing distally truncate to varying extent in frontal view because of overhanging calcification of ovicell; width 1 • 1-1 -2 times that of orifice of non-ovicellate zooid. Oral spines 5 in non-ovicellate zooid, 4 in ovicellate zooid; jointed at base and generally with inwardly directed spiniform apophysis near base. Apertural bar with low tubercle each side of median suture and pelmatidium on each tubercle. (Pelmatidia not observed elsewhere on shield.) Single median sub-oral lacuna between tubercles of apertural bar and first row of intercostal pores, about same size as an intercostal pore, frequently with pair of denticles on proximal margin; lacuna occasionally confluent with adjacent intercostal pores. Each autozooid with 5-8 distolateral pore chambers; uncalcified external windows of pore chambers, seen in zooids on edge of colony (Fig. 67), relatively large (considerably larger than an intercostal pore). Ovicell of category A or occasionally B (p. 4) (Fig. 70), length 0-7-0-9 times that of frontal shield of autozooid; frontal surface non-punctate, frequently with up to 4 ridges in more or less radiating pattern, with median suture. Avicularium absent in holotype, infrequent in other material (Figs 65-67), interzooidal, smaller than an autozooid, nearly always seen on periphery of colony (exceptions occurring at sites of apparent disruption to colony growth in BMNH 1936.12.30.331 A). Palate elongate-triangular, of somewhat variable length, often curving down towards substrate distally. Mandible elongate- triangular, sometimes with very acute tip produced beyond (? damaged or incomplete) palate as narrow down-curved extension (seen in dried material). Cystid of avicularium with 2 or 3 pore chamber windows; length of proximal gymnocyst about equal to that of palate plus frontal non-calcified area. Frontal non-calcified area sometimes occupied by cribrimorph frontal shield after regeneration as kenozooid (Fig. 69). Kenozooid (Fig. 68) infrequent, generally smaller than an autozooid, with cribrimorph frontal shield of costae in radiating pattern. 40 J. D. D. BISHOP & B. C. HOUSEHAM Ancestrula (Fig. 71) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Erect spines jointed at base and with at least 2 subsequent segments. Median proximal spine with broad, sometimes slightly bifid tip. Most distal pair of spines each with short, blunt, inwardly directed apophysis near base. Remaining eight erect spines each with long, slender, inwardly directed spiniform apophysis (sometimes with minutely bifid tip) near base. Ancestrula sometimes regenerated as kenozooid (Fig. 72). Table 9 Puellina bifida (d'Hondt): measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 43^9 45-7 4-21 9 Lf 198-246 222 7-02 12 b 39-53 47-3 7-85 10 199-252 216 740 9 c 52 — — 1 204-208 206 1-54 2 d 52-59 54-7 3-89 14 187-295 233 12-0 14 e — — — 0 195-271 227 9-02 16 a lo 54-62 58-2 4-97 9 Lfg 251-332 287 9-17 12 b 59-73 65-3 7-16 8 243-318 276 10-9 8 c 71 — — 1 252-253 253 0-314 2 d 61-73 65-9 6-15 14 227-426 327 19-1 14 e — — — 0 262-407 310 15-8 15 a loo 65-72 68-9 4-11 7 If 149-179 162 6-30 15 b 79-81 80-0 1-99 2 142-191 173 8-26 11 c 75-78 76-5 3-12 2 169-187 178 8-04 2 d 68-79 74-0 4-96 6 158-221 188 10-7 18 e — — — 0 157-200 175 7-39 14 a Lov 148-184 167 9-28 6 Lr 0 b 155-173 162 5-33 4 144-148 146 2-18 2 c — — — 0 136 — — 1 d 146-198 175 12-7 5 142 — — 1 e 154-177 164 6-15 5 — — — 0 a lov 178-205 195 6-06 6 Lav 0 b 168-218 188 12-7 4 228-233 231 1-72 2 c — — — 0 229 — — 1 d 170-206 188 7-53 5 240 — — 1 e 184-212 197 5-64 5 — — — 0 a = lectotype (LBIMM-BRY-11603); b = BMNH 1936.12.30.331A;e = BMNH 1911. 10.1. 1182B. 1985.1 1.20.103a; c = BMNH 1936.12.30.331B; d = BMNH Figs 62-66 Puellina bifida (d'Hondt): (62) ovicellate autozooid, lectotype (LBIMM-BRY-1 1 603), off Finistere, x 170; (63) non-ovicellate autozooid, BMNH 1985.1 1.20.71a, Channel Stn 252, x240; (64) abraded autozooids at edge of colony showing pore chambers, arrows indicate intact maternal autozooids (that had presumably been ovicellate), BMNH 1936.12.30.331 A, no locality, x91; (65) avicularium, BMNH 1985.1 1.20.103a, Channel Stn 261, x260; (66) avicularium, regenerated non- ovicellate autozooid (on left) with abnormal orifice and maternal autozooid (on right) with ovicell not yet developed, BMNH 1936. 12.30.33 IB, no locality, x 150. BRITISH PUELLINA 41 42 J. D. D. BISHOP & B. C. HOUSEHAM i BRITISH PUELLINA 43 REMARKS. This form has apparently been overlooked since its description as a subspecies of P. innominata. The latitude of the type locality given by d'Hondt (1970:233) is erroneous, and he provided (pers. comm. to JDDB, 1986) the correct position off Finistere quoted here. The type series comprised 'quelques colonies' but no holotype designation for these was published. The specimen received on loan from LBIMM, which was labelled 'holotype', is designated here as lectotype. It is not known whether this is the specimen figured by d'Hondt. The original account and figure clearly refer to the form represented by the lectotype, but are misleading in a number of respects. The supposedly bifid oral spines described by d'Hondt are in fact incomplete, broken just above the level of their inwardly directed apophyses. A minute apical avicularium on the ovicell, mentioned by d'Hondt in both the description and diagnosis and apparently shown in his figure 12a, is not to be seen in the lectotype, nor in other material examined here. The frontal shields of adjacent zooids are not separated by broad, smooth areas as apparently shown in figure 12a. Non-ovicellate zooids with 4, rather than 5, oral spines have not been observed. P. bifida is extremely similar to the Badenian (Middle Miocene) fossil P. kollmanni (David & Pouyet, 1974) from the Vienna Basin (material of P. kollmanni examined by SEM: holotype and paratype, both NMV 1848/38/76; BMNH D44968 and D55589; all from Eisenstadt). The only appreciable difference between the two forms is that the avicularian palate of P. kollmanni is generally more elongate and slender than that of P. bifida, and often almost parallel sided. As mentioned above, the length of the palate in P. bifida is variable, and in some examples the mandible appears to pass beyond the palate as a narrow extension (Fig. 67). If it could be shown that in the latter cases the rostrum was broken short distally, the shape of the palate when complete would be inferred from the mandible to resemble that seen in P. kollmanni, and an argument might then be made for treating the fossil and Recent forms as conspecific. Puellina bifida is also similar to the Recent Mediterranean species P. minima (Harmelin, 1984) of which four colonies of the type series (LBIMM, no registration number, Cassis, France) were examined by SEM. In P. minima, however, the ovicellate zooid has 2 oral spines, not 4. As noted by Harmelin (1984), the ovicell of P. minima may rest on a kenozooid (Fig. 94; ovicell category B of the present paper). This condition is also sometimes seen in P. bifida (Fig. 70). The material from the Canaries recorded as C. minima by Aristegui (1984) has 4 oral spines in the ovicellate zooid, and seems to be closer to P. bifida than to Harmelin's species. P. bifida is known to occur off Finistere (depth 1 30 m); in the English Channel off Boulogne, Normandy, Brittany, Guernsey and Cornwall (at depths, where known, of 29-102 m); in the Irish Sea; and off Antrim. The presence in the Canaries of a very similar, possibly conspecific, form recorded by Aristegui (1984) as Cribrilaria minima suggests that P. bifida may also be found considerably further south than the above-mentioned localities. Puellina (Cribrilaria) arrecta n.sp. (Figs 73-8 land Table 10) whole or part Cribrilaria flabellifera: Harmelin, 1970; 94-96. part Cribrilaria flabellifera: Harmelin, 1978; 186, 187 (material from Stn 72); pi. 2 figs 8, 9, 12. Cribrilaria flabellifera: Hayward & Ryland, 1978; 146. Holotype: BMNH 1985.1 1.20.108h (Channel Stn 314, 49°12-4'N 03°14-8'W, 73 m). Paratypes (examined by SEM): BMNH: 191 1.1 0.1. 737 A (Guernsey); 1985. 11.20.76e-g (Channel Stn 253); 1985. 11.20.82b (Channel Stn 256); 1985.1 1.20.103f (Channel Stn 261); 1985.1 1.20.1 lOc (Channel Stn 314). Figs 67-72 Puellina bifida (d'Hondt): (67) avicularium with mandible intact, BMNH 1985. 1 1 .20.71a, Channel Stn 252, x 166; (68) kenozooid, BMNH 1936.12.30.331A, no locality, x260; (69) avicu- larium regenerated as kenozooid, BMNH 1985.1 1.20.103a, Channel Stn 261, x 200; (70) autozooid with category B ovicell, BMNH 1936.12.30.331A, no locality, x 102; (71) ancestrula, BMNH 1985.11.20.72d, Channel Stn 252, x230; (72) ancestrula regenerated as kenozooid, BMNH 1936.12.30.331B, no locality, x 250. 44 J. D. D. BISHOP & B. C. HOUSEHAM Fig. 73 Puellina arrecta n.sp.: stereo pair of holotype (BMNH 1985.11.20.108h), Channel Stn 314, xlll. OTHER MATERIAL. BMNH: 1985.11.20.35 (northern Bay of Biscay); 1985.11.20.49 and 1985.1 1.20.50a (Channel Stn 247); 1985. 11.20.57b (Channel Stn 249); 1985.1 1.20.63b and 1985.1 1.20.64h (Channel Stn 250); 1985.11.20.69g,h and 1985. 11. 20.72k (Channel Stn 252); 1985.1 1.20.73b, 1985.11.20.75, 1985.11.20.79 and 1985.1 1.20.81b,c (Channel Stn 253); 1985.11.20.84 (Channel Stn 256); 1985.11.20.85a and 1985.11.20.86 (Channel Stn 257); 1985.1 1.20.1 02c (Channel Stn 260); 1985.1 1.20.1 08ij (Channel Stn 314). J. G. Harmelin, personal collection: three colonies from Cassidaigne (Marseille), 350-450 m. DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material pluriserial to multiserial, with over 100 autozooids. Autozooid of variable shape in frontal Figs 74-77 Puellina arrecta n.sp.: (74) whole colony, BMNH 1985. 1 1 .20. 1 lOc, Channel Stn 314, x 20; (75) edge of colony viewed at angle, BMNH 1 985. 1 1 .20.76e, Channel Stn 253, x c. 65; (76) avicularium (mandible intact) and distal part of non-ovicellate autozooid, holotype (BMNH 1985.11.20.108h), Channel Stn 314, x250; (77) avicularium, ovicellate and non-ovicellate autozooids, BMNH 1 985. 11.20.76e, Channel Stn 253, x 157. BRITISH PUELLINA 45 P 46 J. D. D. BISHOP & B. C. HOUSEHAM view, generally sub-oval (longer than broad). Lateral and proximal gymnocyst of autozooid narrow or of moderate width, but frequently with proximal or proximo-lateral extensions between neighbouring zooids. Frontal shield ( = pericyst) of autozooid heavily calcified (especially near midline), with 8 or 9 costae (rarely 7 or 10). Each costa with short, steeply ascending basal portion at periphery of frontal shield leading to less steep subsequent portion contributing to convex central region of shield. Costal ridges prominent, each with tubercle (sometimes with minute pelmatidium) towards periphery of shield; ridges frequently fused to form strongly raised median ridge or umbo, especially towards distal end of shield. Intercostal pores sub-rectangular or reniform, relatively uniform in size throughout shield, often with denticles adjacent to intercostal suture; usually 2 or 3 pores between successive costae. Lipped papilla pore between bases of successive costae, slightly larger than adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, very clearly broader than long, proximal edge straight. Orifice of ovicellate zooid sometimes appearing distally truncate to varying extent in frontal view because of overhanging calcification of ovicell; width 1 -0-1-1 times that of orifice of non-ovicellate zooid. Oral spines 6 in non-ovicellate zooid, 4 in ovicellate zooid; jointed some distance above base; most proximal spine on each side generally with inwardly directed spiniform apophysis above joint; remaining spines with inwardly directed half-circle of denticles above joint. Apertural bar sloping upwards from proximal margin of orifice to first row of intercostal pores, with median suture; often with minute pelmatidium on very low tubercle each side of midline. Single median sub-oral lacuna of variable shape, larger than an intercostal pore, frequently with pair of denticles on proximal margin. Each autozooid usually with 6-8 distolateral pore chambers; uncalcified external windows of pore chambers (seen in zooids on edge of colony: Fig. 79) of moderate size but considerably larger than an intercostal pore. Ovicell of category A (p. 4), length 0-6-0-9 times that of frontal shield; frontal surface non- punctate, with up to 3 low ridges in more or less radiating pattern; median suture occasionally discernible. Avicularium (Fig. 76) frequent, interzooidal, total length (excluding mandible) approximately equal to that of frontal shield of autozooid. Palate flared at each end with relatively narrow middle portion; margins smooth (rarely with irregular minute crenulations). Mandible clearly longer than palate, broadening laterally beyond palate into angular-ovate or triangular blade c. 1-5 times as wide as frontal non-calcified area of cystid. Frontal non-calcified area sometimes occupied by cribrimorph frontal shield after regeneration as kenozooid (Fig. 78). Cystid of avicularium with 2 or 3 pore chambers. Length of proximal gymnocyst about equal to that of palate plus frontal non-calcified area. Kenozooid (Fig. 73) fairly frequent, up to size of autozooid, with cribrimorph frontal shield of costae in radiating pattern. Ancestrula (Fig. 80) tatiform, with median proximal spine overarching frontal surface plus 5 pairs of erect spines around edge of gymnocyst (i.e. total of 1 1 spines). Erect spines jointed, with at least 2 segments above base. Median proximal spine with broad or bifid tip. Most distal pair of spines each with short, inwardly directed apophysis near base. Remaining eight erect spines each with inwardly directed spiniform apophysis near base. Ancestrula sometimes regenerated as kenozooid (Fig. 81). ETYMOLOGY. L. arrectus — upright, steep; referring to the form of the frontal shield in this species. REMARKS. This species is most closely comparable to P.flabellifera (Kirkpatrick, 1 888) with which it has been confused and which is redescribed below. The autozooids have more extensive exposed gymnocyst in Puellina arrecta n.sp. than in P.flabellifera. There are considerable differences in the form of the frontal shield between the two species. The frontal shield of P. arrecta is markedly convex whereas that of P.flabellifera is almost flat. There are fewer costae in P. arrecta. The costae all bear a ridge and/or tubercle in the newly described species, whereas only the most distal pair have distinct ornamentation (ridges) in P.flabellifera. The median umbo of P.flabellifera (when BRITISH PUELLINA 47 Figs 78-81 Puellina arrecta n.sp.: (78) kenozooid and avicularium regenerated as kenozooid, BMNH 1985.1 1.20.1 lOc, Channel Stn 314, x 230; (79) maternal autozooids (numbered 1 and 2) and develop- ing autozooid (numbered 3) with incomplete category A ovicell for zooid 1, holotype (BMNH 1985.1 1 .20.108h), Channel Stn 314, x 155; (80) ancestrula and first autozooid (developed to right of midline of ancestrula because of foreign body distal to ancestrula), BMNH 1985.1 1.20. 76g, Channel Stn 253, x 200; (81) ancestrula regenerated as kenozooid, holotype, x 250. 48 J. D. D. BISHOP & B. C. HOUSEHAM Table 10 Puellina arrecta n.sp.: measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 36-45 38-8 11-0 5 Lf 163-209 187 8-85 10 b 43-47 44-7 2-90 7 187-204 193 3-58 8 c 41 — — 1 179-193 184 2-90 8 d 40-45 42-6 448 11 154-210 185 7-72 18 a lo 57-65 61-0 5-30 5 Lfg 203-291 253 14-2 9 b 65-75 70-0 6-79 5 244-312 276 8-69 8 c 58 — — 1 213-291 248 11-0 6 d 59-71 65-3 6-39 7 233-325 272 11-1 16 a loo 63-65 64-0 249 2 If 148-178 163 6-60 11 b 64-70 66-7 4-96 3 144-181 159 8-21 10 c — — — 0 135-151 146 4-23 7 d 69-74 71-0 2-77 5 135-178 155 7-21 16 a Lov 143-163 151 6-22 4 Lr 109-114 112 3-57 2 b 115-146 133 10-2 5 113-132 122 6-86 5 c 125-150 138 848 4 — — — 0 d 128-157 147 7-08 6 91-127 115 13-0 6 a lov 155-203 181 12-5 4 Lav 208-220 214 447 2 b 163-193 182 6-59 5 201-262 224 11-2 5 c 170-184 177 4-31 3 — — — 0 d 165-180 171 443 6 198-296 253 16-2 5 a = holotype (BMNH 1985.ll.20.110c. 1985.11.20.108h); b = BMNH 1985.11.20.76e; c = BMNH 1911.10.1.737A; d = BMNH present) is restricted to the region of the first pair of costae, but that of P. arrecta generally extends proximally as a median ridge. The sub-oral lacuna is relatively smaller in P. arrecta, and the orifice wider but shorter. The margins of the avicularian palate are generally smooth in P. arrecta rather than crenulate as in P.flabellifera. The mandible of the avicularium is relatively small in P. arrecta, and broadens beyond the palate to less than twice the width of the frontal non-calcified area of the avicularian cystid, as opposed to c. three times the width in P.flabellifera. BMNH 1985.11.20.35 is part of the material recorded as Cribrilaria flabellifera by Hayward & Ryland (1978). Material from the Marseille region at a depth of 350-450 m recorded as C. flabellifera by Harmelin (1970) was examined on loan, and is also referred to P. arrecta n.sp. Harmelin's material from 105-1 10 m in the same region was not examined. P. arrecta is known from the English Channel off Brittany (depths 64-106 m) and Guernsey, from the northern Bay of Biscay at a depth of 180-1 82m, and from the Mediterranean near Marseille (depth 350-450 m). Puellina (Cribrilaria) flabellifera (Kirkpatrick, 1888) (Figs 92, 93 and Table 11) Cribrilina radiata var. flabellifera Kirkpatrick, 1888; 75, pi. 10 fig. 4. Lectotype (selected here): BMNH 1 888.1. 25. 10A (Mauritius). DESCRIPTION OF LECTOTYPE. Fragment of unilaminar colony, zooids distinct, c. 90 autozooids and c. 50 avicularia. Autozooid of variable shape in frontal view, generally sub-oval (longer than BRITISH PUELLINA 49 broad). Exposed lateral gymnocyst of autozooid narrow or absent (except at edge of colony); proximal gymnocyst narrow, but occasionally with short proximal extension(s) between neighbouring zooids. Frontal shield ( = pericyst) of autozooid moderately calcified, without differential median thickening apart from distal umbo (see below), generally with 14 or 15 costae (occasionally 16). Each costa with short, steeply ascending basal portion at periphery of frontal shield leading to nearly horizontal subsequent portion, of relatively constant width, contributing to almost flat central region of shield. Most distal pair of costae forming shallow V in frontal view, ridges prominent and frequently forming truncate median umbo. Remaining costae low, lacking distinct ornamentation. Intercostal pores sub-rectangular or reniform, relatively uniform in size throug- out shield, frequently with denticles adjacent to intercostal suture; usually 4 or 5 pores between successive costae, but fewer between short proximal costae. Lipped papilla pore between bases of successive costae. Orifice of non-ovicellate autozooid D-shaped, slightly broader than long, proximal edge straight. Orifice of ovicellate zooid generally appearing distally truncate in frontal view because of overhanging calcification of ovicell; width 1 • 1-1 -2 times that of orifice of non-ovicellate zooid. Oral spines 6 in non-ovicellate zooid, 4 in ovicellate zooid; jointed some distance above base, apparently without apophyses (but undamaged spines not observed). Apertural bar sloping slightly upwards from proximal margin of orifice towards first pair of costae; without tubercles; sometimes with minute pelmatidium each side of median suture. (Pelmatidia not observed elsewhere on shield.) Single median sub-oral lacuna very large (considerably larger than an intercostal pore), of variable shape, frequently with denticles on lateral and proximal margins. Characters of pore chambers not discernible. Ovicell of category A (p. 4), length generally 0-6-0-7 times that of frontal shield of autozooid; frontal surface non-punctate, frequently with up to 4 ridges in more or less radiating pattern (or occasionally with short median ridge), with median suture. Avicularium frequent, interzooidal, total length excluding mandible approximately equal to that of frontal shield of autozooid. Palate flared at each end, more so distally, with relatively narrow middle portion; margins crenulate. Mandible clearly longer than palate, broadening abruptly beyond palate to c. 3 times width of frontal non-calcified area; further details of shape not discernible because of damage. Length of proximal gymnocyst of avicularium about equal to that of palate plus frontal non-calcified area. Cystid with pore chamber windows, 3 in only example counted. Kenozooid infrequent, smaller than an autozooid, with cribrimorph frontal shield of costae in radiating pattern. Colony origin, including ancestrula, missing. REMARKS. Kirkpatrick (1888: pi. 10 fig. 4) illustrated the undamaged avicularian mandible broadening rapidly beyond the palate into a sub-triangular blade. Neither Kirkpatrick's account nor the relevant BMNH accessions book indicates how many colonies were originally referred to Table 11 Puellina flabellifera (Kirkpatrick): measurements on lectotype (BMNH 1888. 1.25. 10A) in microns Range Mean C.V. n Range Mean C.V. n Lo 45-53 48-6 6-32 13 Lf 184-271 213 12-6 10 lo 51-65 58-2 7-08 10 Lfg 237-340 283 12-7 11 loo 64-70 67-3 4-92 3 If 152-208 184 8-77 15 Lov 122-167 149 12-4 5 Lr 119-156 138 11-2 15 lov 145-171 157 7-44 5 Lav 240-316 269 8-36 15 50 J. D. D. BISHOP & B. C. HOUSEHAM Cribrilina radiata var. flabellifera. Two specimens from Mauritius, stored separately but both registered as BMNH 1888.1.25.10, are the only available material referable to the original description. The specimen on a microslide, clearly the form described by Kirkpatrick and labelled Type', is presumably that examined by Harmelin (1978); it is here selected as lectotype (now BMNH 1 888.1. 25. 10A). It is not possible to identify the lectotype positively as the specimen figured by Kirkpatrick. However the second, boxed, specimen (now 1888. 1.25.1 OB) apparently grew on a different substrate from the lectotype, and is not the same species but a form with a slender elongate-triangular avicularian mandible, probably P. africana (Hayward & Cook, 1983). Kirkpatrick's variety was elevated to species rank by Harmelin ( 1 970). In addition to the original description and another record from the Indian Ocean (Thornely, 1912), the name flabellifera has been applied to material from the Indonesian region (Harmer, 1926; Canu & Bassler, 1929a; Winston, 1986), the central Pacific (Harmelin, 1978), the eastern Pacific (Soule, 1959), the western Caribbean (Banta & Carson, 1978; Winston, 1984), the Mediterranean (Harmelin, 1970) and the Fig. 82 Puellina corbula n.sp.: stereo pair of holotype (BMNH 1985.1 1.20.80), Channel Stn 253, x 62. Figs 83-87 Puellina corbula n.sp.: (83) ovicellate and non-ovicellate autozooids at edge of colony, holotype (BMNH 1985.11.20.80), Channel Stn 253, x 133; (84) edge of colony viewed at angle, holotype, x 72; (85) avicularium recumbent on neighbouring autozooid, holotype, x 350; (86) semi- erect avicularia at edge of colony, with ovicellate and non-ovicellate autozooids, BMNH 1 985. 1 1 .20. 1 OOb, Channel Stn 260, x 1 58; (87) kenozooid, ovicellate and non-ovicellate autozooids at edge of colony, BMNH 1911.10.1.735B, Belfast, x 112. BRITISH PUELLINA 51 52 J. D. D. BISHOP & B. C. HOUSEHAM eastern Atlantic (Harmelin, 1978; Hayward & Ryland, 1978; d'Hondt, 1979; Aristegui, 1983; 1984). It seems probable that a species group, rather than a single species, is involved. At least part of Harmelin's (1970) Mediterranean material, part of the NE. Atlantic material recorded by the same author ( 1 978), and the record by Hayward & Ryland ( 1 978) from the northern Bay of Biscay, are referred to P. arrecta n.sp. above. Puellina (Glabrilaria) corbula n.sp. (Figs 82-9 land Table 12) part Cribrilaria pedunculata: Harmelin 1978; 188, 189 (material from Stn 72); pi. 1 fig. 5; not pi. 1 fig. 6. Holotype: BMNH 1985.1 1.20.80 (Channel Stn 253, 48°52-2'N 03°54-0'W, 79 m). Paratypes (examined by SEM): BMNH: 1911.10.1.735B (Belfast); 1985.11.20.72c (Channel Stn 252); 1985.1 1.20.1 00a,b,c (Channel Stn 260); 1985.1 1.20. 108a and 1985.1 1.20.1 09a (Channel Stn 314). OTHER MATERIAL. BMNH: 1911.10.1.731 (British); 1985.11.20.52a (Channel Stn 248); 1985.11.20.59a (Channel Stn 249); 1985.1 1.20.69a and 1985.1 1.20.72a,b (Channel Stn 252); 1985.1 1.20.74a,b (Channel Stn 253); 1985.1 1.20.99 and 1985. 1 1 .20. 101 (Channel Stn 260). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colonies in available material pluriserial to multiserial, with over 200 autozooids. Autozooid of variable shape in frontal view, generally round or sub-oval (longer than broad). Lateral and proximal gymnocyst of autozooid narrow to broad, but generally of moderate width and frequently with proximal and proximo- lateral extensions between neighbouring zooids. Frontal shield ( = pericyst) of autozooid lightly to moderately calcified, without umbo or any other median thickening, with 9-14 costae (usually 10-12). Each costa with broad, steeply ascend- ing basal portion at periphery of frontal shield leading to narrower, tapering, nearly horizontal subsequent portion contributing to almost flat central region of shield, and with tubercle or short ridge (frequently bearing pelmatidium) near base. Intercostal pores sub-rectangular or reniform, generally slightly larger in first intercostal row (i.e. just proximal to orifice) than in remainder of shield, frequently with denticles adjacent to intercostal suture; intercostal pores present along entire length of costae (i.e. not restricted to nearly horizontal region of shield as in many other species: Fig. 83); usually 6-8 pores between successive costae, but occasionally fewer between short proximal costae. Papilla pore between bases of successive costae inconspicuous, about size of adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, slightly broader than long, proximal edge straight. Orifice of ovicellate zooid occasionally appearing slightly truncate in frontal view because of overhanging calcification of ovicell; width 1-2-1 -3 times that of orifice of non-ovicellate zooid. Operculum sometimes with central tubercle (Fig. 89, partially obscured). Oral spines 7 in non- ovicellate zooid, 4 in ovicellate zooid; jointed some distance above base with 2 (possibly more) subsequent segments, without inwardly directed apophyses. Distal portion of apertural bar fre- quently sloping downwards from proximal margin of orifice towards sub-oral lacunae; proximal portion ascending or flat; apertural bar with pelmatidium each side of midline, without tubercles. One or 2 (rarely 3) median sub-oral lacunae between proximal margin of orifice and first row of intercostal pores, of variable shape, about same size as a large intercostal pore, occasionally divided in two by median denticle. Each autozooid with c. 1 distolateral pore chambers; uncalcified external windows of pore chambers (seen in zooids on edge of colony: Fig. 84) of moderate size but considerably larger than an intercostal pore. Ovicell of category C (p. 4), length 0-7-0-8 times that of frontal shield; frontal surface non- punctate, frequently with 4-6 low ridges in more or less radiating pattern; median suture not observed. Avicularium (Figs 85, 86) minute, infrequent to fairly frequent; palate elongate-triangular, inclined at steep angle to substrate (except one example in holotype, with rostrum recumbent on BRITISH PUELLINA 53 Figs 88-91 Puellina corbula n.sp.: (88) whole colony, holotype (BMNH 1985. 1 1 .20.80), Channel Stn 253, x 36; (89) non-ovicellate autozooid, holotype, x 270; (90) ancestrula and first autozooid, BMNH 1985.1 1.20.72c, Channel Stn 252, x 191; (91) ancestrula regenerated as kenozooid, holotype, x 230. 54 J. D. D. BISHOP & B. C. HOUSEHAM neighbouring autozooid: Fig. 85); rostrum curved at extreme tip to form very small beak-like projection. Mandible elongate-triangular, slightly shorter than palate. Cystid of avicularium forming short, sub-conical, semi-erect support for rostrum; with 1 or 2 pore chamber windows near base. Kenozooid (Fig. 87) infrequent, with cribrimorph frontal shield of costae in radiating pattern. Ancestrula (Fig. 90) tatiform, with 5 pairs of spines plus unpaired proximal and distal median spines (i.e. total of 12 spines) around edge of gymnocyst. Median proximal spine overarching frontal surface, with tapering, pointed tip; remaining spines erect, without inwardly directed (or other) apophyses. Details of jointing of spines not clearly seen, but at least some erect spines jointed at base and with at least 2 subsequent segments. Ancestrula sometimes regenerated as kenozooid (Fig. 91). Table 12 Puellina corbula n.sp.: measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 51-62 56-6 6-21 14 Lf 198-263 225 7-72 17 b 57-69 62-4 4-23 8 214-257 235 5-92 13 c 58-67 62-6 5-01 10 229-270 245 4-82 17 d 54-61 57-4 4-50 8 201-243 222 5-45 20 a lo 57-68 62-7 4-69 14 Lfg 261^50 311 15-2 16 b 68-73 70-3 1-80 8 264-349 304 9-10 13 c 66-80 70-5 6-50 8 262-373 324 11-0 16 d 57-66 60-7 5-50 15 245-341 278 8-92 20 a loo 69-78 73-5 9-74 2 If 172-225 193 6-99 17 b 78-89 83-8 4-78 4 187-243 217 7-95 12 c 83-88 86-0 3-34 3 195-243 226 5-75 17 d 71-80 76-8 5-46 4 179-233 204 7-38 19 a Lov 152-153 153 0-522 2 Lr 114 1 b 145-159 152 4-99 3 — — — 0 c 151-172 160 4-50 6 — — — 0 d 174-182 178 2-31 4 — — — 0 a lov 186-192 189 2-53 2 Lav 149 1 b 192-199 196 1-95 3 — — — 0 c 182-198 194 3-32 6 — — — 0 d 196-222 211 5-50 4 — — — 0 a = holotype(BMNH 1985.1 1.20.80); b= 1985.11.20.100b;c= 1985.11.20.100c;d = BMNH 191 1.10.1.735B. ETYMOLOGY. L. corbula — a little basket; referring to the form of the frontal shield in this species. REMARKS. Puellina corbula n.sp. is most closely comparable with P. africana (Hayward & Cook, 1983), of which two paratype colonies from the E. coast of South Africa (BMNH 1983.11.5.75) have been examined by SEM. The steeply ascending basal portion of each costa is relatively much shorter in P. africana than in P. corbula so that the frontal shield is considerably less raised overall. Unlike P. corbula, the costae of P. africana are without distinct ornamentation (the small tubercles noted in the original description not being apparent in the material examined). The avicularian rostrum is much more elongate in P. africana, and avicularia commonly occur closely applied to the sides of ovicells, a position they rarely occupy in P. corbula. BRITISH PUELLINA 55 P. corbula is also similar to the Mediterranean species P. pedunculata Gautier, 1956, with which it has been confused (and which is redescribed below). In Gautier's species the frontal shield is more evenly convex, and the costae lack any trace of the tubercles or short ridges commonly seen near their base in P. corbula. Non-ovicellate autozooids of P. corbula have 7 oral spines, but there are only 6 in P. pedunculata. The avicularium of P. corbula has a relatively broad base (with 1 or 2 pore chamber windows) generally seen to rest on the substrate, and tapers from the base towards the rostrum, whereas that of P. pedunculata is fully erect and broadens slightly from a narrow base (without pore chamber windows) approximately equal in area to the uncalcified external window of the pore chamber from which it was budded. Unlike P. pedunculata, no area of kenozooidal costate frontal shield is seen distal to the ovicell in P. corbula (category C rather than category B ovicell as defined on p. 4). The ovicell of P. corbula is larger than that of P. pedunculata, both in absolute terms and relative to the respective autozooids, and generally has several ridges in a more or less radiating pattern rather than a single longitudinal ridge. P. corbula is known from the English Channel off Brittany at depths of 73-106 m, and also from a single specimen (BMNH 191 1 . 10. 1 .735B) labelled simply 'Belfast'. Specimens very similar to P. corbula n.sp. also occur in the Mediterranean near Marseille (J. G. Harmelin, pers. comm. to JDDB, 1986). In Harmelin's material tubercles or short ridges near the base of the costae are poorly developed or absent, and the frontal shield has a more evenly arched appearance than is usual in the material described here. However, the avicularium and ovicell are typical of P. corbula, and it seems probable that the Atlantic and Mediterranean forms are conspecific. This record allows the possibility that P. corbula is merely a geographical variant of P. pedunculata to be discounted, since the type locality of Gautier's species is near Marseille. Puellina (Glabrilaria) pedunculata Gautier, 1956 (Figs 95-97 and Table 13) Puellina pedunculata Gautier, 1956; 203, fig. 20. Colletosia pedunculata: Prenant & Bobin, 1966; 596, 597, fig. 207III. not Colletosia pedunculata: Harmelin, 1968; 1 199, fig. 3(5). Cribrilaria pedunculata: Harmelin, 1970; 93, 94, fig. lg,h, pi. 2 fig. 6. not Cribrilaria pedunculata: Harmelin, 1978; 188, 189 (material from Stn 72), pi. 1 fig. 5 ( = Puellina corbula n.sp.). MATERIAL EXAMINED BY SEM. Type series: LBIMM-BRY- 10384 (Marseille, two colonies on separate substrates; lectotype selected below). BMNH: 1970.6. 1.1 3 (Marseille). OTHER MATERIAL. BMNH: 1975.1. 12.185 (Chios). DESCRIPTION. Colony encrusting, unilaminar, zooids distinct; largest colony in available material with c. 55 autozooids. Autozooid of variable shape in frontal view, generally round or sub-oval (longer than broad). Lateral and proximal gymnocyst of autozooid narrow or of moderate width, frequently with proximal and proximo-lateral extensions between neighbouring zooids. Frontal shield ( = pericyst) lightly to moderately calcified, with little or no differential median thickening and no umbo, with 10-15 costae (usually 1 1-13). Each costa with short moderately to steeply ascending basal portion leading to tapering or parallel-sided nearly horizontal subsequent portion contributing to slightly convex central region of shield. Costae lacking distinct ornamen- tation; pelmatidia not observed. Intercostal pores sub-rectangular, relatively uniform in size throughout shield, sometimes with indistinct denticles adjacent to intercostal suture; 3-7 pores between successive costae. Lipped papilla pore between bases of successive costae, larger than adjacent intercostal pore. Orifice of non-ovicellate autozooid D-shaped, clearly broader than long, proximal edge straight. Orifice of ovicellate zooid sometimes appearing distally truncate in frontal view because of over- hanging calcification of ovicell (BMNH 1970.6.1.13); width c. 1-1 times that of orifice of non- ovicellate zooid. Oral spines 6 in non-ovicellate zooid, 4 in ovicellate zooid, jointed some distance 56 J. D. D. BISHOP & B. C. HOUSEHAM BRITISH PUELLINA 57 above base; further details of spine morphology not seen. Distal portion of apertural bar fre- quently sloping steeply downwards from proximal margin of orifice; proximal portion of apertural bar sloping slightly upwards to first pair of costae; without pelmatidia or tubercles, but median suture often discernible. Usually 2 median sub-oral lacunae, each about size of an intercostal pore, sometimes with indistinct denticles on proximal margin. Each autozooid with c. 10 distolateral pore chambers; uncalcified external windows of pore chambers (seen in zooids on edge of colony) of moderate size but considerably larger than an intercostal pore. Ovicell of category B (p. 4), with relatively small area of kenozooidal frontal shield (Fig. 95), length 0-5-0-7 times that of frontal shield of autozooid; frontal surface non-punctate, with low median ridge along entire length, frequently with central umbo. Avicularium (Figs 95, 97) minute, frequent, often budded from kenozooid. Palate elongate- triangular, inclined at steep angle to substrate. Mandible elongate-triangular, slightly shorter than palate. Cystid erect, arising from very narrow base, an elongate, inverted cone, without pore chamber windows. Kenozooid (Fig. 96) frequent, of variable size and shape but always smaller than autozooid, with strongly convex cribrimorph frontal shield of costae in radiating pattern. Each ovicell invariably associated with a kenozooid, as noted above. Ancestrula obscured in available material. Table 13 Puellina pedunculata Gautier: measurements in microns Specimen Range Mean C.V. Range Mean C.V. a Lo 47-50 48-8 2-74 4 Lf 194-233 213 5-51 14 b 40-55 45-2 13-3 5 167-203 187 6-47 12 c 43^6 44-5 5-37 2 199-241 215 5-29 13 a lo 66-70 67-5 2-73 4 Lfg 230-323 271 9-07 14 b 60-64 61-6 2-86 5 193-297 243 12-4 13 c 63-71 67-0 4-29 6 235-312 266 8-56 10 a loo 67-80 72-1 5-35 9 If 168-256 204 10-4 13 b 64-69 66-7 4-09 3 150-187 164 8-55 11 c 69-79 73-7 7-40 3 160-205 175 7-98 13 a Lov 103-115 109 3-96 9 Lr 66-85 73-0 11-2 6 b 105-118 108 5-46 5 67-77 70-1 5-00 7 c 120-152 132 10-1 6 66-72 68-6 3-99 5 a lov 115-153 140 7-74 9 Lav 156-158 157 1-01 2 b 126-135 131 2-66 6 — — — 0 c 148-160 155 4-48 3 — — — 0 a = lectotype (LBIMM-BRY- 10384); b = paralectotype (LBIMM-BRY-10384); c = BMNH 1970.6. 1.13. Figs 92-95 (92) Puellina flabellif era (Kirkpatrick): part of colony with avicularia, ovicellate and non- ovicellate autozooids, lectotype (BMNH 1888.1.25. 10A), Mauritius, x74; (93) P. flabellif era (Kirkpatrick): avicularia (mandibles damaged) and non-ovicellate autozooids, lectotype, x 145; (94) Puellina minima (Harmelin): autozooid with category B ovicell and non-ovicellate autozooid at edge of colony, type series (LBIMM), Cassis, France, x 118; (95) Puellina pedunculata Gautier: category B ovicell with distal avicularium, BMNH 1970.6.1.13, Marseille, x 280. 58 J. D. D. BISHOP & B. C. HOUSEHAM Figs 96-99 (96) Puellina pedunculata Gautier: part of colony with avicularia, kenozooids, ovicellate and non-ovicellate autozooids, lectotype (LBIMM-BRY-10384), Marseille, x87; (97) Puellina pedunculata Gautier: distal part of autozooid with category B ovicell (kenozooidal frontal shield largely obscured) and avicularia, lectotype, x 220; (98) Puellina scripta (Reuss): avicularium (centre, near top) and non-ovicellate autozooids, holotype (NMV 1848/38/75), Badenian, Eisenstadt, x 54; (99) Puellina venusta Canu & Bassler: avicularia, ovicellate and non-ovicellate autozooids, BMNH D55588, Badenian, Hartl-lucke, Eisenstadt, x 56. BRITISH PUELLINA 59 REMARKS. Three craniid brachiopod valves were sent on loan from LBIMM as substrates of the type series of P. pedunculata. One of these was apparently unencrusted, but the other two each bore a colony of P. pedunculata. The colony illustrated in Figures 96 and 97 is selected here as lectotype. Neither of the available specimens could be identified as that illustrated by Gautier (1956). The constant association of the ovicell with an underlying kenozooid in this species does not seem to have been remarked on before. However, a similar arrangement was noted by Harmelin (1978) in his description of the closely related species Cribrilaria cristata. Gautier (1956, 1962) and Prenant & Bobin (1966) indicated pores on the ovicell of P. pedunculata. These are not in fact present. Discussion The use of SEM has been of considerable assistance in the initial discrimination of morphologically similar species of Puellina. However, the routine identification of the great majority of colonies of the species dealt with here can be undertaken with a suitable light microscope. Table 14 gives characters by which this may be achieved. Determination of the ovicell category (p. 4) is perhaps the best starting point for an identification. The present account significantly increases the diversity ofPuellina known from the British Isles. In the most recent systematic treatment of the British cribrimorph fauna, Hayward & Ryland (1979) recognised four species of Puellina and Cribrilaria. Of these, P. setosa was included only tentatively and is not in fact of proven British occurrence (see above). It is now necessary to add P. modica n.sp., P. praecox n.sp., P. bifida, P. arrecta n.sp. and P. corbula n.sp. to the list, although P. arrecta qualifies only through a record from Guernsey; a sixth species, P. directa n.sp., occurs in the English Channel off Normandy and Brittany but has not so far been recorded from the Channel Isles. P. arrecta and P. corbula were recorded from the English Channel by Harmelin (1978) as Cribrilaria flabellifera and C. pedunculata respectively. The record of C. flabellifera from the English Channel by d'Hondt (1979) may similarly refer to P. arrecta. Available material of Recent British Puellina has a marked southerly and westerly bias, with the North Sea unrepresented. This is probably not entirely a reflection of collection bias. Shell sub- strates from the western English Channel may have a remarkably high diversity of encrustation, with 6 (BMNH 1985.1 1.20.69) or 7 (BMNH 1985.1 1.20.72) species ofPuellina on a single small shell. All nine of the species reported here from British and adjacent waters may be found in the western English Channel, although P. gattyae was not recorded on shell material offshore in this area. Four species found in British and adjacent waters also occur in the Mediterranean: P. gattyae, P. venusta, P. arrecta n.sp. and P. corbula n.sp. In addition, P. innominata is evidently very closely related to Cribrilaria innominata form B in the sense of Harmelin (1970) from the Mediterranean, and P. bifida bears considerable resemblance to the Mediterranean species P. minima (Harmelin, 1 984). P. modica n.sp., P. praecox n.sp. and P. directa n.sp. are not known from the Mediterranean. P. setosa, P. radiata (Moll, 1803), P. cassidainsis (Harmelin, 1984), P. pedunculata Gautier, 1956 and P. pedunculata in the sense of Harmelin (1968) (which is considered distinct from Gautier's species) occur in the Mediterranean but have not been found in the vicinity of the British Isles. Harmelin (1984) treated Puellina and Cribrilaria as separate genera, and laid particular stress on the form of the ovicell in differentiating between them. Conversely, Gordon (1984) treated Cribri- laria as a subgenus of Puellina, citing the presence of setiform papillae, a sub-oral lacuna and a pinnate pattern of costae as similarities which outweighed any differences between the two groups. Whilst Gordon's argument is not regarded as conclusive, his arrangement is adopted here, at least in part because difficulties have been encountered with the application of the criterion of ovicell type. Two of the newly described species, P. praecox and P. directa, have ovicells at variance with their other characters and hence fall between Puellina and Cribrilaria. Thus P. praecox has ovicells of category C (p. 4), yet in many other respects resembles a species of Cribrilaria. P. directa has ovicells of category A but otherwise, at least superficially, resembles a species of Puellina. The U 1 o o -— oo 1 1 a S s« o c IU E '5 '^- c 3 _g 11 rs •^ S Ji^>- u 2 m tr u! I'll M c « u i> C "" ^ w S "^ "S 2 *3 u cr . *3 *•• .5 CO a- 0 U Z eo *i s§ ij 111 1 11 t- rr 1 ^c S <'s s ^ U ^ •g u ^ c t8l .> 8 Q t> "^ c £ o u u 3 C 'O c >> > il & t« TD U U oo a — u .2P .2r » o u c ^ ci.^ ca ~2 00^ ** "^ x •S S ^ 13 3 C S ,8 M s 1- U a> s y 0 ^ &> 13 '" S " u CT 13 " -2P C S < I Si§ ^J IK u. "O oj O Z 6 ^ V) Tj- I-J i ^C 5 J* U o.S 'i w w • T5 Q O ..-a* 3 "o 0 c 3 "o c | ^^x 2 60 ^ >~ T3 £ 2 § S" ^ "2 i> o § •S £? ei « > o — • W)^- 1- — t "O • JJ ^ O ^ * < .3 S § 8 = i 5g|^ Z I •^ TT ,3 £ 1 II f a sJ1 b • _, ^— HI 2 13 So O 1 >, •S 9 ° 22?o C o S g 2 u i • fi M =3 o^2 «- 0 tfl J < 111 S § § — O z Z 1 ^ ^ 5 £ o u u 2 .S a icent waters. . praecox u o dges steep; ;ntre slightly mvex 2s O "7" 1 & il J= M T3 tflli «>0 B3 j- J2 * " arrow "rt >"3.« g M ~ « c <- -a ib'.S ^ ^'5 •o1 ft, U Z U V u wSS (2 2*0 Z ^ Tt V3 Z 3 S 8 > S rt Q &) u •- •- .1 in & "Hb x ^ ^ *^^ . ^i O •^ ^ — j^l Ij'S •fc •£; c o ^ - ^^ Q S ™ is *rt CO .— (U "TI3 c *^ O Jrt Cd o; o < Z 5 8 t^i ^0 O -w O wi 1 ^ rs III Q-- «« t S3 o « a 1 _>, •1 f *- 2 £= Q w u G 2-8 ^ O ^> ! a; u o U Z Moderat convex .1 III •o 9 o c 0 "« "O a 'e « >^ 0 « C «> U. Q Q _C .0 «T _ B o <_, 8 \ > !_ « ^ .fa « E g - 2 11 U Z 1 0 J> « _E « c — G — . '$> btf to u o z "3 M a u racters fo J c JS >>-^ x — M u §| | oo , — , ^ ON 0 J^ u O •o 8 o ffi "w g 0 Z 2 £.2 a c rf 8SJ I «>S £•£, S 1 .S a a£ successiv costae RJ 4= t^_ 1 ^ W U **H *u c * 5 o 13 - ^ § ll ^ JJ "C G a GO 1.1 C/3 2 I — « ^ ^ .S e E 3 _ >»^^ 3 Ji c Table 14 — "* c ^ "w a -2 * T3 Hill 13 o oi ci: Costae: i number E^ cd Q i's Proxima gymnoc; 1 •- 1 '> ^ ^ 0 o 0 S W o Avicular 1 s il < 8 BRITISH PUELLINA 61 occurrence in P. bifida and P. minima of ovicells of categories A and B in the same colony, and the evident resemblance of P. praecox to P. bifida, suggests that the conversion of one ovicell type into another may not be a particularly difficult evolutionary step. It is possible to envisage the trans- formation of category A into category B by the development of a distal kenozooid rather than an autozooid. Reduction in the size of the kenozooidal costate frontal shield until none was retained would then yield category C (in which the ovicell would nevertheless be regarded as kenozooidal in nature). Such a transformation of category A through B to C, which is considerably easier to envisage than the reverse pathway, may have occurred more than once within the Puellina clade. If so, the category C pattern of ovicell development would be a homoplasous character within the genus. In most species with ovicells of category A, ontogenetically complete maternal autozooids are commonly seen on the edge of the colony before the budding of the distal autozooid with proximal ovicell commences. In this condition, maternal autozooids may be recognised by their reduced number of oral spines (compared with non-maternal autozooids) which are closely spaced near the proximo-lateral corners of the orifice, and by the large U-shaped distal pore (? pore chamber window) in place of the smaller sub-oval pore chamber window(s) of non-maternal zooids (Figs 48, 55, 66 and 79). It may therefore be inferred that there is a distinct pause between the development of the maternal zooid and its corresponding ovicell, caused by the discontinuous growth of the colony. Although it is perhaps more usual for an ovicell of category A to be produced just one autozooid generation after the maternal zooid, a delay greater than one autozooid generation may be inferred in some cases. For instance in Figure 79, zooid 1 is a maternal zooid which contributed to the budding of completed zooid 2, and both 1 and 2 are contributing to the budding of 3. The ovicell within which an embryo produced by zooid 1 would be brooded is not yet fully developed, being a proximal component of zooid 3, two autozooid generations later than 1 . It is apparent that the U-shaped distal pore (? pore chamber window) of the maternal zooid is responsible for the production of an ovicell during the budding of a later zooid, so that the ovicell always lies immediately distal to the orifice of the maternal zooid. If the maternal zooid is involved in the budding of more than one subsequent zooid, it is the daughter zooid to which the U-shaped distal pore contributes that bears the ovicell. Thus in Figure 79 maternal zooid 1 has contributed to the budding of zooid 2 through one or more distolateral pore chambers and to the budding of zooid 3 through the U-shaped distal pore; accordingly, the ovicell is part of zooid 3. Although the ovicell usually lies at the proximal end of the distal zooid in a more or less median position, it may be displaced some distance from the proximal end and lie to one side or other of the frontal shield if the spatial packing of the zooids dictates this (compare the two ovicells in Figure 43). In species with ovicells of category C, ontogenetically complete maternal autozooids are rarely if ever seen on the edge of the colony without a developing or complete ovicell. Thus any pause between the development of the maternal autozooid and the ovicell must be very short, and the process of budding of the autozooid and ovicell is more or less integrated. In this way ovicells of category C are produced earlier, in terms of autozooid generations, than those of category A, and brooding may presumably commence correspondingly sooner. Acknowledgements We thank P. D. Taylor, P. J. Chimonides, P. L. Cook and S. Lidgard for their comments on the manuscript. E. Buge, J. L. d'Hondt, M. V. Hounsome, W. Schedl and N. Vavra kindly arranged loans of specimens. J. G. Harmelin and J. Aristegui provided encouragement and advice, and JGH lent material from his own collections. The International Commission on Zoological Nomenclature were consulted on the date of establishment of Cribrilaria. We are grateful to the staff of the Electron Microscopy Unit and the Photographic Unit at the BMNH for considerable help. References Aristegui, J. 1983. Estudio Faunistico y Ecologico de los Briozoos Quilostomados (Ectoprocta, Cheilostomata) del Circalitoral de Tenerife, 1-266. Memoria de Licenciatura, Facultad de Biologia, Universidad de La Laguna. 62 J. D. D. BISHOP & B. C. HOUSEHAM 1984. Briozoos Quilostomados (Ectoprocta, Cheilostomata) de Canarias: Estudio Sistematico, Faunistico y Biogeografico, i-iii and 1-524. Tesis Doctoral, Facultad de Biologia, Universidad de La Laguna. Banta, W. C. & Carson, R. J. M. 1978. Bryozoa from Costa Rica. Pacific Science 31: 381^24. Barroso, M. J. 1 9 1 9. Notas sobre briozoos espanoles. Boletin de la Real Sociedad Espanola de His tor ia Natural 19: 340-347. Bishop, J. D. D. 1986. The identity of Cribrilaria innominata (Couch, 1844) (Bryozoa, Cheilostomata). Bulletin of the British Museum (Natural History), Zoology Series 50: 93-102. Busk, G. 1854. Catalogue of Marine Polyzoa in the Collection of the British Museum. Part 2. Cheilostomata (part.), i-viii and 55-120. London. 1859. A Monograph of the Fossil Polyzoa of the Crag, i-xiii and 1-136. London. Canu, F. & Bassler, R. S. 1925. Les bryozoaires du Maroc et de Mauritanie. ler Memoire. Memoires de la Societe des Sciences Naturelles (et Physiques) du Maroc 10: 1-79. & 1928. Les bryozoaires du Maroc et de Mauritanie. 2e Memoire. Memoires de la Societe des Sciences Naturelles (et Physiques) du Maroc 18: 1-85. & 1929. Bryozoaires eocenes de la Belgique conserves au Musee Royal d'Histoire naturelle de Belgique. Memoires du Musee Royal d'Histoire Naturelle de Belgique 39: 1-68. • & 1929a. Bryozoa of the Philippine region. Bulletin of the United States National Museum 100 (9): i-xi and 1-685. Couch, R. Q. 1844. A Cornish Fauna. Part 3. The Zoophytes and Calcareous Corallines, i-xvii and 1-164. Truro. David, L. & Pouyet, S. 1 974. Revision des bryozoaires cheilostomes miocenes du Bassin de Vienne — Autriche. Documents des Laboratoires de Geologie de la Faculte des Sciences de Lyon 60: 83-257. Eggleston, D. 1969. Marine fauna of the Isle of Man: revised lists of phylum Entoprocta ( = Kamptozoa) and phylum Ectoprocta ( = Bryozoa). Report of the Marine Biological Station at Port Erin 81: 57-80. Gautier, Y. V. 1956. Resultats scientifiques des campagnes de la Calypso. 2. Etudes sur Pilot du Grand Congloue (Archipel de Riou, pres Marseille). 5. Bryozoaires. Annales de I'lnstitut Oceanographique 32: 189-225. 1962. Recherches ecologiques sur les bryozoaires chilostomes en Mediterranee occidentale. Recueildes Travaux de la Station Marine d'Endoume 38: 1-434. Gordon, D. P. 1984. The marine fauna of New Zealand: Bryozoa: Gymnolaemata from the Kermadec Ridge. New Zealand Oceanographic Institute Memoir 91: 1-198. Grant, A. & Hayward, P. J. 1986. Bryozoan benthic assemblages in the English Channel. Pp. 115-124 in: Nielsen, C. & Larwood, G. P. (Eds). Bryozoa: Ordovician to Recent. Fredensborg. Harmelin, J. G. 1968. Bryozoaires recoltes au cours de la campagne du Jean Charcot en Mediterranee orientale (aout-septembre 1967). 1. Dragages. Bulletin du Museum National d'Histoire Naturelle (2e Serie) 40:1179-1208. 1970. Les Cribrilaria (bryozoaires chilostomes) de Mediterranee; systematique et ecologie. Cahiers de Biologic Marine 11: 77-98. 1978. Sur quelques cribrimorphes (Bryozoa, Cheilostomata) del' Atlantique oriental. TethysS: 173-192. 1984. Bryozoaires des peuplements sciaphiles de Mediterranee: revision de Cribrilina setosa Waters et description de deux nouvelles especes de Cribrilaria (Bryozoa, Cheilostomata). Zoologica Script a 13: 81-88. Harmer, S. F. 1926. The Polyzoa of the Siboga Expedition. Part 2. Cheilostomata Anasca (with additions to previous reports). Siboga- Expeditie 28b: i-viii and 181-501. Hayward, P. J. & Cook, P. L. 1983. The South African Museum's Meiring Naude Cruises. Part 13. Bryozoa 2. Annals of the South African Museum 91: 1-161. & Ryland, J. S. 1978. Bryozoa from the Bay of Biscay and Western Approaches. Journal of the Marine Biological Association of the United Kingdom 58: 143-159. & 1979. British ascophoran bryozoans. Synopses of the British Fauna (New Series) 14: 1-312. Heller, C. 1867. Die Bryozoen des adriatischen Meeres. Verhandlungen der Zoologisch-Botanischen Gesellschaft in Wien 17 (Abhandlungen): 77-136. Hincks, T. 1880. A History of the British Marine Polyzoa 1: i-cxli and 1-601; 2: 83 pi. London. 1883. Report on the Polyzoa of the Queen Charlotte Islands. Annals and Magazine of Natural History, Series 5 11: 442-451. Holme, N. A. 1966. The bottom fauna of the English Channel. Part 2. Journal of the Marine Biological Association of the United Kingdom 46: 401-493. d'Hondt, J. L. 1970. Campagne d'essais du Jean Charcot (3-8 decembre 1968). 5. Bryozoaires. Bulletin du Museum National d'Histoire Naturelle (2e Serie) 42: 232-256. BRITISH PUELLINA 63 1979. Bryozoen aus dem Nordatlantik und dem Englischen Kanal. 'Meteor' For schungsergebnisse, Reihe D28: 1-5. Johnston, G. 1847. A History of the British Zoophytes (2nd edition) 1: i-xvi and 1-488; 2: 74 pi. London. Jullien, J. 1 886. Les Costulidees, nouvelle famille de bryozoaires. Bulletin de la Societe Zoologique de France 11:601-620. Kirkpatrick, R. 1 888. Polyzoa of Mauritius. Annals and Magazine of Natural History, Series 6 1: 72-85. Landsborough, D. 1852. A Popular History of British Zoophytes, or Corallines, i-xii and 1-404. London. Levinsen, G. M. R. 1909. Morphological and Systematic Studies on the Cheilostomatous Bryozoa, i-vii and 1-431. Copenhagen. Lidgard, S. 1985. Zooid and colony growth in encrusting cheilostome bryozoans. Palaeontology 28: 255-291 . Manzoni, A. 1871. Supplemento alia fauna dei bryozoi Mediterranei. la Contribuzione. Sitzungsberichte der Akademie der Wissenschaften. Mathematisch-Naturwissenschaftliche Classe. Wien. Abt. 1 63: 73-82. Norman, A. M. 1903. Notes on the natural history of East Finmark. Annals and Magazine of Natural History, Series 7 12: SI-US. Osburn, R. C. 1950. Bryozoa of the Pacific coast of America. Part 1, Cheilostomata-Anasca. Allan Hancock Pacific Expeditions 14: 1-269. Prenant, M. & Bobin, G. 1966. Bryozoaires. 2. Chilostomes Anasca. Faune de France 68: 1-647. Reuss, A. E. 1 848. Die fossilen Polyparien des Wiener Tertiarbeckens. Naturwissenschaftliche Abhandlungen. Wien2\ 1-109. Simpson, G. G., Roe, A. & Lewontin, R. C. 1960. Quantitative Zoology, i-vii and 1-440. New York. Soule, J. D. 1959. Results of the Puritan-American Museum of Natural History expedition to western Mexico. 6. Anascan Cheilostomata (Bryozoa) of the Gulf of California. American Museum Novitates 1969: 1-54. Thornely, L. R. 1912. The marine Polyzoa of the Indian Ocean, from H.M.S. Sealark. Transactions of the Linnean Society of London, 2nd Series, Zoology 15: 137-157. Waters, A. W. 1 879. On the Bryozoa (Polyzoa) of the Bay of Naples. Annals and Magazine of Natural History, Series 5 3: 28-43. 1899. Bryozoa from Madeira. Journal of the Royal Microscopical Society 1899: 6-16. Winston, J. E. 1984. Shallow-water bryozoans of Carrie Bow Cay, Belize. American Museum Novitates 2799: 1-38. 1986. Bryozoans from Bali, Lombok and Komodo. American Museum Novitates 2847: 1-49. Manuscript accepted for publication 23 September 1986 Peter R. Colston & Kai Curry-Lindahl For evolution and speciation of animals Mount Nimba in Liberia, Guinea and the Ivory Coast is a key area in Africa representing for biologists what the Abu Simbel site in Egypt signified for archaeologists. No less than about 200 species of animals are endemic to Mount Nimba. Yet, this mountain massif, entirely located within the rain-forest biome, is rapidly being destroyed by human exploitation. This book is the first major work on the birds of Mount Nimba and surrounding lowland rain- forests. During 20 years (1962-1982) of research at the Nimba Research Laboratory in Grassfield (Liberia), located at the foot of Mount Nimba, scientists from three continents have studied the birds. In this way Mount Nimba has become the ornithologically most thoroughly explored lowland rain-forest area of Africa. The book offers a comprehensive synthesis of information on the avifauna of Mount Nimba and its ecological setting. During the 20 years period of biological investigations at Nimba this in 1962 intact area was gradually opened up by man with far-reaching environmental consequences for the rain-forest habitats and profound effects on the birds. Therefore, the book provides not only a source of reference material on the systematics, physiology, ecology and biology of the birds of Mount Nimba and the African rain-forest, but also data on biogeography in the African context as well as conservation problems. Also behaviour and migration are discussed. At Nimba a number of migrants from Europe and/or Asia meet Afrotropical migratory and sedentary birds. Professor Kai Curry-Lindahl has served as Chairman of the Nimba Research Laboratory and Committee since its inception in 1962. Peter Colston is from the Subdepartment of Ornithology, British Museum (Natural History), Tring, and Malcolm Coe is from the Animal Ecology Research Group, Department of Zoology, Oxford. 1986, 129pp. Hardback. 0 565 00982 6 £17.50. Titles to be published in Volume 53 Puellina (Bryozoa; Cheilostomata; Cribrilinidae) from British and adjacent waters. By J. D. D. Bishop & B. C. Househam The genera of pelmatochromine fishes (Teleostei, Cichlidae). A phy logenetic review. By P. H. Greenwood The cranial muscles and ligaments of macrouroid fishes (Teleostei, Gadiformes); functional, ecological and phylogenetic inferences. By G. J. Howes Certain Actiniaria (Coelenterates, Anthozoa) from the Red Sea and Tropical Indo- Pacific Ocean. By K. W. England Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset Bulletin of the British Museum (Natural History) Miscellanea Zoology series Vol53 No 2 29 October 1987 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), 1987 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Mr J. F. Peake Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISBN 0 565 05034 6 ISSN 0007-1498 Zoology series Vol53 No. 2 pp 65-1 38 British Museum (Natural History) Cromwell Road London S W7 5BD Issued 29 October 1 987 Miscellanea Contents Page Notes on Atlantic and other Asteroidea. 5. Echinasteridae. By Ailsa M. Clark 65 Observations on the marine nematode genus Spirina Gerlach, 1963 (Desmodoridae: Spiriniinae) with descriptions of two new species. By J.W. Coles 79 Caleupodes, a new genus of eupodoid mite (Acari: Acariformes) showing primary opisthosomal segmentation. By A. S. Baker 103 The Barbus perince-Barbus neglectus problem and a review of certain Nilotic small Barbus species (Teleostei, Cypriniformes, Cyprinidae). By K. E. Banister 115 x^^^l^ r/^t&j™^ ' 290CTW87 \* Notes on Atlantic and other Asteroidea. 5. Echinasteridae Ailsa M. Clark* Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Introduction In the course of reviewing the asteroids of the Atlantic, deciding on the best subdivision of the taxonomically difficult family Echinasteridae has posed numerous problems. Echinasterids are very variable and tend to diversity into local forms, some of them sufficiently well marked and constant to be worthy of recognition as named taxa. Instances of this are provided by Fisher (1911), Hayashi (1940) and Dyakonov (1950) when dealing with the asteroids of three overlapping areas: the whole North Pacific, the vicinity of Japan and the USSR. The three of them described respectively: 8 species and 5 subspecies, of which 9 were new, 17 species of which 10 were new and 19 species with 8 new, all in the genus Henricia. With regard to the various Atlantic taxa, Dr F. Jensenius Madsen (pers. comm.) has recently undertaken a revision of the genus Henricia in connection with the series 'Marine Invertebrates of Scandinavia' but this still leaves Echinaster, which is well represented in the western tropical Atlantic, to be reassessed. The present study is based partly on examination of the relevant types and partly on a large collection from the Pillsbury, Gerda, Oregon and other vessels. The taxa of Echinaster (Othilia) are the main concern, especially two new species now described, but one new species of Henricia is also distinguished. Diagnoses for Echinaster (Echinaster) and Echinaster (Othilia) are given and a tabular key is provided to the Atlantic species of Echinaster, utilizing both old and newly recognized characters now thought to be of taxonomic significance. Because of superficial similarities between species, it must be emphasized that few reliable conclusions can be reached about the identity of most echinasterids without first removing the skin from an arm with bleach to reveal the precise form and arrangement of the underlying skeletal plates. In the case of Henricia, preparations of the spinelets are also invaluable. Systematic Account Family ECHINASTERIDAE Verrill Echinasteridae Verrill, 1870: 343; Fisher, 1911: 258-260; Downey, 1973: 82-83; Clark & Courtman-Stock, 1976:88. Only three genus-group names are involved in the Atlantic fauna, namely Echinaster Muller & Troschel, 1840, Henricia Gray, 1840 and Othilia Gray, 1840. The main problem is the status of Othilia which was revived as a distinct genus from the synonymy of Echinaster Muller & Troschel, 1840, by Tortonese & Downey (1977) on the basis of the presence of patches of crystal bodies (or glassy tubercles) on many of the primary plates in the type species of Othilia, Asterias echinophora Lamarck, 1816. Such structures have not been observed in Asterias seposita Retzius, 1783, the type species of Echinaster (see Tortonese & Madsen, 1978). Unfortunately, discovery by Clark & Tortonese (1986) that specimens from West Africa considered to be only subspecifically distinct from E. sepositus also have patches of crystal bodies, though ill-defined and restricted to some proximal plates, devalues this character as being of generic weight. However, there is support from two other characters which could be of generic or, at least, subgeneric weight. Firstly, the con- sistent enlargement to some degree of the primary carinal and adradial series of dorsolateral abactinal plates so as to frame two distinct alternating longitudinal series of skeletal meshes along the median part of each arm, coupled with further longitudinal arrangement of abradial dorso- *Present address: Gyllyngdune, South Road, Wivelsfield Green, Haywards Heath, Sussex RH17 7QS Bull. Br. Mus. not. Hist. (Zool.) 53(2): 65-78 Issued 29 October 1 987 66 A. M. CLARK laterals with transverse links between the adradials and the superomarginals. This is restricted to the species included in Othilia. Although Echinaster callosus von Marenzeller from the Indo-West Pacific has large, longitudinally arranged adradial plates, these are linked by diagonal chains of plates and the carinals are not linked in a continuous series. E. callosus probably merits supra- specific distinction for itself but that is beyond the scope of the present study. E. sepositus and the other species of Echinaster have an irregular reticulum of finer abactinal plates, though in some individuals the abradial plating may tend to form oblique longitudinal and short transverse lines converging on the superomarginal series. Only E. modestus Perrier from the west tropical Atlantic may show some enlarged adradial meshes on the proximal half of the arm suggesting the condition found in the species of Othilia. Secondly, there is a complete absence of actinal plates in all the species of Othilia, the inferomarginals bordering directly on the adambulacrals. (A superficial appearance of supposed actinal spines in some individuals is due to development of isolated spines on either the adradial ends of inferomarginals or the abradial ends of adambulacrals.) In E. sepositus and E. modestus at least one partial series of actinal plates is developed proximally. However, it must be mentioned that some non-Atlantic species currently included in Echinaster also lack actinal plates. Pending a full review of all of these, it seems best to treat Othilia as a subgenus of Echinaster. Among the more useful specific characters of this subgenus now recognized are some resulting from differential development of the secondary abactinal plates developed between the adradials and the superomarginals during increase in girth of the arms — namely the abradial plates (Figs la,b, 3A,B). In most species of Echinaster (Othilia) transverse series of abradial plates initially formed are subsequently linked longitudinally or obliquely by elongate plates and it is usually these that come to bear the abradial spines. However, in some cases the spines are borne preferentially on the transverse plates. Another important character is provided by the nature of the articulations between the two kinds of plates. Usually the longitudinal ones are superimposed on the transverse plates and stand out to some extent, especially medially if the area bearing the spine is swollen, but sometimes the plating is integrated so that the longitudinal plates lie between the transverse bars flush with the surface. Surprisingly, a convex madreporite proved to be consistent throughout a wide geographical range in one new species. The shape and armament of this unique plate may be significant to a lesser extent. Campbell & Turner (1984) have also utilized the regularity of the gyri as a specific character. ECHINASTER (ECHINASTER) Miiller & Troschel Echinaster Muller & Troschel, 1840: 102; Tortonese & Downey, 1977: 829-830, RhopiaGray, 1840:281. Verrillaster Downey, 1973: 89. TYPE SPECIES. Asterias seposita Retzius, 1 783, by subsequent designation by Fisher, 1913 under the name of Asterias sagena Retzius, 1805; validation of A. seposita by Tortonese & Madsen, 1979. A subgenus of Echinaster in which the primary abactinal skeleton is not distinct; the plating forms an irregular reticulum medially, though sometimes resolves itself abradially into oblique longitudinal and nearly transverse lines meeting the superomarginal series; few species have a marked distinction between primary and secondary plates. The armament consists of small, usually blunt, slightly spaced spines, most often 0-75-1 -0 mm long, rarely exceeding 1 -5 mm; these may be limited to the nodal plates. The surface of the plates is usually smooth but exceptionally the matrix of certain proximal plates shows small inconspicuous areas of well embedded crystal bodies, usually only distinguishable when denuded. The marginal plates are in fairly regular rows and more or less enlarged in comparison with the abactinal plates, though never conspicuous. Both series are mainly aligned below the ambitus or widest past of the arm but interradially the early supermarginals may be displaced upwards to some extent if the wedge of intermarginals is widened. The inferomarginals are separated from the adambulacrals proximally at least in the Atlantic species, by at least one partial series of actinal plates. Most of the adambulacral spines are aligned in an irregular line transverse to the furrow into which one or more modified spines are inset. ECHINASTERIDAE 67 Fig. 1. a. Echinaster (Othilia) guyanensis sp. nov. Holotype. Lateral but slightly dorsal view of proximal part of arm (interradius to left) showing some carinal plates in profile (C), adradials (D), transverse and superimposed abradials (B) and superomarginals (S). b. E. (O). paucispinus sp. nov. Holotype. Lateral view, letters as in (a) with addition of inferomarginals (I). Glands (cross-hatched) and papular pores (black) shown in only a few areas, c. E. (O.) sentus (Say) half cross-section of proximal part of arm, plates lettered as in (a) and (b), the carinal shown offset from midline (arrowed), the adradial partly in section (cross-hatched), spines behind the plane of section shown by discon- tinuous lines. d,e. Echinaster (Echinaster) downeyae sp. nov. Gerda st. 311: d. abactinal spinelets, e. cross-section through furrow showing inset furrow spines (F) of two opposite adambulacral plates with a pair (P,P) of larger slightly inset spines, one behind the other, superficial adambulacral spines and also (on left only) actinal spinelets (A) of one plate, sutures omitted. The scale = 2 mm for a-c, 0-25 mm for d, 0-5 mm for e. 68 A. M. CLARK ECHINASTER (OTHILIA) Gray Othilia Gray, 1840: 281; Fisher, 191 1: 260 (in key), 261; Tortonese & Downey, 1977: 830. Echinaster: Verrill, 1915: 35-36; Downey, 1973: 85-86; Walenkamp, 1976: 17-18; de Avila-Pires, 1983: 431- 436. TYPE SPECIES. Asterias spinosa Retzius, 1805, non Pennant, 1777, =A.echinophora Lamarck, 1816, by subsequent designation by Fisher, 1911. A subgenus of Echinaster in which the primary abactinal skeleton is distinct or even conspicuous in adult specimens, there being a pentagon of linked primary radials and interradials on the disc around the anal area and a large pentagonal or triangular midradial skeletal mesh distal to each primary radial opposite the base of each arm; this is followed along the arm by two series of alternating adradial meshes delimited by the median longitudinal carinal series of plates with an adradial series of dorsolateral plates each side. These primary nodal plates are often linked by one or more secondary intercalary plates in adults, enlarging the meshes; other secondaries link the adradials to the superomarginals, the abradial meshes being usually smaller and subdivided by a distally incomplete longitudinal series of elongate secondary abradials; there is sometimes also a shorter second abradial series in very large specimens (R > c. 60 mm) of some species, or else a longitudinal series is lacking in others where the abradial meshes are broad rectangular. The armament ranges from single large spines on some nodal primary plates, exceeding 1-5 mm in length, to smaller spaced spines or coarse spinelets on most plates including the intercalaries, sometimes multiple on some primaries; the spines of the abactinal and marginal plates form more or less regular longitudinal series totalling 7-13, most often 11. Patches of crystal bodies (glassy tubercles) are embedded in the matrix of many larger plates, especially primaries, and are visible when denuded (or dried unless the skin is very thick); the patches are occasionally reduced in extent or absent, especially when there are large single spines. The marginal plates are in regular horizon- tal rows, slightly enlarged, usually rhombic; the superomarginals are aligned at or just below half the height of the arm, along the ambitus or widest part of the circumference, but diverge from the inferomarginal series basally when the intermarginal plating is well developed; the inferomarginals are aligned ventrally and are contiguous with the adambulacrals throughout their extent, actinal plates being completely absent; the two inferomarginal series of adjacent rays meet interradially distal to the oral plates with usually a single interradial plate. The adambulacral spines are few and coarse, tending to be aligned in an irregular row transverse to the furrow; one or sometimes two or three spines are inset on to the furrow face of the plate, the innermost at least reduced in size, compressed and curved. Echinaster (Othilia) guyanensis sp. nov. Figs la, 2A,B Echinaster sentus: John & Clark, 1954: 140 (listed). [Non E. sentus (Say, 1825)]. lEchinaster echinophorus: Cherbonnier, 1959: 171-172, fig. 3C,D; ?Zoppi de Roa, 1967: 281. [Non E. echinophorus (Lamarck, 1816)] Echinaster brasiliensis: Downey, 1973: 88, pi. 42, figs C,D; Walenkamp, 1976: 78-81, pi. 18, figs 3,4; 1979: 72-74, pi. 15, figs l^t, pi. 16, figs 1^. [Non E. brasiliensis Muller & Troschel, 1842]. MATERIAL. Oregon st. 4190, 05°48'N, 52°53'W, off French Guyana (Guiana), 45m, holotype (USNM). Also from numerous other Oregon and Pillsbury stations between Honduras and the Guyanas as well as in offshore parts of the Caribbean from Pedro Cays, S of Jamaica and from S of Haiti. The westernmost record is from the SW side of Cape Tres Puntas, Guatemala, collected by D. Nestor, via Professor T. Hopkins. HOLOTYPE. R/r 40/9mm = 4-4/l; Br/br (proximal breadth to distal breadth measured at 10% R from tip) 10-0/5-0 mm = 2-0/1; R/br = 8-0/l. The five arms are fairly broad, even at the tip, somewhat tapering in the distal half, appearing rather flattened; the skin is thick enough to obscure the details of the plating but not the regular pattern; the primary pentagon on the disc is conspicuous, also the primary adradial skeletal meshes ECHINASTERIDAE 69 alternating along the arms. The plate area is approximately equal to the mesh area. Some inter- calary secondary plates link the nodal primaries both longitudinally and transversely. The adradials are linked to the superomarginals laterally by chains of one to three transverse plates on which are superimposed a well-developed, almost continuous, longitudinal series of elongate abradial plates, extending nearly to the tip of the arm, most of them convex medially, while 14 of those on one side of the arm which was cleaned bear an abradial spine. None of these secondary abradials bears a patch of crystal bodies (or glassy tubercles) in contrast to many of the primary carinals and adradials on which the patches usually extend for about half the exposed area of the plate. Carinal spines are fairly well developed but are reduced or absent on some nodal plates; spines are better developed on the adradials, numbering 12-14 in a longitudinal series. All these spinose primaries have the spine mounted on a rounded elevation, the larger ones inset from the distal end of the plate. The spines are single, conical, taper to a point and measure up to 1-5 mm long. The larger proximal adradial skeletal meshes include c. 6 papulae and 8-10 dermal glands. The superomarginal plate series is aligned just below half the height of the arm and the plates are rhombic in shape. They number c. 27 on one side of the arm cleaned; of these c. 1 5 bear a single upper or median spine and have smaller patches of crystal bodies than the spineless plates which usually alternate with them. The distal spines are inconspicuous but the more proximal ones are the largest of all the spines. Intermarginal plating is restricted to the base of the arm and there are few intermarginal spines. Thus the number of longitudinal spine series on one arm is 9 rather than the more usual 1 1 of this subgenus, that is one carinal and two each of adradial, abradial, supero- and inferomarginals. The inferomarginal plates number c. 36, of which 29 have spines in one series, only sporadic plates being spineless. Actinal plates are lacking, the inferomarginals abutting directly on to the adambulacrals. Most adambulacrals bear 4 spines of which two are inset into the furrow but there may be an occasional fifth spine either at the abradial (lateral) end or slightly inset into the furrow, especially on some proximal plates. The spines form a fairly straight transverse row and are webbed together. Most of the oral plates bear 3 marginal spines and one suboral; there are probably also a few inset spines near the distal end hidden in the furrow. The madreporite is 2-7 mm in diameter and almost circular; it is very unusual for an Echinaster in being distinctly convex; the fairly straight ridges which separate the gyri are studded with spinules all along their lengths. OTHER SPECIMENS. The maximum R is 67 mm; R/r ranges from 4-1-4-9/1, mean 4-5/1 in 14 speci- mens with R > 30 mm; R/br ranges from 8-0-12-0/1 , the degree of taper being rather varied and the arms narrower at the tips in some specimens. The primary skeleton is always fairly regular and conspicuous, the meshes showing dark in contrast in the dried specimens. The reticulum may be more delicate and open than in the holotype and some of the meshes may be partially confluent. The frequency of the spines and their size is often least on the carinal series but both increase towards the superomarginals. The size of the patches of crystal bodies often exceeds half the exposed area of those primary plates that bear them, the patches in the holotype being smaller than usual. The number of adambulacral spines is usually 5 on most plates of the proximal half of the arm, sometimes even 6, with 3 spines often inset into the furrow. DISCUSSION. Out of a total of > 1 50 specimens taken between Guatemala and French Guyana, only four have R > 50 mm. The distinctly convex madreporite appears to be an almost constant feature, only three specimens studied having the plate virtually flat. The body form with arms appearing slightly flattened and more tapering distally is also remarkably consistent in the material as dried. The main variations are the relative coarseness of the reticulum and the degree of development of the spines. There seems to be some local tendency in one direction or another, for instance in Venezuelan specimens the skeleton seems to be unusually delicate and the spines smaller. Further to this, the collection of the British Museum (Natural History) includes three very old specimens from Puerto Cabello, W of Caracas, Venezuela (c. 68°W), at least two of them collected by Brandt, the third coming via the Wyville Thomson collection and possibly also from the same source, collectors in this area of South America prior to 1850 being few. These specimens have a very delicate, almost lacelike, reticulum, a few plates in the abradial areas being longitudinally aligned and slightly superimposed; no spines or articulations for spines can be seen (though admittedly the 70 A. M. CLARK Table 1. Tabular key to the Atlantic species of Echinaster (Othilia) above and Echinaster (Echinaster) below the line 8 echinophorus 20-57 3-5-4-8/4-4 c EF r ua C 7(9) SP C c guyanensis 29-67 4-1-4-9/4-5 bn OC ru C 9(11) SP KD)(C) n(i) sentus 27-63 3-8-4-8/4-4 b(c) OEF rux O 11 SP 1C n spinulosus 35-90 4-1-5-9/5-0 ns OF rx D(C) 11 S(0)B(P) D i graminicola 15-27 3-3 b=c OF ip C 9(11) SB D i serpentarius 41,48 4-8,5-3 s(?b) EF ia c* 11 MB D i brasiliensis 35-66 4-5-5-3/4-8 b OF ri ax c 9-11 S(M)B I,D n paucispinus 38-63 4-5-5-8/5-3 n(sb) OF rx(u) c 11(12) SB D d(i) sepositus 50-90 5-0-8-0/6-5 db(a) OF (r)p N X SOB I n sep. madseni 44-80 6-8,7-1 a(n) OF f r) p N X SO(B) I i modestus 33-78 4-0-6-1/4-9 sa(c) EF r p J* X 0(S)B(P) I i reticulatus c.70 4-7, 5-0 c OF (r)p N X S(O)P — i 1 . Range of R in specimens examined: 2. Range and mean of R/r: 3. Shape of arms (of limited use because of variability); c — short, stout, cylindrical, only slightly tapering, d — digitate, relatively long and fingerlike, fairly broad at tips, br (breadth at 10%R from tip) usually 5-7 mm at R c. 40 mm. b — broad throughout but slightly tapering, tips still fairly broad (Fig. 2A). n — tapering evenly but markedly to fairly narrow tips, br 2-5-5-0 mm (Fig. 2C). s — slender, tapering more basally than beyond, a — long and tapering to attenuated tips. 4. Shape of madreporite: O — almost circular. E — distinctly elongate, interradial diameter at least 1 -25 x tangen- tial diameter. C — surface convex. F — surface flat or slightly concave. 5. Surface of madreporite: p — with peripheral encircling spi- nules only. a — with additional spinules near the periphery. x — with extra spinules even centrally but peripheral ones larger, u — with uniform spinules all over the ridges, or peripheral ones appearing smaller, i — gyri irregular, r — gyri radiating fairly regularly, separated by almost straight ridges. 6. Distinctness of primary pentagon outlining anal area on disc and of enlarged double series of primary adra- dial skeletal meshes along arms, at least in intact dried specimens. C — conspicuous (Fig. 2A,C). D — distinct. J — the proximal adradial meshes at least on arms just distinguish- able. O — obscured by thick skin. N — no distinct longitudinal arrange- ment. * — pattern not or less distinct on disc than on arms. 7. Number of longitudinal series of abactinal and marginal plates or spines proximally: i.e. carinals, adradial and abradial dorsolateral, supero-, inter- and inferomarginals (multiple marginal spines counted as one; this is an odd number unless the carinal series is so zig-zag as to appear as two series). X — only limited longitudinal arrangement or complete irre- gularity. 8. Abactinal spines: number, shape and alignment: M — multiple on all or most plates. 0 — often multiple but with rarely more than two spines on one plate. S — predominantly or entirely single. P — pointed. B — blunt, often bullet shaped. C — arising from apex of large con- vexity. 1 — inset from distal end, sometimes central, convexity smaller. D — arising near distal edge of plate from only a minimal elevation (spine often appressed distally in preservation). 9. Abactinal and superomarginal spines: size at R 40-50 mm, largest spines only. c — conspicuous, at least 1 -5 mm. n — noticeable, 1-1-1 -4 mm. i — inconspicuous, 0-75-1-0 mm. d — diminutive, c. 0-5 mm. 10. Crystal bodies (glassy tubercles): occurrence and patch size. + 1 -- present or absent. e — extending over at least half the exposed area of many primary plates. r — restricted to compact patches less than half the plate area (Fig. f — few and appearing deeply inset in the matrix or eroded, in very small patches. ECHINASTERIDAE 71 10 12 13 14 15 16 17 18 + r(c) >( = ) U(S) a(D 15-27/30-65 R 0 3(4) 0 2(1) T + e(r) < = > S 1 22-36/31-73 R 0 4,5 0 3,2 T + r(f) > M(S) pl 24-35/47-82 E(R) 0 3(4) 0 2(1) T(0) + e(r) S 1 30-^8/87-100 E o 4(3) 1 1 O -1- e > S 1 20-23/C.73 (E) o 2,3 0(1) 1 — + r S 1 — — E o 5,4 0 2(1) T -1- r < F t(a) 22-32/61-100 E 0 5,4 0 1(2) T + r <(=) F lo 29^0/73-97 E o 4(3,5) 0,1 2 T(0) -> R(0) u R s,m 3(2) 0 1 T + f > R u — — R s,m 3 0 1 T > R(0) u — — R(E) m 5(4-7) 1 2(1) (T) — — =? R u — — R s 3(4) 0 1 ? 1 1 . Area of proximal primary abacti- nal plates relative to adradial mesh area (estimated approximately by eye). > — greater. = — equal. < — less. 12. Abradial dorsolateral plating: U — undeveloped, only short series of one or two intercalary plates linking adradials with supero- marginals. F — more extensive, all plates nearly flush with the surface, consisting of chains of trans- verse rectangular plates, some- times bifurcating, often linked by oblique or longitudinal plates set between the trans- verse ones, that is integrated with them, all these plates lack- ing patches of crystal bodies (Fig. IB). S — consisting of longitudinal elon- gate and more or less convex spinose plates standing out from the surface, their ends be- ing superimposed on the trans- verse chains, often forming continuous longitudinal series (Fig. la). M — of mixed form, some plates becoming modified into poly- gonal shapes with the addition of extra lobes and sometimes also developing a patch of crystal bodies or bearing a spine, though other proximal ones are elongate and super- imposed while distal ones are all polygonal with crystal bodies, resembling the adrad- ial plates. O — tending to form oblique linear series towards the superomar- ginals, sometimes forming a partial trelliswork. R — forming an irregular reticulum without consistent linear arrangement. 13. Abradial dorsolateral spines: a — absent or few, limited to arm bases. 1 — all or most on longitudinally- aligned or oblique plates. p — all or most on polygonal plates. t — all or most on transverse plates, even if oblique or longitudinal plates are present. o — all on transverse plates, which are the only ones present. u — uniformly on most abradial plates. 14. Spinose supermarginal plates: total number of plates in series and per- centage with a main upper or distally middle spine at R 40-60 mm. 1 5 . Occurrence of intermarginal plates: E — extending beyond half the arm length (measured laterally from interradius to arm tip). R — restricted to less than half the arm length. 16. Actinal plates: m — in multiple series proximally. s — in a single series, even on disc. o — lacking altogether, inferomar- ginals contiguous with adam- bulacrals throughout (note that isolated lower inferomarginal spines or lateral adambulacral ones may simulate actinal spines. 17. Adambulacral spines: total number on one plate; number on abradial or 'subambulacral' part of the plate; number inset into furrow and appear- ing smaller than the main furrow mar- gin spine, its tip at a higher level. 18. Arrangement of adambulacral spines: T — in a transverse, almost straight series, though the spine follow- ing the main furrow margin spine may have its tip aligned distally. O — this same spine more obviously offset distally. 72 A. M. CLARK Fig. 2. A. B. Echinaster (Othilia) guyanensis sp. nov. C. E. (O.) paucispinus sp. nov. Holotypes. All xl-5. ECHINASTERIDAE 73 preservation is poor). Thomson gave his specimen an MS name derived from the last character but further material from the area is needed to establish if such an extreme variant still exists. Unfortunately, all Zoppo de Roa's Venezuelan samples (1967) appear to be from localities east of Caracas or from offshore islands to the north and east. She referred her specimens to E. echinophorus and described them as having prominent spines ('espinulas conspicuas') 1 + mm long in 7-9 series; they were commonly found among mangroves at l-5m. Surprisingly, despite abundant collections, neither Engel (1939) nor Ummels (1963) recorded any specimens of Echinaster from the Netherlands Antilles to the north of Venezuela. On the northern coast of Brazil to the east of the Amazon outflow, which might form an ecological barrier, a few other specimens of Echinaster were taken by the Oregon which resemble E. guyanensis in some characters but are otherwise more like E. echinophorus, the common species on the east coast of Brazil north of c. 20°S. The taxonomic status of these specimens is uncertain; they will be discussed further by Clark & Downey (in prep). The fact that such a common species of Echinaster with a fairly wide geographical range has previously escaped recognition can be attributed not only to the confusion as to specific limits in the genus in the tropical west Atlantic which has prevailed, but also to the superficial resemblance between E. guyanensis and E. brasiliensis, which latter species was only recently shown by de Avila Pires (1982), in a fine review of Echinaster in Brazil, to be limited in that country to the southern half, giving way to E. echinophorus in the northern part of the east coast. Due to lack of material, she was unable to establish the specific identity of any specimens from the north coast of Brazil. RANGE. From Guatemala east and south along the Atlantic coasts of Central America and northern South America, at least to French Guyana, possibly also to westernmost north Brazil; 13-106 m; from coarse calcareous sand; shell gravel; muddy sand; sand and shells; muddy sand and gravel; or 'muddy' or 'sandy' substrates. Echinaster (Othilia) paucispinus sp. nov. Figslb,2C,3A,B Thyraster serpentarius: Gray, Downey & Cerame- Vivas, 1968, fig. 29 [?non text]. Echinaster sp. B, Downey, 1973: 89, pi. 41, figs C,D. MATERIAL. Hourglass Project st. D. 27°37'N, 83°58'W, c. 65 miles W of Egmont Key, Florida, 55m, holotype (USNM), paratype (BM(NH) no. 1986.1.2.1) and second paratype (Florida Department of Natural Resources, St Petersburg Reference Collection no. 1 6816). Other samples from Hourglass sts E, also W of Egmont Key, 73 m, J, K, L and M, W of Sanibel Island, 18, 37, 55 and 73 m (31 specimens). Also from St Petersburg Laboratory collection: Topsoil Bluff, Walton County, NW Florida, 30°22'N, 86°21-5'W, 12m (2); off Panama City, NW Florida, 30°05'N, 85°46'W, (5). University of Alabama no. 8114-0157, off Panama City, c. 30m; no. 8114-0147, 28°32'N, 84°19'W, c. 30m; no. 8114-0153, 27°23-5'N, 84°08'W, 85m (8). Continental Shelf Associates, Inc., via Dr R. Turner, NW of Dry Tortugas, 25°12-9'N, 83°22-7'W, 67 m (1). HOLOTYPE. R/r 55/5-5 mm = 5-8/1; Br/br (proximal breadth to distal breadth at 10% R from tip) 10-5/4-0 = 2-7/1; R/br= 13-8/1. The five arms are relatively slender and taper evenly to rounded tips. The skin is thin and the regular pattern of the primary skeleton conspicuous, the meshes appearing dark in dried specimens in contrast to the plates; the meshes appear about equal to the plates in total area but may slightly exceed them. Single secondary intercalary plates link most primary carinals and adradials both longitudinally and transversely on the proximal parts of the arms. The adradials are linked to the superomarginals by short transverse chains of secondary abradial plates some of which are linked longitudinally by other abradials integrated with them, that is inset on the same level, not super- imposed. Most of the abradial spines are borne on these longitudinal plates. None of these second- ary abradial plates bears a patch of crystal bodies (or glassy tubercles) embedded in the surface matrix, in contrast to nearly all the primary carinals and adradials. These patches are compact and relatively small, occupying less than half the surface area of the plate. This is correlated with the remarkably small size and low frequency of the spines, most of which are little more than 0-5 mm Fig. 3. A, B. Echinaster (Othilia) paucispinus sp. nov. Holotype (A) and paratype (B) Lateral views of proximal part of arm. C-E. Henricia downeyae sp. nov. Holotype. E. Lateral view of arm. A,B x4, C,D x 1-5, E x6. ECHINASTERIDAE 75 long; however, they are rubbed and poorly preserved in this specimen; in the paratypes they are tapering and blunt-tipped but still inconspicuous, at most 0-75 mm long, and arise at the distal ends of the plates, except for those on the longitudinal abradials. The larger proximal adradial meshes include a total of c. 12 mixed papulae and glands in approximately equal numbers but the number of papulae declines towards the lower surface, only single glands occurring between some inferomarginals and adambulacrals. The superomarginal plate series is aligned just below half the height of the arm. The individual plates are 4-lobed and mostly contiguous but distally some of them are slightly separated or linked by secondary plates; most bear a diminutive spine just above and distal to the patch of crystal bodies. The intermarginal plating extends for more than half the lateral arm length but is irregular, few plates bearing spines, so that there are only 9 longitudinal series of spines: one carinal and two each of adradial, abradial, supero- and inferomarginals, but all the spines are very inconspicuous. The inferomarginal plates are similar to the superomarginals in shape; on the lower side they are directly contiguous with the adambulacrals, no actinal plates being developed. The adambulacrals bear 4 or occasionally 5 spines of which two are more or less reduced in size and inset into the furrow on most plates of the proximal half of the arm; the abradial part of the plate may bear a spine when there are five but is more often bare. The madreporite is oval, 2-7 x 2-2 mm, flat and with radiating ridges; the spinules are somewhat rubbed but there appear to be some inset from the periphery in addition to the peripheral ones. The paratype with R c. 59 mm probably has the armament intact and shows that spinules are absent from the central part; it has c. 15 peripheral spinules. OTHER SPECIMENS. The paratype just mentioned also has fairly slender arms though the tips are slightly broader. It differs from the holotype mainly in lacking longitudinal abradial plates (Fig. 3B) though a few of the transverse chains of proximally three, distally two, abradials fork; the uppermost (or middle one when three) often bears an abradial spine. The plating of the wide intermarginal areas tends to follow the same pattern as that of the abradial areas and is more regular than in the holotype. Another 'Hourglass' specimen has particularly stout arms, R/r being 51-52/1 1-5 mm = only 4-5/1, with Br/br 13/6 mm and R/br only 8-5/1. DISCUSSION. This species resembles most closely E. brasiliensis M tiller & Troschel, known from the southern half of Brazil and northern Argentina. They share reduced development of longitudinal abradial plates which are integrated, rather than superimposed, when present, coupled with the relatively slender skeletal reticulum and small size of the crystal body patches. However, in E. brasiliensis the abradial spines are almost invariably borne on the transverse plates, even when longitudinal or oblique ones are present and the size of these and of all the other spines is consider- ably greater than in E. paucispinus. The two are separated geographically by the ranges of E. guyanensis and E. echinophorus , the latter being the common Echinaster (Othilia) of the northern part of the east coast of Brazil, with sporadic records through the northern Antilles to Nicaragua, while two small specimens have just been recognized from localities in the southern Florida keys, where it overlaps with the common coral reef species E. (Othilia) sentus (Say). RANGE. The west Florida shelf in the eastern Gulf of Mexico, NW and W of Florida, c. 30-5°N, 86-5°W, S to nearly 25°N; 12-85 m; associated with sponges usually in combination with coral, shell, bryozoa, algae, sand or (in one case only) rock. Henrida downeyae sp. nov. Figs ld,e, 3C-E ?Henricia sp. Downey, 1973: 85, pi. 38, figs C,D. MATERIAL. Pillsbury st. 1187, 18°17'N, 75°03'W, off Haiti, 1033m, holotype (USNM). Gerda st. 311, 25°44'N, 79°32-5'W, off Miami, Florida, 805-787 m (2 specimens); st. 354, 25°39'N, 79°32'W, 805-832m (4); st. 830, 25°41-5'N, 79°59'W (1). Pillsbury st. 689, 08°14'N, 57°38'W, off Surinam, 1370-1444 m (3). The holotype has R/r 30/7 mm = 4-3/1; Br/br (proximal to distal arm breadths, distal breadth measured at 10% R from tip) 8-0/2-5 mm = 3-2/1. 76 A. M. CLARK The five arms are swollen basally constricting the interradii. The abactinal reticulum is fairly open, mesh area exceeding plate area, at least when fully expanded; the plates have 2, 3 or 4 lobes and are similar in size; their contours are simply convex without well defined ridges. The spinelets of the holotype are badly rubbed and incomplete but those of the other specimens, when intact, are consistently mainly trifid with three glassy flanges ending in slightly flared points on about the same level (Fig. Id); a few spinelets have one or two additional flanges and points; the central perforated core is narrow. The skeletal meshes of the holotype are sunken in drying but a few positive papulae, as distinct from dermal glands, are recognizable. The superomarginal plates form conspicuous regular horizontal subambital series except inter- radially where those of the first 3-4 mm run obliquely upwards nearly to the ambitus. Each plate is 4-lobed but the upper lobe is shorter, making the plate appear longer than broad. The upper of the two short intermarginal series has c. 1 2 plates extending for c. 1 mm, the lower for only 3 mm on the arm cleaned. The intermarginal meshes are rounded, well defined and similar in shape to the abactinal meshes. The inferomarginal plates resemble the superomarginals in shape but the upper lobes are better developed giving a slightly larger average size. The actinal plates also have 4 lobes but the adradial one is usually very short or even undeveloped. The area of each just exceeds that of the corresponding inferomarginal. The main (adradial) actinal series extends for c. 23 mm or about 2/3 R on the arm cleaned but the distal plates are very small. A second series of c. 1 plates extends for 4-8 mm proximally but is occluded interradially by the incurving of the first inferomarginal plates. Distal to the adradial series the rounded meshes or sunken areas continue to the terminal plate between the inferomarginals and the adambulacrals but the single organ in each may be a dermal gland rather than a papula. The adambulacral plates bear numerous spines about three abreast across the abradial (sub- ambulacral) part of the plate graduated to the several relatively coarse but not much longer spines on the furrow margin. Inset deep in the furrow is a single small spine with 2 (occasionally 3) much larger other spines on about the same level just superficial to it (see Fig. le), which were probably initially furrow margin spines but have become inset during growth. The madreporite is concealed by the folding of the interradii in preservation. The best preserved of the other specimens shows a single madreporite, its irregular ridges covered with scattered spinelets not much smaller than the abactinal spinelets. OTHER SPECIMENS. These too are relatively small, R up to only 25mm, R/r 4-7-5-0/1. The best preserved, from Gerda st. 31 1, has R 19 mm; the abactinal reticulum is more compact than in the holotype, the plate and mesh areas being approximately equal; the superomarginal plate series is less distinct than that of the holotype but the inferomarginals are well developed; the intermarginal plates number only about 8 but the main actinal series extends for c.\2 mm (over ^ R). There are no skeletal meshes adjacent to the adambulacrals, the only ones on the lower surface being between the inferomarginals and the actinals, but some of these appear to have papulae in them. At this small size and with the dried condition of the material the distinction between papulae and glands is not clear. Breakage of an arm shows that superficial to the deeply inset small curved furrow spine there are two larger spines arising on the same level, one proximal and one distal to it, also somewhat inset but so long that they project further than the next spines which are on the furrow margin itself (Fig. le). If these were initially furrow margin spines but have become inset during growth, as seems likely, it is possible that larger specimens with R > 35 mm might have additional spines inset on to the furrow face of some adambulacrals. Downey described 'about six short, stumpy spines' (presumably foreshortened as viewed ventrally) 'roughly in two vertical rows of three each' on the furrow faces of the adambulacrals of the two specimens she designated as 'Henricia sp.', probably from the larger one, which was from the vicinity of Porto Rico and had R 62 mm. If this specimen is conspecific with the holotype of H. downeyae, it would support the theory of adambulacral spines shifting relative to the contours of the plates during growth. However, close scrutiny of the photographic negative of this specimen from which her plate 38C was made has shown that it has two madreporites in spite of only having five arms, while she herself found three madreporites in her smaller specimen from SW of Florida. Accordingly, both of these could be potentially fissiparous and so possibly conspecific not with H. downeyae but with the ECHINASTERIDAE 77 fissiparous H. sexradiata (Perrier), which also has multiple spines inset into the furrow. In a specimen of H. downeyae from Surinam with the characteristic trifid spinelets, some of the adambulacral plates have two compressed curved inset furrow spines vertically one above the other and then the next two spines are almost superficial, demonstrating another variation. AFFINITIES. Although the trifid form of the spinelets in Henricia downeyae agrees with that in H. pertusa (O. F. Miiller) from the bathyal NE Atlantic, as being distinguished by Madsen (pers. comm.), that species has a more simple adambulacral armament with only a single inset spine. There is probably greater affinity with the partly sympatric West Indian H. sexradiata (Perrier), which shares the complex furrow armament. H. sexradiata differs most obviously in having six or seven arms, coupled with multiple madreporites and a fissiparous habit, also relatively more slender arms. The abactinal reticulum is probably coarser, though this character, like the occurrence of papulae on the lower side, varies in most species of the genus. The spinelets, though mainly glassy in both, have one prolonged point in H. sexradiata, approximating in this character to the other West Indian species H. antillarum (Perrier), in which the arms are consistently much more attenuated than in H. downeyae. RANGE. From the Florida Strait east and south to Surinam; 342-1370 (71444) m. Acknowledgements Thanks are due primarily to Professor Thomas H. Hopkins, University of Alabama, Dauphin Island Sea Laboratory, for valuable consultation about taxonomic characters and for provision of specimens. Abundant other material was provided by Miss M. E. Downey, National Museum of Natural History, Smithsonian Institution, Washington, DC, Mr David K. Camp, Bureau of Marine Research, Florida Department of Natural Resources, St Petersburg, Florida and Dr Richard L. Turner, Florida Institute of Technology, Melbourne, Florida. I am also indebted to Dr F. Jensenius Madsen, Universitetets Zoologiske Museum, Copenhagen and Professor E. Tortonese, Istituto Zooprofilattico, Genoa, for consultation. References Avila-Pires, T. C. S. de 1982. Contribuicao estudo do genero Echinaster Miiller & Troschel, 1840 (Echinodermata: Asteroidea) no literal Brasiliero. Anais Academia brasileira de Ciencias 55(4): 43 1-448, 1 3 figs. Campbell, D. B. & Turner, R. L. 1984. Echinaster graminicola, a new species of spinulosid sea star (Echinodermata: Asteroidea) from the west coast of Florida. Proceedings of the Biological Society of Washington 91(\): 167-1 78, 4 figs. Cherbonnier, E. 1959. Echinodermes de la Guyane francaise. 2. Bulletin du Museum national d'Histoire naturelle, Paris 31: 168-172, figs 3, 4. Clark, A. M. & Court man-Stock, J. 1976. The echinoderms of southern Africa. Publications of the British Museum (Natural History) No. 776: 1-277, 276 figs. & Tortonese, E. 1986. Echinaster sepositus madseni (Retzius) n. subsp. from West Africa (Echinodermata: Asteroidea). Atti della Societa Italiano di Scienze Naturale, Milano, 127(1-2): 65-71, 2 figs. Downey, M. E. 1973. Starfishes from the Caribbean and the Gulf of Mexico. Smithsonian Contributions to Zoology No. 126: 1-1 58, 48 pis. Dyakonov, A. M. 1950. [Sea stars of the USSR] Opredeliteli po Faune SSSR. No. 34: 1-199, 23 pis. [Translation to English; Israel program for Scientific Translations Ltd, Jerusalem 1968, 183 pp.] Engel, H. 1939. Echinoderms from Aruba, Curacao, Bonaire and northern Venezuela. Capita Zoologica 8(4): 1-12. Fisher, W. K. 1911. Asteroidea of the North Pacific and adjacent waters. 1. Phanerozonia and Spinulosa. Bulletin of the United States National Museum 76(1): 1-419, 122 pis. Gray, I. E., Downey, M. E. & Cerame-Vivas, M. J. 1968. Seastars of North Carolina. Fishery Bulletin Fish & Wildlife Service, US 67: 127-163. Gray, J. E. 1840. A synopsis of the genera and species of the class Hypostoma (Asterias, Linn.). Annals & Magazine of Natural History 6: 175-184; 275-290. 78 A. M. CLARK Hayashi, R. 1940. Contributions to the classification of the sea-stars of Japan. 1. Spinulosa. Journal of the Faculty of Science, Hokkaido Imperial University (6)1: 107-204, 7 pis. John, D. D. & Clark, A. M. 1954. The Rosaura Expedition. 3 The Echinodermata. Bulletin of the British Museum (Natural History). (Zool.)2(6): 139-162, 12 figs, 1 pi. Mil llrr. J. & Troschel, F. H. 1840. [Untitled] Monatsberichte der Koniglichen preussischen Akademie der Wissenschaften, Berlin 1840: 100-106. Tortonese, E. & Downey, M. E. 1977. On the genera Echinaster Miiller & Troschel and Othilia Gray, and the validity of Verrillaster Downey (Echinodermata: Asteroidea). Proceedings of the Biological Society of Washington 90(4): 829-830. & Madsen, F. J. 1979. On the specific name Echinaster sepositus, its validity and authorship (Echinodermata Asteroidea). Natura, Milano 70(4): 291-294. Ununels, F. 1 963. Asteroids from the Netherlands Antilles and other Caribbean localities. Studies on the fauna of Curacao 15:72-101. Verrill, A. E. 1870 [1867-71]. Notes on the Radiata in the Museum of Yale College, with descriptions of new genera and species. 1-3. Transactions of the Connecticut Academy of Arts & Sciences 1(2): 247-351. [The date of this work is usually cited as 1867 but the entire printing was destroyed by fire (see Bell, 1891. Ann. Mag. not. Hist.: 25) and Verrill himself (1915) cites the date for all three parts as 1870, though Sladen and Fisher both give 1871.] 1915. Report on the starfishes of the West Indies, Florida and Brazil. Bulletins from the Laboratories of Natural History, State University of Iowa 7: 1-232. Walenkamp, J. H. C. 1976. The asteroids of the coastal waters of Surinam. Zoologische Verhandelingen. Leiden. No. 147: 1-91, 16 figs, 18 pis. 1979. Asteroidea (Echinodermata) from the Guyana Shelf. Zoologische Verhandelingen. Leiden. No. 170: 1-97, illust. Zoppi de Roa, E. 1967 [71968]. Contribucion al estudio de los Equinodermos de Venezuela. Acta Biologia Venezuela 5(17): 267-333, 29 figs. Manuscript accepted for publication 26 March 1986 Observations on the marine nematode genus Spirinia Gerlach, 1963 (Desmodoridae: Spiriniinae) with descriptions of two new species J. W. Coles 15 Rushfield Road, Liss, Hampshire GU33 7LW Synopsis Observations on the nematode genus Spirinia Gerlach, 1963 are made including notes on its systematic position. The genus is characterized by the shape of the head, form and position of the amphids, narrow slightly cuticularized buccal cavity, short oesophagus with a pronounced bulb and form of the male genital apparatus. The tail-shape separates the genus into two sub-genera — conical in the subgenus Spirinia and flagellate in the subgenus Perspiria. A list of the valid species is given together with notes of those considered as nomina dubia and a dichotomous key for the determination of species. Descriptions and/or notes are given of the following species: Spirinia (S.) parasitifera (Bastian, 1865): S. (S.) schneideri (Villot, 1875); S. (S.) hopperi sp. nov.; S. (S.) laevis (Bastian, 1865); S. (S.) gerlachi (Luc & DeConinck, 1959); S. (Perspiria) hamata Wieser & Hopper, 1967; S. (P.) mokii sp. nov. Introduction This report was prompted by a collection of Spirinia parasitifera (Bastian, 1865) from localities in North America (Canada; Massachusetts and Florida, USA) provided by Mr Bruce E. Hopper of the Canadian Department of Agriculture. Apart from one population in Nova Scotia (Bras d'Or Lake) they show variation in the form of the manubrium (cephalated proximal portion of the spicule) not apparent in S. parasitifera from European waters. As this is the only character in which they differ a taxonomic distinction is not justified. Specimens of S. parasitifera, as well as four other species, including one described as new (S. (S.) hopperi) from British coasts are also examined. A new species belonging to the subgenus Perspiria Wieser & Hopper, 1967 (S. (P.) mokii) from Malaysia is described. Observations are also made on specimens of S. (P.) hamata Wieser & Hopper, 1967 from Florida. Typically Spirinia species live in sandy and muddy habitats and are often found in sand surrounding the roots ofZostera (eelgrass). Twenty-five nominal species have been described, 1 3 of which are currently included in the genus. See Gerlach & Riemann (1973) for nomenclatorial changes and synonymies up to 1972. Since 1972 another species (Spirinia (S.) gnaigeri Ott) has been described from an inter-tidal sand-flat, Bermuda (Ott, 1977). A list and key to all valid species of Spirinia is included (see p. 80). Notes are also given of species considered as nomina dubia. Material and methods The specimens from British coasts were fixed in formalin with sea-water. All specimens studied are deposited in the collection of the British Museum (Natural History). The examination was carried out mainly with the light microscope using transmitted light interference microscopy in some instances. Specimens of S. (S.) parasitifera, S. (S.) schneideri, S. (S.) hopperi and S. (S.) laevis were also examined with a scanning electron microscope for elucidating details of the labial and cephalic sensilla and for the appearance of the cuticle. Only absolute measurements are given (see Tables I-X). The head diameter was taken at the mid-level position of the amphids. Measurements of the arc-length of the spicules are given, not the chord. Bull. Br. Mus. not. Hist. (Zool.) 53(2): 79-101 Issued 29 October 1987 80 J. W. COLES Systematic position of Spirinia The genus (as Spird) was established by Bastian (1865) but first fully characterized by De Man (1890). As Spirina Filipjev, 1918 it was first included in the family Desmodoridae (established as sub-family Desmodorini) by Filipjev (1922), subsequently raised to family by Steiner (1927). Chitwood (1936) subdivided the Desmodoridae into four subfamilies and included Spirina in a tribe Spirinacea of the subfamily Richtersiinae and Wieser (1954) followed this classification in a revision of the genus with a key to species. Gerlach (1963) changed the name Spirina Filipjev, 1918 to Spirinia because the name Spirina was found to be preoccupied by a mollusc. More recently there have been changes in the classifi- cation of the Desmodoridae and Spirinia is currently placed as the type-genus of the subfamily Spiriniinae. See De Coninck (1965), Gerlach & Murphy (1965), Wieser & Hopper (1967), Gerlach & Riemann (1973) and Lorenzen (1981). Wieser & Hopper (1967) erected the subgenus Perspiria for the species of Spirinia having a flagellate and coarsely-striated tail. Spirinia is redefined as follows: Desmodoridae, Spiriniinae. Cuticle with or without transverse striations. Body setae present. Head narrow but not set-off from body, more truncate in some species than in others. Buccal cavity small and narrow, the walls only slightly cuticularized, sometimes with minute teeth. Amphids a prominant single spiral, circular in outline, situated near the anterior end. Oesophagus usually short in comparison with the total body-length (an exception being a species belonging to the subgenus Perspiria) and with a posterior bulb. Testis single and straight (outstretched). Spicules paired, arcuate, of equal length, with a club or hook-shaped expansion (manubrium) at the proximal end. Gubernaculum short, without a posteriorly directed apophysis. Ovaries paired and reflexed. Tail conical in the subgenus Spirinia, flagellate in the subgenus Perspiria. The closest related genus is Chromaspirina Filipjev, 1918, which has a larger and more heavily- cuticularized buccal cavity, often armed with larger teeth than in Spirinia. Also the transverse striations are much coarser compared with those in Spirinia. Cuticular striations are not present in S. Schneider i and S. hopper i. Lorenzen (1981) has noted the similarity of Alaimonema Cobb, 1920 to Spirinia. Although Cobb (1920) gives quite a detailed description, the important differential characters are not clear, and there is only an illustration of the anterior end of the worm. List of valid species of Spirinia Subgenus Spirinia sensu stricto Gerlach, 1963 S. (S.) gerlachi (Luc & De Coninck, 1959) S. (S.) gnaigeriOtt, 1977 S. (S.) hopperisp. nov. S. (S.) laevioides Gerlach, 1963 S. (S.) laevis (Bastian, 1865) S. (S.) parasitifera (Bastian, 1865) S. (S.) schneideri(Vi\\ot, 1875) S. (S.) septentrionalis (Cobb, 1914) S. (S.) tenuicauda (Allgen, 1959) Subgenus Perspiria Wieser & Hopper, 1967 S. (PJflagellataViliello, 1971 S. (P.) hamata Wieser & Hopper, 1967 S. (P.) mokiisp. nov. S. (P.) striaticaudata (Timm, 1962) Dubious species Spirinia (S.) sabulicola (Filipjev, 1918) was described from the Black Sea from female specimens only. Males have not been found since and there are no redescriptions, so this species is unlikely to THE GENUS SPIRINIA 8 1 be recognized with any degree of certainty. It is therefore treated as a nomen dubium. S. (S.) similis (Cobb, 1898) was described without an illustration and also unlikely to be recognized. It is therefore regarded as a nomen dubium. Key to subgenera and species 1 Tail shape conical, of varying length but never filiform (subgenus Spirinia) 2 1' Tail shape filiform, transverse striations of the cuticle more prominent on the tail (subgenus Perspiria) 10 2 Body setae short (up to 10 um) 3 2' Body setae long (greater than 10 um) 7 3 Cuticular striations present 4 3' Cuticular striations absent 6 4 Striations on the cuticle resolvable into dots S. (S.) septentrionalis 4' Striations on the cuticle not resolvable into dots 5 5 Amphids about one-third of the corresponding head diameter in width, tail typically conical . S. (S.) parasitifera 5' Amphids 75% of corresponding head diameter in width. Males twice the length (2-12 mm) of the females (1-35 mm), tail uniformly thinner in its posterior part . . . S. (S.) tenuicauda 6 Tail short (1-1 -5 times anal body diameter in length) S. (S.) Schneider i 6' Tail longer (2-3 times or more than the anal body diameter in length) . . S. (S.) hopperi I Body setae long in cervical region only, body length 2-3 mm. Proximal end of spicules hook-like in shape S. (S.) laevis 1' Body setae long throughout the length of the body, body length less than 1-6 mm ... 8 8 Amphids greater than 50% of the corresponding head diameter . . . S. (S.) gerlachi 8' Amphids less than 50% of the corresponding head diameter .... 9 9 Sub-cephalic setae present anterior to cephalic setae, cervical setae up to 1 5 um in length S. (S.) laevioides 9' Sub-cephalic setae absent anterior to cephalic setae, cervical setae 20-22 um in length . S. (S.) gnaigeri 10 Pre-anal supplements present in male S. (P.)flagellata 10' Pre-anal supplements absent in male 11 I 1 Cervical setae fairly long, other body setae lacking, spicules with only a small club-shaped expansion at the proximal end S. (P.) striaticaudata 11 ' Cervical setae short 12 12 Proximal end of spicules markedly hook-shaped, body length about 2 mm . . S. (P.) hamata 1 2' Proximal end of spicules with club-shaped manubrium, not hook-like, body-length less than 2 mm . S. (P.) mokii Species studied Spirinia ( Spirinia) parasitifera (Bastian, 1865) (Figs la; 2a-f; 3a; 5a^c; 8a-b) Spira parasitifera Bastian 1865; Spirina parasitifera: Filipjev, 1918. For full synonymy see Gerlach & Riemann, 1973. TYPE LOCALITY. Falmouth, Cornwall, England. MATERIAL STUDIED. 4<$d 1$ from Laminaria holdfasts off the west coast of Scotland, near Gallanchmore Farm, situated on the mainland side of the sound of Kerrera, two miles south of Oban, BM(NH) 1961.105-134; 1 £ from sand-flats, Whitstable, Kent BM(NH) 1964.255; 4^, 6?? and 1 juvenile from sand around the roots of Zostera, Hannafore, Looe, Cornwall BM(NH) 1985.1.26-29; 9<$<$, 6?? and 4 juveniles from Bras d'Or Lake, Nova Scotia, Canada (12m coarse bottom) BM(NH) 1985.2.52-54; 1^ from sediment on the south side of Pigeon Key, Florida, USA BM(NH) 1986. 12-5; 16^, 10?? from Thalassiabed, Biscayne Bay, Florida, BM(NH) 1985.2.33, 82 J. W. COLES Fig. 1 Lateral views of head of (a) Spirinia (S.) parasitifera; (b) S. (S.) schneideri; (c) S. (S.) hopperi; (d) S. (S.) laevis. Bar scales: (a) = 12 urn; (b) =20 ^m; (c) = 10 ^m; (d) = 10 \im. 38-40, 43-46; 1? from Diplanthera bed, Biscayne Bay off Virginia Key, Florida BM(NH) 1985.2.34; 4^, 1? from the same locality but from a public beach area BM(NH) 1985.2.35-37, 39; 7 rt - Ov 0 C •* ^^ rO CT\ ^f 5 _, « 00 -1. O C 3 O O O O SI 00 O - 9 - so ^ O *O C r- — , — c 3 — — — 0 v ro 0 <; 3 O O O O I/ t O 00 - ^ *O »o ON ^O c Of O 0 C 3 — — O O r* «— i ro 0 C 3 O O O O rt J r •» | Of CN OO - 00 O C •* \& OO ^5 ^D 3 — — — ' 0 r t> U V c3 rxl? t-~ o <= 3 — — —O f> "a" c rs o c 3 O O O O rt rt '.^ ^^ r 4 rt g OH r~ oo - •^- o c * T^ ^^ ^^ 3 — — 0 !/• O 0 O O 0 0 o 5 'o a & i/ •> "*< C/5 o r- - •< v~i m oo so ON CO """S tVs 00 O C 5 — _ O O O o £ r^ p SO rt O ^t 00 — — —i o 0 ro O § (N O 0 O O O 0 o £ m »/ O so " >o p c •> * so rt t*™^ ^t ON 5 — — ' — O O ro O 1 O O O O 0 O a G G O oo r 5 •1 ON rt O SO o O 0? (N O C r^i O C ) ^H ^ _ O > O O 00 £ 3 C^ S£ 9 •r. rt (*_ o « m r~ — r~ o c - SO Tt O rt 00 > — 1 i— < — O O PO o u M rt *° O 0 O O O o • O — o V •> 3 1 « so O C - SO rt O rt ON > — — ~ o 0 ro O H £ rs o c > O O O O 0 O g O r. V o so — SO O C « •/•> — T« O O — ' en o M rs o c 5 O O O O O o 1 « >y KO O C 1 * rt */"J ON rt OO > — — O O O (N O O. 02 fS O C J 0 O O O O o 3 J3 W) •Q ^ i T3 g rt QJ I- •— 4 >% i^fel ll-s J3 ^? "fl ! o 'i i ^" 0> ^* 2«!-i f 'tn ^» fc ! ^ J3 Q ^ C t^ f| M 3 0 ^ C 0 •- i 0 00 u £ rt 0 ( '6 S c 2^6^ 3 « 8 c 3 P c £ >* ^ T 111 1 BO -7 rt <*- -i ^o 1 » S 5 P [1 jj THE GENUS SPIRINIA 87 •n «n - ^ m ON CO t ^ p 0 o — *— 0 o o o 0 0 0 "Si cs H g co GO _3 3 C "3 f—l ao O ^_J 53 1 c/) 53 -a (~^ o C 53 O. — - > C •a 0 -diameter O 1/5 O 0 "o "GO c nceofner -o c o c 73 diameter :hofspicu :hofgube nee of vul ant e*nr\ •a >< "O si) c II 33 E "rt e3 M« g o n 1 b X o H 3 p C § < J 3SJ 88 J. W. COLES r- 3 0 0 2 2 § 2 C r- O+ O ^J. 00 CO c ti ^H m o 0 T— . »— 0 o r* Of (N O 0 0 o CD o O & g (N CD ^O t-~ o 0 VO 2 0 o o O •~ O ^O CN| O 0 o 0 0 o o o | *2 hk Q c s, o >/~> g o »s ^ O 0 2 2 oo 0 •^- r-> 0 0 rs o 52. I ^0 (N O 0 0 0 CD o o o .O C/3 •S r- *o A o\ Tt 0 0 £ ro S Tt VO 0 0 p "8 ^0 0 ea (S r- oo ^— H l/~J T^~ ON ^f V g "3 Tf O 0 — ^~ 0 f^*s U SC Of o r-- o°\ o rn 0 2 10 OO 0 Tf ON 0 0 CO O *5 t. ,j2" *o o IT) _H rn — oo ^^ ^o fN jX ON o o — >— < o o o O x_ *o ^H 0 0 0 0 o o o 6 a •5 VO .^* r- O IT) ^^ fN fN t~~ ro r~ CO Ci, ON o o ,— 1 1— 1 o o o o *0 o 0 0 0 o o o o 'o | S O 0 0 a fN SO ^ >/~i O 0 0 fN 0 *° ^^ 0 o o 0 o o o 0 t/3 u -• 0 0 0 0 0 0 0 o t SO t^ ro o oo TJ- r- CO 3 < 6 a: c /"> (N vo c n? vo o o — < — < T ° c ^•6006 oo (N O r-~ -^ in n , c 1 O+ O O O — — O O en O O O O O O p E o ». — u int — ~^F~-cn ^Hinoc — i •— i •— ' O O O •S *0 f^OOOO OOOO '5, CM o •* fNt^ — t^O OWiO m ^3 \ooo- — — oo o 1 rnOOOO OOOO CO iremenl O u Q O+ o ^o ^* >n o os ^4" r- p p — i — i p p en O O O O O O VC r ff CO >> u a o _ CO V? TtOO-H^- -HO — o — — w ••i-oooo oooo ^> X Q> W ce H 1 fS — OO — — OOOO TTOOOO OOOO co C u S o ^o — t^ «n o«nt^ (N B ^r1 OOOO — — — O O p a rnOOOO OOOO a co bO Sj 3 C 3 (-< b£) *r^ o •*-» Cd i ^ OS *O -C w £ C rt |l||||il| " 5o<~ S o c. 25 mm SL) Barbus stigmatopygus can be distinguished from spotted speci- mens of Barbus perince by the presence of a few large gill rakers confined to the upper part of the first ceratobranchial, and by the position of the middle spot which is on, or in front of, the vertical from the base of the last dorsal fin ray. Key to the small Barbus species of the Nile (excluding Lakes Victoria and Tsana) This key will only function when the fish have acquired most of their adult characters. Very small specimens are too similar to be keyed out. Delayed barbel development and lateral line tubule formation should be borne in mind. Parallel striae on scales Juvenile large Barbus species Radiating striae on scales No barbels in fish > 1 6 mm SL Barbels in fish < 1 6 mm SL • 5 Lateral line complete • Barbus anema Lateral line incomplete -4 Lateral line scales much deeper than long • Barbus pumilis Lateral line scales about as deep as long, dark spot on dorsal fin . . Barbus leonensis 2 barbels in fish > 20 mm SL • Barbus tongaesis 4 barbels in fish > 20 mm SL 6 First ceratobranchial with papillate tissue ventrally, few large gill rakers dorsally, flanks with up to 5 spots (NB juveniles have 0 or 2 barbels) .... . Barbus stigmatopygus No such papillate tissue, gill rakers ventrally on first ceratobranchial . Last unbranched dorsal fin ray smooth ........... Last unbranched dorsal fin ray serrated .... 9 Body deep ( > 30% SL at 35 mm SL), sometimes 3 spots on flanks, silvery . . Barbus perince Body shallow ( < 25% SL at 35 mm SL), irregular longitudinal stripe .... Barbus yeiensis 138 K. E. BANISTER 9 No large spots or blotches on flanks . ........ . Barbus kerstenii* Large spots or blotches on flanks 10 10 Pelvic fin origin in advance of vertical from dorsal fin origin .... Barbus pellegrini* Pelvic fin origin more or less on the vertical from dorsal fin origin . . . Barbus neumayeri* *within the Nile system limited to the Lakes Albert and Edward drainages. Acknowledgements I am especially grateful to Dr Roland Bailey for so generously donating his collection from the Sudd to the BM(NH) and for giving me access to his field observations. My colleague Gordon Howes kindly drew Figures 3 and 9. Drs P. H. Greenwood and R. G. Bailey offered constructive advice on the manuscript. References Balinsky, B. I. 1948. On the development of specific characters in Cyprinid fishes. Proceedings of the Zoological Society of London 118: 335-344. Banister, K. E. 1980. A new species of Barbus (Pisces, Cyprinidae) from Africa. Bulletin of the British Museum (Natural History) (Zoology) 38 (3): 145-150. Boulenger, G. A. 1903. Diagnoses of four new species of Barbus from the Nile. Annals and Magazine of Natural History (7) 12: 532-533. 1905. On a second collection of fishes made by Mr S. L. Hinde in the Kenya district East Africa. Proceedings of the Zoological Society of London 1905 (1): 62-64. 1907. Zoology of Egypt: the fishes of the Nile, ii, 577 pp, 95 pis. Published for the Egyptian Govt. by Hugh Rees, London. 1911. Catalogue of the freshwater fishes of Africa in the British Museum (Natural History) Vol II, xii + 529 pp. 382 text figs. London. 1921 . Liste des publications ichthyologiques et herpetologiques (1877-1920). Annales de la Societe royal zoologique et malacologique de Belgique 52: 1 1-88. Daget, J., Gosse, J-P. & Thys van den Audenaerde, D. F. E. (Eds) 1984. Checklist of the freshwater fishes of Africa. Vol. 1, xviii+410 pp. ORSTOM. Paris and MRAC Tervuren. Greenwood, P. H. 1963. A collection of fishes from the Aswa drainage system Uganda. Proceedings of the Zoological Society of London 140(1): 61-74. Gunther, A. C. 1868. Catalogue of the fishes in the British Museum 7: 1-512. - 1869. Appendix C. In: Pethrick, J. & B. H. Travels in Central Africa and explorations of the Western Nile tributaries, Vol. II. Jubb, R. A. 1967. Freshwater fishes of southern Africa. A. A. Balkema, Cape Town. Kortmulder, K. & van der Poll, R. J. 1981. The juvenile and adult pigment patterns of Barbus lateristrigata Cuv. & Val. 1842, B. titteya (Deraniyagala 1929) and B. narayani Hora 1937 (Pisces, Cyprinidae) and their taxonomic value. Netherlands Journal of Zoology 31 (2): 453-465. Mori, Y. 193 la. On the transformation of ordinary scales into lateral-line scales in the goldfish. Journal of the Faculty of Science, Imperial University of Tokyo (4) 2 (3): 185-194. . On the regeneration of the lateral line scales and lateral line organs in the goldfish. Journal of the Faculty of Science, Imperial University of Tokyo (4) 2 (3): 195-200. Neave, F. 1940. On the histology and regeneration of the teleost scale. Quarterly Journal of Microscopical Science 81: 541-568. Poll, M. 1939. Poissons. Exploration du Pare National Albert Mission G. F. de Witte 24: 1-82. Rendahl, H. 1935. Einige neue fische aus dem weissen Nil. Annales Zoologici Societas Vanamo 2 (2): 1 1-18. Riippell, E. 1837. Neuer Nachtrag von Beschreibungen und Abbildungen neuer Fisch, im Nil Entdeckt. Museum Senckenbergianum 2: 1-28. Sandon, H. 1950. An illustrated guide to the fishes of the Sudan. Sudan Notes and Records, 61 pp. McCorquodale & Co. Schultz, L. P. 1957. The generic names Barbus and Puntius. Tropical Fish Hobbyist 5: 14-15, 29-31. Weber, M. & de Beaufort, L. F. 1916. The Fishes of the Indo- Australian Archipelago 3. E. J. Brill. Leiden. Manuscript accepted for publication 29 May 1986 British Museum (Natural History) The birds of Mount Nimba, Liberia Peter R. Colston & Kai Curry-Lindahl For evolution and speciation of animals Mount Nimba in Liberia, Guinea and the Ivory Coast is a key area in Africa representing for biologists what the Abu Simbel site in Egypt signified for archaeologists. No less than about 200 species of animals are endemic to Mount Nimba. Yet, this mountain massif, entirely located within the rain-forest biome, is rapidly being destroyed by human exploitation. This book is the first major work on the birds of Mount Nimba and surrounding lowland rain- forests. During 20 years (1962-1982) of research at the Nimba Research Laboratory in Grassfield (Liberia), located at the foot of Mount Nimba, scientists from three continents have studied the birds. In this way Mount Nimba has become the ornithologically most thoroughly explored lowland rain-forest area of Africa. The book offers a comprehensive synthesis of information on the avifauna of Mount Nimba and its ecological setting. During the 20 years period of biological investigations at Nimba this in 1962 intact area was gradually opened up by man with far-reaching environmental consequences for the rain-forest habitats and profound effects on the birds. Therefore, the book provides not only a source of reference material on the systematics, physiology, ecology and biology of the birds of Mount Nimba and the African rain-forest, but also data on biogeography in the African context as well as conservation problems. Also behaviour and migration are discussed. At Nimba a number of migrants from Europe and/or Asia meet Afrotropical migratory and sedentary birds. Professor Kai Curry-Lindahl has served as Chairman of the Nimba Research Laboratory and Committee since its inception in 1962. Peter Colston is from the Subdepartment of Ornithology, British Museum (Natural History), Tring, and Malcolm Coe is from the Animal Ecology Research Group, Department of Zoology, Oxford. 1986, 129pp. Hardback. 0 565 00982 6 £17.50. Titles to be published in Volume 53 Puellina (Bryozoa; Cheilostomata; Cribrilinidae) from British and adjacent waters. By J. D. D. Bishop & B. C. Househam Miscellanea The genera of pelmatochromine fishes (Teleostei, Cichlidae). A phylogenetic review. By P. H. Greenwood Certain Actiniaria (Coelenterates, Anthozoa) from the Red Sea and Tropical Indo- Pacific Ocean. By K. W. England Printed in Great Britain by Henry Ling Ltd.. at the Dorset Press, Dorchester, Dorset Bulletin of the British Museum (Natural History) The genera of pelmatochromine fishes (Teleostei, Cichlidae). A phylogenetic review Peter Humphry Greenwood Zoology series Vol 53 No 3 26 November 1987 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.) © British Museum (Natural History), 1987 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : MrJ. F. Peake Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISBN 0565 05035 4 ISSN 0007- 1 498 Zoology series Vol53 No. 3 pp 139 -203 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 26 November 1 987 The genera of pelmatochromine fishes (Teleostei, Cichlidae). A phylogenetic review Peter Humphry Greenwood Department of Zoology, British Museum (Natural History), Cromwell Road, London SW7 5BD Contents Introduction 139 Materials and methods 141 A review of Thys' supraspecific pelmatochromine groups 142 I Pelmatochromines with microbranchiospines and without a pharyngeal hanging pad -142 Pelmatochromis Steindachner 142 Pterochromis Trewavas 154 II Pelmatochromines without microbranchiospines, and with a visor-like hanging pad in the pharynx 157 Thysia Loiselle & Welcomme 160 ChromidotilapiaBoulenger 165 Pelvicachromis Thys van den Audenaerde 169 Parananochromis new genus 1 74 Nanochromis Pellegrin 180 Limbochromis new genus 188 The phylogenetic relationships of the pelmatochromine taxa 191 The intrarelationships of the chromidotilapiines 194 Intrafamilial relationships of the chromidotilapiines Pelmatochromis and Pterochromis . . . . . . . . .195 The distribution of Pterochromis, Pelmatochromis and the chromidotilapiines . . 199 Summary and conclusions 200 Acknowledgements ............. 200 References . .201 Introduction The main objectives of this paper are three-fold. First, to revise and review, at the generic level, a major group of west African taxa often referred to as the Pelmatochromines since all were once classified in the genus Pelmatochromis. Second, to determine whether these genera are closely related phylogenetically, and third, to examine their intrafamilial relationships. It is hoped that the information gained in this analysis will be a contribution towards any ultimate investigation of intrafamilial phylogeny in the Cichlidae as a whole. The term 'pelmatochromine' used here has no implied or established taxonomic or phyletic implications. It is employed simply as a convenient reference to the various supraspecific groups recognised by Thys van den Audenaerde (1968a) in his revision of the speciose west African genus Pelmatochromis Steindachner, 1894. Thys' 1 paper marks an important change in our taxonomic understanding of these fishes, especially since it is the first attempt to break down Regan's (1922) manifestly polyphyletic Pelmatochromis generic concept into a number of monophyletic lineages. 'Following current convention, Dr Thys van den Audenaerde's name will be abbreviated to Thys in all future references to his work cited in the body of the text. Bull. Br. Mm. not. Hist. (Zool.) 53(3): 139-203 Issued 26 November 1987 139 140 P. H. GREENWOOD Although Thys retained his various lineages within a nominal Pelmatochromis, he did suggest the recognition of three subgenera (Pelmatochromis, Chromidotilapia Blgr. 1898, and Pelvicachromis Thys, 1 9680). He also included, as a subgenus of Pelmatochromis, Pellegrin's ( 1 904) Nanochromis which Regan (1922) had regarded as generically distinct from that genus but closely related to it. In addition to these four subgeneric groups, Thys included in Pelmatochromis, but as taxa incertae sedis with regard to their subgeneric placement, a number of species treated as members of Pelmatochromis by Regan (1922), together with others which had been described since 1 922 (see Thys, 1 9680, and pp. 1 54 & 1 60). Later in the same year Thys (19686) placed the nominate subgenus Pelmatochromis (Pelmatochromis) in Tilapia Smith, a move contested by Trewavas (1973) who resurrected the taxon as a genus, and thus by implication raised the other Pelmatochromis subgenera to full generic rank. Trewavas (op.cit.) also included in Pelmatochromis one of the species (P. ocellifer Blgr, 1 899; but see Trewavas op.cit.) which Thys (19680) had treated as being of uncertain subgeneric affinity, and resurrected another species (P. nigrofasciatus [Pellegrin, 1900]) which he had synonymised with P. ocellifer Blgr, 1899. The following year Trewavas (1974) explicitly returned the subgenus Chromidotilapia to its former generic rank, thus, again by implication, recognising the subgenera Pelvicachromis and Nanochromis as genera. Of the pelmatochromine taxa which Thys (19680) treated as of uncertain infra-generic relation- ships, Pelmatochromis ruweti Poll & Thys (1965) has been placed in Tilapia (Thys, 19680), P. congicus Blgr ( 1 897) has become the monotypic taxon Pterochromis Trewavas ( 1 973), P. ansorgii was made the type species of the genus Thysia Loiselle & Welcomme (1972), P. cerasogaster (Blgr, 1899) was included in Hemichromis Peters by Loiselle (1979), as was P. exul Trewavas (1933) by Trewavas (1973), and P. thomasi (Blgr, 1915) has become the type species of the currently monotypic genus Anomalochromis Greenwood (19856). Since Pterochromis and Thysia would, at least superficially, seem to qualify for inclusion in any review of the pelmatochromines, both are considered below. There seems no doubt that Pelmatochromis ruweti is indeed a species of Tilapia (see Trewavas, 1983), and that Hemichromis and Anomalochromis are not closely associated, phylogenetically, with the pelmatochromines (see Greenwood, 1 9850 & b\ also p. 1 95). Thus, little more will be said about the two latter taxa except in a general context. The pelmatochromines, as here restricted, are a moderately large group comprising some 35 species, to which more will undoubtedly be added when the various genera are subject to revisions at the species level. Geographically speaking, the assemblage is a west African one whose range extends from Senegal in the north to the Zaire river and several of its tributaries in the south. Within that area some of the constituent genera have relatively circumscribed distributions, and many of the species would appear to occupy even more localised areas. No pelmatochromines occur in any of the African Great Lakes (see below) and none has been recorded in other habitats within what is considered to be eastern Africa (i.e. east of longitude 29°E). Within west Africa no pelmatochromines have been recorded outside the region bounded by latitudes 18°N and 9°S. The supposedly unique pelmatochromine representative in east Africa, Pelmatochromis exul Trewavas from Lake Turkana, is now known to be based on the misidentification of three small Hemichromis bimaculatus specimens (Trewavas, 1973). In itself that corrected record is of interest since it represents the most easterly occurrence of H. bimaculatus, a species belonging to an otherwise essentially west African genus (see Loiselle, 1979; Greenwood, 19850). Most of the pelmatochromine species occupy fluviatile habitats, with some seemingly restricted to fast flowing, even torrential waters, as is the case with many Nanochromis species (see Roberts & Stewart, 1976; Stewart & Roberts, 1984). In many respects the pelmatochromines are the ecological counterparts of several haplochro- mine taxa (sensu Greenwood, 1980) in east Africa, particularly with regard to species of the genera Astatotilapia, Thoracochromis, Ctenochromis and Orthochromis (Greenwood, 1979). It is this apparent haplochromine-pelmatochromine equivalence which, in part, aroused my interest in the pelmatochromines, particularly from the viewpoint of possible phyletic relationships between either individual genera or the two groups as a whole. PELM ATOCHROMINE FISHES 1 4 1 Thys' (19680) analysis of the pelmatochromines was based on what are now called orthodox or evolutionary taxonomic lines. A cladistic (or Hennigian) approach is used here, both to test the hypothesis of intragroup relationships put forward by Thys (his subgenera) and to erect an hypo- thesis of interrelationships amongst those lineages, a problem only touched upon by Thys (19680; fig. 17). Using the same methodology, I shall also attempt to establish the wider, intrafamilial relationships of these various lineages, especially to test Thys' (19686) hypothesis that the phyletic relationship of at least some pelmatochromines lies with the genus Tilapia (sensu Trewavas, 1983). Thys' inclusion of two Pelmatochromis (Pelmatochromis) species in Tilapia was strongly con- tested by Trewavas (1973), and I would support her views, albeit from a different standpoint. That difference raises questions about the validity, in phylogenetic arguments, of the characters used in previous discussions, and questions the current definition of the genus Tilapia itself and of the tilapiines sensu Trewavas (1983) as a whole, a lineage which includes certain pelmatochromine genera. Materials and methods Materials. The entire spirit collection of pelmatochromine species in the British Museum (Natural History) was examined. Those specimens from which new skeletal preparations were made are listed separately, together with those used for dissection; often a dissected specimen was later prepared as a skeleton. Dissections used for previous studies (see Greenwood, 1978; 1983; 19850 & b and 1986), and named in those papers, are not detailed here. Most of the British Museum (Natural History) collection of cichlid skeletal material (both dry and alizarin preparations) was examined or has recently been examined (see Greenwood, 19850 and 1986). For extra-familial out-group comparisons, the Museum's skeleton collections of the Serranidae, Centropomidae, Sparidae, Labridae and Pomacentridae was examined (see Greenwood, 1976; 19850), and other anatomical data were derived from various specimens detailed in Greenwood (19850 & b and 1986). All the specimens mentioned below are listed under the generic names in use before this revision, that is, following Thys (19680) and Trewavas (1973 and 1974). Additional material examined for superficial characteristics is as follows: Museum of Comparative Zoology, Harvard University: Nanochromis splendens, holotype (MCZ 50476) and paratype (MCZ 50311); N. parilius holotype (MCZ 50474) and 8 paratypes (MCZ 50340); N. consortius, holotype (MCZ 50551); N. minor, paratype (MCZ 50310); Pelmatochromis buettikoferi (MCZ 35465, 60 & 38 mm SL); Chromidotilapia guntheri (MCZ 48071; ca 70 mm SL); Chr.cf. guntheri (MCZ 48135, ca 60 & 45mmSL, and MCZ 32524, ca 70mmSL). Nanochromis longirostris (MCZ 35398; 10 specimens 45-75 mm SL, and MCZ 35389, one specimen ca 60 mm SL). Thysia ansorgii (MCZ 48070, one specimen ca 54 mm SL). Musee de PAfrique Centrale, Tervuren: Pelmatochromis nigrofasiatus (det. E. Trewavas; registered as P. ocellifer), 3 specimens (from MCA 1 18090-99); Pelmatochromis nigrofasciatus 2 specimens (MCA 52454- 55); Nanochromis cavalliensis, holotype (MCA 168574), and 4 paratypes (MCA 1963-471). California Academy of Sciences: Nanochromis robertsi, holotype (SU 63143) and 2 paratypes (SU 67173). British Museum (Natural History: Myaka myaka (1973.5.14:717-23). Konia dikume (1978.5.14:643-59); K. eisentrauti (1961.10.18:11-14). Pungu maclareni (1973.7.18:137-140). Stomatepia mariae (1973.5.14: 984^988); S. mongo (1971.10.20:24-33). Dissections, partial dissections, and osteological preparations (see above) involved the following specimens (all from BM [NH] collections unless indicated otherwise): Chromidotilapia batesii (1912.6.29:4; skeleton), Chr. finleyi (1973.5.14:544; dissection and skeleton); Chr. guntheri (1973.5.14:615-625 and 1934.8.31: 179-188, one specimen of each: dissection and skeleton), Chr. kingsleyae (1867.5.3:2; 1912.4.1:526 and 1 908.5.30: 1 86; both skeletons). Iranocichla hormuzensis (paratypes, 198 1 . 1 . 12: 1-2; partial dissection of one). Nanochromis cavalliensis (holotype, MAC 168574; partial dissection of pharyngeal region); N. parilius (para- types, MCZ 50475; 4 specimens as alizarin preparations, and BMNH 1977.1.1 1:33-34, paratype; one speci- men as an alizarin preparation and partial dissection); N. nudiceps (1963.10.22:9, and one unregistered specimen, from below Stanley Pool, collected by Mr Owen Clark; both partial dissections). Oreochromis macrochir (J. L. B. Smith Institute, RUSI 22134; dissection); O. mossambicus (RUSI 19290; dissection); O. 142 P. H. GREENWOOD niloticus (1981.2.17:601-620; one specimen partially dissected); Pelmatochromis buettikoferi (syntype, 1911.5.31:47; a skeleton, and 1981.6.19:103 and 1983.1.25:38; dissected and skeletonised); P. longirostris (1903.7.28:77-88; one specimen, an alizarin preparation, another dissected and skeletonised), Pelvicachromis humilis (19 15. 4. 13:441; dissected and skeletonised, and 1972.3.16:8-10; one specimen dissected and skeleton- ised); P. kribensis (syntype, 1912.6.29:28; skeleton); P. pulcher (syntypes, 1901.1.28:13-20; one specimen dissected, and one already a skeleton; 1908.12.9:68; a skeleton); P. subocellatus (1888.12.13:8-10; one speci- men skeletonised). Thysia ansorgii (1977.11.8:436-455; five specimens dissected and some skeletonised). Tilapia busumana (1934.8.31:189-199; one specimen skeletonised, and 1942.12.30:38-42; one specimen dissected and skeletonised); T. ruweti (1969.3.28:1-6; one specimen dissected and skeletonised; another specimen, ex Okavango Swamps, Botswana, unregistered, treated similarly); T. sparrmanii(\9Q7.3.\ 5:45-47; one specimen dissected and skeletonised, and RUSI 23538, dissected). T. zillii (1973. 3.27:7-8; one specimen dissected). Tristramella simonis (1949.9. 16:399-444; one specimen dissected and skeletonised). A list of the Tilapia, Oreochromis and Sarotherodon species examined (and some partially dissected) in connection with the nature of the pharyngeal teeth and the arrangement of the pharyngeal bones (see pp. 148 & 198), and of the non-tilapiine taxa also examined is available in the Fish Section of this Museum. Methods. The methodology employed has already been discussed in relation to that of Thys (19680) and Trewavas (1973, 1974); no further comment is required (see p. 141 above). Since the species-level taxonomy of the pelmatochromines, especially the genera Pelvicachromis and Chromidotilapia, is in a rather unsatisfactory state at present, I have concentrated my atten- tions on the type-species of the various genera. This is particularly so with regard to anatomical data and, where information is available, on the natural history of the species. Nevertheless, other species have not been neglected (where material permitted) and, as far as possible, are included when intergeneric comparisons are made. Likewise, particular attention has been paid to the correct specific identification of the specimens used. Whenever sample size allowed, these have been taken from amongst the syntypes of a species. A review of Thys' supraspecific pelmatochromine taxa As Thys (1968a) recognised, the pelmatochromines can be divided into two groups on the basis of at least one feature, namely the nature of the pharyngeal roof immediately anterior to the upper pharyngeal bones. In one group, to which the majority of taxa belong, this tissue is organised into a distinct, visor-like pad projecting downwards and forwards for a short distance in front of and medial to the epibranchial of the first gill-arch, and is clearly circumscribed from the roof of the bucco- pharyngeal cavity by a distinct transverse groove (see Trewavas, 1974:389-391; Greenwood, 1983: 265-267). Associated with the presence of this pad (but probably not in a functional sense, see Trewavas, 1974; Greenwood, 1983) is an absence of microbranchiospines on any gill-arch. The second group has no visor-like pad. The comparable area of the bucco-pharynx (as in the majority of cichlids) is merely a little more pachydermatous than the surrounding tissues. In this group, too, as in the majority of cichlids, microbranchiospines are present on the lateral aspects of gill-arches 2-4. The absence of a hanging pad, and the presence of microbranchiospines are taken to be the plesiomorphic condition in cichlids. As was noted above (p. 140) Thys' supraspecific groups, some a little modified, are now recog- nised as genera, and some of his taxa, unallocated to a group, have recently been placed in newly described genera. It is as genera that they will be reviewed below. I Pehnatochromines with microbranchiospines but without a pharyngeal hanging pad PELMATOCHROMIS Steindachner, 1894 TYPE SPECIES: Pelmatochromis buettikoferi Steindachner, 1 894 (subsequent designation by Regan, 1922:252). PELMATOCHROMINE FISHES 143 SYNONYMY. Pelmatochromis (part): Steindachner, 1894 (the species buettikoferi only); Pellegrin, 1904 (the species buettikoferi [with which ocellifer was synonymised] and nigrofasciatus); Boulenger, 1 9 1 5 & 1 9 1 6 (the species buettikoferi, ocellifer, and nigrofasciatus [in part]); Regan, 1922 (the species buettikoferi, corbali, nigrofasciatus [in part], longipinnis [now a synonym of nigrofasciatus; see Trewavas, 1973]); Thys van den Audenaerde, 1968a (the species buettikoferi, corbali and ocellifer, with which nigrofasciatus was synonymised — but see Trewavas, 1973 for a correction of this misidentification); Trewavas, 1973 (the species buettikoferi, nigrofasciatus and ocellifer). Paratilapia (in part): Boulenger, 1915 (the species corbali only). Tilapia (in part): Thys van den Audenaerde, 19686 (the species buettikoferi and ocellifer (with which was included nigrofasciatus). The Pelmatochromis generic concept has had a checkered history. Originally created by Steindachner for two species (one of which, P.jentinki, is now referred to Tylochromis; see Regan, 1922), the genus was expanded by Pellegrin (1904) to include 18 species. Boulenger (1915) con- tinued this expansion by raising the number to 36. Regan's (1922) far more critical review of the 'generic' characters then recognised, reduced the number of species to 24. Thys' (19680) review of Regan's Pelmatochromis concept actually increased the species number to 27. But, unlike previous revisers, he recognised the polyphyletic nature of the assemblage, breaking it down into a number of what he called 'natural groups' (Thys, 19680:382). As an interim measure he designated those groups as subgenera. The nominate subgenus in Thys' scheme contained but two species, P. buettikoferi and P. corbali, which, however, he thought were probably conspecific although he did not formally synonymise the two taxa. The most recent revision (Trewavas, 1973) recognises three species (P. buettikoferi, P. nigrofasciatus and P. ocellifer, the latter being one of the species treated by Thys (19680) as a Pelmatochromis of uncertain subgeneric affinity, but which he later [19686] transferred, with the other two, to Tilapia). Trewavas (op.cit.) makes no reference to P. corbali, presumably accepting Thys' (19680) view that it is a synonym of P. buettikoferi. One reason for this conceptual instability, at least when seen in retrospect, would seem to be the persistent lack of any critical definition for the genus. That is to say, a definition based on derived characters uniquely shared by all the constituent species. In every definition to date, the characters used are either plesiomorphic ones or, if apparently derived ones (like the unicuspid jaw teeth and cycloid scales), are widely distributed amongst other taxa. Those other taxa, however, are each definable on the basis of their own and uniquely shared synapomorphies, suggesting that the apomorphies shared with Pelmatochromis are either homoplasies or features indicative of relation- ship at a very distant level. In other words, Pelmatochromis is, and always has been defined and categorised on the absence rather than the presence of defining features, a sort of monophyly by default. Its real monophyly has yet to be established. Thys (19680) defined the subgenus Pelmatochromis (Pelmatochromis) on the basis of its mem- bers possessing microbranchiospines (absent in members of all his other named Pelmatochromis subgenera, but present in one of his unnamed groups) and not having a Tilapia-spot, but instead a dark blotch (the so-called Pelmatochromis-spot) on the flank immediately below the soft dorsal fin. The presence of microbranchiospines cannot be considered a derived feature within the Cichlidae (see Stiassny, 1981). The absence of a Tilapia-spot might, on the other hand, seem to be an apomorphy since its presence is apparently a plesiomorphy amongst African cichlids (Oliver, 1984; personal observations). Furthermore, it would seem that a Pelmatochromis-spot could be interpreted as a derived condition of the plesiomorphic Tilapia-spot. However, a Tilapia-spot is present in two of the three Pelmatochromis species (see Trewavas, 1973:12), and in one of these (P. nigrofasciatus) it sometimes occurs in combination with an incipient Pelmatochromis-spot (op.cit.: 7). On the basis of those data, the presence or absence of a Tilapia-spot and, or, the presence of a Pelmatochromis-spot would be of doubtful value for establishing the monophyly of what is now the genus Pelmatochromis. Three other seemingly derived features of Pelmatochromis (cycloid scales, unicuspid jaw teeth, and the shape of the gill-rakers) must also be considered. On the basis of outgroup comparisons amongst perciform fishes, ctenoid rather than cycloid 144 P. H. GREENWOOD Fig. 1 Anterior premaxillary teeth (frontal view) of Pelmatochromis buettikoferi. Magnification x 16 (BMNH 191 1.5.31:47). scales should be considered plesiomorphic. The situation within the Cichlidae is, in my view, less readily interpreted (see discussions in Greenwood, 1979, and Oliver, 1984). If, as Oliver (1984) argues, cycloid scales are to be treated as apomorphic (i.e. derived) features in the Cichlidae, then I would agree with his further suggestion that they '. . . arose two or more times in the family'. The value of cycloid scales for establishing the monophyly of Pelmatochromis can, however, be discounted since the scales in all pelmatochromine taxa are predominantly of that type. Large specimens of all three Pelmatochromis species have, in the outer row of both jaws, teeth which are relatively slender unicuspids, their crowns somewhat attenuated and slightly compressed (Fig. 1). Such unicuspid teeth can, in my opinion, be considered an apomorphic feature within the cichlids, even though unicuspids occur commonly amongst percoid outgroups. My argument for considering unicuspid teeth in cichlids as a derived feature is as follows: i. Unicuspid teeth in such basal percoids as the Serranidae and Centropomidae are small, very numerous and form a relatively dense felt on the alveolar surface of the dentigerous jaw bones, with no clear differentiation between outer and inner tooth rows. In cichlids with unicuspid teeth, those in the outer row are clearly larger than the inner row teeth, are usually separated from the latter by a distinct gap, are well separated from one another and, relative to the basal percoid type, are larger, ii. Whenever it has been possible to study the post-larval ontogeny of jaw teeth in cichlids, a definitive unicuspid dentition in the outer tooth rows is preceded by one of essentially bicuspid or weakly bicuspid teeth in which the minor cusp forms a sort of shoulder to the major cusp (Green- wood, 1974; 1982). Such early bicuspids may be interpreted as an interim dentition since, in the few examples studied, the earliest recognisable jaw teeth present in buccal larvae are very fine, needle- like objects (see Greenwood, 1956:228 for haplochromines; Fishelson, 1966:596-7 and fig. 28 for Tilapia tholloni where the teeth are described as tricuspidate but the photograph indicates other- wise; and Balon, 1977:165 for Labeotropheus species). The occurrence of some bicuspid teeth in P. nigrofasciatus was noted by Thys (19680:370; fig. 1 6), but at that time this species was misidentified as P. ocellifer (see Trewavas, 1973) and was thus not included in his subgenus Pelmatochromis (Pelmatochromis). These bicuspid teeth are discussed in detail by Trewavas (1973:6; fig. 4) who showed that distinctly bicuspid (i.e. 'notched') teeth in P. nigrofasciatus are found only in small fishes, and are less numerous than the slender conical ones PELMATOCHROMINE FISHES 145 with which they occur. The sole specimen with notched teeth available to Trewavas is 22 mm SL; in the next larger specimen she examined (38 mm SL) notched teeth are absent and are replaced either by unicuspids or by very weakly bicuspid teeth in which the minor cusp is reduced to no more than a shoulder continuing the base of the major cusp towards the margin of the tooth (op.cit., fig. 4). Such shouldered teeth, unlike the unicuspids which replace them, have flattened and not conical crowns. I have re-examined the small P. nigrofasciatus studied by Trewavas and can confirm her obser- vations. Furthermore, I find that in two larger individuals (100 and 109 mm SL, MRAC 52454-5 identified by Trewavas) some outer teeth situated posterolaterally in the dentary have faint indications of a shoulder and, like definitely shouldered teeth, have an obliquely cuspidate and compressed, not conical, major cusp. Similar shadow-shouldered teeth occur in a 65 mm SL speci- men (from MRAC lot number 118090 -99), which also has some more obviously bicuspid (i.e. notched) teeth amongst the unicuspids. In both these specimens the non-unicuspid teeth are not restricted to any particular region of the jaw. Shadow-shouldered teeth, identical with those in P. nigrofasciatus, occur amongst the slender, conical-crowned unicuspids in the outer tooth rows of both jaws in the holotype of P. ocellifer, a specimen of ca. 65 mm SL. Many teeth are missing from this now poorly preserved specimen but it seems likely that the shadow-shouldered forms were confined to the anterior and anterolateral regions of the jaws. Thus, pace Trewavas (1973:14) it seems that barely recurved, weakly shouldered teeth, quite unlike the definitive conical-cusped and moderately recurved unicuspids, occur in small specimens of at least two Pelmatochromis species, and that some may persist in larger individuals. Trewavas, on the contrary, believed that in P. buettikoferi and P. ocellifer the '. . . outer teeth are curved and conical at all stages from a standard length of 28 mm'. Teeth in the inner rows, at least in P. nigrofasciatus (Trewavas, 1973:12) also undergo an ontogenetic change in form, being unicuspid in fishes of 68 mm SL and above, but shouldered in smaller individuals. However, the inner teeth in the smallest P. ocellifer and P. buettikoferi examined are all small and slender unicuspids. The ontogenetic changes leading to a unicuspid adult dentition in the jaws of two Pelmato- chromis species would seem to reinforce the assumption that the definitive dentition in this taxon represents a derived condition. A unicuspid definitive dentition, of course, occurs in many cichlid genera (including other pelmatochromines), taxa which, in most cases, can be defined on the basis of their own autapomorphic features, or which have other apomorphic characters not present in Pelmatochromis. Thus, in itself the unicuspid dentition of Pelmatochromis could as well be a synapomorphic character indicative of that taxon's membership in some larger phyletic assem- blage, rather than an autapomorphy for the genus alone. Without the support of other congruent autapomorphies it does not establish the monophyly of the genus. The third apparently derived feature shown by Pelmatochromis concerns the morphology of the gill-rakers (Fig. 1 1). Although there is a very wide range of gill-raker form, size and arrangement in African and in Neotropical cichlids, the rakers in the outer row on the first gill-arch in all three Pelmatochromis species are of a type rarely encountered in the family. They are also quite unlike those found amongst other members of the Labroidei (the suborder to which the cichlids belong) or in the majority of percoid fishes. On those grounds I would consider that the rakers in Pelmatochromis are indeed apomorphic. But again, there is the problem of whether or not the character is an autapomorphic one, and thus indicative of the taxon's monophyly (see below). The outer ceratobranchial gill-rakers in Pelmatochromis are moderately long, pachydermatous structures, so arranged on the greater part of the arch as to be directed obliquely forward across its surface and not, as is the usual condition, pointed aborally away from the arch. In outline each raker is approximately crescentic (the concave face uppermost), deep over most of its length but with the distal part drawn out rather abruptly into a fine, bluntly pointed, horn-like tip. Proximally at least, the membrane covering the bony skeleton of the raker is continued medially where it meets a prominent ridge of tissue running a little mediad to the midline of the arch. Between each of the rakers of the inner and outer rows this ridge is dra\vn-up into a prominent but low papilla. Over the concave upper surface of each raker (but most noticeably in those on the middle section of each 146 P. H. GREENWOOD arch) its covering tissue is thrown into numerous, contiguous, low tubercles, so that the raker comes to resemble the chela of a crab's claw (albeit a rather compressed one). The inner rakers of this arch are shorter than their outer counterparts, and are subconical in shape; as in the outer rakers, the aboral surface is tuberculate. Inner and outer row rakers are virtually continuous, the gap between them being filled by the low papilla noted above. There is apparently some interspecific variation in the morphology of the outer rakers on the first arch, both in the extent to which the distal 'horn' is produced (and thus is more or less obvious and horn-like) and in the degree to which the aboral face is tuberculated. Since, however, these features are formed in soft tissue it is possible that both could be influenced by fixation and preservation, by the time for which the specimen has been preserved, and by prefixation post- mortem changes. Fig. 2 Pelmatochromis ocellifer; from the original drawing by J. Green. Scale bar in mms. Both rows of gill-rakers on the epibranchial part of the first arch are simple, slender and relatively long, and are implanted at right angles to the arch's surface. Although the ceratobranchial gill-rakers in Pelmatochromis can be ranked as an apomorphic feature, it is difficult to decide whether they should be considered an autapomorphy for the genus rather than a synapomorphy shared with most other pelmatochromine taxa except Pterochromis (for which see p. 155). Allowing for possible influences of preservation and other chance factors, gill-raker morphology in these other taxa seems merely to be an intensification of certain details (tuberculation, thicker epithelial cover, and greater attenuation of the horn-like tip) found in Pelmatochromis. On the present evidence I would withold a final judgement on that point, but would note here that the evidence suggests synapomorphy rather than autapomorphy (see p. 191). Since none of the three presumed apomorphic features of Pelmatochromis can be established unequivocally as an autapomorphy, the monophyly of the genus still cannot be established. Its status as a discrete taxon is based only on its possessing a suite of characters, both apo- and plesiomorphic which, taken in combination, differentiate it from other superficially similar taxa. PELMATOCHROMINE FISHES 147 Fig. 3 Pelmatochromis buettikoferi. Neurocranium in: A. Dorsal; B. Left lateral view. C. Neurocranial apophysis for the upper pharyngeal bones. P = parasphenoid. Scale bar in mms (BMNH 191 1.5.3:48). 148 P. H. GREENWOOD Interestingly, Trewavas (1973:19) commented that '. . . the adult dentition of P. nigrofasciatus is itself so distinct as to suggest that Pelmatochromis is diphyletic', but did not develop her argument further nor, in that particular context, did she indicate the nature of the distinction. Presumably she is referring to the size discrepancy between inner and outer row teeth in this species as compared with the condition in other Pelmatochromis species, and to the slight posterior broadening of the inner tooth rows of the dentary. In the absence of any other seemingly autapomorphic morpho- logical features in P. nigrofasciatus (its reproductive isolation is assumed), or derived features shared uniquely by the two other species, the possibility of the genus being non-monophyletic can no more be established than can its monophyly. The interrelationships of Pelmatochromis are discussed on p. 192 below. DIAGNOSIS. Several of the features noted here are discussed in more detail on pp. 149-153. Pelmatochromis is distinguished from other genera of African cichlids by the following characters taken in combination: Body moderately deep (Fig. 2), not fusiform. Neurocranial apophysis for the upper pharyngeal bones formed from the parasphenoid alone B Fig. 4 Occlusal view of the upper pharyngeal bones, in situ, of: A. Parananochromis longirostris (BMNH 1903.7.28:77-83), and B. Tilapia zillii (BMNH 1973.3.27:7-8). Scale bar in mms. PELMATOCHROMINE FISHES 149 (i.e. of the Tilapia-type; Fig. 3C); supraoccipital crest prominent (Fig. 3B); supraethmoid not contacting the vomer (at least in P. buettikoferi). Palatine with a two-point neurocranial contact, anteriorly through an articulation in the region of the ethmovomerine suture, and posteriorly with a facet on the ventral face of the lateral ethmoid (see Greenwood, 19850); a calyx present between the metapterygoid and the hyomandibula (see Greenwood, 19850; 1986). Prepharyngeal pad in the bucco-pharyngeal roof prominent but not developed into a visor-like hanging pad clearly demarcated from the surrounding tissue. Posterior facing margins of the major upper pharyngeal bones (the fused pharyngobranchials 3 and 4) so aligned as to form a broad and shallow V whose apex is directed anteriorly (see Fig. 4A). No 'kukri'-shaped teeth on the lower pharyngeal bone (see p. 197 and Fig. 30), the teeth of which all have the tip of the major cusp pointing forwards and lying a little anterior to, or in the same vertical as the tooth's posterior margin. Anterior blade of the lower pharyngeal bone short, its length contained 3-4 times in the overall length of the bone. Infraorbital bones forming a complete lower and posterior margin to the orbit, the first infra- orbital (lachrymal) longer than deep and with 4 openings to its laterosensory canal (Fig. 5A). Outer row jaw dentition, at least in fishes > 25 mm SL, composed predominantly of slender and slightly recurved unicuspids with attenuated conical crowns (Fig. 1); in smaller specimens some weakly cuspidate teeth with notched or shouldered crowns also occur (see p. 144); inner tooth rows are composed mainly of slender unicuspids. Ascending process of the premaxilla 30-35 percent of head length. Anguloarticular with roofed canal for the mandibulo-preopercular laterosensory canal. Lower jaw length less than half head length. Scales cycloid, 25-30 pored scales in the lateral-line series; upper lateral-line clearly separated from the dorsal fin base over the line's entire length, and by at least 1\ or 3 scales at its highest point (ca the 8th scale in the series), and by 1^ or 2 scales over the last pored scale; 16 scales around the caudal peduncle. Cheek with 3 scale rows. Chest scales not noticeably smaller than those situated anteriorly on the ventral and ventrolateral aspects of the flanks and belly. Gill-rakers in the outer row on the ceratobranchial of the first gill-arch number 14-16, are transversely orientated across and close to the arch, are long, pachydermatous and crescentic, have the concave upper surface tuberculate, and the distal tip produced and horn-like. These rakers are continuous with a prominent projection developed from the membranous ridge extending down the middle of the gill-arch (see p. 145). Gill-rakers on the epibranchial of the first arch number 3-8, are simple, relatively slender and elongate, and are arranged in two rows. Microbranchiospines present on the outer face of gill-arches 2-4. Modal total number of vertebrae 25; there are 13-15 (mode 14) abdominal and 10-12 (mode 1 1) caudal vertebrae (all counts exclude the fused PU1 and Ul centra). Spondylophysial apophysis for the retractor pharyngeal muscles well-developed and situated on the 3rd centrum. A single supraneural (predorsal) bone. Dorsal fin with 13-15 spines, anal fin with 3 spines. Tilapia-spoi present in at least the young of two species, 'replaced' ontogenetically in one of those species by a Pelmatochromis-spot which is also present in a third species that never develops a Tilapia-spot (see Trewavas, 1973 and p. 143 above). All species are substrate guarders. NOTES ON THE ANATOMY AND OTHER FEATURES. Anatomical and osteological studies have been made principally on the type species, P. buettikoferi, but certain osteological features were checked from radiographs and by superficial dissections of P. ocellifer and P. nigrofasciatus. In most features the cranial and pharyngobranchial myology, osteology and arthrology of P. buettikoferi depart but slightly, if at all, from what seemingly can be taken as the modal and plesiomorphic condition for African cichlids. In an intrafamilial context there are few derived features (see Barel et al., 1976; Cichocki, 1976; Anker, 1978; Stiassny, 19810 & 1982; Greenwood, 1983 & 1985a). Osteology. The neurocranium (Fig. 3) has the plesiomorphic Tilapia type of apophysis for the upper pharyngeal bones (Greenwood, 1978) in which the entire articular surface is contributed by the 150 P. H. GREENWOOD parasphenoid. The skull (Fig. 3) is relatively foreshortened and broad interorbitally, and the ethmovomerine region is somewhat decurved, sloping downwards at an angle of ca. 50° to the horizontal. The ethmovomerine length (measured directly from the anterior tip of the vomer to the lateral tip of the lateral ethmoid) is short and is contained about 3$ times in the neurocranial length (measured directly from the anterior tip of the vomer to the ventral rim of the basioccipital facet). At least in P. buettikoferi (the other species could not be examined) the supraethmoid does not contact the vomer, an unusual feature (see Trewavas, 1973; Greenwood, 1985a) but one of as yet indeterminable phylogenetic significance. In P. buettikoferi (again the only species which could be checked) the lateral commissure of the pars jugular is is of the Haplochromis-type, and there is a stout precommissural bridge between the Fig. 5 Right infraorbital bones of: A. Pelmatochromis buettikoferi (lachrymal only); B. Chromidotilapia guntheri (lachrymal + 4 others); C. Parananochromis longirostris (lachrymal + 2 others); D. Pelvicachromis pulcher (lachrymal + 2 others); E. Thysia ansorgii (lachrymal + 2 others). Scale bar inmms. PELMATOCHROMINE FISHES 1 5 1 prootic and pterosphenoid bones (see Greenwood, 1986). An Haplochromis-type lateral com- missure is a plesiomorphic feature, but the significance of a precommissural bridge is still unknown (Greenwood, op.cit.). The supraoccipital crest is well-developed and relatively long basally, its anterior point lying above about the mid-point of the orbit. The infraorbitalbone series (Fig. 5 A) is complete, with almost half the second ossicle overlapping a posterior, plate-like flange developed below and behind the posterior lateral-line canal opening of the lachrymal. There are four lateral-line canal openings in the lachrymal, the presumed plesimorphic condition in cichlids (Trewavas, 1983; Greenwood, 19850). Few characters associated with the pharyngobranchial skeleton in P. buettikoferi require comment. There is a very broad and elongate, near-rectangular cartilaginous expansion of the second epibranchial's anterior margin (see fig. 1 1 in Trewavas, 1973). This process is partly lodged within the body of the prepharyngeal pad. From my own observations on a number of African cichlids there is apparently little or no correlation between the extent to which the cartilage is developed and the degree to which the pad is developed, unless it is of the hanging pad type. For instance, in both Orthochromis machadoi and Hemichromis bimaculatus the prepharyngeal pad is very well-developed and approaches the visor-like hanging pad found in several pelmatochromine taxa (but not Pelmatochromis). Yet, in O. machadoi and H. bimaculatus the cartilage is very poorly developed (Greenwood, 19840 & 19850). It seems, however, that in species with a fully-developed hanging pad the cartilage is always extensive (personal observations on various pelmatochromine species, and in those members of the Ophthalmotilapia assemblage from Lake Tanganyika which have a hanging pad; see Greenwood, 1983). Unfortunately, this cartilage is invariably shrivelled in dry skeletons, and to date few alcian-blue counterstained alizarin preparations are available for study. Thus the validity and significance of my few observations cannot be tested. It should be noted that in the Tilapia busumana dissected for this purpose, the cartilage is as extensive as that in the Pelmatochromis nigrofasciatus illustrated in fig. 1 1 of Trewavas (1973), and does not resemble the smaller and very differently shaped cartilage that author shows for T. busumana. It is difficult to tell, however, whether these differences are attributable to intraspecific variability or to the techniques used for preparing the specimens. The quadrangular area of epibranchial 4 in P. buettikoferi is deep and near-rectangular, and the shank spine is ridge-like; the bone thus closely resembles that of Chromidotilapia kingsleyae illustrated by Greenwood (19850; fig. 15B). The lower pharyngeal bone in all species (Fig. 6A; and figs 5A-D in Trewavas, 1973) has a short anterior blade (ca. 3^-4 times in the overall length of the bone, measured as the vertical between the tip of the blade and the tips of the articular horns). Its posterior margin has a deep median indentation so that the outline of the bone is distinctly arrow-head shaped. There are some, slight, interspecific differences in the lower pharyngeal dentition (see Trewavas, 1973), but no major modifications are apparent in tooth shape or cusp form, which are essentially of the generalised African cichlid type (see p. 197 for a description and discussion of the 'kukri'-shaped teeth in tilapiines). The upper pharyngeal teeth are likewise of a generalised type. The upper pharyngeal bones (pharyngobranchials 2 and 3 + 4, the former lying anterior and closely apposed to the latter) are so arranged that the posterior facing medial margins of the fused third and fourth pharyngo- branchials of each side form a broad and shallow V. The apex of the V, the second pharyngobran- chials, is directed anteriorly (Fig. 4A). In outline the fused third and fourth pharyngobranchials are narrowly sub-ovoid. A broadly V-shaped arrangement of the upper pharyngeal bones is the plesiomorphic condition for both the African and Neotropical cichlids (see discussion on p. 198). Little comment is required on the caudal fin and axial skeleton, except to note that the single supraneural bone is well-developed, that there is (at least in P. buettikoferi) a stout and prominent spondylophysial apophysis on the third centrum, and that the number of vertebrae is low, both in total and in separate counts of abdominal and caudal elements. Again, this is the plesiomorphic condition. In all three species there are 13-15 (mode 14) abdominal vertebrae and 10-12 (mode 1 1) caudal elements, giving total counts of 24-26 (mode 25); the fused PU1 and Ul centra are not included so that the total is lower than that given by Trewavas (1973). 152 P. H. GREENWOOD No outstanding features were noted in the skeleton of the jaws, or in the suspensorium and palatoquadrate arch. A distinct calyx formed from the metapterygoid and hyomandibula, and open medially, is present (see Greenwood, 1985#), and the mandibulo-preopercular laterosensory canal passes through a roofed canal in the anguloarticular bone (see Greenwood, 1985« & b). There are five Fig. 6 Lower pharyngeal bone, in occlusal (left) and ventral (right) views, of: A. Pelmatochromis buettikoferi; B. Thysia ansorgii; C. Chromidotilapiafinleyi. Scale bar in mms. PELMATOCHROMINE FISHES 153 laterosensory openings in the dentary and seven in the preoperculum; that is, the usual number amongst African cichlids and in the majority of Neotropical taxa as well. Both these conditions are plesiomorphic for cichlids, but the calyx is a widespread apomorphy in the family. Squamation. The scales in all Pelmatochromis species are predominantly cycloid ones, with a few weakly granular scales sometimes occurring as well. Scales on the chest are not noticeably smaller than those on the belly and ventrolateral aspects of the flanks, and what size change there is takes place gradually. The cheek is completely scaled, being covered by three rows at its deepest point. The upper lateral-line is, at its highest point, separated by at least 2| scale-rows from the dorsal fin base, and there are at least 2^ scale-rows above the last pored scale in the series. There are 1 6 scales around the caudal peduncle, the commonest and probably plesiomorphic condition in African cichlids. Fins. No sexual dimorphism is apparent in the shape of the caudal fin, which has a gently rounded margin. Neither is there a marked dimorphism in pelvic fin shape. Here the first unbranched ray is clearly the longest in both sexes although, relatively, that ray may be longer in males than females; however, the small sample sizes available for all species, and the difficulty of sexing many of the specimens, require that observation to be checked. Other external features and the dentition are discussed on pp. 143-145. Myology andarthrology of the jaws and palatoquadrate arch. No really outstanding features were noted in either system of P. buettikoferi; both closely approximate to the generalised condition amongst African cichlids (see Anker, 1978; Stiassny, 1 98 la & b, & 1982; Greenwood, 1985a). Tendon A 1)9 of the adductor mandibulae I muscle is very broad, and fuses completely with the tendon of adductor muscle Aw (i.e. like the condition in Hemichromis but unlike that of the other taxa examined; see Stiassny, 198 \a & b; Greenwood, 1985a and unpublished observations). As in Hemichromis a discrete division of A I/? tendon inserts on the nipple process of the anguloarticular. Muscle A3 of the adductor mandibulae complex is poorly developed, its anterior half being entirely tendinous. The adductor arcus palatini muscle has an extensive insertion on the palatine bone (see Greenwood, 19850; 169). The dorsal gill-arch muscles of P. buettikoferi, like the musculature of the jaws and palatoquad- rate arch, are essentially similar to those of other generalised cichlids. They closely resemble Hemichromis (Greenwood, 19850) in their general lay-out, although the relative size of certain muscles is different. The major association of the levator externus 4 with the horn of the lower pharyngeal bone is exactly as in Hemichromis. Also as in Hemichromis, only a small slip of the levator posterior muscle inserts on the horn, the major part of the muscle being inserted on the fourth epibranchial bone (see Greenwood, 1985a: 159-1 62). Muscles derived from the circumoesophageal series, and extending forward between the upper pharyngeal bones to insert on the 2nd and 3rd pharyngobranchials (Anker's [1978] 'sub-epithelium muscular tissue'), are well-developed. They originate as a single longitudinal band which bifurcates anteriorly at a point about two-thirds of the way along the length of the third pharyngobranchial. Ventral gill-arch musculature. The rectus ventralis and the obliquus ventralis muscles have the same relationship with the semicircular ligament as do the muscles in Hemichromis and a number of other African and Neotropical species I have dissected (see Greenwood, 19850; also unpub- lished observations, and p. 163 below). The transversus ventralis anterior muscle is interrupted medially by the anterior blade of the lower pharyngeal bone, the common condition in African taxa (see Stiassny, 1982:442; fig. 13(a)). Relationships. The virtual absence of unequivocally autapomorphic features amongst the characters studied makes it impossible to corroborate the assumed monophyly of Pelmatochromis (see p. 146 above). Likewise, the absence, except for one possible character, of synapomorphic features with a relatively restricted distribution amongst other cichlids, makes it impossible to hypothesise, with moderate certainty, a sister-group for the genus, or to suggest that its contained taxa be transferred to some other genus or genera. Certainly there is no evidence to support Thys' (19686) idea of placing two species in the genus Tilapia (see also Trewavas, 1973), a problem discussed further on p. 195. The single derived and possibly synapomorphic feature noted above is the morphology of the gill-rakers. It would certainly argue against a relationship with Tilapia, but does suggest 154 P. H. GREENWOOD Pelmatochromis could be the plesiomorphic sister taxon to all the other pelmatochromines except Pterochromis (see p. 191). Included species. Pelmatochromis buettikoferi Steindachner, 1894; P. ocellifer Boulenger, 1899; and P. nigrofasciatus (Pellegrin), 1900. For synonymies of these species see Trewavas (1973); in that work no mention is made of P.corbali(B\gr), 1913, which Thys( 19680) thought to be'. . .very nearly related and probably identical' with P. buettikoferi. PTEROCHROMIS Trewavas, 1973 TYPE SPECIES. Pelmatochromis congicus Boulenger (1897) SYNONYMY. Pelmatochromis (in part): Pellegrin, 1904; Boulenger, 19 15; Regan, 1922;Thys, 1968a. In all, only the species congicus is involved. Tilapia (in part): Thys, \968b (the species congicus only). Fig. 7 Pterochromis congicus. From Boulenger, 1915; about natural size. GENERIC STATUS AND RELATIONSHIPS. Thys (1968a) treated congicus as a species of Pelmatochromis but did not include it in any of his named subgeneric divisions. He did, however, group it with P. ocellifer ( = P. nigrofasciatus, see Trewavas, 1973) and P. ruweti (now Tilapia ruweti), and indicated a possible relationship of the group with Tilapia sparrmanii (Thys, 19680:381 and fig. 17). Later that year, Thys (19686) formally included P. ocellifer and P. congicus, together with P. buettikoferi as a subgenus (Pelmatochromis) of Tilapia; P. ruweti was placed in the nominate subgenus Tilapia, together with T. sparrmanii. Trewavas (1973) disputed Thys' reasoning (except that for including P. ruweti in Tilapia) and resurrected Pelmatochromis as a genus for the species buettikoferi, ocellifer and nigrofasciatus (see p. 143 above). Since, in her view, the species congicus could not be grouped with the other three species, and was not congeneric with any other taxon, she erected the genus Pterochromis for it alone. I would concur both with Trewavas' view on the need to separate congicus from the other three PELM ATOCHROMINE FISHES 1 5 5 Pelmatochromis species, and with the generic status she accorded it. But, I would not support her contention that Pterochromis congicus ' ... is a specialised form' derived from Pelmatochromis nigrofasciatus, a relationship originally proposed by Thys (19680; but see Trewavas [1973:10] regarding Thys' confusing P. nigrofasciatus with P. ocellifer, the latter name being used when he suggested the relationship of P. congicus). Discounting the near impossibility of identifying an actual ancestral species, Pterochromis congicus and Pelmatochromis nigrofasciatus (or for that matter any Pelmatochromis species) have no shared derived features which might indicate their shared common ancestry. The resemblances rest almost entirely on symplesiomorphic characters, as for example the low vertebral counts, the Tilapia-typc neurocranial apophysis, the presence of a Tilapia-spot in young fishes, a sessile pre- pharyngeal pad, microbranchiospines being present, the position of the upper lateral-line relative to the dorsal fin base, the size of the scales (27-28 in a lateral series), the elongate first pelvic fin ray, and the sexually monomorphic caudal fin shape. Apart from its autapomorphic features (see below), Pterochromis has two characters, cycloid scales and unicuspid outer row jaw teeth, which could be considered apomorphic, and which are also shared with Pelmatochromis (and other taxa outside the pelmatochromine assemblage as well). The polarity and significance of these characters are discussed on pp. 143-145. If it be accepted that both are apomorphies, then their widespread occurrence amongst African cichlids reduces their value as indicators of recent shared common ancestry, especially when no other and congruent synapomorphies are shared by the taxa under review. In the case of Pterochromis and Pelmatochromis there are apparently no such congruencies. The presence of unicuspid teeth and cycloid scales in Pterochromis could as well indicate relationship with Hemichromis as with Pelmatochromis if comparisons are limited to west African taxa, and could even indicate affinity with taxa in other parts of the continent. But in every instance there are no features to support that relationship, and several characters to refute it. Some of the characters used by Trewavas (1973) to distinguish Pterochromis from Pelmato- chromis, namely the long, slender gill-rakers, the long ascending processes of the premaxillae and the elongate lower jaw, would at first glance seem to be apomorphic features. Chief amongst these are the gill-rakers. Long, slender, untoothed, and numerous gill-rakers in the outer row of the first ceratobranchial are of uncommon occurrence in the Cichlidae, both African and Neotropical. They are not found in those African taxa which, on the basis of their few derived features, are taken to represent a plesiomorphic morphotype; for example Heterotilapia multispinus (see Oliver, 1984), Pelmatochromis (see p. 153) and Astatotilapia (see Greenwood, 1979 & 1980). However, long and slender rakers are common in other percoid groups, including mem- bers of the Labroidei, the suborder in which cichlids are currently classified. In most of these species, however, the rakers are tooth-bearing, as they are in the neotropical cichlid Cichla (whose derived or primitive status is disputed; see Stiassny, 1982). Thus, the gill-rakers in Pterchromis could be interpreted as a plesiomorphic feature, but one showing partial derivation since the teeth are lost. Alternatively, the gill-rakers could be neo- morphs which, in most respects, mimic the primitive condition. For the moment I see no means of resolving that ambiguity other than by applying the principle of parsimony, and concluding that the feature is indeed a plesiomorphic one. Likewise I find difficulty in assessing the polarity of the probably correlated features of long ascending processes on the premaxillae, and a long lower jaw. Both are characters which occur in African cichlids and in out-group percoids as well. The upwardly inclined mouth (a feature not necessarily correlated with the elongate lower jaw and long premaxillary process) would, on the basis of both out- and intragroup comparisons, seem to be an apomorphy, albeit one which has evolved on several occasions. One group of characters, the nature of the lower jaw dentition and dental arcade, not used by Trewavas (1973) in her generic diagnosis but noted in the generic description, is clearly an autopomorphy for Pterochromis. In that jaw the outer teeth situated anteriorly and immediately anterolaterally are several times larger and coarser than other teeth in the row, and are somewhat procumbently implanted. Their distinction is enhanced by the shape of the dental arcade which is noticeably angled mediad for a short section of its length immediately behind the enlarged anterior 1 56 P. H. GREENWOOD teeth. Indeed, the small unicuspid teeth in that sector of the outer row seem to be continuous with the outermost row of inner teeth. This impression is enhanced both by the rather irregular serial arrangement of the inner teeth, and by those teeth being restricted to a transverse arc across the anterior and anterolateral part of the dentary. Once again, as is so often the case with cichlid fishes, the characters which serve to distinguish a taxon give no indications of its phyletic relationships. DIAGNOSIS. As for Pelmatochromis (p. 148) except that: i. The unicuspid outer-row jaw teeth are relatively coarser in the upper jaw, and have only slightly recurved tips. In the dentary, the outer teeth situated anteriorly and immediately anterolaterally are much larger and coarser than the succeeding teeth in the row, and are implanted so as to slope forwards; the anterior teeth are somewhat isolated from the others since anterolaterally the row has a noticeable, and rather abrupt, mediad inclination, thus appearing to become continuous with the outermost row of the inner tooth-series (see fig. 8 in Trewavas, 1973). ii. The mouth (Fig. 7) is inclined upwards at a noticeable angle (ca. 45° to the horizontal). The ascending processes of the premaxillae are long (42-44 percent of head length), their tips reaching a point just before the anterior tip of the supraoccipital and lie above about the centre of the orbit; lower jaw also long (50-55 percent of head length). iii. Upper lateral-line, at its highest point, separated from the dorsal fin base by 3^ or 4 scales; posteriorly there is one large and one very small scale between the last two pored scales and the dorsal fin base. iv. Gill-rakers in the outer row of the first ceratobranchial are numerous (17-19), long, smooth and fine (i.e. flattened, bristle-like structures), and are implanted at approximately right angles to the surface of the arch. Rakers on the first epibranchial are slender, but are shorter proximally; arranged in two rows, with 9-11 in the outer row. v. 1 3 abdominal and 1 2 caudal vertebrae (total count 25). vi. Dorsal fin with 13-14 spines. vii. Tilapia-spol on the dorsal fin but no indications of a Pelmatochromis-spot (see p. 143). viii. Breeding habits unknown. NOTES ON THE ANATOMY AND OTHER FEATURES. Because so little study material is available, few observations could be made on the soft anatomy or the osteology of Pterochromis congicus (the sole representative), other than those which can be gleaned from superficial dissections and from radiographs. Judging from the latter, the neurocranium has an outline very similar to that of Pelmatochromis buettikoferi (see p. 149, and Fig. 3). The ethmovomerine region is short and moderately decurved, and the supraoccipital crest is high, with its anterior part extending to a point over about the middle of the orbit. The apophysis for the upper pharyngeal bones, examined by dissection, is of the Tilapia- type (Greenwood, 1978). The infraorbital bone series is complete, the lachrymal a long and relatively shallow bone with four openings to its laterosensory canal; the second infraorbital barely overlaps the posterior flange of the lachrymal. No details of the palatoquadrate arch and the suspensorium could be seen from the radiographs, but a limited dissection showed a double palato-ethmoid articulation of the common cichlid kind (see Greenwood, 19850:133-134 and 137-140). There are seven laterosensory canal openings on the preoperculum. Theyaws, apart from the long premaxillary ascending processes, and the peculiar dental arrange- ment in the dentary (see p. 155), provide no obvious points of comment which could be detected from a radiograph or which are visible externally. The passage of the laterosensory canal through the anguloarticular is roofed. There are 1 3 abdominal and 1 2 caudal vertebrae (excluding the fused PU 1 and U 1 centra) in the two specimens radiographed; in one of these a large, well-developed spondylophysial apophysis is visible on the third centrum. As seen radiographically, the caudal skeleton has no distinctive features. PELMATOCHROMINE FISHES 157 Little is known about the branchial skeleton, including the pharyngeal jaws. The lower pharyn- geal bone is figured by Trewavas (1973: fig. 7). The upper elements have the same shape and spatial relationships as were described for Pelmatochromis (p. 151), and must be considered of the generalised and plesiomorphic cichlid type. Squamation. Body scales in Pelmatochromis congicus are large and obvious; those on the chest have a gentle size-gradient with the scales on the ventrolateral aspects of the flanks, and although smaller are not noticeably so. Anteriorly, the upper lateral-line at its highest point (over the 8th pored scale) is separated from the dorsal fin base by 3| or 4 scales; posteriorly, the last two pored scales are separated from the fin base by one large and one very small scale. Save for a small naked area anteroventrally, the cheek is covered by three rows of scales. Fins. There is apparently no sexual dimorphism in the shape of the caudal fin margin, which is very gently rounded. The first pelvic fin ray is clearly the longest. With so few specimens known, it is impossible to be certain about the absence of sexual dimor- phism in caudal and pelvic fin shapes, or of the significance to be attached, in that context, to the filament-like elongation of certain posterior rays in the dorsal and anal fins. Relationships. There are no features which indicate that Pterochromis is closely related to any other pelmatochromine taxa, or on current evidence, to any other African lineage. Phylogenetically speaking it must be considered incertae sedis. Included species: the type and only species Pterochromis congicus (Blgr) 1897. II Pelmatochromines without microbranchiospines, and with a visor-like hanging pharyngeal pad All taxa included in this pelmatochromine subdivision have a typical, visor-like and well- circumscribed projection of the pharyngeal tissues situated immediately anterior to the upper pharyngeal bones of each side (see p. 142; also Trewavas, 1974; Greenwood, 1983). As conceived here, this subgroup also contains an assemblage which Thys (19680:365, his species 1 7-2 1 ) maintained have ' . . . a feebly developed or even no hanging pad on the roof of the pharynx' . I cannot agree with Thys on this point; specimens of all the named species he mentions (and one unnamed species, no. 21) have a distinctly visor-like pad. The pad is seemingly subject to the vagaries of preservation and curation (and possibly by the condition of the live fish too), and it can show intraspecific variability in its prominence. But, in its basic form and clear circumscription Fig. 8 Dentary teeth (right) of Pelvicachromis pulcher to show characteristic shape of chromidotilapiine teeth; seen in near lateral aspect. Magnification x29(BMNH 1901.11.28:13-20; a syntype). 158 P. H. GREENWOOD Fig. 9 Right dentary of Pelvicachromis humilis to show the characteristic chromidotilapiine feature of the posteriorly directed teeth situated anterolaterally on that bone (BMNH 1972.3.16:8-10). Magnification x 14. from the surrounding pharyngeal tissue, it is always apparent in the specimens and species I have examined. Microbranchiospines are absent in all members of group II pelmatochromines, a condition which is not necessarily correlated with the presence of a hanging pad since both structures are present in other cichlid genera (see Greenwood, 1983:267). Both the presence of a hanging pad, and the absence of microbranchiospines can be ranked as synapomorphic features for this pelmatochromine group (see p. 191). Two other characters can be treated as congruent synapomorphies. Throughout the size-range of specimens available to me, the outer row jaw teeth are charac- teristically shaped unicuspids. Unfortunately, those size ranges cover few small specimens of any species, and thus it has proved possible to trace the ontogenetic history of tooth form in only one species. No published information on that point is available either. In the exceptional species, Chromidotilapia kingsleyae (see p. 168), specimens 30-35 mm SL do have some teeth in which a very weak shoulder (see p. 145) is present, suggesting that at least weakly bicuspid teeth may precede the definitive and characteristic unicuspids of larger individuals. The characteristic group feature of the definitive outer row jaw teeth concerns the shape of the crown. Whereas in Pelmatochromis and Pterochromis the crown is an attenuated cone (see Fig. 1), in this group it can be described as a unilaterally compressed cone, the compression being manifest on the labial aspect of the crown (Fig. 8). That is to say, in cross-section its labial face, relative to the lingual one, is flattened, although its surface is still clearly curved. As a consequence of this differential flattening, the anterior and posterior margins of the crown, especially near its tip, are relatively acute and ridge-like, not rounded as they would be in a uniformly conical one, and the crown comes to resemble the tip of a surgical suture needle. As compared with the crown of the teeth in Pelmatochromis, and especially that in Pterochromis, this part of the tooth in all group II pelmatochromines is more strongly recurved. The second derived feature concerns the orientation, on both sides of the dentary, of the first few (1-4) teeth occupying an anterolateral position in the outer tooth row of that bone. In these teeth the curvature of the crown is directed posteriorly and not buccally as it is in all other teeth of the row (figs 9 & 10). Occasionally, in one species, the atypically orientated teeth occur unilaterally, PELMATOCHROMINE FISHES 159 and there are indications in Thysia that such teeth appear late in the ontogenetic succession of tooth forms since posteriorly directed teeth are found only in specimens more than 95 mm SL. A third derived feature, the form of the gill-rakers on the first ceratobranchial, is present in group II fishes, but its status as an autapomorphy is questionable; it could be a synapomorphy shared Fig. 10 Right dentary (in part) of Pelvicachromis pulcher to show posteriorly directed anterolateral teeth and the general morphology of chromidotilapiine teeth (same specimen as Fig. 8). Magnification x!4. Fig. 1 1 Partially schematic drawing to show the general morphology of the gill-rakers on the first arch of a chromidotilapiine species. Modified after Trewavas (1974). 1 60 P. H. GREENWOOD with Pelmatochromis (see p. 145). Basically, gill-raker morphology in both groups is similar (Fig. 11), the rakers being markedly pachydermatous, crescentic in shape, aligned transversely across and close to the arch, their tips produced into a horn-like projection, and the concave upper surface of the raker thrown into a number of low projections. Also common to both groups are the projections developed from the median membrane, and which are positioned between the rakers of the inner and outer rows. Again as in Pelmatochromis, the shape of the epibranchial rakers, as compared with the lower elements, is distinctly different, being dagger- or spear-shaped and inserted almost perpendicularly to the arch. However, in group II pelmatochromines the epibranchial rakers are reduced to a single row over much of that bone. Presumably the presence of a hanging pad contiguous with the proximal part of the arch inhibits the development of rakers in that region because a short inner row is present distally on the epibranchial. In general, the outer row ceratobranchial rakers of group II species are thicker (i.e. more pachydermatous) than in Pelmatochromis, the upper surface is usually more tuberculate, and the distal tip appears more protracted and horn-like. One other but polythetic feature may be an autapomorphy for group II pelmatochromines, and is discussed on p. 191. On the basis of all constituent taxa sharing the apomorphic features discussed above (but with great reservation about the autapomorphic status of the gill-rakers), I would treat the group II species as a monophyletic assemblage. The possible relationships of that assemblage to Pelmatochromis, and the interrelationships of its constituent genera are discussed later (p. 191). With one exception (the species ansorgii, see below) all the taxa treated here as members of group II, were placed by Thys (1968a) in one or other of three named subgenera, viz. Chromidotilapia, Pelvicachromis or Nanochromis, taxa raised to generic level by Thys' and other's subsequent revisions (see Thys, 19686; Trewavas, 1973, and p. 140 above). Certain species described after 1968 have been placed by their authors (Thys & Loiselle, 1971; Trewavas, 1973 & 1974; Roberts & Stewart, 1976; Stewart & Roberts, 1984) in one or other of these genera. Pelmatochromis ansorgii Blgr, the taxon of uncertain subgeneric affinity in Thys' (19680) revision was chosen as the type species for a new genus described by Loiselle & Welcomme in 1972, and will be considered first. THYSIA Loiselle & Welcomme, 1972 TYPE SPECIES. Pelmatochromis ansorgii Boulenger, 1901 SYNONYMY. Pelmatochromis (in part): Pellegrin, 1 904 (ansorgii only); Boulenger, 1915 (ansorgii, annectens and arnoldi); Regan, 1922 (ansorgii, annectens and arnoldi); Thys, 1968a (ansorgii with which arnoldi, annectens and maculifer were synonymised). Tilapia (in part): Ahl, 1939 (the species maculifer, see Thys, 1968a). A description of the two Thysia species, Thysia ansorgii (Blgr) and Th. annectens (Blgr), together with extensive notes on their biology, is given by Loiselle & Welcomme (1972), although few anatomical details are discussed. DIAGNOSIS. Thysia is distinguished from other African taxa by the following characters taken in combination. Body moderately deep, not fusiform (Fig. 12). Neurocranial apophysis for the upper pharyngeal bones formed from the parasphenoid only (i.e. of the Tilapia-type; Fig. 13B) supraoccipital crest prominent; supraethmoid and vomer in contact. Palatine with a two-point neurocranial contact, anteriorly through an articulation in the region of the ethmovomerine suture, and posteriorly with the ventral face of the lateral ethmoid (see Greenwood, 1985a); a calyx present between the metapterygoid and the hyomandibula. A well-developed and circumscribed, visor-like hanging pad developed from the buccopharyn- geal roof immediately anterior to the upper pharyngeal bones. Posterior facing margins of the major upper pharyngeal bones (the fused 3rd and 4th pharyngobranchials) forming a broad, shallow V, its apex directed anteriorly (Fig. 4A). No 'kukri'-shaped teeth on the lower pharyngeal bone (see p. 1 97 and Fig. 30), the teeth of which all have the tip of the major cusp pointing forwards PELMATOCHROMINE FISHES 161 Fig. 12 Thysia ansorgii. One of the syntypes (BMNH 1901.1.28:9-12); scale bar in mms. Drawn by Gordon Howes. and lying a little anterior to, or in the same vertical as the tooth's posterior margin. Anterior blade of the lower pharyngeal bone short, its length contained about 4 times in the overall length of the bone. Infraorbital bones forming a complete lower and posterior margin to the orbit, the first infra- orbital (lachrymal) with 4 openings to its laterosensory canal. Outer row jaw teeth, at least in specimens over 40 mm SL, unicuspid and with crowns of the compressed-conical type (see p. 158; Fig. 8); in fishes over 95 mm SL, a few (1-4) teeth situated anterolaterally in the lower jaw are aligned so that the crown points posteriorly, not bucally like the other teeth (see p. 158; Fig. 8). Lower jaw less than 50 percent of head length. Anguloarticular with a roofed canal for the passage of the mandibulo-preopercular laterosensory canal. Scales cycloid, 25-30 in the lateral-line series. Upper lateral-line clearly separated from the dorsal fin base over its entire length, and by 2^-3 scales at its highest point (i.e. the 8th or 9th pored scale in the series), and by one small and one large scale over the last 2-4 pored scales. Cheek almost entirely covered by 3 or 4 rows of scales, but with a small naked area anteroventrally. Chest scales not noticeably smaller than those situated anteroventrally on the flanks and belly; 16 scales around the caudal peduncle. Gill-rakers on the ceratobranchial of the first arch pachydermatous, elongate and crescentic, orientated almost transversely across the arch, and closely applied to its surface; upper surface of the rakers tuberculate; distal tip produced and horn-like (see p. 160); 8-1 1 in number. A single row of dagger-like rakers on the upper part of the epibranchial. Microbranchiospines absent on all arches. Total vertebral counts 25 or 26, comprising 13 abdominal and 12 or 13 caudal elements (the fused PU1 and Ul centra excluded from all counts). A single, well-developed supraneural (predorsal) bone is present. Dorsal fin with 13-16 spines, anal with 3. Sexual dimorphism in the shape of the caudal fin margin, whose upper half projects further posteriorly in adult males than in females. Pelvic fin with the first ray longest in both sexes, and greatly produced in sexually active males. No Tilapia-spot; substrate guarder (Loiselle & Welcomme, 1972). 1 62 P. H. GREENWOOD NOTES ON THE ANATOMY AND OTHER FEATURES. Anatomical and osteological studies were made on the type species, Th. ansorgiionly, but comments on dental features include Th. annectens as well. Like Pelmatochromis, most anatomical features in Thysia depart in no significant way from the modal condition in other African species. B Fig. 13 Neurocranium of Thysia ansorgii in A. Dorsal; B. left lateral view. Scale bar in mms (BMNH 1977.11.8:430-^55). The neurocranium (Fig. 1 3) is generally similar to that in Pelmatochromis buettikoferi(p. 149). The supraoccipital crest is tall, the neurocranial apophysis for the upper pharyngeal bones is of the Tilapia-type, as is the lateral commissure (see Greenwood, 1986) but in Thysia there is no pre- commissural bridge although well-developed precommissural spurs are present on the prootic. The skull of Thysia differs, however, in having a slightly broader interorbital region (width con- tained about twice in the neurocranial length, as compared with 2^ times in P. buettikoferf), and the anterior tip of the supraoccipital extends further forward to a point above the anterior third of the PELMATOCHROMINE FISHES 163 orbit; also, the ethmovomerine skull region is a little shorter in Thysia than in Pelmatochromis (3|times in neurocranial length cf. 3^ times) and slopes less steeply (ca. 30° to the horizontal, cf. 50° in P. buettikoferi), and the supraethmoid is suturely united with the vomer. The infraorbital series is complete, and the lachrymal has four openings to its laterosensory canal system. Somewhat more than the anterior third of the 2nd infraorbital overlaps the posterior shelf of the lachrymal (see Fig. 5E). In the pharyngobranchial skeleton, the fourth epibranchial has a well-developed shank ridge which is almost spinous, and its quadrangular area is elongate and near-rectangular (see Greenwood, 1985a). There is a prominent and expansive cartilaginous extension from the anterior margin of the 2nd epibranchial; in one of the two specimens dissected the projection is almost square, in the other it is anvil-shaped. The upper pharyngeal bones, and their arrangement, are similar to those in Pelmatochromis, and the lower bone closely resembles that of P. buettikoferi in its shape, proportions and dentition. In most respects the drawing of the lower pharyngeal bone in Loiselle & Welcomme (1972: fig. 4, which cf. Fig. 6B) is inaccurate, particularly with regard to the depiction of a deep median depression in the posterior margin, short and triangular articulating horns, and the somewhat 'waisted' appearance of the dentigerous surface. The vertebral column is made up of 13 abdominal and 12 or 13 caudal vertebrae (the fused PU1 and Ul centra excluded); there is a single and slender supraneural bone. No outstanding features were noted in the skeleton of the caudal fin. Likewise the suspensorium, palatoquadrate arch and the jaw skeleton show no obvious departure from the usual African cichlid condition. There is a well-developed calyx, open medially, formed from the metapterygoid and hyomandibular bones, and the palatine has a two-point articulation with the ethmovomerine region of the skull (see Greenwood, 19850). In the preoperculum and the dentary there are, respectively, 7 and 5 openings to the laterosensory canals; the anguloarticular has a roofed passage for the mandibulo-preopercular canal. Myology and arthrology of the jaws and palatoquadrate arch. Once again, there are no basic differences between Thysia and Pelmatochromis, although the relative sizes of certain muscles differ a little in the two taxa, and there is a smaller area of fusion between the A I/? and Aw tendons in Thysia (see p. 1 58). All in all, however, the conditions are those of the generalised African cichlid type. Dorsal and ventral gill-arch musculature is essentially like that of Pelmatochromis, but in Thysia no part of the levator posterior muscle inserts on the articular horn of the lower pharyngeal bone, and the insertion of the levator externus 4 muscle is merely a thin, tendinous thread. In the two latter characters Thysia shows the plesiomorphic condition for these muscles amongst cichlids, but it is not alone in so doing. As in Pelmatochromis (and many other African cichlids; see Anker, 1978; Greenwood, 1985a) the rectus ventralis muscle inserts entirely on the semicircular ligament, whilst obliquus ventralis 3 has the greater part inserted on the third hypobranchial, only a thin posterolateral section joining the ligament. The transversus ventralis anterior muscle is interrupted medially by the anterior blade of the lower pharyngeal bone. Squamation. Between the highest point in the upper lateral-line (i.e. about the 8th and 9th pored scale) and the dorsal fin base there 1\ or 3 scales, and between the last 2-4 pored scales and the fin there is one large and one very small scale. The chest is fully scaled, with the scales in that region not noticeably smaller than those on the venter. A small naked area occurs anteroventrally on the cheek which is otherwise fully scaled. As in the majority of African cichlids there are 16 scales around the caudal peduncle. Fins. The lower rays in the upper half of the caudal fin are produced in adult males, with the result that the fin's posterior margin is irregular; in females, although the margin is regular, it is slightly extended around its midpoint. In both sexes the first pelvic ray is the longest (albeit only slightly longer than the second ray in a few of the juveniles and adults examined). Adult males have the first ray greatly produced and filament-like (see fig. 1 in Loiselle & Welcomme, 1972). Coloration. Loiselle & Welcomme (1972) give a detailed colour description, and there are good 1 64 P. H. GREENWOOD colour photographs in Voss (1977; 1980) and Linke & Staeck (1981). A noteworthy feature in sexually active females is the presence of a prominent silver spot on the venter immediately above the ano-genital region. Biology. Various aspects of Thysia biology, including its breeding behaviour and colour changes, are described by Loiselle & Welcomme (1972) and Voss (1977; 1980). Relationships. The relationship of Thysia amongst the pelmatochromine fishes will be discussed in detail later (p. 191). Its possible close relationship with Hemichromis and ' Pelmatochromis' thomasi (now Anomalochromis thomasi; see Greenwood, 19856) was suggested by Loiselle & Welcomme (1972); that idea is discussed, and discounted, in Greenwood (1985# & b). In their paper Loiselle & Welcomme also argued that Thysia was not '. . . closely related to either the Nanochromis- Pelvicachromis complex or to Chromidotilapia9, a conclusion very much at odds with that arrived at here (see p. 191). Several of the features in Thysia which Loiselle & Welcomme used in their argument (head shape, degree of jaw prostrusibility, position of the mouth, colo- ration) are not applicable to the question of overall relationship, but are concerned more with autapomorphic specialisations and thus with a low (e.g. generic or specific) level of classification. They do, however, comment on two characters (gill-raker morphology, and the pharyngeal pad) which I would consider to be of importance in establishing wider levels of interrelationship. Regarding gill-rakers, Loiselle & Welcomme (1972:54) indicate that those of Chromidotilapia differ from the rakers in Thysia, a conclusion with which I would not agree, especially since the feature they stress (tuberculation) is one most susceptible to the effects of preservation (p. 146). In my view, the overall gill-raker morphology, and indeed its more detailed aspects, are very similar in the two taxa, and can be considered as a unifying synapomorphy at the suprageneric level for these and other taxa (p. 1 9 1 ). Likewise, I do not agree with Loiselle & Welcomme's idea that, as compared with Thysia the hanging pad in Nanochromis is '. . . rudimentary' (op.cit.:55). Again, there is the possibility of preservation artefacts, but in none of the Nanochromis material I have examined would I describe the pad as rudimentary or in any way less well-developed than it is in Thysia. CHROMIDOTILAPIA Boulenger, 1898 TYPE SPECIES: Chromidotilapia kingsleyae Boulenger, 1898 (by original designation). SYNONYMY. Pelmatochromis (part): Pellegrin, 1904 (the species kingsleyae, guntheri and pellegrini [a junior synonym of guntheri; see Boulenger, 1915]); Boulenger, 1915 (kingsleyae, guntheri, and nigrofasciatus in part [the species batesii, treated as a synonym]); Regan, 1922 (the species included by Boulenger, 1915 above). Pelmatochromis (Chromidotilapia) in part: Thys, 1968a (the species guntheri, kingsleyae, batesii and schoute- deni). Chromidotilapia (part): Boulenger, 1898 (the species kingsleyae only); Trewavas, 1974 (kingsleyae, guntheri [with which loennbergi was included as a subspecies], andfinleyi). The genus Chromidotilapia was first described by Boulenger (1898) for his new species kingsleyae, (Fig. 14) although he did include, tentatively, another species, C.frederici (Castelnau). The latter, under a nomen novum, is now considered to be a Serranochromis (see Bell-Cross, 1975; Greenwood, 1979:304). Interestingly, among the diagnostic features for Chromidotilapia, Boulenger (1898) noted that some of the lower jaw teeth had their unicuspid crowns directed posteriorly and not buccally, a feature now known to characterise all member taxa of the group II pelmatochromines (see p. 1 58). Some six years after Boulenger erected this genus, Pellegrin (1904) synonymised it with Pelmatochromis, but gave no reasons for so doing. Regan (1922) followed Pellegrin, but again without comment or explanation. Thys (1968a), however, resurrected the taxon as, at least temporarily, a subgenus of Pelmatochromis. With Thys' (1968Z?) inclusion of the nominate subgenus of Pelmatochromis in Tilapia, the subgenus Chromidotilapia regained its former generic status, at least by implication, and was formally confirmed therein by Trewavas (1974:388-389). PELMATOCHROMINE FISHES 165 Fig. 14 Chromidotilapia kingsleyae. Holotype; from Boulenger, 1898. Scale bar in mms. Thys' diagnosis for Chromidotilapia is based on (1) The well-developed hanging pad in the pharynx; (2) The broad and tuberculate gill-rakers in the outer row of the first gill-arch (which rakers he considered to have all tubercles of about the same size); (3) No sexual dimorphism in the shape of the caudal fin, which has a subtruncate posterior margin in both sexes; (4) The absence of microbranchiospines; (5) The first pelvic fin ray is the longest in both sexes. Trewavas (1974) provided some data on intraspecific variability in gill-raker morphology, but did not comment on its relevance to Thys' diagnostic features for the genus, which she accepted without comment. I have reservations about the significance and value of the gill-raker tuberculation character used by Thys since there is considerable interspecific and some intraspecific variability in this feature amongst the Chromidotilapia species I have examined. There is also the problem associated with post-mortem changes and with preservation induced variation (see p. 146). Questions relating to autapomorphies and hence generic status amongst the pelmatochromines, especially those involving group II taxa, will be discussed later. For the moment it will suffice to note that none of those used by Thys can be considered autapomorphic for the genus. Parenthetically it should be noted here that Trewavas (1974:395) mentions a species, C. caudifasciatus, which she considers is close to Chromidotilapia batesii. Nowhere else in her paper does she mention C. caudifasciatus, and I can only assume that the reference is to the species caudifasciatus of Boulenger (1913), which Thys (19680:367 and 380) included in his subgenus Nanochromis of Pelmatochromis. That species is considered later in this paper (p. 174), and certainly does not appear to be a Chromidotilapia. DIAGNOSIS. As for Thysia (p. 160) except for: (1) Lachrymal with 5 openings to its laterosensory canal in all but one species (where there are 4). (2) Although posteriorly directed outer teeth occur anterolaterally in the lower jaw, their presence is apparently restricted to large individuals; it is not known at what size they first appear. Similarly, it is not known at what size the typical compressed- conical teeth are first present (see p. 1 58). (3) Lateral-line series with 25-29 scales; at its highest point (about the 8th scale) the upper lateral-line is separated from the dorsal fin base by two large and one very small scale, or by only two large scales. The last five or six pored scales are separated from the 166 P. H. GREENWOOD B Fig. 15 Neurocranium of Chromidotilapia kingsleyae in: A. Dorsal; B. left lateral view. C. Neuro- cranial apophysis for the upper pharyngeal bones. P = parasphenoid. Scale bar in mms. (BMNH 1867.5.3:2). PELMATOCHROMINE FISHES 167 fin base by one large and one very small scale. In some species the chest scales are slightly, but distinctly, smaller than those on the anteroventral aspects of the belly and flanks, but the size gradient between scales in the two regions is gradual; chest scales in other species as in Thysia. Cheek with 3 or 4 scale rows; in some species a narrow naked area anteroventrally. (4) Gill-rakers on the outer row of the first ceratobranchial are, in most species, more obtusely tuberculate than those in Thysia. (5) Caudal fin without any sexual dimorphism in the shape of its posterior margin, which varies from subtruncate to gently rounded. Dorsal fin with 14-17 spines, anal with 3. (6) Total vertebral count 24 or 25, comprising 12 (rare) — 14 abdominal and 12 or 13 caudal centra (fused PU1 and Ul centra excluded). (7) Mouth brooder, at least during the later stages of brood care (see Myrberg, 1965; Thys, 1968a; Peters & Berns, 1978; Linke & Staeke, 1981). (8) Adult females with an extensive metallic (silvery to coppery) flush on the dorsal fin; no silvery blotch above the vent in females. NOTES ON THE ANATOMY AND OTHER FEATURES. Dissections and osteological studies were made on C. kingsleyae, C. guntheri and C.finleyi. There is some intrageneric variation in the overall morphology of the neurocranium Figs 1 5-17). Fig. 16 Neurocranium of Chromidotilapia guntheri loenbergi in left lateral view. Scale bar in mms. (BMNH 1973.5.14:615-621). Fig. 17 Neurocranium of Chromidotilapia finleyi in left lateral view. Scale bar in mms. (BMNH 1 973 . 5 . 1 4: 544; one of the paratypes). 1 68 P. H. GREENWOOD Chromidotilapia kingsleyae and C. guntheri have a more elongate and narrower skull than does C. finleyi', however, even in the comparatively foreshortened neurocranium of C. finleyi, the ethmovomerine region is relatively longer than in Pelmatochromis species (ca 3| times in neurocra- nial length cf. 3| times), and it is also somewhat narrower. This ethmovomerine attenuation is most extreme in C. kingsleyae and C. guntheri, where the region's length is contained about 2\ to 3 times in neurocranial length. In both these species the ventral profile of the ethmovomerine skull region is noticeably elongate and decurved when compared with Pelmatochromis; even in C. finleyi the curvature and elongation is apparent. Skull width in C. kingsleyae and C. guntheri is less than in C. finleyi and Pelmatochromis. In all three Chromidotilapia species the supraethmoid and vomer are suturally united. The lateral commissure can be of either the Haplochromis-lype (C. guntheri and C. finleyi) or of an intermediate Tilapia-Haplochromis type in C. kingsleyae. A precommissural bridge is not developed in any species, but there are prominent precommissural spurs present ventrally on the prootic, and less obvious ones dorsally on that bone (Greenwood, 1986). The infraorbital series (Fig. 5B) is complete. About two-thirds of the second infraorbital bone overlaps the posteriorly directed shelf of the lachrymal. The latter bone has a characteristic apron- like shape in all three species; its overall proportions are clearly rectangular in C. kingsleyae and C. guntheri, but are more nearly square in C. finleyi. The latter species also differs in having only four and not five openings to its laterosensory canal. Four openings are the commonest condition amongst pelmatochromines, and are thought to represent the plesiomorphic state in cichlids (see Trewavas, 1983; Oliver, 1984; Greenwood, 1985a). In the vertebral column there are 12 (rarely) to 14 abdominal vertebrae (mode 13), and 12 (mode) or 13 caudal elements (excluding the fused PU1 and Ul centra), giving a total count of 24 (rare) to 26. In the two skeletons of C. kingsleyae available, and in the single skeleton of C. batesii, there is a strong and deep spondylophysial apophysis on the third abdominal centrum. The myology, arthrology and osteology of the jaws and pharyngobranchial region in Chromido- tilapia are, with few minor exceptions, like those in Thysia (see p. 1 63). The dorsal gill-arch muscula- ture, examined in C. guntheri and C. finleyi, however, is more like that of Pelmatochromis than Thysia since there is a small musculose, rather than tendinous, slip of the fourth levator externus muscle inserting onto the fourth epibranchial, and a musculose attachment of the levator posterior muscle to the horn of the lower pharyngeal bone. In those respects, Chromidotilapia is more derived than is Thysia. Due to inadequate material, little could be learnt about the extent to which the anterior carti- laginous flange on the second epibranchial is developed. In C. finleyi the flange is large and rectangular in outline. Judging from the available skeletal material the pharyngobranchial skeleton in Chromidotilapia shows no outstanding features (see also Greenwood, 19850). In the lower jaw, the tendon A I/? of the adductor mandibulae I muscle is completely free from that of the Aw tendon, the usual condition in cichlids. The ascending process of the premaxilla is, relative to the alveolar arm of the bone, longer in Chromidotilapia than in Thysia, and markedly longer than in Pelmatochromis. The dentition in adult specimens of all species examined has both the compressed-conical uni- cuspid teeth typical of group II pelmatochromines, and the equally typical posteriorly directed anterolateral teeth in the dentary (see p. 158). Smaller fishes, however, have outer row teeth of a slightly different kind. In these, the crown has a less obviously flattened labial surface and conse- quently has less distinct sharp-edged margins to the crown. Smaller specimens also lack the posteriorly directed anterolateral dentary teeth. Furthermore, some 'shadow-shouldered' teeth or even distinctly 'shouldered' teeth (see p. 145) are present in both jaws, at least in specimens of C. kingsleyae less than 60 mm SL, and in C. finleyi of less than 50 mm SL. In both these species the figures given refer to the largest of the 'small' specimens available for study in which such teeth were observed, and thus may not truly reflect the minimum sizes above which a fully 'adult' dentition is present. Posteriorly directed anterolateral teeth are absent in the dentary of most C. kingsleyae specimens less than 50 mm SL, but are present in one fish of only 40 mm SL. The size range of the samples from which these data were derived are 3 1-74 mm SL, and 29-46 mm SL for C. kingsleyae and C. finleyi respectively. PELMATOCHROMINE FISHES 169 As in Thysia, the first pelvic fin ray is the longest, and is greatly produced in adult males. Unlike Thysia, there is no sexual dimorphism in the shape of the caudal fin margin which, in Chromidotilapia, ranges from subtruncate to gently rounded. Coloration. As far as I can tell from the literature, especially the excellent colour photographs in Linke & Staeck (1981), all known species of Chromidotilapia have an unusual form of sexually dimorphic coloration. That is, an extensive metallic sheen, ranging from coppery to silver in colour, on the dorsal fin of sexually active females. Although a metallic band is developed on the dorsal fin of adult females in many Pelvicachromis species, it is, in these, far less extensive an area, and is a much more clearly defined band than is the case in Chromidotilapia. Again, judging from the literature, the Chromidotilapia-type of female dorsal fin colouring has not been recorded in other taxa of African cichlids, but our knowledge of live colours in these fishes is extremely limited. A possible exception, for example, may be Nanochromis caudifasciatus (see p. 1 74 for its new generic placement); in Linke & Staeck's (1981) figure of that species there is seemingly a silvery sheen on the female's dorsal fin which closely approximates to that of Chromidotilapia in area and appearance. In at least one species (C.finleyi, Trewavas, 1974) young fishes have a spot on the dorsal fin closely resembling a Tilapia-spot. Breeding habits. Chromidotilapia is unique amongst the pelmatochromines in being a mouth- brooder. The breeding biology of C. guntheri is described in some detail by Myrberg (1965), there are general accounts of other species in Linke & Staeck (1981), and the mouth-brooding habits of C. batesii and C. guntheri are reviewed in the broad context of the cichlids as a whole by Peters & Berns(1978). Included species. It is clear from the literature (Thys, 1968a; Trewavas, 1974 and personal communications; Linke & Staeck, 1981), that species level taxonomy for Chromidotilapia is in a very unsatisfactory state. Currently, seven nominal species are recognised, viz.: C. kingsleyae Blgr. 1898;C.g««//zen(Sauvage), 1882, with two subspecies (see Trewavas, 1974), C. batesii (Blgr), 1901; C.finleyi Trewavas, 1974; C. linkei, Staeck, 1980; C. schoutedeni (Poll & Thys), 1967, and C. busumtwensis Paulo, 1979. It seems very likely from Linke & Staeck's (198 1) account of supposedly intraspecific colour variation, that a number of sibling species probably have yet to be recognised arid described. Trewavas (cited by Linke & Staeck [1981:89] as a personal communication) has suggested that Pelvicachromis humilis (Blgr), 1916 may be closely related to, if not actually a member of, Chromidotilapia. Certainly the general facies of that species are distinctive amongst the other Pelvicachromis, and do resemble those of Chromidotilapia. However, P. humilis does not share with Chromidotilapia any of the critical autapomorphies defining that genus, but it does share with other Pelvicachromis species the apomorphic features of that taxon (see. p. 174). PEL VICACHROMIS Thys van den Audenaerde, 1968 TYPE SPECIES. Pelmatochromis pulcher Boulenger, 1901 (by original designation). SYNONYMY. Pelmatochromis (part): Pellegrin, 1904 (the species pulcher, subocellatus and taeniatus); Boulenger, 1915 (the species pulcher, kribensis, subocellatus and taeniatus); Boulenger, 1916 (the species humilis); Regan, 1922 (as for Boulenger, 1915 and 1916). Pelmatochromis (Pelvicachromis): Thys, 1968a (the species pulcher, humilis, roloffi, subocellatus, taentiatus [with which is synonymised kribensis], and a number of informally named taxa). Like the other subgenera defined by Thys (1968a), Pelvicachromis was raised to generic rank by his referral (Thys, 19686) of the nominate subgenus of Pelmatochromis to Tilapia, and by Trewavas' (1974) subsequent recognition of that subgenus as a full genus. Thys (19680) defined and diagnosed Pelvicachromis on a single feature, the unusual sexual dimorphism exhibited by the pelvic fins. In Pelvicachromis females the second or third branched pelvic ray is the longest, thereby giving the fin a distinctly rounded apex to its distal margin; in males, the first pelvic ray is clearly the longest and is often further produced in sexually active males. Judging from extra- and intrafamilial out-group comparisons, this characteristic should be treated as an autapomorphy for the genus. However, it is closely approached by most, if not all 1 70 P. H. GREENWOOD species of Nanochromis (sensu Thys, 1968a; but see below where the taxon is subdivided into two related genera). Pelvic fins in which the second and third rays are the longest also occur in at least one other African lineage, Orthochromis (Greenwood, 1979:297). In Orthochromis, however, this fin character is congruent with other apomorphies which would not indicate a close phyletic relationship with either Pelvicachromis or Nanochromis; furthermore, in Orthochromis the shape of the pelvic fins is identical in both sexes. As note above, in Nanochromis, as in Pelvicachromis, the fin is sexually dimorphic, but, unlike Pelvicachromis, Nanochromis females have the second ray only very slightly longer than the first; despite that, the apex of the fin's distal margin has a rounded rather than a pointed appearance. Again, there are congruent apomorphic features which indicate that Nanochromis is a distinct lineage. At least as far as morphological features are concerned, the characteristic pelvic fin dimorphism of Pelvicachromis would seem to be the sole autapomorphic feature for the genus. That character, combined with the rather steeply decurved dorsal head profile, the short snout and the relatively elongate body-form in all species, make Pelvicachromis an easily recognisable member of the pelmatochromine group. DIAGNOSIS. Pelvicachromis is distinguished from other genera of African cichlid fishes by the following characters taken in combination: Body form relatively elongate and moderately compressed, dorsal head profile decurved (Fig. 1 8). Neurocranial apophysis for the upper pharyngeal bones usually of the Tilapia-lype, that is formed from the parasphenoid only, but in at least some individuals of one species it is of the Tropheus-typQ, with small contributions from the basioccipital bones (see Greenwood, 1978). Supraoccipital crest relatively low, and not rising to a high peak posteriorly (Fig. 1 9). Supraethmoid and vomer in contact in at least two species. Palatine bone with a two-point neurocranial contact, anteriorly with an articulation in the region of the ethmovomerine suture, and posteriorly with a facet on the ventral face of the lateral ethmoid. A visor-like, clearly circumscribed hanging pad anterior to the upper pharyngeal bones. Posterior facing margins of the major upper pharyngeal bones (the fused 3rd and 4th pharyngo- branchials of each side) aligned so as to form a broad and shallow V with an anteriorly directed apex. No 'kukri'-shaped teeth (see p. 197) on the lower pharyngeal bone. Anterior blade of the lower pharyngeal bone short, contained 3-4 times in the overall length of the bone. Infraorbital series of bones complete, the first infraorbital (lachrymal) usually with four open- ings to its laterosensory canal, but possibly in some species (or individuals) with five openings (see p. 171). Outer row teeth in both jaws unicuspid and of the compressed-conical type (see p. 158), some teeth situated anterolaterally in the dentary have the cusp directed posteriorly and not buccally (Figs 8 & 9). Inner row teeth arranged in 1-3 rows, unicuspid or with weak and bilateral shoulders. Lower jaw length less than 50 percent of head length. Premaxillary ascending processes as long as, slightly longer or slightly shorter than the alveolar arms. Anguloarticular with a roofed passage for the mandibulo-preopercular laterosensory canal. Mouth horizontal or very slightly inclined. Scales cycloid. 25-30 in the lateral-line series; lateral-line clearly separated from the dorsal fin base over its entire length, by 1| or two scales at its highest point (above the 8th or 9th pored scales) and by one large and one very small scale above the last 4-6 pored scales. Cheek with 2 (rare) or 3 rows of scales; a small naked area anteroventrally sometimes present. Chest scales distinctly smaller than those on the belly and antero ventral aspects of the flanks, the size intergradation being rather abrupt. Sixteen scales around the caudal peduncle. Gill-rakers of the outer row on the first ceratobranchial long, pachydermatous and crescentic, aligned transversely and close to the arch, the concave upper face variously tuberculate, the distal tip produced and horn-like; 9-12 in number. Epibranchial rakers of the first arch slender and flattened, sword- or dagger-shaped, arranged in a single row except near the epi-ceratobranchial junction. Microbranchiospines are absent. Abdominal vertebrae 13-15 (mode 1 4) in number, caudal elements 10-12 (mode 1 2), giving total counts of 23-27 (all counts exclude the fused PU1 and Ul centra). A single, well-developed supraneural (predorsal) bone. Dorsal fin with 14-16 spines, anal with 3. PELMATOCHROMINE FISHES 171 Fig. 18 Pelvicachromis pulcher, one of the types. From Boulenger (1901); Proceedings of the Zoological Society of London. Scale bar in mms. Caudal fin rounded, but showing sexual dimorphism in some species where males have the median rays, or the lower rays of the upper lobe, produced. Pelvic fins sexually dimorphic; females with the second ray (sometimes both the 2nd and 3rd rays) longer than the first ray, thus producing a rounded apex to the fin's distal margin; in males the first ray is noticeably the longest, and the fin has a pointed apex. Body coloration in breeding individuals with a marked sexual dimorphism. No true Tilapia- spot, but in some species the females have a dark spot approximately in the position of the Tilapia- spot. Apparently all species are cave spawners and breeders. NOTES ON THE ANATOMY AND OTHER FEATURES. In its overall morphology, the neurocranium in Pelvicachromis pulcher , P. subocellatus and P. taeniatus (the only species for which adequate study material was available) closely approximates to that of Pelmatochromis buettikoferi (cf. Figs 3 & 19). The supraoccipital crest, however, is much lower and slopes gently rather than steeply. The fronto-parietal crests in the Pelvicachromis species examined are better developed than in P. buettikoferi, and unlike that species, the vomer and supraethmoid are suturally united in Pelvicachromis, which thus shows the common African cichlid condition. In both genera the ethmovomerine region of the skull is short (ca. 3^-3| times in neurocranial length) and slopes upwards fairly steeply at an angle of about 50° to the horizontal. The available neurocrania of Pelvicachromis pulcher and P. taeniatus have a typical Tilapia-type apophysis for the upper pharyngeal bones, but in the single P. subocellatus examined there is, bilaterally, a very small contribution by the basioccipital, the apophysis thus closely resembling the Tropheus-type (Greenwood, 1978). In all three species the lateral commissure is of the Haplochromis-type, and a well-developed precommissural bridge is present in P. pulcher and P. subocellatus, but is absent in P. taeniatus, as are any indications of precommissural spurs on the prootic (see Greenwood, 1986). The infraorbital bone series is complete (Fig. 5D). The first bone (lachrymal) has four openings to its laterosensory canal system; almost half of the second bone overlaps the posteriorly directed flange on the lachrymal. Trewavas (1983) records that some (but unnamed) Pelvicachromis species are variable with respect to the number of openings in the lachrymal laterosensory canal system, 172 P. H. GREENWOOD B Fig. 19 Neurocranium of Pelvicachromis pulcher in: A. Dorsal, and B. Left lateral view; the basi- sphenoid bone is missing in this specimen. Scale bar in mms (BMNH 1901.1.28:13-20). having either 4 or 5; that situation was not encountered in any of the species or specimens I examined. The vertebral column is composed of 23-26 vertebrae (excluding the fused PU1 and Ul centra), comprising 13 or 14 (mode) abdominal and 10-12 (mode) caudal elements (figures obtained from radiographs of four species, viz: P. pulcher type series; P. taeniatus holotype and one other speci- men, and specimens identified as P. subocellatus and P. roloffi). Counts were also obtained from radiographs of P. humilis holotype, and are noticeable for the higher number (15) of abdominal centra (total count 27); the taxonomic status of P. humilis is discussed later (p. 1 74). A single supraneural (predorsal) bone is present in all the species and specimens radiographed. In other osteological features examined the skeleton of Pelvicachromis showed no seemingly significant differences from that of the other pelmatochromines considered so far. The pharyngo- branchial skeleton, for example, is like that in Pelmatochromis, as is the lay-out of its pharyngeal bones. There is an expansive cartilaginous projection from the anterior margin of the second epibranchial bone (as in Pelmatochromis), although the shank ridge on the fourth epibranchial is better defined and spine-like in Pelvicachromis (see p. 151). The lower pharyngeal bone and its dentition are similar in both genera (see Fig. 6). PELMATOCHROMINE FISHES 173 The gill-rakers show some interspecific variation in the extent to which the upper surface is tuberculated, and in the extent to which the median membrane is developed between the inner and outer raker series on the first ceratobranchial. However, possible differences in the quality of preservation (and the length of time in preservative) render difficult any assessment of the signifi- cance that can be attached to these differences. Be that as it may, the rakers in all the species examined are clearly of the derived pelmatochromine type discussed on p. 159. Myology and arthrology. Only P. pulcher, P. subocellatus and P. roloffi have been dissected. In none does the myology or the arthrology of the jaws show any outstanding features. Both are essentially of the type described for Pelmatochromis (p. 1 53), except that, as in most African species dissected, there is no fusion between the tendons A I/? and Aw in the adductor mandibulae series. Likewise, the dorsal pharyngobranchial musculature is like that of Pelmatochromis, and is thus of the basic cichlid type. Ventrally, and unlike the other pelmatochromines discussed so far, the obliquus ventralis 3 muscle in P. pulcher and P. subocellatus has no insertion on the semi- circular ligament, but inserts entirely on the third hypobranchial. Pelvicachromis roloffi, on the other hand, has a few fibres of the muscle attaching to the ligament, as is the case in the other pelmatochromines. All three species have the typical African cichlid condition of the transversus ventralis anterior muscle being interrupted medially by the anterior blade of the lower pharyngeal bone (see Stiassny, 1982). Squamation. Little need be added to the details given in the diagnosis (p. 1 70), except to note that there are few scales between the highest point of the upper lateral-line and the dorsal fin base, namely 1| or 2, and a single scale, or one large and one very small scale, between the fin base and the last 4-6 pored upper lateral-line scales. Chest scales in Pelvicachromis are distinctly smaller than those on the belly and anteroventral aspects of the flanks, and there is a fairly abrupt change in size between these scales and those on the chest. The degree of abruptness shows some interspecific variation. Fins. The nature of the sexual dimorphism in the pelvic fins was noted above (p. 169), where the differences between Pelvicachromis and members of the Nanochromis group (sensu Thys, \96Sa) were also discussed. The phyletic significance of those similarities, and differences, will be considered later (p. 194). As Thys (1968a:357;399) noted, there are two distinct but sympatric groups of Pelvicachromis species. In one there is an obvious sexual dimorphism in caudal fin shape, with males having the central caudal fin rays (or the lower few rays in the upper lobe of the fin) somewhat extended so as to disrupt the outline of an otherwise rounded-subtruncate to clearly rounded margin (the latter being the usual female condition). The second group shows no sexual dimorphism in caudal fin outline, which is either rounded or rounded-subtruncate in both sexes. Dentition. In the Pelvicachromis species I have examined, the dentition is typically that of all group II pelmatochromine taxa (see p. 1 58). It has proved impossible to determine at what size the anterolateral, backward-pointing teeth appear in the dentary, or to gather any information on the nature of the teeth in small specimens. Inner row teeth of both jaws are always slender and somewhat smaller than those of the outer row, and show a variety of crown shapes; some have simple attenuated-conical crowns, others are very similar to the compressed-conical teeth of the outer row, or may have compressed-conical crowns with faint traces of shoulders flanking the tip bilaterally. Again, because of insufficient material covering small specimens, it is impossible to tell if there are size correlated changes in the nature of the inner teeth. Coloration. A true Tilapia-spot has not been recorded in any Pelvicachromis species, and it has yet to be discovered whether or not the similarly placed ocellus (or conjoined ocelli) which occur in the adult females of some species is the homologue of a Tilapia-spot (see colour figures, pages 91, 1 15 & 1 18 in Linke & Staeck, 1981). Breeding. All the species whose breeding behaviour is known are substrate spawners and guarders, often using holes and small cave-like irregularities of the bottom (Linke & Staeck, 1981: 82-87). Aquarist literature relating to this genus can be misleading since there are considerable difficulties involved in species identification, and because several 'trade' binomia have been con- 1 74 P. H. GREENWOOD cocted by dealers (see Thys, 19680). However, the accounts given by Linke & Staeck (1981) are both critical and carefully researched, and provide a good summary of Pelvicachromis breeding biology. Included species. Species level taxonomy for Pelvicachromis is in a far from satisfactory state, as Thys (19680) clearly showed in his critical review of the genus. Currently, P.pulcher (Blgr) 1901 , P. humilis (Blgr) 1916, P. roloffi(Thys) 1968, P. taeniatus (Blgr) 1901 and P. subocellatus (Gnthr) 1871 are recognised, together with a number of informally designated 'forms' discussed by Thys ( 1 9680), and by Linke & Staeck (1981); for detailed synonymies of these species reference should be made to both those publications. Uncertainties about the correct generic placement of Pelvicachromis humilis were noted above (p. 1 69) in connection with Trewavas' suggestion (pers. comm., mentioned by Linke & Staeck, 1 98 1 ) that the species could be a member of the Chromidotilapia lineage. Certainly in its general facies P. humilis does approach the Chromidotilapia type, particularly in having a straight and sloping dorsal head profile rather than a rounded, decurved one. It also differs in certain osteological characters from the modal Pelvicachromis condition; the supra- occipital crest is lower, the ethmovomerine skull region is slightly longer and slopes at a somewhat lower angle (ca. 40° cf. 50° to the horizontal), the interorbital region is narrower, and the vertebral count (27) is higher as a consequence of there being 15 rather than 13 or 14 abdominal centra. However, in other antatomical and osteological features, P. humilis shows nothing which would suggest that it is more closely related to Chromidotilapia than to Pelvicachromis. Since P. humilis apparently shares none of the autapomorphic features characterising Chromidotilapia (it is not, for instance, a mouth-brooder), it would seem best to treat it, for the time being, as a somewhat atypical member of the genus Pelvicachromis. Its relationships within that genus may be clarified when a detailed and wide-ranging species-level revision of Pelvicachromis has been carried out. Parenthetically, note can be made of what seems to be an individual aberrancy in one of the two P. humilis specimens I have dissected. In that fish, besides the adductor mandibulae 1 muscle having a tendon to the maxilla, and its ventrally directed A 1/7 tendon to the lower jaw, there is a third and short tendon linking the head of the muscle with the coronoid process of the anguloarticular bone (Fig. 20). The second specimen examined has the usual two-tendon arrangement. PARANANOCHROMlS&n. nov. TYPE SPECIES: Pelmatochromis longirostris Boulenger, 1903. SYNONYMY. Pelmatochromis (part): Pellegrin, 1904 (the speices longirostris only); Boulenger, 1915 (the species longirostris and caudifasciatus). Pelmatochromis (Nanochromis) in part: Thys, 1968a (the species longirostris and caudifasciatus) Nanochromis (part): Trewavas, 1975 (the species gabonicus) Thys (19680:379-381) included Pelmatochromis longirostris and P. caudifasciatus in his subgenus Nanochromis (a taxon originally described and ranked as a genus by Pellegrin, 1904). Thys' reasons for so placing these two species was that, unlike most of the other pelmatochromines he dealt with, but like the species he referred to Nanochromis, both P. longirostris and P. caudifasciatus have only 12 (and not 16) scales around the caudal peduncle. He also expressed the belief that these two species, together with certain unnamed material '. . . seem to constitute a continuous evolutionary line' with Nanochromis nudiceps (the type species of Pellegrin's genus). At the time Thys' paper was published (1968), Nanochromis (sensu Pellegrin) was rather poorly defined, the chief character given to distinguish it from Pelmatochromis (sensu Regan, 1922) being the extreme dorsad position of the upper lateral-line; Regan (op.cit.) diagnosed the genus on the basis of its having the 'Upper lateral line contiguous to dorsal fin', a slightly inaccurate description since almost the anterior half of that line is separated from the dorsal fin base by at least 1 scale row. Thys (19680) expressed doubts about the validity of the lateral line position as '. . . a good generic character' but did not elaborate further on that point, nor on what a 'good' generic character might be. He also expressed some misgivings about the significance of the reduced circumpeduncular PELMATOCHROMINE FISHES 175 TA1 TAA TA1b Fig. 20 An unusual arrangement of the tendons associated with the adductor mandibulae 1 muscle in an aberrant specimen of Pelvicachromis humilis (BMNH 1972.3.16:8-10). Left lateral view. Scale bar in mms. Pal = palatine bone; TA1 = maxillary tendon of the first adductor muscle; TAA = anomalous tendon inserting on the anguloarticular bone; TA 1 b = tendon from first adductor muscle to the lower jaw. scale count since, although uncommon amongst African cichlids (where 1 6 scales are usual), it does occur in a few other species which, because of this feature, he had not assigned to a particular genus or subgenus. Nevertheless, and presumably because caudifasciatus and longirostris fell into what he con- sidered 'a continuous evolutionary line', Thys included the two species, with those of Nanochromis, as a distinct subdivision (i.e. subgenus) of Pelmatochromis. In a phylogenetic context, I cannot agree with Thys on the supposed insignificance of the lateral- line position in Nanochromis (see p. 183), the more so since in that genus (sensu Pellegrin) it is congruent with other apomorphic features recently noted by Stewart and Roberts (1984). In my view, the partial but extreme contiguity of the dorsal fin base and the upper lateral-line, together with the other synapomorphic features in Nanochromis (especially the nature of the infraorbital bone series; see p. 181) indicate an immediate common ancestor for the species showing them. That ancestor was not shared by longirostris and caudifasciatus (or by the species gabonicus of Trewavas, 1975; see synonymy above), all taxa in which the Nanochromis synapomorphies are not present. However, Parananochromis (which I have erected for the three latter species) and Nanochromis (sensu Pellegrin) do share, uniquely amongst the pelmatochromines, two apomorphic characters, namely 1 2 scales around the caudal peduncle, and a Haplochromis-type neurocranial apophysis for the upper pharyngeal bones (the latter a feature not noticed by Regan [1922], Thys [19680], or Trewavas [1975]). On the basis of those synapomorphies I would postulate a more distant, but still close, common ancestry for the two genera. It might be argued that the relationship of the lineages could best be indicated by ranking them as subgenera of Nanochromis. I would agree with that argument but for the possibility of future research showing the existence of two lineages within Nanochromis as presently construed (see p. 187). For that reason I prefer to treat Parananochromis as a genus, and rely on a cladogram (Fig. 29) to show its sister-group relationship with Nanochromis. In the sense that Nanochromis and Parananochromis are sister taxa, I would agree with Thys' idea that together 176 P. H. GREENWOOD Fig. 21 Parananochromis longirostris; holotype. From the original drawing by J. Green. Scale bar in mms. they form a continuous evolutionary line, with Nanochromis representing the more derived portion of the lineage. Thys (19680) notes that in the species he referred to his subgenus Nanochromis there is '. . . a feebly developed or even no hanging pad in the roof of the pharynx' and that the ventral fins are rounded in both sexes. The species which I have examined, and which I place in Parananochromis, all have, pace Thys, a well-developed and typically circumscribed, visor-like hanging pad, and although the pelvic fins are distinctly or almost rounded in females, they are pointed in males. Thys & Loiselle (1971) described two new Nanochromis species (sensu Thys, 1968a), N. robertsi and N. cavalliensis. Superficially, both seemingly would belong to Parananochromis since neither shows the defining features of Nanochromis (sensu Pellegrin, see p. 181). However, in both species the neurocranial apophysis for the upper pharyngeal bones is of the Tilapia-type, and both have 1 3 or 14 scales around the caudal peduncle (cf. 12 in Parananochromis and Nanochromis, and 16 in other group II pelmatochromines). On those grounds I exclude robertsi and cavalliensis from both Parananochromis and Nanochromis (see p. 188). DIAGNOSIS. Parananochromis is distinguished from other genera of African cichlids by the follow- ing characters taken in combination: Body form moderately elongate and compressed; dorsal head profile sloping steeply but somewhat decurved (Fig. 21). Neurocranial apophysis for the upper pharyngeal bones of the Haplochromis-type, with sub- stantial contributions from the basioccipital. Supraethmoid not contacting the vomer in the one species examined (P. longirostris). Supraoccipital crest relatively low, and not rising to a high peak posteriorly. Palatine with a two-point neurocranial contact, anteriorly through an articulation in the region of the ethmovomerine suture, and posteriorly with the ventral face of the lateral ethmoid (see Greenwood, 1985a). A calyx is developed between the metapterygoid and the hyomandibula. Infraorbital bone series complete, the first infraorbital (lachrymal) with four openings to its laterosensory canal (Fig. 5C). Lower jaw length less than 50 percent of the head length; anguloarticular with a roofed canal for the passage of the mandibulo-preopercular laterosensory canal. Outer row jaw teeth of the typical group II pelmatochromine type (see p. 158); inner row teeth unicuspid, arranged in from 1-3 rows. A well-developed and clearly circumscribed, visor-like hanging pad developed from the pharyn- geal roof immediately anterior to the upper pharyngeal bones. Posterior facing margins of the fused 3rd and 4th pharyngobranchials aligned so as to form a broad and shallow V, its apex PELMATOCHROMINE FISHES 177 Fig. 22 Neurocranium of Parananochromis longirostris in A. Dorsal; B. Left lateral view. C. Neurocranial apophysis for the upper pharyngeal bones. Scale bar in mms. P = parasphenoid; Bo = basioccipital (BMNH 1903.7.28:77-83; one of the syntypes). 1 78 P. H. GREENWOOD directed anteriorly (Fig. 4A). Anterior blade 3y-4 times in the overall length of the bone. No 'kukri'-shaped teeth on the lower pharyngeal bone (see p. 197). Scales cycloid. Lateral-line with 27-30 scales; upper lateral-line separated at its highest point (ca. the 8th pored scale) from the dorsal fin base by 1^ scales, and by one scale over the last 3-5 pored scales. Chest scales noticeably smaller than those on the belly and anteroventral aspects of the flanks, the size transition between these scales and those of the chest abrupt. The two or three median rows of scales on the chest are larger than those situated more laterally. There are 1 2 scales around the caudal peduncle. Gill-rakers in the outer row of the first gill-arch of the typical group II pelmatochromine type (seep. 160). Microbranchiospines absent. Vertebral column with 12-14 abdominal and 12-13 caudal elements (excluding the fused PU1 and Ul centra), giving a total count of 25 or 26. A well-developed supraneural (predorsal) bone present. Caudal fin with slight sexual dimorphism in at least one species where the lower 3 or 4 rays in the upper lobe are produced; otherwise the caudal fin margin is slightly subtruncate to rounded in both sexes. Pelvic fins with the first ray produced in males, the second ray as long as or but slightly shorter than the first in females. Dorsal fin with 14-16 spines, anal with 3. Apparently some sexual dimorphism in body coloration; at least one species (P. caudifasciatus) with a fairly extensive silvery sheen on the dorsal fin in sexually active females. No Tilapia-spot, or any markings resembling one, is developed. In other features (as in many of those listed) Parananochromis resembles Pelvicachromis. NOTES ON THE ANATOMY AND OTHER FEATURES. A shortage of material for deep dissections or for making osteological preparations has restricted the number of observations which can be made on the anatomy of Parananochromis; radiographs, however, are available for all three species. The neurocranium in P. longirostris closely resembles that of Pelvicachromis pulcher (cf. Figs 22 & 19), except that the supraoccipital crest is, relatively, a little higher in the former species. The crest has a gently curved dorsal margin so that, as in Pelvicachromis pulcher, there is no obvious posterior peak; the anterior tip of the supraoccipital lies above the midpoint of the orbit. The length of the ethmovomerine region is contained about 3| times in the neurocranial length, and slopes at an angle of about 65° to the horizontal. Unlike Pelvicachromis pulcher, the pharyngeal apophysis in all three Parananochromis species is of the typical Haplochromis-type, with clearly demarcated and moderately large basioccipital facets. Thys (1968a) makes no comments on the nature of this apophysis in either P. longirostris or P. caudifasciatus. Trewavas (1975:233), referring to P. gabonicus, notes that '. . . although the base of the skull was not examined in this unique specimen, the other characters indicate that it belongs to the group in which the apophysis ... is formed by the parasphenoid alone'. Presumably she was influenced, in making that remark, by her considering the species to be a Nanochromis (sensu Pellegrin), a genus which Regan (1922) placed in his division of cichlids having a Tilapia-type apophysis. I have examined the apophysis in Trewavas' specimen of gabonicus, and find that, in fact, it is of the typical Haplochromis-type. Stewart & Roberts (1984) were the first to record, but without further comment, that the apophysis in at least some true Nanochromis species is of the Haplochromis- and not the Tilapia-type. I have now been able to examine that apophysis in all described species of Nanochromis, and can confirm that in each it is of the Haplochromis-type. The lateral commissure in Parananochromis longirostris is of the Haplochromis-type (Greenwood, 1986) and a precommissural bridge is present. As far as can be told from radiographs, the shape of the neurocranium in Parananochromis caudifasciatus is like that of P. longirostris, but that of P. gabonicus has a lower supraoccipital crest, probably lower even than in species of Pelvicachromis. The infraorbital series in all three species is complete (cf. that in Nanochromis; see p. 1 86), and the first infraorbital bone (lachrymal) has four openings to its laterosensory canal; in P. longirostris, the only species checked, about half of the second infraorbital overlaps the posteriorly directed flange of the lachrymal (Fig. 5C). The vertebral column is composed of 25 or 26 vertebrae, comprising 12-14 abdominal and 12 or 13 caudal elements (the fused PU1 and Ul centra excluded). A single well-developed supraneural PELMATOCHROMINE FISHES 1 79 (predorsal) bone is present (it is absent or greatly reduced in size amongst species of Nanochromis, seep. 187). Gill-rakers on the cerato- and epibranchial of the first gill-arch are like those described for Chromidotilapia (see p. 160). There is, apparently, some interspecific variability in the extent to which the median membrane between the inner and outer rows of gill-rakers is developed. The membrane in P. longirostris is very low and barely produced into a tubercle between each pair of inner and outer rakers, whereas in P. caudifasciatus and P. gabonicus it is high, with prominent tubercles; in the latter species the tip of each outer-row raker on the ceratobranchial is noticeably produced and horn-like. The pharyngobranchial skeleton, except for the pharyngeal bones, has been examined only in P. longirostris. In all three species the lower pharyngeal bone has a deep median indentation of its posterior margin and a short anterior blade (its length contained ca. 3| to 4 times in the overall length of the bone), features which impart a distinctly arrow-head outline to the bone. The lower pharyngeal teeth are somewhat widely spaced, few in number, and the outermost row is implanted very close to the margin of the bone. The upper pharyngeal bones are subovoid in outline and, as in all other pelmatochromines (and the majority of African cichlids), are orientated so that their posteriorly directed margins form a broad and shallow V whose apex points anteriorly (see Fig. 4A). Fig. 23 Right epibranchial 2 of Parananochromis longirostris in dorsal view, to show the extensive cartilaginous flange (coarsely stippled) associated with its anterior border. Scale bar in mms. (BMNH 1903.7.28:77-83; one of the syntypes). In P. longirostris, the only species dissected in depth, there is a very expansive, anvil-shaped cartilage developed on the anterior margin of the second epibranchial (Fig. 23); it almost fills the anterior part of the hanging pad, and extends to its anterior border. Squamation. An unusual feature of Parananochromis, one which I would consider a possible synapomorphy for the lineage (but see p. 190), is the presence on the chest of 3 or 4 median rows of relatively but distinctly larger scales within the otherwise small scales of this area. The size tran- sition between the small chest scales, and those on the ventrolateral aspects of the flanks and belly is abrupt. Also characteristic of Parananochromis, but a feature shared with Nanochromis, is the reduced number (12) of circumpeduncular scales. Cheek squamation patterns are interspecifically variable. In P. gabonicus there are two rows, with a deep naked strip ventral to the second row, and an extensive naked area anteroventrally; P. caudifasciatus has a large anteroventral naked area, but otherwise the cheek is covered by 3 or 4 rows of scales; P. longirostris has 2 or 3 scale rows, with a naked strip (about one scale row deep) along its entire ventral margin, and a large naked area anteroventrally. The upper lateral-line is high on the body, with 1| scales separating it from the dorsal fin base at 1 80 P. H. GREENWOOD its highest point (about the 8th or 9th pored scale), and one scale separating the fin base from the last 4 or 5 pored scales. The scales on the nape and nuchal region are not noticeably reduced in size (cf. Nanochromis; p. 183). Fins. Only an adult female of P. gabonicus is available, but in the other two species adult males have the first branched pelvic ray clearly longer than the second. In the female P. gabonicus the second ray is as long as the first, and in female P. longirostris and P. caudifasciatus the second ray is either as long as the first or is but marginally shorter, and consequently the apex of the fin is more nearly rounded than pointed. The margin of the caudal fin is rounded to subtruncate in females of all three species, and, apparently, also in adult male P. caudifasciatus. Sexually active males of P. longirostris, however, have the margin of the upper caudal lobe, at about its mid-section, produced into a short, broadly triangular projection (see figure in Linke & Staeck, 1981:66). Judging from the figure of P. gabonicus in that publication (op.cit.:65), there is a slight indentation at about the mid-point of the posterior margin, but I have not been able to confirm this from preserved specimens. Dentition. None of the three specimens has any particular character of note in its dentition, which is typically that of a group II pelmatochromine (see p. 1 58). Since only one small specimen is available (a 21 mmSL P. caudifasciatus) it is not possible to make any generalisations about possible ontogenetic changes in tooth form. In this small fish the outer teeth are unicuspid, but some teeth have faint traces of a shoulder below the major cusp; the crowns are but weakly curved, and no teeth with posteriorly directed crowns are present in the dentary. No inner row teeth are visible, even at a magnification of 30X. Myology and arthrology of the jaws and gill-arches. Only P. longirostris has been dissected, and no outstanding features were noted in the musculature of the jaws. In the dorsal pharyngobranchial musculature no part of the levator posterior muscle inserts on the horn of the lower pharyngeal bone. Except for a small tendinous slip inserting on the 4th epibranchial bone, all of the levator externus 4 muscle inserts on that horn. This condition is more like that in Thysia (see p. 163) than that in the other pelmatochromines. In those, a small part of the levator posterior inserts on the horn, and the insertion of the levator externus 4 on the epibranchial is musculose rather than tendinous. In its ventral branchial musculature, Parananochromis longirostris is like the majority of pelmat- ochromines (and probably the other African cichlids; see Greenwood, 1 9850: 1 63) since a large part of the obliquus ventralis 3 muscle inserts on the semicircular ligament (Pelvicachromis is unusual in this respect, because in two of the three species dissected the muscle inserts entirely on the third hypobranchial bone; see p. 173). Coloration. No information is available on the live colours of P. gabonicus (see Trewavas, 1975 for the preserved coloration of a female fish), and for P. longirostris the data are from a recently dead male (see Linke & Staeck, 1981:66). There is obvious sexual dimorphism in the coloration of P. caudifasciatus (see Linke & Staeck, 1981:58-59), although this may be less pronounced than in other pelmatochromines; these authors record a shiny, metallic band on the spinous dorsal fin of sexually active females (see p. 169). Breeding biology. No information is available for either P. gabonicus or P. longirostris, but P. caudifasciatus is known to be a substrate spawner (Linke & Staeck, 1981:60). Included species. At present only three named species can be referred to Parananochromis; that is P. longirostris (Blgr) 1 903, P. caudifasciatus (Blgr) 1913 and P. gabonicus (Trewavas) 1 975. Judging from Thys' (1968a) comments, however, more species will doubtless be added. Two species, Nanochromis robertsiand N. cavalliensis, both described by Thys & Loiselle (1971) have a close superficial resemblance to the species of Parananochromis, but on other and derived characters appear to belong to a lineage distinct from both Nanochromis and Parananochromis (see p. 188). NANOCHROMIS Pellegrin, 1904. TYPE SPECIES: Pseudoplesiops nudiceps Blgr, 1 899 (by original designation). SYNONYMY. Pseudoplesiops (non Bleeker): Boulenger, 1899 (the species nudiceps and squamiceps). PELMATOCHROMINE FISHES 1 8 1 Nanochromis: Pellegrin, 1904, replacement name for Pseudople slops Blgr, 1899 (the species nudiceps and dimidiatus); Boulenger, 1915 (the species nudiceps, dimidiatus and squamiceps); Regan, 1922 (none listed); Stewart & Roberts, 1984 (the species nudiceps, dimidiatus, squamiceps, transve stilus, consortius, parilius, minor and splendens). Pelmatochromis (Nanochromis), in part: Thys, 1968a (the species nudiceps, dimidiatus and squamiceps). Since its original description (under the preoccupied name Pseudople slops) the genus has been defined and diagnosed principally on the high position of its upper lateral-line. In all but two species a little more than the posterior half of the upper lateral-line is contiguous with the base of the dorsal fin, and fewer than five scales anterior to the point of contiguity are separated from the fin base by more than one, or one and a half, often very small scales. The exceptional species, N. dimidiatus and TV. squamiceps, have only the last 3 to 5 scales of the upper lateral-line contiguous with the fin base, the greater part of the line having one very small scale interposed between it and the fin base. The supposed distinctness of these two species has been queried by Matthes (1964); however, in Thys' opinion (19680:307) they are not only distinct but may not even be closely related. The inclusion of dimidiatus and squamiceps in Nanochromis rather than in Parananochromis (see above, p. 176) is based on their having an interrupted infraorbital bone series in which only the lachrymal and one other canal bearing bone is present, the latter not curving around, and closing, the posterior margin of the orbit (see below). Pellegrin (1904), when proposing the new name Nanochromis for the taxon (see synonymy), noted the presence of a clearly defined hanging pad in the pharynx, and was the first author to comment on the close similarity between Nanochromis and Pelmatochromis (then conceived in its broadest sense [see Introduction] but probably used by Pellegrin with particular regard to the species now placed in the genus Pelvicachromis). Regan (1922), too, commented on Nanochromis being 'Scarcely different from Pelmatochromis'. Neither author, however, gave any reasons for their views. Thys (1968a:380) broadened the earlier concept of Nanochromis to include certain species here referred to the new genus Parananochromis, but reduced the status of Nanochromis to a subgenus of Pelmatochromis. His definition of Pelmatochromis (Nanochromis) was based, principally, on all its constituent species having a reduced number of scales (12) around the caudal peduncle since he expressed doubts about the significance of the high lateral-line position as a group character. In my view (p. 194) the high lateral-line is a good apomorphy indicating the monophyletic nature of the species possessing it, and also serves as one of the synapomorphic characters defining the taxon Nanochromis. The first really detailed taxonomic account of Nanochromis is Stewart & Roberts' (1984) paper, a follow-up of their earlier work describing a number of new species from the Zaire River (Roberts & Stewart, 1976). Although I would question some of the supposed autapo- morphies used by Stewart & Roberts (1984) to establish the monophyly of Nanochromis, the other characters used by them appear to be good, and synapomorphic, features for that purpose (see discussion below). That Nanochromis is a member of the group II pelmatochromine lineage is established by the nature of the dentition, the morphology of its gill-rakers, and by the well-developed, visor-like pharyngeal hanging pad. Before presenting a diagnosis for Nanochromis some consideration must be given to the four supposed synapomorphic features which, in addition to the high upper lateral-line position, Stewart & Roberts (1984:82-83) used to establish the monophyly of the genus. These characters, and comments upon them, are: (1) Only two infraorbital bones, a lachrymal with two lateralis canals extending to the lower rim and an elongate second infraorbital that projects posteriorly rather than follows the orbital rim; the infraorbital lateralis canal terminates posteriorly at a pore on the cheek. Reduced and, or, variously interrupted infraorbital series occur in at least three other African cichlid taxa, viz. Lamprologus (sensu lato; see Colombe & Allgayer, 1985), Telmatochromis and Julidochromis. None of these, however, has the characteristic lachrymal plus one elongate, handle- like bone seen in Nanochromis, and all have various apomorphic features which are not shared with Nanochromis. Furthermore, none shares with Nanochromis the visor-like hanging pad and peculiar 1 82 P. H. GREENWOOD dental features of a group II pelmatochromine (see p. 157), or the high-positioned upper lateral- line of that genus. The nature of the infraorbital series in Nanochromis would therefore seem to be an autapo- morphy. Stewart & Roberts' statement about the number of openings to the lachrymal latero- sensory canal could, at first sight, be misleading since there are four openings on that bone, the plesiomorphic condition in cichlids. Presumably the authors exclude from their count the anterior and posterior openings (which are shown in fig. 3 of Stewart & Roberts, 1984) since these could be interpreted as not reaching the bone's lower (i.e. ventral) rim. The elongate second infraorbital in N. squamiceps and N. dimidiatus, unlike that in the other species, does follow the curvature of the eye, and thus does not project straight posteriorly in a handle-like fashion. However, as in the other species, the infraorbital bones in N. squamiceps and N. dimidiatus do not form a complete bony rim to the orbit, and the infraorbital canal (or at least its lower portion) terminates posteriorly as a pore on the cheek. Interestingly, both these species, as will become apparent below, tend not to show certain 'Nanochromis' characteristics in the fully- developed manner seen in the remaining species. (2) No predorsal bones. The absence of predorsal bones (supraneurals) is a derived feature in cichlids, where one bone is the usual condition and two are found amongst African cichlids only in such plesiomorphic taxa as Heterochromis and Tylochromis; two predorsals are of more frequent occurrence in Neotropical genera (Cichocki, 1976:182; Oliver, 1984:29). A single predorsal, usually small and comma-shaped, is visible in radiographs of N. squamiceps, N. dimidiatus and some N. splendens (the radiographed samples included some paratypes of all three species). It is absent, nevertheless, in all the examined material of the other Nanochromis species. Clearly, there is an apomorphic trend within Nanochromis for the predorsal bone to be lost since even when present it is greatly reduced in size, and in at least one species (TV. splendens) it is more often absent than present. (3) Cheek and nape scaleless. Reduced squamation, or its complete suppression in those particular areas of the body, occurs in other cichlids, and is usually associated with torrenticolous habits. But, as was the case with reduced infraorbital bones, the phenomenon in these other taxa is either linked with apomorphic features not shared by Nanochromis, or with the plesiomorphic condition of characters which are apomorphic in Nanochromis. As Stewart & Roberts (1984) note, nape and cheek squamation is not greatly reduced in two species, TV. dimidiatus and TV. squamiceps, taxa which seem to constitute a plesiomorphic line within the genus. Thus, a scaleless nape and cheek, like the absence of a supraneural bone, could be interpreted as a synapomorphic condition uniting certain species within the genus, but not as one for the genus as a whole. (4) Five lateralis pores on the mandible and seven on the preopercle, including one at its dorsal tip. Such numbers for the openings in these canals are apparently universal amongst African cichlids. The character cannot therefore be considered a synapomorphy for Nanochromis. In brief, of the four characters listed by Stewart & Roberts (1984), only the peculiarities exhibited in the infraorbital series can be considered an apomorphy for Nanochromis as a whole. It thus joins the extreme dorsal position of the upper lateral line as a defining synapomorphy for the genus. DIAGNOSIS. Nanochromis is distinguished from all other African taxa by the following characters taken in combination: Body attenuated and slender, the dorsal head profile decurved or, if straight, sloping steeply (Fig. 24). Neurocranial apophysis for the upper pharyngeal bones of the Haplochromis-type. Supraoccipi- tal crest low, without a well-defined peak. Pronto-parietal crests weak. Supraethmoid not suturally united with the vomer. Palatine with a two-point neurocranial contact, anteriorly through an articulation in the region of the ethmovomerine suture, and posteriorly with a facet on the ventral face of the lateral PELMATOCHROMINE FISHES 183 Fig. 24 Nanochromis nudiceps; one of the syntypes. Drawn by Gordon Howes. Scale bar in mms. (BMNH 1899. 11. 27:64). ethmoid (see Greenwood, 1985a): A calyx is formed between the metapterygoid and the hyomandibula. Infraorbital bone series reduced to the lachrymal and an elongate second bone closely articu- lated with the lachrymal from which it protrudes in a handle-like manner; there is a prominent pore in the cheek at the posterior end of the second infraorbital, coincident with the last lateralis pore in that bone. Lachrymal with four openings to its laterosensory canal system. Lower jaw less than 50 percent of the head length. Ascending process of the premaxilla from slightly shorter to slightly longer than the alveolar arm. Anguloarticular with a roofed canal for the mandibulo-preopercular laterosensory canal. Jaw teeth of the typical group II pelmatochromine type (see p. 158); the first appearance of posteriorly directed teeth situated anterolaterally in the lower jaw is correlated with the size of the fish (see p. 187). A well-developed, clearly circumscribed, visor-like pad developed in the pharynx immediately anterior to the upper pharyngeal bones. Posterior facing margins of the major upper pharyngeal bones aligned so as to form a broad and shallow V, its apex directed anteriorly. Anterior blade of the lower pharyngeal bone relatively short, its length contained 3-4 times in the overall length of the bone. No 'kukri'-shaped teeth on that bone (see p. 197), the teeth of which all have the tip of the major cusp lying a little anterior to, or in the same vertical as the posterior margin of the tooth. Scales cycloid, 23-28 in the lateral-line series. Cheek naked or partly scaled; nape and chest seemingly naked in most species (but very small, scattered and deeply embedded scales sometimes present), obviously scaled in others where the nape scales are neither noticeably small nor deeply embedded, but the chest scales are small and confined to the upper part of that region. Upper lateral-line situated high on the body with, in most species, slightly more than its posterior half contiguous with the dorsal fin base, and with not more than one or one and a half scales between about the fifth pored scale in the series and the first pored scale which is contiguous with the dorsal fin base. Both lateral-line series in all species are often interrupted by the intercalation of several non-pored scales. Twelve scales around the caudal peduncle. Gill-rakers of the group II pelmatochromine type, but the tubercles are often small or difficult to detect, and the epibranchial rakers vary from short and broad to relatively slender; 6 to 9 rakers in the outer row of the first ceratobranchial, and 3-5 in a single row on the first epibranchial. Microbranchiospines are absent. Total number of vertebrae 25-28, comprising 12-14 abdominal and 12-14 caudal elements (excluding the fused PU1 and Ul centra). Predorsal (supraneural) bone absent in five species, present or, more frequently, absent in another, and present but very small and comma-shaped in two others. 184 P. H. GREENWOOD Caudal fin apparently showing no sexual dimorphism in the shape of its posterior margin. In all but one species the first pelvic ray in adult males is produced. In females the second ray is as long as or slightly longer than the first ray (a condition also seen in the adult males of N. transvestitus). Dorsal fin with 16-19 spines, anal with 3. Marked sexual dimorphism in the coloration of adults (see p. 187). A dark mark resembling a Tilapia-spot occurs in at least two species. Where breeding habits are known, the species are substrate guarders. NOTES ON THE ANATOMY AND OTHER FEATURES. The small adult size attained by most Nanochromis species, combined with a paucity of material and its often indifferent state of preservation, has severely limited the amount of anatomical information which could be gained. Certain osteologi- cal data were obtained from radiographs of all described species, and certain soft anatomical features and osteological characters were studied from partial dissections of N. nudiceps, the type species, and TV. parilius (one specimen of each). The neurocranium of N. nudiceps (Fig. 25) has a very low supraoccipital crest whose dorsal margin slopes gently upwards at a low angle; the anterior point of the supraoccipital lies in a vertical passing through the middle of the orbit. Judging from radiographs, the situation is similar B Fig. 25 Neurocranium of Nanochromis nudiceps in: A. Dorsal, and B. Right lateral view. Scale bar in mms. (Unregistered specimen; see text, p. 141). Bo = basioccipital. PELMATOCHROMINE FISHES 185 B Fig. 26 Lower pharyngeal bones (dentigerous surface left, and ventral surface right) of: A. Limbochromis robertsi; holotype. B. Nanochromis nudiceps (unregistered specimen, see text, p. 141). in all other species. The fronto-parietal crests are low and narrow. The ethmovomerine region, whose length is contained about 3| times in the neurocranial length, slopes at an angle of ca. 30° to the horizontal, its tip lying well below the level of the parasphenoid and the apophysis for the upper pharyngeal bone. That apophysis is of the Haplochromis-type, with substantial contributions by the basioccipital to its articular surface. The lateral commissure is broad and of the Haplochromis- type; no precommissural bridge is present (see Greenwood, 1986). Apart from the low supraoccipital crest, the weaker fronto-parietal crests, and a narrower interorbital region, the neurocranium of N. nudiceps more closely resembles that in most Pelvicachromis species than it does the neurocranium in Parananochromis longirostris (see pp. 172 & 177 respectively). What little information that could be obtained on other aspects of the cranial osteology (jaws and suspensoriuni) and of the branchial skeleton, suggests that no features depart significantly from those of other pelmatochromine taxa. However, it must be remembered that only one partial dissection was possible. In all Nanochromis species the lower pharyngeal bone has a deep indentation in its posterior margin, and a short anterior blade which is contained ca. 3-4 times in the overall length of the bone, features which impart an arrowhead-like outline to the bone (Fig. 26B). Teeth on this bone are fine (except for the slightly larger teeth in the posterior transverse row, and those posteriorly in the median rows), are weakly cuspidate, and are fairly widely spaced. 186 P. H. GREENWOOD The shape and orientation of the upper pharyngeal bones in all species are of the common cichlid type, with the near-ovoid upper pharyngeal bones of each side forming a deep and shallow V, its apex directed anteriorly. No other details are available for the pharyngobranchial skeleton. The peculiar and particular arrangement of the two infraorbital bones present (the lachrymal and an elongate second bone) has been touched upon above (p. 1 8 1 ). In all species except N. dimidiatus and N. squamiceps the upper margin of the elongate second bone is virtually straight, and so does not conform closely to the posteroventral margin of the eye-ball. In outline, the lachrymal and the second bone resemble a simple drawing of a saucepan (Fig. 27). La Max Dent Fig. 27 The two infraorbital bones of Nanochromis nudiceps shown in situ. Scale bar in mms. (Same specimen as in Fig. 25). La = lachrymal; 2io = second infraorbital bone; Max = maxilla; d = dentary. The two exceptional species have a second infraorbital whose upper margin is distinctly curved and so follows closely the eye's lower contours; the 'saucepan' outline is thus less obvious. It has proved impossible to determine whether, ontogenetically, the second infraorbital is a single or a compound element. Some intraspecific variability exists in the number of laterosensory canal pores in this bone, with three being the usual condition but with only two pores- sometimes present, either uni- or bilaterally. Four openings are invariable in the lachrymal. The squamation in Nanochromis shows several unusual features. The high position of the upper lateral-line in all species, and the contiguity of somewhat more than its posterior half with the base of the dorsal fin in all but two of those, is one of the diagnostic apomorphies for the genus (see p. 1 83). Even in N. squamiceps and N. dimidiatus, the species with only the last 3-5 pored scales touching the dorsal fin base, more than the posterior two-thirds of the line is separated from the fin base by only a single scale. Nanochromis squamiceps and N. dimidiatus also differ from their congeners in having the nape fully scaled (naked or with a few, small, scattered, and deeply embedded scales in the other species), and in the chest being partly scaled, but with the ventral and ventrolateral aspects naked. The chest is completely naked in all other Nanochromis. In these species too, and unlike N. squamiceps and N. dimidiatus, the ventral face of the belly is naked from the pelvic fin insertions to the vent. The cheek in N. squamiceps is partially covered by three rows of scales, but there is an extensive naked area anteroventrally; this extends from the posteroventral margin of the lachrymal downwards to the preoperculum, and posteriorly almost to below the mid-point of the orbit. In the N. dimidiatus type specimens examined, the naked area is even larger, with only one or two scales present below the second infraorbital bone; these speci- mens, however, are now in a very poor condition so some scales could have been lost. The cheek is entirely naked in all other Nanochromis species. PELMATOCHROMINE FISHES 187 Gill-rakers in Nanochromis are basically of the group II pelmatochromine type discussed on p. 159. There are 6-9 rakers in the outer row of the first ceratobranchial. Departure from the modal group II type of raker is manifest in the very small tubercles developed on the concave upper surface of the first ceratobranchial rakers; these may appear smooth when examined at a low magnification despite the large number of tubercles present. Some, and apparently interspecific, variation occurs in the degree to which the median mem- brane between the inner and outer rows of rakers is developed on the first arch. But, as noted before, this could be a reflection of the specimens' state of preservation. Only a single row of 3-5 epibranchial rakers is present on the first arch, and these vary interspeci- fically, from short and broad to slender and elongate. As in other group II pelmatochromines, there are no microbranchiospines. The dentition of Nanochromis shows no marked departure in tooth form or pattern from that of other group II pelmatochromines (see p. 158). Because the size range of available specimens is so limited it has not proved possible to determine at what standard length the posteriorly orientated anterolateral dentary teeth first appear, or whether the definitive tooth form is preceded by one of a different kind (see pp. 144 & 1 58). In the sample of N. transvestitus examined, the smallest individual with posteriorly directed teeth in the dentary has a standard length of 24 mm, and the smallest N. parilius with such teeth is 31 mm SL. The vertebral column is composed of 12-14 abdominal, and the same number of caudal elements (excluding the fused PU1 and Ul centra), giving total counts of 25-28. In the single syntypical specimen of TV. squamiceps, and the four syntypes of N. dimidiatus radiographed, a small, comma-shaped predor sal bone is present, as it is in two of the 10 N. splendens paratypes examined. In the specimens of all other species no trace of a predorsal bone is visible in the radiographs. There apparently is no sexual dimorphism in the shape of the caudal fin margin, which is rounded in both sexes. The first branched pelvic ray of adult males is clearly the longest, although in some individuals the second ray can be almost as long as the first. Females have the second branched pelvic ray as long as, or slightly longer than the first, as is apparently the case in both sexes of N. transvestitus. (For a discussion of this character, see p. 169). There are 16-19 spinous rays in the dorsal fin and 3 in the anal. The little information gained from a partial dissection of a N. nudiceps specimen showed no outstanding features in the myology and arthrology of the jaws. No data are available for the pharyngeal and branchial musculature. Coloration. All species of Nanochromis have a marked sexual dimorphism in adult coloration (see Roberts & Stewart, 1976, and especially the colour photographs in Linke & Staeck, 1981); one species, N. transvestitus is very unusual amongst African cichlids in having the female more brightly coloured than the male (Stewart & Roberts, 1984). At least two species, N. splendens and N. dimidiatus, seem to have a Tilapia-spot on the soft dorsal fin of some individuals (Roberts & Stewart, 1976:29; and Linke & Staeck, 1981: 63, for the species respectively). Nanochromis dimidiatus females, when adult, are also interesting in that a silver blotch is developed on the flank above the genital opening. A similar silver mark is otherwise known only in Thysia (see Loiselle & Welcomme, 1972; colour photograph in Linke & Staeck, 1981:135), and in specimens identified by Voss (1980:23, 25 & 27) as Pelvicachromis subocellatus. Such a mark has not been recorded, as far as I am aware, in other African cichlids. Breeding biology. All species of Nanochromis are substrate guarders; for detailed accounts of spawning habits see Linke & Staeck (1981). Included species: Nanochromis nudiceps (Blgr) 1899; N. consortius Roberts & Stewart, 1976; N. dimidiatus (Pell.) 1904; N. minor Roberts & Stewart, 1976; N. parilius Roberts & Stewart, 1976; N. splendens Roberts & Stewart, 1976; N. squamiceps (Blgr) 1902; TV. transvestitus Stewart & Roberts, 1984. Intrageneric relationships If it be accepted that the interrupted and reduced series of infraorbital bones (see p. 1 8 1 above), and the high-set upper lateral-line with at least the last 3-5 pored scales contacting the dorsal fin base, 188 P. H. GREENWOOD are unifying synapomorphies for Nanochromis, then two groups of species can be recognised within the genus. The smaller group, comprising the species N. squamiceps and N. dimidiatus is characterised by having only the last 3-5 pored scales of the upper lateral-line contiguous with the dorsal fin base, a completely scaled belly and nape, a partially scaled chest and cheek, and by the presence of a predorsal bone (albeit reduced and comma-shaped) in all individuals. The second and larger group, contains all the remaining species. It is characterised by having the posterior half (or slightly more) of the upper lateral-line contiguous with the dorsal fin base, a naked nape, cheek and belly, and by the absence of a predorsal bone (at least in the majority of individuals, see p. 187 above). In all these features members of the group exhibit a more derived state than is shown by either N. squamiceps or N. dimidiatus. For the moment it would seem preferable to treat the two groups informally and not recognise them as subgenera. That decision, however, may well be revised when more is known about the soft anatomy, the osteology and the breeding habits of all Nanochromis species. LIMBOCHROMIS gen. nov. TYPE SPECIES: Nanochromis robertsi Thys van den Audenaerde & Loiselle, 1971 . SYNONYMY. Nanochromis (non Pellegrin, 1904): Thys & Loiselle, 1971 (the species robertsi and cavalliensis). ETYMOLOGY. From the Latin limbus, an edge, and chromis; limbus in this particular context is used more in the spirit of its common English form of 'in limbo'. This new genus represents a taxonomic defeat. It is established for two species which, although clearly members of the group II pelmatochromine assemblage, cannot be included in any of the genera recognised here, nor can it be identified as a sister group to any of those taxa. Fig. 28 Limbochromis robertsi; holotype (SU 63 143). Scale bar in mms. Drawn by Gordon Howes. Thys & Loiselle (1971:20), with obvious misgivings, referred their two new species to Nanochromis (sensu Thys, 1968a) simply because both have a reduced number of circumpeduncu- lar scales, namely 12-14. The authors noted that both species differed from the concept of Nanochromis which they were using in having the caudal fin margin produced into a pair of streamer-like points, and that the geographical distribution of both species lay outside the range of other species referred to that genus (then treated as a subgenus) by Thys (1968a). PELMATOCHROMINE FISHES 189 As Nanochromis is now restricted to species having part of the lateral-line contiguous with the dorsal fin base, only two infraorbital bones, and a Haplochromis-type of neurocranial apophysis for the upper pharyngeal bones (see pp. 1 82-1 84), neither robertsi nor cavalliensis can be considered members of that genus. In both species the upper lateral-line is completely separated from the dorsal fin base (by at least one scale at its nearest point of approach to the fin), the infraorbital series is complete, and the neurocranial apophysis is of the plesiomorphic Tilapia-lype. Furthermore, neither species has, in either sex, the length of the second branched pelvic ray equal to, or slightly longer than the first ray. The pelvic fins and the plesiomorphic condition of the neurocranial apophysis in robertsi and cavalliensis serve to exclude the species from membership of the genus Parananochromis, a taxon which, in their general habitus, they resemble more than they do any species of Nanochromis. Again, the nature of the pelvic fins excludes robertsi and cavaliensis from Pelvicachromis (see p. 1 69), as does the reduced number of circumpeduncular scales ( 1 2 or 1 3 in cavalliensis, 1 4 in robertsi, compared with 16 in all Pelvicachromis species). The plesiomorphic pharyngeal apophysis in both Pelvicachromis and the two species under review, is, of course, valueless for establishing any possible relationship between them. It is for these various reasons that I have erected the genus Limbochromis for the species robertsi and cavalliensis. Its monophyly would seem to be attested by a reduced number of scales around the caudal peduncle occurring in species which do not share, with other pelmatochromines having that feature, the derived Haplochromis-type pharyngeal apophysis on the skull or an elongate second pelvic fin ray in females. The question then arises as to whether or not the reduced peduncular scale count in Limbochromis indicates its relationship, as the plesiomorphic sister group, to Parananochromis and Nanochromis combined. Here one is faced with a dilemma which I do not think can be settled until more information is available about the four taxa which must be involved in solving that problem, viz. Limbochromis, Pelvicachromis, Parananochromis and Nanochromis. The three latter taxa share one apparently synapomorphic feature, namely, pelvic fins in females which have the second branched ray longer than the first. Parananochromis and Nanochromis, however, have the derived Haplochromis-type pharyngeal apophysis but Pelvicachromis has the plesiomorph Tilapia-type; the two former species have a reduced (i.e. derived) number of scales around the caudal peduncle but Pelvicachromis does not. Thus, using their one synapomorphic feature (pelvic fins), Pelvicachromis could be interpreted as the plesiomorphic sister taxon of Parananochromis and Nanochromis. If that argument is accepted, then the reduced circumpeduncular scale count, the only derived character shared by Limbochromis and the Nanochromis-Parananochromis generic pair, is not a synapomorphy but a homoplastic feature. As a counter argument one could propose that the reduced scale count is the true synapomorphy (shared by Limbochromis and the Nanochromis-Parananochromis pair) and that the pelvic fin feature is the homoplastic one independently derived in Pelvicachromis. In this case, Limbochro- mis, would be the plesiomorphic sister taxon of the Nanochromis-Parananochromis generic pair. In the absence of other and differentially congruent synapomorphic features which might refute one or other of those arguments, either solution would seem equally parsimonious (see also p. 194). DIAGNOSIS. Since so little material of either species is available for study, the diagnosis is, per- force, based essentially on superficial characters. Limbochromis (Fig. 28) closely resembles Parananochromis except that: (1) The neurocranial apophysis for the upper pharyngeal bones is of the Tilapia-type, i.e. formed entirely from the parasphenoid. (2) There are 12-14 (modally 12 or 13) scales around the caudal peduncle. (3) The first branched pelvic fin ray in adult females is clearly longer than the second ray; dorsal fin with 15 or 16 spines. (4) Bilateral naked patches are present on the chest in one species. (5) The upper lateral-line at its highest point (the 8th pored scale) is separated from the dorsal fin 190 P. H. GREENWOOD base by two large and one very small scale, or by 1^ scales; posteriorly, the last 3 or 4 pored scales are separated from the fin base by 1^ scales, or by one large and one very small scale; lateral-line with 28 or 29 scales; chest scales moderately small, with a fairly abrupt size transition between them and those on the ventrolateral aspects of the flanks and belly. (6) Vertebral column with 26 vertebrae (excluding the fused PU 1 and U 1 centra) comprising 1 3 or 14 abdominal and 12 or 13 caudal elements. (7) Caudal fin in adult males of one species (L. robertsi) with two, symmetrically produced streamers; probably with two shorter projections in females. Nature of caudal fin margin in adult males unknown for the second species (L. cavalliensis}, but adult females have a short projection; caudal margin subtruncate to rounded in young. (8) No data available on live colours in L. cavalliensis, but there is pronounced sexual dimorphism in L. robertsi (see Linke & Staeck, 1981: 74-75). NOTES ON THE ANATOMY AND CERTAIN OTHER FEATURES. Because no deep dissections could be made on the few available specimens, and since no osteological preparations exist, virtually nothing can be said about the musculature and the skeleton ofLimbochromis. Those characters which could be investigated by superficial dissection, and from radiographs, do not depart greatly from the usual pelmatochromine group II condition or, indeed, from that of the pelmatochromines as a whole. The dentition and gill-rakers are typically those of group II pelmatochromines, as is the absence of microbranchiospines (see p. 1 57). Nothing is known about size-correlated changes in tooth form, or at what size the posteriorly directed anterolateral teeth in the dentary first appear. The shape of the lower pharyngeal bone (Fig. 26A) and the nature of the lower pharyngeal teeth show no outstanding features, and the shape, dentition and orientation (a broad, shallow V) of the upper pharyngeal bones are like those in other group II pelmatochromine species. One unusual feature, with respect to most other pelmatochromines, is the occurrence, in one species, of five rather than four openings to the lachrymal laterosensory canal system. The occasional occurrence of five openings in Pelvicachromis has been reported (but not confirmed) by Trewavas (1983) and five openings are usual in all but one species of Chromidotilapia (see p. 168). It is obvious that some sort of sexual dimorphism occurs in the shape of the caudal fin margin (see Thys & Loiselle, 1971), but its exact nature is still undetermined. The seemingly symmetrical development of streamer-like prolongations from both the upper and lower lobes of the caudal fin's posterior margin in one species (L. robertsi; see Fig. 28) is unusual amongst African cichlids, but seems to be repeated, albeit weakly, in the other species ofLimbochromis. More information is needed about this feature, as are detailed accounts of live colours in sexually active fishes. Judging from the radiographs, neurocranial shape in Limbochromis is like that in Paranano- chromis. A well-developed, dagger-like predorsal (supraneural) bone is present. The vertebral column has 13 or 14 (mode) abdominal, and 12 (mode) or 13 caudal centra (excluding the fused PU1 and Ul centra), with a total count of 26. As in some other group II pelmatochromines, the upper lateral-line in Limbochromis has a relatively high position on the body, with its highest point (the 8th pored scale) separated from the dorsal fin base by 1^ to 2 large and one very small scale; posteriorly, the last 3 or 4 pored scales are separated from the fin base by 1| scales or one large and one very small scale. The cheek is partly covered by 3 or 4 rows of scales, with a large naked area anteroventrally, and in one species, a narrow, naked strip below the scaled area. Chest scales are small, and have a relatively abrupt size transition with those on the ventrolateral aspects of the flanks. In L. robertsi, but not in L. cavalliensis, the scales on the medioventral aspect of the chest are somewhat larger than those occurring laterally (i.e. as in Parananochromis, see p. 179); L. cavalliensis, but not L. robertsi, has a small and bilateral naked patch on the anterolateral aspect of the chest. What information there is on the breeding biology of the two Limbochromis species (see Thys & Loiselle, 1971: 198, 199, and 202-203), does not indicate that either is a mouth-brooder. Included species. Limbochromis robertsi (Thys & Loiselle) 1971, and Limbochromis cavalliensis (Thys & Loiselle) 1971. PELMATOCHROMINE FISHES 1 9 1 The phylogenetic relationships of the pelmatochromine taxa One of the objects of this paper is to investigate whether the pelmatochromine fishes of Thys ( 1 968# , see p. 139) are a distinct and monophyletic lineage within the African Cichlidae, and if that is so, what is their relationship to other African lineages. For the moment, Thys' (19686) transfer of his subgenus Pelmatochromis ( Pelmatochromis ) to Tilapia will be ignored; his action was challenged and rejected by Trewavas (1973), and my reasons for rejecting it will be discussed later (p. 1 95). Thus the pelmatochromine genera concerned are those originally defined by Thys (19680) as subgenera of Pelmatochromis, and subsequently modified by Loiselle & Welcomme (1972), Trewavas (1973 & 1974) and in this paper. The genera involved are: Pelmatochromis Steindachner, Thysia Loiselle & Welcomme, Chromidotilapia Blgr, Pelvicachromis Thys, Parananochromis gen. nov., Nanochromis Pellegrin, and Limbochromis gen. nov. For the purposes of this discussion, a monophyletic lineage is recognised by all its constituent taxa uniquely sharing one or more derived characters which would suggest their having stemmed from a recent common ancestor with that or those features. Thus, although all pelmatochromines have at least two supposedly derived features, cycloid scales and unicuspid oral teeth when adult (see p. 143), these are shared with a large number of other taxa: I would interpret that level of synapomorphy as possibly indicative of a remote rather than a recent shared ancestry, if it is not an example of homoplasy. As indicators of recent common ancestry I have chosen those synapomorphic features which are shared by fewer species, using the extent of their occurrence amongst different taxa to build up various hierarchical levels of relationship. For example, a well-developed, visor-like hanging pad developed from the pharyngeal roof (p. 157) is a derived feature shared by several pelmatochro- mines, and is congruent with four other apomorphic features (see pp. 157-160). In brief, they involve the absence of microbranchiospines, the morphology of the oral teeth, and the morphology of the gill-rakers on the first gill-arch. The apomorphic status of all these characters is based on outgroup comparisons within the family, the suborder Labroidei and other percoid fishes as well. A sixth character may also indicate a common ancestry for this group, but it occurs mosaically and its derived status has still to be ascertained with certainty. This feature is the presence of a well- defined silver spot or blotch situated immediately above the ano-genital region of sexually active female fishes. It has been recorded in Thysia, one Pelvicachromis species, and in Nanochromis dimidiatus (see respective generic descriptions for references). Unfortunately, live colours are known for relatively few cichlids, thus making it impossible to determine, with reasonable certainty, whether this seemingly derived characteristic is indeed an apomorphy. The five principal synapomorphies noted above occur together in Thysia, Chromidotilapia, Pelvicachromis, Parananochromis, Nanochromis and Limbochromis, and are taken to indicate the monophyletic status of that assemblage, the individual genera of which are themselves defined on the basis of each having one or more of its own autapomorphic characters. The two remaining pelmatochromines, Pelmatochromis and the monotypic Pterochromis, share none of the five features discussed above; the relationship of Pterochromis can be dealt with rapidly, but regrettably, rather unsatisfactorily. Pterochromis shares no derived features with either Pelmatochromis or, as noted above, with the Thysia-Limbochromis complex, except of course for those suggesting relationship at a much more inclusive level than we are concerned with here. The unicuspid teeth of Pterochromis, relatively stout and with conical crowns, are of a general- ized type amongst unicuspids, and are quite unlike those in the Thysia-Limbochromis group. Its long, slender and fine gill-rakers on the first arch are a derived feature, but are unique amongst the taxa under discussion (including Pelmatochromis). Similar rakers do occur within the tilapiine cichlids (sensu Trewavas, 1983) but Pterochromis does not share a single derived feature with any tilapiine taxon. The gill-rakers, therefore can be considered an autapomorphic feature for Pterochromis. Likewise, the elongate lower jaw, long premaxillary ascending processes, and the steeply inclined 192 P. H. GREENWOOD mouth of Pterochromis must be considered autapomorphies since none is congruent with other derived features occurring in taxa sharing those particular oral specialisations. Thys (1968a) and Trewavas (1973) suggest that Pterochromis is a specialised form derived from Pelmatochromis ocellifer, but neither author discussed that idea in any detail; the proposed relationship would seem to be based on the shared plesiomorphic characters of the two taxa. Certainly no unique synapomorphies are shared by them, thus making it impossible to establish such a phylogenetic history. The inadequate study material of Pterochromis has severely restricted investigations of its anatomy and osteology, and almost nothing is known about the biology of the species. Thus, for the moment, and based only on the superficial features discussed here and on page 155,1 can suggest no testable hypothesis of relationship for the genus, and would exclude it from even an informal association with the other pelmatochromine taxa. Pelmatochromis provides a rather different sort of problem. Like Pterochromis it shows few derived features, has well-developed microbranchiospines, and lacks a visor-like pharyngeal hanging pad, thus excluding it from actual membership of the Thysia-Limbochromis group (see p. 157). Indeed, there seems to be not a single unequivocal autapomorphy on which to define the genus (see p. 146). Its status as a discrete taxonomic entity is therefore based simply on its having a suite of characters, both apo- and plesiomorphic which, taken in combination, distinguish it from other and superficially similar taxa. Pelmatochromis, like Pterochromis, has unicuspid outer row jaw teeth, and as in the latter genus, their cusp form is of a simple conical type, differing from the compressed-conical cusps in members of the Thysia-Limbochromis group (see p. 157); also, unlike members of that group, Pelmatochromis has no posteriorly directed teeth situated anterolaterally in the lower jaw. The outer row gill-rakers on the first ceratobranchial of Pelmatochromis species are very unlike those in Pterochromis and, indeed, those in the majority of African cichlids. Their closest morpho- logical counterparts are to be found in members of the Thysia-Limbochromis lineage. They could, in fact, be described as somewhat simplified versions of the rakers in that lineage (see p. 146). Gill-rakers, then, are the sole derived features in Pelmatochromis which might indicate its relationship. With only that character available I would, at best, suggest an informal and very tentative status for Pelmatochromis as the putative and plesiomorphic sister group of the Thysia- Limbochromis lineage. Certainly, and unlike Pterochromis, there are no apomorphic features Fig. 29 Diagram showing the distribution of various derived features amongst the chromidotilapiine cichlids (see pages 194-195); this figure can be used as a key to the genera. (1) Visor-like hanging pad in the pharynx (p. 157). (2) Outer teeth in both jaws unicuspid, with the labial surface of the crown flattened and the lingual aspect convex, resulting in the crown having well-defined and acute margins (p. 1 58 & Fig. 8). (3) A few teeth situated anterolaterally in the lower jaw so orientated that the crowns point posteriorly and not buccally (Fig. 9). (4) No microbranchiospines. (5) Ceratobranchial gill- rakers in the outer row of the first gill-arch pachydermatous, transversely aligned, with a tuberculate and concave upper surface and a protracted distal tip (p. 159 & Fig. 1 1). (6) Two large scales and one very small scale, or fewer scales (1| the modal number) between the dorsal fin base and the highest point of the upper lateral line (the 8th or 9th pored scale). (7) Five openings to the laterosensory canal system in the lachrymal (1 st infraorbital) bone (but one species in each genus with only four openings). (8) 12-14 scales around the caudal peduncle. (9) Mouth brooders. ( 1 0) The 2nd (sometimes the 2nd and 3rd) branched pelvic fin ray of females longer than the 1st branched ray, giving the distal tip of the fin a rounded rather than a pointed appearance. (11) Small to very small scales on the chest (which is entirely naked in some species, and has naked areas in a few others). (12) Neurocranial apophysis for the upper pharyngeal bones formed by the parasphenoid and basioccipital bones (i.e. of the Haplochromis-type), see p. 1 78. (13) Twelve scales around the caudal peduncle. (14) Only two bones in the infraorbital series (the lachrymal and one other bone articulating with it; see p. 186). (15) At least the last 3 or 4 pored scales of the upper lateral line contiguous with the fin base in most species. (16) Predorsal (supraneural) bone absent in many species, small and reduced in the others; see p. 187. (17) Supraoccipital crest low. PELMATOCHROMINE FISHES 193 •14-17 '12-13 '11 .10 8 •1-5 shown by Pelmatochromis which would preclude its representing a morphotype from which the less derived members of the Thysia-Limbochromis line could have evolved. In brief, out of the original, if somewhat informally recognised pelmatochromine group of Thys (1968fl), one monophyletic assemblage can be recognised on the basis of various synapomorphies. It comprises the genera Thysia, Chromidotilapia, Pelvicachromis, Limbochromis , Parananochromis and Nanochromis. As an informal epithet for that lineage, I propose the term 'chromidotilapiines'. The two remaining genera, Pterochromis and Pelmatochromis are currently of indeterminable phyletic affinity, although there is some suggestion that Pelmatochromis might be the sister group of the chromidotilapiines (see above). 194 P. H. GREENWOOD The intrarelationships of the chromidotilapiines A tentative intragroup phylogeny for the chromidotilapiines is set out in Fig. 29. The first dichotomy is based on the number of scales between the highest pored scale in the upper lateral-line (usually the 8th or 9th) and the base of the dorsal fin. Thysia, having the highest count (2| or 3 scales) and sharing no other derived features with the other species, except the group synapomorphies discussed earlier, is taken to be the plesiomorphic sister taxon to all the others combined. The derived characteristic of these other taxa lies in their having the dorsal fin base separated from the high-point of the lateral-line by no more than two large and one very small scale, or, as is modal for the group, one and a half or one scale. The first division of this group is in the form of an unresolved polychotomy caused by Limbochromis (see discussion on p. 189). That genus shares some derived features with the more apomorphic members, but its apomorphies are combined with certain plesiomorphic features which are represented apomorphically in the other taxa. Pelvicachromis occupies, as it were a 'mirror-image' position with respect to the same characters. In addition to the possible relationships discussed on p. 1 89, a fully resolved cladogram could be produced if two other assumptions were made. First, that the apparent plesiomorphy of the high circumpeduncular scale count in Pelvicachromis actually represented a secondary, that is derived, return to the plesiomorphic condition from the low counts in Limbochromis, Parananochromis and Nanochromis. Second, that the derived number of laterosensory pores (5 cf. 4) in certain Chromidotilapia and Limbochromis species represents a homoplastic rather than a synapomorphic feature. Such a cladogram, based on those assumptions and employing the other characters used in Figure 29, would make Limbochromis the plesiomorphic sister group of Pelvicachromis, Parananochromis and Nanochromis combined, and Pelvicachromis the plesiomorphic sister taxon of Parananochromis and Nanochromis. However, since the two assumptions are untestable at present, this solution is no more satisfac- tory than that discussed before (p. 189). Thus, for the moment, a polychotomous arrangement would seem to be the only acceptable one. Within the polychotomy, Chromidotilapia and Limbochromis have retained the plesiomorphic condition of the pelvic fin in females, with the first branched ray longer than the second, as it is in males also. Chromidotilapia has slightly more scales (2 plus one very small scale) above the high- point of the lateral-line and the dorsal fin base than does one species of Limbochromis (the other has the same number as Chromidotilapia). Chromidotilapia also retains the plesiomorphic feature of large chest scales (small, sometimes with naked areas in the other taxa). No data are available on the breeding habits of Limbochromis, but Chromidotilapia, uniquely amongst the chromidotila- piines, is a mouth brooder. Another derived feature in that genus is the presence of five latero- sensory canal openings in the lachrymal (but one species retains the plesiomorphic number of four); however, one of the two Limbochromis species also has five openings. Both Limbochromis species show the derived condition of a reduced scale count around the caudal peduncle, a feature not shared with either Pelvicachromis or Chromidotilapia, but one shared with two members of the fourth branch to the polychotomy, namely Nanochromis and Parananochromis. Both Chromidotilapia and Limbochromis, like Pelvicachromis, have the plesiomorphic Tilapia-type pharyngeal apophysis on the skull base. Pelvicachromis together with the Parananochromis-Nanochromis generic pair, are characterised by having the derived pelvic fin condition where, in females, the second or second and third branched rays are longer than the first ray. Other synapomorphies shared by the three taxa (and approached by Limbochromis) are their small to very small chest scales abruptly size-demarcated from the scales on the ventrolateral aspects of the flanks, and a further reduction in the number of scales (1 j, rarely 2) between the high-point of the upper lateral-line and the dorsal fin base; again a characteristic found in one Limbochromis species. A major dichotomy in this assemblage is that between Pelvicachromis with its plesiomorphic Tilapia-type pharyngeal apophysis, and Parananochromis and Nanochromis which have the PELMATOCHROMINE FISHES 195 derived Haplochromis-type. Both the latter genera are further linked by sharing a reduced number of scales around the caudal peduncle, 1 2 as compared with 1 6 in Pelvicachromis. The dichotomy between Parananochromis and Nanochromis is based on three autapomorphic features of the latter taxon, viz. only two bones in the infraorbital series (see p. 186), at least the posterior quarter and usually the posterior half of the upper lateral-line contiguous with the dorsal fin base, and the predorsal bone either absent or, in a few species, very greatly reduced in size. Furthermore, in all but two Nanochromis species the scales on the chest are also greatly reduced in size, often with large naked areas, and in some species are completely absent (see p. 186). In most respects this hypothesis of relationships within the chromidotilapiines is both inad- equate and unsatisfactory. It can only be considered as a target for further attempts based on more detailed anatomical studies and the use of other and non-morphological characteristics. From an evolutionary point of view the various chromidotilapiine genera, from Thysia at one extreme to Nanochromis at the other, show a clear cut reductional trend in a number of characters. These involve the overall body shape (from deep and barely compressed to slender and elongate), a decrease in the number of scales between the lateral-line and the dorsal fin base, a reduction in the number of scales around the caudal peduncle, and a reduction in the size of the scales on the chest (leading to partial or complete suppression in some taxa). The contrast between the end-points of these trends, as represented by Thysia and Nanochromis, is very marked, yet within the assemblage the trend is certainly a very gradual one with the various stages represented by species of different genera. That phenomenon provides an interesting parallel with a similar pattern of change, but one involving different characters, amongst the endemic haplochromine species of Lake Victoria, and for which the term cladistic gradualism was coined (Greenwood, 1981; 19846). Intrafamilial relationships of the chromidotilapiines, Pelmatochromis and Pterochromis The phyletic integrity of the chromidotilapiines seems well-established on the basis of various synapomorphic characters (see above). Two of those characters, the absence of microbranchio- spines and the presence of a visor-like hanging pad in the pharynx, occur in other African cichlids. The phyletic significance of that fact cannot be properly evaluated in the present state of our knowledge about suprageneric affinities in the Cichlidae. What information we do possess cer- tainly does not suggest a recent common ancestry for the taxa having those features. Rather it would suggest their repeated evolution in different lineages (see Greenwood, 1983). Earlier suggestions that, through Thysia, the chromidotilapiines could be related to the Hemichromis-Anomalochromis lineage (see Loiselle & Welcomme, 1972) cannot be substantiated on the basis of uniquely shared apomorphies (see Greenwood, 1985« & b). The Hemichromis- Anomalochromis group is another of those cichlid lineages easily defined on the basis of its autapo- morphic characters, but otherwise only showing a suite of widely occurring and plesiomorphic features which give no indication of its nearest relatives. Outside the chromidotilapiines, Pelmatochromis provides a particularly intriguing phylogenetic puzzle. Although its supposed close relationship to the chromidotilapiines (Thys, 1968a) cannot be satisfactorily substantiated (see p. 192), its generalised oral dentition and the derived form of its gill-rakers would seem to favour the possibility of it being the plesiomorph sister taxon to that lineage (see above, p. 193). In complete contrast to that possible relationship is the other suggestion put forward by Thys (1968fl & b), that Pelmatochromis is closely related to, or even a member of the genus Tilapia. Thys (19680) expressed the view that Pelmatochromis nigrofasciatus (which he mistakenly identified as P. ocellifer; see Trewavas, 1973) and Pterochromis congicus (then included in Pelmatochromis) seemed to bridge the morphological gap between Pelmatochromis and Tilapia. He also included in that bridging group Pelmatochromis ruweti, a species now included, and properly so, in Tilapia (see Thys, 1968«). As a member of that genus, T. ruwetiis excluded from the discussion which follows. Reasons for considering Pelmatochromis nigrofasciatus and Pterochromis congicus being close to, or actually members of Tilapia are, according to Thys, the common possession of a Tilapia-spot 196 P. H. GREENWOOD on the dorsal fin, the presence of microbranchiospines, lanceolate and not tuberculate gill-rakers, pointed pelvic fins in both sexes, 16 scales around the caudal peduncle, no hanging pad in the pharynx and '. . . all stages (with regards to mouth shape and dentition) of a gradual transition to Tilapia sparrmanii, type species of the genus Tilapia . . .' (Thys, 1968#: 369). Of the characters he lists, all but the dentition and gill-rakers in both Pelmatochromis and Pterochromis, and the oral ones in Pterochromis, are plesiomorphic features and thus of no value in establishing relationships. Jaw morphology in all three species of Pelmatochromis is very close to that in Tilapia sparrmanii (and other Tilapia species) and therefore cannot be considered as bridging the gap between the Tilapia condition and the far more extreme, predator-like one characterizing Pterochromis (see p. 155). Nor can I agree with Thys' comment on the gill-rakers as applied to Pelmatochromis. As has been frequently mentioned in this paper, the gill-rakers in that genus closely resemble those in the chromidotilapiines, and are quite unlike the relatively simple, near- vertically orientated structures in T. sparrmanii and indeed of other Tilapia species (the concept of Tilapia used here being that of Trewavas, 1983). Admittedly there is some resemblance between the gill-rakers of Pterochromis congicus and those of certain tilapiine species, but there are no other features to suggest either a close relationship between that species and any Tilapia or any Pelmatochromis species (see above, pp. 155-6). Thus, the only characters which might be of relevance are the dental ones. Since the teeth of Pterochromis are all unicuspid and conical, and bear no resemblance to those occurring as the principal dental elements in any Tilapia species (see Trewavas, 1973: 12), Pterochromis can be dismissed immediately from this aspect of the discussion. Trewavas (1973) deals at length with the presumed resemblances between Pelmatochromis and Tilapia teeth. From her survey she concluded that 'The only species of Pelmatochromis showing any evidence of derivation from an ancestor with a Tilapia-\ike dentition is P. nigrofasciatus. The fact that the teeth are notched only at such an early stage that they can hardly be functional as such suggests that they are a vestigial feature derived from an ancestor that had indeed a Tilapia-\ike dentition' (op. cit.: 19). Broadly speaking, I would agree with Trewavas' views on the teeth being a vestigial feature, especially since there are traces of a bicuspid juvenile dentition in another Pelmatochromis species (see p. 145). But even though, in the present state of our knowledge, P. nigrofasciatus has the most clearly bicuspid juvenile teeth, I would not consider them to be strictly Tilapia-like in detail. Rather, I would consider that they share with those of Tilapia only the feature of a bicuspid crown. The shared possession of teeth with bicuspid crowns, however, is too vague a character to use critically when attempting to unravel phyletic relationships, the more so since simple bicuspid crowns would seem to be a primitive adult condition amongst African cichlids (Greenwood, 1979; 1980), and there are many and derived variants on the bicuspid theme. In short, none of the features listed by Thys (1968a) would support either a conclusion that Pelmatochromis and Pterochromis be treated as members of the genus Tilapia, or that the two taxa are closely related to Tilapia and might be derived from a Tilapia busumana or T. buttikoferi-like ancestor as was suggested by Thys (19680: 377). Thys' suggestions led me to check carefully on the published diagnostic features for Tilapia and to examine in some detail specimens of T. busumana, T. sparrmanii, T. ruweti and T. zilli. The object of that study was not only to see if there are synapomorphic features shared by Tilapia and Pelmatochromis, but to find autapomorphic features which would define Tilapia and the group tilapiini as a whole (sensu Trewavas, 1983). Neither search proved successful. Trewavas' (1983) detailed account of Tilapia in relation to its sister taxa Oreochromis, Sarotherodon and Danakilia provided only one synapomorphic feature for those genera, namely a long gut. The other derived features apparent from her account are autapomorphies for the various genera other than Tilapia, which seemingly has no defining autapomorphies, at least in its morpho- logical characteristics. My studies on the cranial and pharyngobranchial anatomy of Tilapia also failed to produce a single autapomorphy. One feature not previously noted, however, would seem to provide an additional synapomorphy for the tilapiines as a whole. Namely, the shape, in lateral profile, of those pharyngeal teeth situated PELMATOCHROMINE FISHES 197 Fig. 30 Two 'kukri' shaped lower pharyngeal teeth from the anterior dental field in Oreochromis variabilis. Magnification x 100(BMNH 1911.10.28:14). in the anterior dental field of the lower pharyngeal bone. In the majority of cichlids, both African and Neotropical, these teeth are simple unicuspids or are clearly bicuspid. In either type, the tip of the cusp lies in the vertical passing through the body of the tooth or but slightly behind it. The tilapiines, however, have anterior field teeth with a very distinctive cusp form; the upper part of the tooth is angled posteriorly, and the vertical through its tip lies well behind the body of the tooth. When viewed laterally such teeth have an outline recalling the blade of a 'kukri', the Nepalese knife much favoured as a weapon by Gurkha soldiers (see Fig. 30). I have checked several species in each of Trewavas' (1983) tilapiine genera Tilapia, Oreochromis, Sarotherodon, and the monotypic Danakilia and Iranocichla, and find that 'kukri' teeth are present in all. Interestingly, none was recognizable in the endemic tilapiine genera of the Cameroonian lake Barombi Mbu (Myaka, 1 98 P. H. GREENWOOD Konia, Stomatepia and Pungu), while in the Levantine Tristramella, teeth closely approaching the 'kukri' type are present, but apparently restricted to the posterolateral dental fields. 'Kukri' teeth do not occur in any of the pelmatochromine genera. 'Kukri'-shaped teeth, or teeth closely similar to them, also occur in at least some Lake Malawi species of the genus Pseudotropheus, for example, P. tropheus, P. auritus and P. livings toni (not all species of this large genus were examined). That record, coupled with the general absence of 'kukri' teeth in other African and Neotropical taxa suggests that such teeth have evolved at least twice within the family since there is no evidence to suggest that Pseudotropheus and the tilapiines are closely related. Another apomorphic feature in the pharyngeal jaws of all but one species of the Tilapia, Oreochromis, Sarotherodon, Danakilia, Iranocichla group, concerns the shape and spatial relation- ships of the upper pharyngeal bone (fused 3rd and 4th pharyngobranchials) of each side. Whereas in Pelmatochromis, Pterochromis and all the chromidotilapiines, as well as in the majority of African and Neotropical taxa, the two bones form a broad and shallow V (see p. 149, and Fig. 4A), in the tilapiine genera the bones are closely apposed along their medial faces, and together have a distinctly cardiform outline (Fig. 4B), one enhanced by their relatively broader and less ovoid toothed surfaces. The exceptional species is Tilapia busumana. Here the bones are narrower and form the generalized V-shape. Tilapia busumana is, of course, the species which Thys (1968a:377) indicated as being close to the stem of the pelmatochromine assemblage. Arguments for not accepting a close relationship between Pelmatochromis (or any pelmatochromine) and Tilapia have been given above, and are not affected by the two taxa sharing this plesiomorphic feature. That it is a plesiomorphy lost in, apparently, all other tilapiines, would, however, seem to indicate the possibility of T. busumana being a primitive tilapiine. A cardiform arrangement of the upper pharyngeal bones is by no means confined to the tilapiine genera. It has been found in Labrochromis and Astatoreochromis from Lake Victoria, in Tylochromis and Cardiopharynx of Lake Tanganyika, in Cyathochromis and Petrotilapia from Lake Malawi, and in at least four Neotropical species as well. A condition approaching the cardiform one more closely than the V-shaped arrangement occurs in the supposedly primitive Asiatic Etroplus oligacanthus and in the Malagasian Ptychochromis. Arrangements intermediate between the cardiform and V-shaped types, but rather closer to the former, are found in Simochromis and Xenotilapia from Lake Tanganyika and Paratilapia polleni from Madagascar, and a reversed cardiform arrangement (with the 'auricular' part of the 'heart' directed anteriorly rather than towards the oesophagus) occurs in the Malagasian Paretroplus dami. With such a range of variability, and of distantly related taxa in which the cardiform arrange- ment occurs, it would seem to be a feature which has evolved several times, and independently, within the Cichlidae, including some of its members generally accepted as having a plesiomorphic level of organization (e.g. Tylochromis and the etroplines of Asia and Madagascar). That it is an apomorphic feature seems probable from its absence in other labroids and in the percoid outgroups examined. In the majority of species with a cardiform arrangement, the teeth of the upper pharyngeal bones are either enlarged and molariform, or fine, numerous and densely packed, a situation which may be correlated with the bones themselves having a larger surface area relative to the subovoid bones associated with the typical V-shaped arrangement. The absence of 'kukri'-shaped teeth on the lower pharyngeal bones, and the persistence of the plesiomorphic V arrangement of the upper pharyngeal elements in all pelmatochromine taxa, provide yet other reasons for refuting the idea that Pelmatochromis and Pterochromis are members of the tribe Tilapiini sensu Trewavas (1983). That assemblage besides Pelmatochromis and Pterochromis, contains the genera Tilapia, Oreochromis, Sarotherodon, Danakilia, Iranocichla, Tristramella, the endemic genera of Barombi Mbu, and possibly Steatocranus and Gobiochromis. Its only unifying group-character cited is the Tilapia-type apophysis, which as Trewavas (op. cit. and 1986) recognises, is plesiomorphic for the cichlids. As such it cannot be used to establish the monophyly of the Tilapiini (see also Greenwood, 1978, and 1987). Establishing the monophyly of Trewavas' Tilapiini, even after the removal of Pelmatochromis and Pterochromis, is a task that remains to be done. The use of upper pharyngeal bone arrange- PELMATOCHROMINE FISHES 199 ment, and the presence of 'kukri'-shaped teeth may prove of some value in that task, but neither feature can yet be considered a fully corroborated synapomorphy for the group. Of the two taxa which can be eliminated from the Tilapiini, Pelmatochromis may be the plesiomorph sister taxon to all the chromidotilapiines (see p. 192), but the relationships of Pterochromis are at present unknown (see p. 192). Its principal diagnostic phenotypic characters are all autapomorphic ones that could be derived, by differential growth, from a Pelmatochromis- like ancestor which did not have the derived gill-raker morphology of the extant Pelmatochromis species. Other changes from such a presumed ancestor would involve slight modifications to the detention and dental patterns. All that is speculation, and a reflection of the fact that such a proto-Pelmatochromis could well provide a morphotype from which a number of extant lineages, including Tilapia and its close relatives, evolved. If that is so, their unifying synapomorphies have still to be identified. In conclusion, what can be said about the intrafamilial relationships of chromidotilapiines, Pelmatochromis and Pterochromis! In a positive sense, the answer must be 'Nothing'. Previously suggested relationships can be refuted, and one is left with a number of apparently monophyletic lineages whose relationships at higher phyletic levels are presently indeterminable. The distribution of Pterochromis, Pelmatochromis and the chromidotilapiines The chromidotilapiines as a whole (Thysia, Chromidotilapia, Pelvicachromis, Limbochromis , Parananochromis and Nanochromis) have a fairly extensive latitudinal range from about 10°N to 5°S. They are, however, with the exception of Nanochromis and one Chromidotilapia species, confined to a relatively narrow strip, some 200-300 miles wide, extending along the western margin of Africa. This strip lies mainly within the lowland forest zone, but in places it extends into the areas of moist woodlands and savannah. The range of Nanochromis, however, extends further eastwards into the central regions of the Zaire river and its tributaries, although the greater number of its species is found in the fast-flowing waters of the river's lower reaches. The exceptional Chromidotilapia species, C. schoutedeni occurs outside the marginal zone and is found in the lower parts of the Lualaba river. Pelmatochromis has a wide but scattered distribution which includes the central Zaire area, Senegal, Guinea, Liberia and Sierra Leone, while Pterochromis is apparently restricted to Central Zaire in the region of Kisingani. For detailed distributions of the various species see Thys, 1968#; Thys & Loiselle, 1971; Loiselle & Welcomme, 1972; Trewavas, 1973 and 1974; Roberts & Stewart, 1976; Linke & Staeck, 1981; Stewart & Roberts, 1984. There is remarkably little distributional overlap between the chromidotilapiines, Pelmato- chromis and Pterochromis on the one hand, and any of the non-tilapiine fluviatile species of eastern and middle Africa on the other. The non-tilapiine species in question are represented by members of the haplochromine genera Astatotilapia, Ctenochromis , Thoracochromis, Astatoreochromis and Orthochromis (see Greenwood, 1979). In geographical terms, and to a certain extent in their trophic specializations as well, the chromidotilapiines, plus Pelmatochromis and Pterochromis, together with Hemichromis and Anomalochromis (see Greenwood, \9S5a & b) seem to be the far western equivalents, and ecologi- cal counterparts, of the eastern haplochromines mentioned above. It is, of course, difficult to quantify degrees of morphological differentiation, but the two geographical entities seem to have evolved and differentiated to about the same extent. A noticeable difference, however, is the absence of any far western species in which the pharyngeal mill is hypertrophied. Indeed, the morphological details involved in the differentiation of the two groups are generally quite differ- ent. For example, no eastern species has evolved a dentition like that in the chromidotilapiines, none has a visor-like hanging pad in the pharynx, none has gill-rakers with the complex form of those in the chromidotilapiines, and all have retained microbranchiospines. However, in superfi- cial features the eastern and western rheophilic taxa, Orthochromis and Nanochromis respectively, are, perhaps not unexpectedly, remarkably similar. 200 P. H. GREENWOOD Summary and conclusions The pelmatochromine cichlids (sensu Thys, 1968a) of western Africa can be divided into a mono- phyletic lineage (informally categorized as the chromidotilapiines) comprising the genera Thysia, Chromidotilapia, Pelvicachromis, Nanochromis, Parananochromis and Limbochromis (the two latter genera described in this paper), and two other monophyletic taxa, Pelmatochromis and Pterochromis. The intrarelationships of the chromidotilapiines cannot be satisfactorily resolved; it is suggested that Thysia is the plesiomorphic sister taxon to the other five genera combined, and that within this combination Parananochromis and Nanochromis are sister genera, and are the most derived members of the chromidotilapiines. There are indications that two of the chromidotilapiine genera, although monophyletic assem- blages themselves, should be subdivided. In Nanochromis, for example, one group of species has the posterior half of the upper lateral-line in contact with the dorsal fin base, whilst in the other, only the last four or five scales of the lateral-line are contiguous with the fin base. In Pelvicachromis, P. humilis shows certain derived features which suggest that it may represent an off-shoot of the main species assemblage. Both these problems may be clarified when more material is available, and further anatomical and osteological studies can be made. Ethological and other aspects of the species' natural history may also prove of great value in unravelling the still outstanding problems. Although Pelmatochromis apparently shares one derived character, the nature of its gill-rakers, with the chromidotilapiines, no further and congruent synapomorphies could be discovered to support its possible sister-group relationship with those species. Equally, the absence of other derived features in Pelmatochromis does not give any indication of other taxa to which it might be related. Pterochromis exhibits several autapomorphic features but, like Pelmatochromis, it seems to show no derived features positively indicative of its relationships within the family. The suggestion (Thys, 1968a & b) that Pelmatochromis might be closely related to, or even a member of the genus Tilapia is reexamined and found to be untenable both for the reasons given by Trewavas (1973) and on the basis of other characters as well. Its suggested membership of a tribe containing Tilapia, Oreochromis, Sarotherodon and several other genera (see Trewavas, 1983) is also rejected. In that connection, the phylogenetic basis (as opposed to one based on overall similarity) of the Tilapiini is reconsidered. Two possibly apomorphic characters, for at least some members of the tribe, are identified; neither is present in any of the pelmatochromine taxa, thus providing further evidence against their possible close relationship with Tilapia. Neither the chromidotilapiines nor Pelmatochromis and Pterochromis are closely related to any of the fluviatile haplochromine lineages from eastern Africa, and there is very little geographical overlap between the two groups. The essentially far western distribution of the chromidotilapiines, Pelmatochromis and Pterochromis contrasts strongly with the wider geographical distribution of the eastern haplochromine lineages (see above, p. 199, and Greenwood, 1979). Like the various haplochromine lineages of Lake Victoria (see Greenwood, 1981; 1984), the chromidotilapiines exhibit a gradual trend in the evolution of certain morphological characters, with each stage in the morphocline represented by one or several species. Amongst the Lake Victoria haplochromines this display of cladistic gradualism principally involves trophic specializ- ation. In the chromidotilapiines it involves traits associated with an increasingly rheophilic way of life. Little is known about the feeding habits of the chromidotilapiines but judging from the oral and pharyngeal dentition, and from associated structures, there is probably little intragroup trophic specialization and differentiation. Acknowledgements I am extremely grateful to my colleagues, past and present, in the Fish Section who, at some time or other, have been involved with me in arguments and discussions about the pelmatochromine cichlids, or have helped with various curatorial problems. PELMATOCHROMINE FISHES 201 In particular, I thank most warmly Gordon Howes for the meticulous and painstaking art work he has produced, and for the unstinting help he has given me in so many other ways. For their expertise in producing radiographs, I would warmly thank Bernice Brewster and Mandy Holloway without whose help many anatomical details could not have been studied. I am very grateful as well to John Spratt of the Museum's E.M. Unit who produced the scanning E.M. photographs. For the loan of study material I am indebted to Dr Guy Teugels of the Museum National d'Histoire Naturelle, Paris; Prof. D. F. E. Thys van den Audenaerde of the Musee Royal d'Afrique Centrale, Tervuren; and to Dr William Eschmeyer, California Academy of Sciences. At the Museum of Comparative Zoology, Harvard University, my best thanks go to Prof. Karel Liem and Mr Karsten Hartel for not only giving me access to the material in their care, but for providing several specimens for dissection. Also at the MCZ, I must thank Dr Melanie Stiassny who spent much time discussing cichlid phylogeny with me and was so generous in sharing her unpublished results and ideas. Finally, to Prof. M. N. Bruton of the J. L. B. Smith Institute of Ichthyology, and to Prof. Tom Hecht of the Department of Ichthyology and Fisheries Science, Rhodes University, both of Grahamstown, South Africa, go my warmest thanks for their providing me both with study material and with such congenial working conditions on several visits. References Anker, G. Ch. 1978. The morphology of the head muscles of a generalized Haplochromis species: H. elegans Trewavas, 1933 (Pisces, Cichlidae), Netherlands Journal of Zoology, 28: 234-271. Balon, E. K. 1977. Early ontogeny of Labeotropheus Ahl, 1927 (Mbuna, Cichlidae, Lake Malawi), with a discussion on advanced protective styles in fish reproduction and development. Environmental Biology of Fishes, 2(2): 147-176. 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). Netherlands Journal of Zoology, 26: 163-265. Bell-Cross, G. 1975. A revision of certain Haplochromis species (Pisces: Cichlidae) of Central Africa. Occasional Papers of the National Museum and Monuments of Rhodesia; Series B.5 (7): 405-464. Boulenger, G. A. 1 897. Descriptions of new fishes from the Upper Congo — II. Annals and Magazine of Natural History (6), 20: 422^27. 1898. A revision of the African and Syrian fishes of the family Cichlidae. Part I. Proceedings of the Zoological Society of London, 1898: 132-152. 1899. A revision of the African and Syrian fishes of the family Cichlidae. Part II. Proceedings of the Zoological Society of London, 1899: 98-143. 1915. Catalogue of the freshwater fishes of Africa. 3. London. British Museum (Natural History). 1916. Catalogue of the freshwater fishes of Africa. 4. London, British Museum (Natural History). Cichocki, F. 1976. Cladistic history of cichlid fishes and reproductive strategies of the American genera. Ph.D. thesis, University of Michigan, Ann Arbor. Colombe, J. & Allgayer, R. 1985. Description de Variabilochromis, Neolamprologus, et Paleolamprologus genres nouveaux du Lac Tanganyika, avec redescription des genres Lamprologus Schilthuis, 1891 et Lepidolamprologus Pellegrin, 1904 (Pisces, Teleostei, Cichlidae). Revue Francois des Cichlidophiles, No. 49: 9-28. Fishelson, L. 1966. Untersuchungen zur vergleichenden Entwicklungsgeschichte der Gattung Tilapia (Cichlidae, Teleostei). Zoologische Jahrbiicher (Anatomic), 83: 571-656. Greenwood, P. H. 1956. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae). Part I. Bulletin of the British Museum (Natural History), Zoology, 4: 223-244. 1974. Cichlid fishes of Lake Victoria, east Africa: the biology and evolution of a species flock. Bulletin of the British Museum (Natural History), Zoology. Supplement 6: 1-134. 1976. A review of the family Centropomidae (Pisces, Perciformes). Bulletin of the British Museum (Natural History) , Zoology, 29: 1-8 1 . 1978. A review of the pharyngeal apophysis and its significance in the classification of African cichlid fishes. Bulletin of the British Museum (Natural History), Zoology, 33: 297-323. 1979. Towards a phyletic classification of the 'genus' Haplochromis (Pisces, Cichlidae) and related taxa. Part I. Bulletin of the British Museum (Natural History), Zoology, 35: 265-322. 202 P. H. GREENWOOD — 1980. Towards a phyletic classification of the 'genus' Haplochromis (Pisces, Cichlidae) and related taxa. Part II. Bulletin of the British Museum (Natural History), Zoology, 39: 1-101. 1981 . Species flocks and explosive evolution; in Chance Change and Challenge — The Evolving Biosphere (eds P. H. Greenwood & P. L. Forey): 61-74. Cambridge University Press and British Museum (Natural History), London. — 1983. The Ophthalmochromis assemblage of cichlid fishes reconsidered. Bulletin of the British Museum (Natural History) , Zoology, 44: 249-290. 1 984cr. The haplochromine species (Teleostei, Cichlidae) of the Cunene and certain other Angolan rivers. Bulletin of the British Museum (Natural History), Zoology, 47: 187-239. — 1 9846. African cichlids and evolutionary theories; in Evolution offish species flocks (eds A. A. Echelle & I. Kornfield): 141-154. University of Maine at Orono Press. 1985a. Notes on the anatomy and phyletic relationships of Hemichromis Peters, 1858. Bulletin of the British Museum (Natural History), Zoology, 48: 131-171. 19856. The generic status and affinities of Paratilapia thomasi Blgr, 1915 (Teleostei. Cichlidae). Bulletin of the British Museum (Natural History), Zoology, 49: 257-272. 1986. The pars jugular is and the intrarelationships of cichlid fishes (Labroidei; Teleostei). Journal of Natural History, 20: 949-974. 1987. The pharyngeal apophysis on the base of the skull in cichlid fishes. A reply to Trewavas (1986). Netherlands Journal of Zoology, 36(4): 562-564. I, hike, H. & Staeck, W. 1981. Afrikanische Cichliden I. Buntbarsche aus Westafrika. Tetra Verlag, Melle. Loiselle, P. V. 1979. A revision of the genus Hemichromis Peters, 1858. Annales du Musee Royal de I'Afrique Centrale (Serie in 8vo), Zoologie, No. 228: 1-24. & Welcomme, R. L. 1972. Description of a new genus of cichlid fish from West Africa. Revue de Zoologie et de Botanique Africanes, 85: 37-58. Matthes, H. 1964. Les Poissons du Lac Tumba et de la region d'Ikela. Annales du Musee Royal de I'Afrique Centrale (Serie in 8vo), Zoologie, No. 126: 1-204. Myrberg, A. A. 1965. A descriptive analysis of the behaviour of the African cichlid fish, Pelmatochromis guentheri(Sauvage). Animal Behaviour, 13: 312-329. Oliver, M. K. 1984. Systematics of African cichlid fishes: determination of the most primitive taxon, and studies on the haplochromines of Lake Malawi (Teleostei: Cichlidae). Ph.D. thesis, Yale University. Pellegrin, J. 1904. Contribution a Tetude anatomique, biologique et taxonomique des poissons de la famille des cichlides. Memoires de la Societe Zoologique de France, 16 (1903): 41-399. Peters, H. M. & Berns, S. 1978. Uber die vorgeschichte der maulbnitenden Cichliden: II. Zwei Typen von Maulbriitern. Aquarien Magazin (Kosmos- Verlag, Stuttgart), No. 7: 324-331. Poll, M. & Thys van den Audenaerde, D. F. E. 1965. Deux Cichlidae nouveaux du Sud du Bassin du Congo. Revue de Zoologie et de Botanique Africaines, 72: 322-333. 1967. Description de Pelmatochromis schoutedeni sp.n. du Congo oriental (Pisces, Cichlidae). Revue de Zoologie et de Botanique Africaines, 75: 385-391 . Regan, C. T. 1922. The classification of the fishes of the family Cichlidae — II. On African and Syrian genera not restricted to the Great Lakes. Annals and Magazine of Natural History (9) 10: 249-264. Roberts, T. R. & Stewart, D. J. 1976. An ecological and systematic survey of fishes in the rapids of the lower Zaire or Congo River. Bulletin of the Museum of Comparative Zoology, 147: 239-317. Staeck, W. 1980. Chromidotilapia linkei n.sp. aus dem Mungo River, Kamerun. Senckenbergiana Biologica, 60(1979): 153-157. Steindachner, F. 1894. Die Fische Liberias. Notes of the Leyden Museum, 16: 1-96. Stewart, D. J. & Roberts, T. R. 1984. A new species of dwarf cichlid fish with reversed sexual dichromatism from Lac Mai-ndombe, Zaire. Copeia, 1984: 82-86. Stiassny, M. L. J. 198 la. The phyletic status of the family Cichlidae (Pisces, Perciformes): A comparative anatomical investigation. Netherlands Journal of Zoology, 31: 275-314. 19816. Phylogenetic versus convergent relationship between piscivorous cichlid fishes from Lakes Malawi and Tanganyika. Bulletin of the British Museum (Natural History), Zoology, 40: 67-101. 1982. The relationships of the neotropical genus Cichla (Perciformes, Cichlidae): a phyletic analysis including some functional considerations. Journal of Zoology, London, 197: 427-453. Thys van den Audenaerde, D. F. E. 1968o. A preliminary contribution to a systematic revision of the genus Pelmatochromis Hubrecht sensu lato (Pisces, Cichlidae). Revue de Zoologie et de Botanique Africaines, 77: 349-391. 19686. An annotated bibliography of Tilapia (Pisces, Cichlidae). Musee Royal de I'Afrique Centrale; Documentation Zoologique, No. 14: xl + 406 pp. PELMATOCHROMINE FISHES 203 & Loiselle, P. V. 1971. Description of two new small African cichlids. Revue de Zoologie et de Botanique Africaines,83: 192-206. Trewavas, E. 1973. On the cichlid fishes of the genus Pelmatochomis with a proposal of a new genus for P. congicus; on the relationship between Pelmatochromis and Tilapia and the recognition of Sarotherodon as a distinct genus. Bulletin of the British Museum (Natural History), Zoology, 25: 3-26. 1974. The freshwater fishes of rivers Mungo and Meme and lakes Kotto, Mboadong and Soden, West Cameroon. Bulletin of the British Museum (Natural History), Zoology, 26: 328-419. 1975. A new species of Nanochromis (Pisces, Cichlidae) from the Ogowe system, Gabon. Bulletin of the British Museum (Natural History), Zoology, 28: 233-235. 1983. Tilapiine fishes of the genera Sarotherodon, Oreochromis anc/Danakilia. British Museum (Natural History), London. 1 986. The pharyngeal apophysis on the base of the skull in cichlid fishes. Netherlands Journal of Zoology, 35 (1985): 7 16-7 19. Voss, J. 1977. Les livres ou patrons de coloration chez les poissons Cichlides africains. Revue Francaise d'Aquariologie, 4 erne Annee. No. 2: 33-80. 1980. Color patterns of African cichlids. T. F. H. Publications, New Jersey. Manuscript accepted for publication 1 1 February 1987 British Museum (Natural History) The birds of Mount Nimba, Liberia Peter R. Colston & Kai Curry- Lindahl For evolution and speciation of animals Mount Nimba in Liberia, Guinea and the Ivory Coast is a key area in Africa representing for biologists what the Abu Simbel site in Egypt signified for archaeologists. No less than about 200 species of animals are endemic to Mount Nimba. Yet, this mountain massif, entirely located within the rain-forest biome, is rapidly being destroyed by human exploitation. This book is the first major work on the birds of Mount Nimba and surrounding lowland rain-forests. During 20 years (1962-1982) of research at the Nimba Research Laboratory in Grassfield (Liberia), located at the foot of Mount Nimba, scientists from three continents have studied the birds. In this way Mount Nimba has become the ornithologically most thoroughly explored lowland rain-forest area of Africa. The book offers a comprehensive synthesis of information on the avifauna of Mount Nimba and its ecological setting. During the 20 years period of biological investigations at Nimba this in 1962 intact area was gradually opened up by man with far-reaching environmental consequences for the rain-forest habitats and profound effects on the birds. Therefore, the book provides not only a source of reference material on the systematics, physiology, ecology and biology of the birds of Mount Nimba and the African rain-forest, but also data on biogeography in the African context as well as conservation problems. Also behaviour and migration are discussed. At Nimba a number of migrants from Europe and/or Asia meet Afrotropical migratory and sedentary birds. Professor Kai Curry-Lindahl has served as Chairman of the Nimba Research Laboratory and Committee since its inception in 1962. Peter Colston is from the Subdepartment of Ornithology, British Museum (Natural History), Tring, and Malcolm Coe is from the Animal Ecology Research Group, Department of Zoology, Oxford. 1986, 129pp. Hardback. 0 565 00982 6 £17.50. Titles to be published in Volume 53 Puellina (Bryozoa; Cheilostomata; Cribrilinidae) from British and adjacent waters. By J. D. D. Bishop & B. C. Househam Miscellanea The genera of pelmatochromine fishes (Teleostei, Cichlidae). A phylogenetic review. By P. H. Greenwood Certain Actiniaria (Coelenterates, Anthozoa) from the Red Sea and Tropical Indo- Pacific Ocean. By K. W. England Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset Bulletin of the British Museum (Natural History) Certain Actiniaria (Cnidaria, Anthozoa) from the Red Sea and tropical Indo-Pacific Ocean K. W. England Zoology series Vol 53 No 4 17 December 1987 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.) © British Museum (Natural History), 1987 The Zoology Series is edited in the Museum's Department of Zoology Keeper of Zoology : Mr J. F. Peake Editor of Bulletin : Dr C. R. Curds Assistant Editor : Mr C. G. Ogden ISBN 0565 05036 2 ISSN 0007-1498 Zoology series Vol 53 No. 4 pp 205-292 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 17 December 1987 IL i Certain Actiniaria (Cnidaria, Anthozoa)jfrqm Red Sea and tropical Indo-Pacific Ocean! K. W. England Department of Pure and Applied Zoology, University of Reading, RG6 2AJ ISSUED Contents Synopsis 206 Type designations ............. '206 Introduction .............. 206 Methods ,207 Reappraisal of certain taxonomic characters 2iO~" Marginal spherules 210 Sphincter muscles 211 Mesenterial arrangements ........... 214 Nematocyst signatures 214 Taxonomic section 215 Suborder Nynantheae 215 Tribe Athenaria 215 Family Edwardsiidae 215 Genus Edwardsia . . . . . . . . . . .215 Edwardsia hantuensis sp.nov 221 Genus Edwardsioides . .......... 224 Genus Edwardsianthus nov. ......... 224 Edwardsianthus pudica comb. nov. ........ 224 Genus Scolanthus 229 Scolanthus armatus comb, nov 229 Tribe Thenaria . ............ 232 Subtribe Endomyaria ........... 232 Family Aliciidae 232 Genus Triactis 232 Triactis producta ........... 233 Family Actiniidae 238 Genus Anthopleura 238 Anthopleura dixoniana 240 Anthopleura handi ........... 245 Anthopleura nigrescens .......... 248 Anthopleura waridi comb. nov. ........ 250 Genus Gyractis 254 Gyract is excavata ........... 255 Genus Mesactinia nov 261 Mesactinia ganensis sp. nov. ......... 263 Genus Neocondylactis nov. ......... 263 Neocondylactis singaporensis sp. nov. ....... 265 Genus Telactinia nov 268 Telactinia citrina comb. nov. 269 Family Stichodactylidae .......... 272 Genus Antheopsis 274 Antheopsis malayensis sp. nov. ........ 274 Genus Entacmaea ........... 276 Entacmaea quadricolor .......... 277 Bull. Br. Mus. not. Hist. (Zool.) 53(4): 205-292 Issued 17 December 1987 205 206 K. W. ENGLAND Subtribe Acontiaria ............ 279 Family Hormathiidae 279 Genus Calliactis ........... 279 Calliac tis polypus ........... 279 Family Sagartiidae 281 Genus Verrillactis . . . . . . . . . . .281 Verrillactis paguri, comparison with ....... 283 Carcinactis ichikawai and C. dolosa Acknowledgements 284 References 285 Synopsis Thirteen species of sea-anemone from the tropical Red Sea and the Indo-Pacific Ocean are described in detail and notes are given on a further three. Certain taxonomic characters are discussed, marginal spherules are re-defined and the concept 'nematocyst signature' is introduced. The genus Edwardsia de Quatrefages, 1 841 , is revised. Details of two nematocyst types found in the nemathybomes are given, newly designated pterotrich and microbasic t-mastigophore. The genus Edwardsioides Danielssen, 1890, is re-established and the related genus Edwardsianthus nov. introduced. Descriptions are given of Edwardsia hantuensis sp. nov., Edwardsianthus pudica (Klunzinger, 1877) comb, nov., and Scolanthus armatus (Carlgren, 1931) comb. nov. Triactis producta Klunzinger, 1877 (Aliciidae) is redescribed. Descriptions are given of the actiniid species Anthopleura dixoniana (Haddon & Shackleton, 1893), A. handi Dunn, 1977, A. nigrescens (Verrill, 1928), and A. waridi (Carlgren, 1900). The genus Gyractis Boveri, 1893, is re-established to accommodate the species G. excavata Boveri, 1 893, and its relation to Aulactinia Verrill, 1 864, is discussed. Three new genera are proposed in the Actiniidae: Mesactinia gen. nov., Neocondylactis gen. nov. and Telactinia gen. nov.; and the included species Telactinia citrina (Haddon & Shackleton, 1893) comb, nov., Mesactinia ganensis sp. nov. and Neocondylactis singaporensis sp. nov. are described. The scope of the family Stichodactylidae is widened to include the genus Entacmaea Ehrenberg, 1834; and the genus Antheopsis Carlgren, 1900, is reinstated. New information is given on the species Entacmaea quadricolor Rueppell & Leuckart, 1828, and Antheopsis malayensis sp. nov. is described. The date of Carlgren's definition of the family Hormathiidae is discussed. A note is given on the inclusion of the conspecific Actinia decorata (Dana, 1846) with Calliactis polypus (Forsskal, 1775). The three species Carcinactis dolosa Riemann-Ziirneck, 1975, Carcinactis ichikawai Uchida, 1960, and Verrillactis paguri (Verrill, 1969a) are compared. Type designations Type material is designated of the nominal species Antheopsis malayensis sp. nov. (p. 274), Edwardsia hantuensis sp. nov. (p. 221 ), Mesactinia ganensis gen. nov., sp. nov. (p. 263) and Neocondylactis singaporensis gen. nov., sp., nov. (p. 265). Priapus polypus Forsskal, 1775, is designated type species of the genus Calliactis Verrill, 19696 (p. 279), family Hormathiidae. Type species are designated of the genera Edwardsianthus gen. nov. (p. 224) and Edwardsioides Danielssen, 1890 (p. 224), family Edwardsiidae; Mesactinia gen. nov. (p. 261), Neocondylactis gen. nov. (p. 265) and Telactinia gen. nov. (p. 268), family Actiniidae. Introduction Little taxonomic work on tropical sea-anemones has been published since Carlgren's (\95Qa,b) papers on the Actiniaria and Corallimorpharia of the Great Barrier Reef, South Queensland and New South Wales. The only substantial publication since 1 950 has been that of Dunn ( 1 98 1 ) on the clownfish sea-anemones. The opportunity was taken, during military service, to collect Actiniaria and Corallimorpharia from a variety of tropical and sub-tropical sites; in Aden and Bahrain in 1966; in Gan, Addu Atoll, Maldives, in 1970 and 1971; and in Singapore and Malaysia (Penang and Pulau Tioman) from 1969 to 1972. The collection comprised some 50 species of which 13 are ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 207 described here and notes on a further 3 included. In addition numerous locality records of anemones not treated taxonomically are included. The study was augmented by material from the British Museum (Natural History) and specimens kindly made available by others from Sri Lanka, Hawaii, the Tuamoto Isles, Hong Kong, Madagascar and Egypt. The new material has been deposited in the British Museum (Natural History). Methods The collecting sites are described according to habitat. A species list from each habitat is given. Not all the species found are described in this paper, but lists of them are included to show the wide variety that occurred in these areas and to provide faunal records. In the areas investigated four distinct habitats were studied as follows. (a) Mangrove swamp. Here Actiniaria and Corallimorpharia were absent, although Stephenson, Stephenson & Tandy ( 1 93 1 : 38) recorded some specimens from a pool near the mangroves of Low Isles in the Great Barrier Reef. Mangrove areas were examined in Singapore and around Pulau Tioman, Malaysia. (b) Intertidal mud with occasional rocks, but not associated with mangroves. In Aden, near the Causeway below Jebel Hadid, one such area was rich in actiniarians. They included Edwardsianthus pudica (Klunzinger, 1877), comb, nov., Antheopsis koseirensis (Klunzinger, 1877), Stichodactyla gigantea (Forsskal, 1775), Heteranthus verraculatis Klunzinger, 1877, found in the mud, and Aiptasia sp., Anthopleura stellula (Ehrenberg, 1834) and Haliplanella luciae (Verrill, 1898) found on small stones, pieces of rope and an old oil drum. Large numbers of hermit crabs (species not determined) were found in this area, possibly attracted by fishermen's offal, but the crabs never had associated symbiotic anemones. Other Cnidaria observed included Cassiopeia sp. (Scyphozoa), Cladonema sp. (Hydrozoa: hydroid stage), and a large cerianthid. At Singapore several similar areas occurred on the northern coast, including Pungol Point and Pasir Ris which were examined. Numerous specimens of Actinothoe sp. were found on several different species of Nassarius spp. (Gastropoda), and Neocondylactis singaporensis gen. nov. sp. nov., Scolanthus armatus (Carlgren, 1931) and a large cerianthid were observed in the mud. Anthopleura dixoniana (Haddon & Shackleton, 1893) comb. nov. was found in small holes in an outcrop of yellow sandstone at Pungol Point and on stones of a broken wall at Pasir Ris, both locations being at Mean High Tide Level. Another muddy habitat at Singapore was an area which has now been reclaimed at Bedok. From here were collected specimens of Paracondylactis sp., Sagartianthus sp. associated with the crab Diogenes diogenes, and Stichodactyla gigantea (Forsskal, 1775). On the south coast of Pasir Panjang and Buona Vista was a diverse area of mud, sand and gravel with patches of the green alga Halophila ovalis. Here was again found Actinothoe sp. on Nassarius spp., and also Edwardsia hantuensis sp. nov., Stichodactyla gigantea (one large and several small specimens with symbiotic shrimp), and a cerianthid. (c) Fringing coral reefs. Three zones were recognized: 1 . The reef flat, usually comprising coarse sand with occasional rocks and coral heads. 2. The reef edge, an area extending back from the outer edge for some ten metres. This area comprised coral debris and sand with large coral heads and occasional hard rock, especially where wave action was high. Shelter from wave action was associated with growth of many living corals and algae. 3. The reef slope, fairly steep and comprised entirely of coral debris, which continued down to depths of more than thirty metres around the islands off Singapore and to greater depths in parts of the reefs at Gan and elsewhere. In other localities it consisted of rocks covered with many species of corals and occasional actinians. On the reef flats surrounding the Singapore islands, particularly Pulau Hantu, the following species were noted in the sand: Actinodendron sp., Edwardsianthus pudica comb, nov., Metapeachia tropica (Panikkar, 1939), Scolanthus armatus comb, nov., Stichodactyla gigantea, S. kenti (Haddon 208 K. W. ENGLAND & Shackleton, 1 893), and several species of Discosomatidae. On the Cyrene Reef, however, the reef flat was covered with S. kenti. On Gan, from similar habitats on the reefs inside the lagoon, E. pudica, S. kenti and Rhodactis sp. were recorded. At the reef edge the number of species was correlated with the amount of wave action. If the edge was composed only of rock and was subject to heavy wave action few actinians or corallimorpharians were present. In Gan the only species found was Gyractis excavata Boveri, 1893, inhabiting cracks and crevices in extremely hard rock, but in Aden it was associated in such places with Phymanthus loligo (Ehrenberg, 1834). Gyractis excavata has been recorded from similar locations in Hawaii (as Actiniogeton sesere by Dunn, 1 9746: 1 8 1-1 88) and Fiji (J.S. Ryland, personal communication). Where wave action was low, as at Singapore and inside the lagoon of Gan, the reef edge was composed of living coral and sand. Many actinians occurred here, attached to coral stems above or below the sand. Edwardsianthus pudica comb, nov., Radianthus main (Haddon & Shackleton, 1 893), Phymanthus muscosus Haddon & Shackleton, 1 893, and a cerianthid occurred in the sand. Among the corals and attached to their stems were Cryptodendron adhaesivium Klunzinger, 1877, Entacmaea quadricolor (Rueppell & Leuckart, 1828), Telactinia citrina comb, nov.; and the corallimorpharians Rhodactis bryoides (Haddon & Shackleton, 1893), R. rhodostoma (Ehrenberg, 1834), Metarhodactis boninensis Carlgren, 1943, and Ricordea yuma (Carlgren, 1900). Reef slopes composed of rock were usually covered with corals, and few actinians were seen. This was so at Gan where no actinians were found during several dives made in the area. The only anemone seen was on the wreck of the oil tanker British Loyalty. If the reef slope consisted of coral debris the surface was mobile preventing animals retaining a hold, and few animals of any species were found. At the foot of such reef slopes, at 30-40 m depth, specimens of Radianthus malu and R. macrodactyla (Haddon & Shackleton, 1893) were occasionally found, but their small number suggests that they may have been displaced from the reef edge. In some areas of Pulau Hantu and Pulau Tioman, outcrops of rock occurred on the reef slope with the tops just below Mean Low Water Springs (MLWS). These locations seemed favoured by Radianthus magnifica (Quoy & Gaimard, 1833). In coral areas that had been devastated by storms, human activity, or the starfish Acanthaster planci (Linnaeus, 1758), it was observed that the area was first reoccupied by Rhodactis sp. This corallimorpharian appeared to colonize rapidly and large areas covering many square metres were soon carpeted. This was noted at Pulau Tioman and again in Gan where the reef corals alongside the jetty had been destroyed by abrasion by ships at low water. (d) Rocky areas, not associated with reefs. In the intertidal zone these were the habitat forAiptasia sp., Anthopleura dixoniana, A. stellula and A. waridi (Carlgren, 1900) comb, nov.; and also for Corynactis sp. In deeper waters actinians were found also on antipatharians: for example, several specimens of Nemanthus sp. were found on the main stems of black corals in Pulau Tioman. Around Singapore Islands there were many specimens of gorgonians and whip corals at depths between 4 and 40 metres. Despite prolonged searching no actinians were found on them. Collection. Specimens were removed with the substrate to which they were attached or were prised carefully from it. Attempts to loosen specimens from the substrate by chemical means rarely worked, the anemones closing up tightly and remaining strongly attached. Soda water and weak solutions of formaldehyde and bleach were tried unsuccessfully. With burrowing anemones such as Edwardsianthus spp. and cerianthids it was found most effective to dig gently downwards to a depth of 40-50 cm but 30-40 cm away from the animal and then to tunnel underneath it before attempting capture. It was not uncommon to lose a specimen, since when disturbed most burrowing species contract rapidly and move downwards. Preservation. Specimens were narcotized in 3-5-4-0% MgSO4.7H2O in sea water. In the field this was not always possible so MgSO4 crystals were added to a vessel of sea water containing the specimen. With small specimens 30 minutes was adequate for narcotization, but larger specimens were left overnight and some were still not completely narcotized by morning. For large specimens, menthol or a combination of menthol and MgSO4 was also used overnight. When an anemone had ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 209 lost tactile sensitivity 40% formaldehyde solution was injected from a pipette or hypodermic syringe into the enteron via the mouth, care being taken to avoid inflating it. 10% formaldehyde solution was then sprayed from a pipette over the outside of the specimen. Finally it was transferred to a fresh container containing 10% formaldehyde-sea water solution and labelled. Specimens were transferred to 70% C2H5OH later on. Storage in formaldehyde-sea water solution for prolonged periods sometimes hinders subsequent staining and the tissues may start to macerate. Other methods of killing and preserving specimens have been used (Lee, 1921: 521), but not always successfully. The use of hot Bouin's fluid on specimens of Anthopleura waridi resulted in the ectoderm becoming detached, with consequent loss of, the nematocysts from the spherules at the margin and in the fosse, and this led to an incorrect determination (England, 1969: 5). Corallimorpharians, such as Rhodactis rhodostoma, were the most difficult to preserve, since they exuded mucus when chemicals were added to the water to narcotize them. The mucus not only prevented narcotization but seemed to hinder fixation of the tissue by formaldehyde, especially internally. Fixing in 70% C2H5OH overcame the difficulty but the specimens tended to be badly contracted. Glutaraldehyde was used with good results. Examination of material. Specimens were examined with a dissecting microscope under alcohol. At this stage it was useful to stain some of the smaller species in bulk with borax carmine to facilitate the observation of certain features, such as sphincter muscles, micro-mesenteries and suckers on the column if present. Alcoholic picro-indigo-carmine was found a better stain than borax carmine for showing micro-mesenteries that hardly break the surface of the endoderm, and which would normally be detected only be means of serial sections. If it was desired to remove this stain differentiation in 70% C2H5OH was used. Continued changes of 70% C2H5OH removed the stain completely from- most specimens, but occasionally picro-indigo-carmine became permanent, depending perhaps on the method of fixation. Alternatively, if acid alcohol (0-5% HC1, 70% C2H5OH) was used the stain was both differentiated and fixed. This method of bulk staining greatly reduced the need for cutting sections. If sections were required selected pieces of the specimen were removed and stained with borax carmine (unless already stained). Histological sections were cleared in cedarwood oil or terpineol and embedded in paraffin wax (m. pt. 55-57°C) or Fibrowax (m. pt. 57°C). Double embedding in celloidin and Fibrowax was also employed. 10 urn sections were cut using a Cambridge rocking microtome. Most of the sections were counter-stained with picro-indigo-carmine (Stephenson, 1928: 55-56), but other staining methods used included Ehrlich's haematoxylin and eosin, and Mallory's triple stain (Peacock, 1966: 421 & 441). Drawings prepared from the stained sections were made without the aid of a camera lucida. Where part of an illustration was left blank it was because the specimen was inadequate at that point (Figs 3d, lOb). Where an illustration is intended to be diagrammatic this is stated. Nematocysts. Squash preparations were made of small portions of preserved tissue, approximately 1 mm2, from tentacles, column, actinopharynx and filaments, and also from acontia, spherules, nemathybomes and vesicles, when present, to identify the types of nematocyst and to measure them. The samples taken from tentacles, column and actinophrynx were obtained by scraping off ectoderm: this avoided contamination with nematocysts from the endodermal epithelium or filaments, or from acontia which, when present, frequently extended from the enteron into the tentacles. The selected tissue was treated with 2% aqueous KMnO4 on a slide for 2 minutes, the solution removed with filter paper and a drop of glycerine placed on the tissue. It was then teased and afterwards gently squashed under a coverslip by finger pressure using a little sideways move- ment. It was necessary to exert only gentle pressure at this stage, before examining the preparations to identify the types of nematocyst present. Large holotrichs and spirocysts were easily damaged or completely destroyed if too much pressure had been applied or too much sideways motion given to the coverslip. It was at this stage still possible to see the nematocysts in situ in the tissue and to confirm their presence in that position. When only a few nematocysts occurred in certain positions it was sometimes difficult to be certain that they belonged there and were not contaminants from 210 K.W.ENGLAND elsewhere. When doubt existed serial sections were examined to confirm their true location. After the types of nematocyst had been determined and any large, easily damaged nematocysts measured, the preparation was subjected to further squashing to break up the tissue surrounding the nematocysts. The majority of nematocysts were thereby clearly displayed, mostly lying flat. Three separate slides of the tissue from each location were examined. Sometimes replicate prep- arations were stained with 1% methylene blue or 1% acid fuchsin. Methylene blue stains atrichs, basitrichs and heterotrichs intensely so that they stand out from the surrounding tissue, but the capsular contents cannot be seen. If basitrichs and atrichs of a similar size range occurred in the same piece of tissue it would not be possible to separate them under these conditions. For general work and identification of the capsular contents of undischarged nematocysts staining with KMnO4 was best. The KMnO4 partially macerated the tissue but did not damage the nematocysts or alter their size. When the nematocyst contents remained indistinct after staining, identification was achieved using a lOOx phase contrast objective. Each preparation was scanned under oil immersion and the largest and smallest capsules of each undischarged nematocyst type were observed and measured (±0-5 urn) with an eyepiece micrometer. It was usual to measure 20-30 capsules of each type from three slides, and the size ranges reported here summarize these measurements. If fewer than 10 nematocysts were found altogether, the number seen was recorded as well as their size range (see Tables). Terminology. The nematocyst nomenclature of Weill (1934) as modified by Carlgren (\94Qa: 3-4) has been used. Each nematocyst type is illustrated under each species. In addition two types of nematocyst originally noted by Carlgren (1940a: 23), the pterotrich and microbasic t-mastigophore, are redescribed and named (p. 219-221). To describe the anatomy and structure of sea-anemones the terminology of Stephenson (1928, 1935) has been used throughout, though modified in places (see below). Reappraisal of certain taxonomic characters Marginal spherules. All blister-shaped protuberances occurring on the margin or in the fosse are termed spherules in this work (Fig. 1). Since they include all three layers of the body wall they are similar in structure to verrucae but the ectoderm in this case is non-adhesive. Some spherules bear verrucae or adhesive spots on their outer faces. Spherules may have a battery of densely packed specific nematocysts, or they may have only a few nematocysts of the same types and size ranges as elsewhere in the column. A nematocyst battery, when present, is an addition to the spherule and often stands out from it, either as a cap completely covering it or as a small projection easily seen in both live and preserved material (Fig. la, c-e). Spherules showing this extent of variation, together with spherules devoid of nematocyst batteries, can occur with a single specimen, for example in Anthopleura handi (p. 245). Many protuberances are armed with a large battery of atrichs and in the literature have been termed either acrorhagi or spherules, but if the battery of nematocysts is absent they have been called pseudo-acrorhagi or pseudo-spherules. The term acrorhagi was proposed by Andres, 1883: 283. Kwietniewski used the term 'Randblaschen' (18976: 33), and others the term 'Randsackchen' (alternatively spelt 'Randsacken') (Lager, 1911: 219). Haddon, Kwietniewski and Lager appear each to have used these terms for both acrorhagi and pseudo-acrorhagi, so that it is not clear to which genus many of the species they described are best referred. It is now considered that acrorhagi are found, for example, in Actinia, Anemonia, Anthopleura and Oulactis, while pseudo-acrorhagi are found in Paracondylactis, Gyractis and in some other genera. Some sea-anemones have marginal spherules which instead of atrichs possess large numbers of nematocysts of other types. Telactinia citrina comb, nov., and to a lesser extent Mesactinia gen. nov., for example, have large batteries of basitrichs (see p. 263, 270). Again, Triactis has spherules both in the fosse and on marginal peduncles armed with micro- and macrobasic amastigophores. The term acrorhagus is no longer useful because it includes at least three varieties of spherule for which precise terms are needed and it is not used in this account. Also, the term pseudo-acrorhagus ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 211 CO Fig. 1 Variation in marginal spherules (diagrammatic), (a) Typical appearance of Anthopleura sp. with large atrichal marginal spherules, (b) A marginal spherule lacking a nematocyst battery, (c) An atrichal marginal spherule completely capped by a nematocyst battery, (d) An atrichal marginal spherule with a small nematocyst battery, bent into the fosse, (e) Same, but upright, (f) Marginal spherule (to right) without a nematocyst battery but with an atrichal spherule in the fosse (as in Anthopleura waridi). (co column, n nematocyst battery, sp spherule, sph sphincter, t tentacle.) (pseudo-spherule) is a misnomer in that a spherule is always present, whether it has a battery of nematocysts or not. It is proposed that the term acrorhagus be dropped in favour of spherule. When spherules are specialized by the presence of specific nematocysts in a definite battery they can be classified as follows: a. spherules armed with atrichs b. spherules armed with basitrichs c. spherules armed with micro- and macrobasic mastigophores Atrichal spherules Basitrichal spherules Mastigophoral spherules The position of a spherule can be on the margin, in the fosse or on the column. A full designation of a spherule of an Anthopleura specimen might be 'atrichal marginal spherule', or of one of Actinia 'atrichal spherule in the fosse'. Spherules are often armed with more than one type of nematocyst so the name applied is that of the more important or most numerous nematocyst present. The term 'nematosphere', used to described the bunches of spherules on stalks between the tentacles of Heterodactyla hemprichi Ehrenberg, 1834, is retained. Sphincter muscle. The type of sphincter has been used as a taxonomic character in the tribe Thenaria at various levels, for example: 212 K. W. ENGLAND Fig. 2 Variation within a single sphincter due to state of expansion of the crown (diagrammatic), (a) A typical 'circumscribed-pinnate' sphincter, (b) A 'circumscribed-diffuse' sphincter (c) A 'diffuse' sphincter. Arrows indicate directions of expansion, dashed lines divide the sphincter into its elements. (Stippled, endoderm; black, mesogloea; white, endoderm.) Subtribe level Genus level Species level Genera with endodermal sphincter Genera with mesogloeal sphincter Endomyaria Mesomyaria To separate genera that might otherwise form one large genus, for example in the Actiniidae: Sphincter circumscribed Epiactis Sphincter diffuse Entacmaea Sphincter absent Phialoba The presence of a strong or a weak sphincter may be used to distinguish species In some species the sphincter is consistently of the same type, and unless the specimen is extremely distorted it is of uniform appearance in transverse section. In other species variation is observed as shown below in Anthopleura dixoniana (Fig. 18) and Gyractis excavata (Fig. 27). The sphincter of A. dixoniana varies from circumscribed to diffuse even within a single specimen. The size and appearance of a sphincter can also differ within a species between one group or colony of individuals and another. Haddon was unaware of this when he proposed two species separated on this character, Actinioides dixoniana and A. papuensis, of which only one proved valid. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 213 The appearance of a sphincter in transverse section depends on its degree of contraction and on the state of expanison of the crown (Fig. 2). As the crown expands its circumference increases and some mesogloeal folds of the sphincter are drawn outwards into the mesogloea of the margin. When a circumscribed sphincter is involved its main stem (see Fig. 2a) becomes assimilated and the branched folds spread out and become smaller until in the extreme case they are assimilated into the body wall. A striking example of this was seen in an Australian species Epiactis australiensis Carlgren, 1950c (Fig. 3). The first sections examined, cut parallel with the directive axis, appeared to have several separate sphincters (Fig. 3a), one large and four small. A transverse section taken at an angle of 45° to the first showed some of the smaller sphincters coalescing with the main muscle (Fig. 3b). A section cut at 90° to the directive axis showed all the smaller sphincters absorbed into the main muscle (Fig. 3c). A further change was seen at the next 45° position where the muscle passed through the mesentery (Fig. 3d). The muscle folds have not been drawn since they were completely anastomosed with those of the mesentery, but there were three separate sphincters. Variation in the appearance of a single sphincter muscle around the margin due to local differences in its degree of contraction suggests that apparently different types of sphincter may prove to be the same. Figure 2 shows how a circumscribed sphincter transforms into a diffuse one. If the circumscribed sphincter is divided into its elements or mesogloeal folds, as indicated by the dashed lines in Figure 2a, the individual parts can be identified in the so-called diffuse sphincter (Fig. 2c). When this model is applied to the sphincters of Anthopleura dixoniana (Fig. 18) and Gyractis excavata (Fig. 27), each species is shown to have a sphincter which assumes several facies in transverse sections. Similar variation in the sphincter musculature is found in many other species, as for example in Anthopleura nigrescens (Fig. 22) and Mesactinia ganensis gen. nov., sp. nov. (Fig. 29). If the model in Figure 2 is applied, however, the sphincter muscles can be used at least at species level as a taxonomic character. •CO Fig. 3 Epiactis australiensis. Vertical sections showing variation of the sphincter in one specimen due to expansion of the crown, (a) At the directive axis, (b) at 45 degrees to the directive axis, (c) at 90 degrees to the directive axis and (d) at 45 degrees to the directive axis where the sphincter passes through the mesentery. (Hatched, ectoderm; black, mesogloea; white, endoderm; co column.) 214 K.W.ENGLAND Mesenterial arrangement. The relation of the number of tentacles, and hence of the number of mesenteries at the disk, to the number of mesenteries at the base is considered to be of generic importance. Three categories are recognized: a. more mesenteries proximally than distally b. same number of mesenteries distally as proximally c. more mesenteries distally than proximally These categories appear to be clear cut but in practice they intergrade. The mesenteries may develop from the base upwards, as for example in some species of Anthopleura, so that there are a few more at the base than at the margin. The reverse is also true, when tentacles and the corres- ponding mesenteries appear at the margin and grow downwards, as in Mesactinia ganensis gen. nov., sp. nov. It is also possible for mesenteries to appear at margin and base simultaneously and grow towards the centre of the column. An equal number of mesenteries may occur distally and proximally in juveniles but later on the number of mesenteries at the base might increase more rapidly than the number of tentacles, resulting finally in a difference of a complete cycle. Some species of sea-anemone have very small mesenteries at the margin and/or at the base, where they are little more than gussets of tissue as for example the microcnemes of Edwardsia. Among acontiate anemones Calliactis polypus (see England 1971: 25) may have minute mesenteries at both margin and limbus, but there are always more tentacles than basal mesenteries. Verrillactis paguri (Verrill, 1869a) has more small mesenteries at the limbus than at the margin. It is possible for the mesenteries to appear normally at the margin and grow downwards, but further down the column to develop only into small elevations of tissue, seen in transverse section, to comprise mesogloea covered with endoderm (as in Mesactinia ganensis gen. nov., sp. nov., p. 263). Only careful observation, for example by bulk staining of the specimens, will detect their presence (p. 209). The presence of just a few more mesenteries at margin or base is not considered to be of generic importance and such specimens can be classified under (b). Only when there are at least half a cycle more mesenteries present at margin or limbus should a specimen be referred to categories (a) or (c). Nematocysts. The types and size-ranges of nematocysts have been used as taxonomic characters from species to family level. Carlgren (1949: 86) used the presence or absence of certain nematocyst types in the acontia to define families, and in many other papers he used their size-range to identify species. Hand (1955, 1956) was the first to illustrate all the cnidome of an individual species. He grouped them according to that part of the specimen from which the different types of nematocysts came (tentacles, column, actinopharynx, filaments and, when appropriate, nemathybomes, spherules and acontia). This method has been adopted here for the following reasons: a. The diagrams show the shape and structure of the nematocysts found, and obviate confusion over the terms used to name them. Several names now exist for each nematocyst type. For example, a single type has been termed basitrich (Weill, 1934), microbasic b-mastigophore (Cutress, 1955: 128), b-rhabdoid (Schmidt, 1969: 298), and spirula (den Hartog, 1980: 7, after Stephenson, 1928: 62-63). b. It provides a 'nematocyst signature' of each species which can be used for rapid comparison with the 'signatures' of others. When species belonging to different families and to some extent to different genera are compared, a clear distinction between 'signatures' is usually apparent. Figures 40-42 illustrate generic differences between three acontiate species. But when congeneric species are compared only small differences in the size ratios of the two or three nematocyst types in one part may separate species. Figures 19, 21, 23 and 25 illustrate the closeness of the 'signatures' in four congeneric species. In fact these 'signatures' are so similar that, in the absence of morphological information, they might have been interpreted as coming from a single species. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 215 Taxonomic section Suborder NYNANTHEAE Carlgren, 1898 Subtribe ATHENARIA (ABASILARIA) Carlgren, 1898 Family EDWARDSIIDAE Andres, 1881 Genus EDWARDSIA de Quatrefages, 1841 Edwardsia de Quatrefages, 1841: 427. Scolanthus Gosse, 1853: 153. Halcampa: Panceri, 1869: 1 (part). NOMENCLATURE. An application to the International Commission on Zoological Nomenclature to conserve the generic name Edwardsia de Quatrefages, 1841, and the family name EDWARDSIIDAE Andres, 1881, was submitted by Williams (1979) and was favourably resolved in Opinion 1294 (Bulletin of 'Zoological Nomenclature 42: 31-33). d Fig. 4 Edwardsia. Development of mesenteries and their relation to the number of tentacles, (a) 'Edwardsia ' stage, (b) 12 tentacle stage, (c) 16 tentacle stage, (d) 24 tentacle stage, (e) 32 tentacle stage, (f) 'pudica' arrangement, 20 tentacles, (ac actinopharynx, dl dorso-lateral exocoel, 1 lateral exocoel, v ventral exocoel. Retractor muscles represented as black rectangles; lumen of pharynx not shown.) 216 K.W.ENGLAND TYPE SPECIES. Edwardsia beautempside Quatrefages, 1842, by designation by Carlgren (1949: 23). DEFINITION. Edwardsiidae (sensu Carlgren, 1949: 22) with column divided into physa, scapus, scapulus and capitulum; physa short, ampullaceous, without nemathybomes or cuticle. Scapus long with nemathybomes sunk in the mesogloea. Nemathybomes always containing two types of nematocyst. Siphonoglyph weak, ventral. Mesenteries comprising eight macrocnemes (perfect) and at least four microcnemes (imperfect) which are very weak and in capitulum only; first four microcnemes always paired with macrocnemes. Gonads, filaments, and parietal and retractor muscles on macrocnemes only. Parietal muscles well developed, retractors strong-diffuse to restricted-reniform. Cnidome: spirocysts, basitrichs, pterotrichs, microbasic amastigophores and t-mastigophores. REMARKS. Fifty-five nominal species were referred to Edwardsia by Carlgren ( 1 949) but only about four more have been described since, and some of the original fifty-five have been shown by Manuel (1977) and Williams (1981) to be invalid. Williams recognized 40 valid species, but much literature searching is still necessary before a single species can be confidently identified. The task can be reduced if certain species are eliminated on conspicuous taxonomic characters and placed in other genera. The two characters in question are the presence of one or two types of nematocyst in the nemathybomes, and the pattern of development of the microcnemes. Carlgren (1921: 28) noticed that one type of nematocyst was present in the nemathybomes of some species of Edwardsia and two types in others. He used this character to differentiate between species, although he did not use it in his later survey of the whole order (Carlgren, 1949). Other authors have noted this character without using it taxonomically , though Williams (1981) listed the nominal species of Edwardsia under these and other headings. In this work species with only one type of nematocyst (basitrichs) in the nemathybomes were separated from those with two types (pterotrichs and microbasic t-mastigophores, p. 219). The presence or absence of a type of nematocyst is usually considered to be of generic or higher significance. The arrangement of the microcnemes is also regarded here as a generic character. Microcnemes develop after the eight macrocnemes in two ways. Normally four arise (Fig. 4a) in the lateral and ventral exocoels and are paired with the ventro-lateral and dorso-lateral couples, and four more arise as two pairs in each dorsal exocoel (Fig. 4b-c)). They correspond to the 16 tentacles. In some species the two pairs of microcnemes do not develop and there are only 12 tentacles instead of 16. The plan of development may continue with four more pairs of microcnemes arising in the lateral and ventral exocoels, corresponding to 24 tentacles (Fig. 4d). In some species development continues with pairs of secondary microcnemes arising in the dorsal exocoels as well, one pair on each side of the first pair (Fig. 4e). Species with this arrangement have up to 32 tentacles. Examples of species with each of these arrangements are: 1 2 tentacles E. duodecemtentaculata Carlgren, 1 93 1 16 tentacles E. longicornis Carlgren, 1921 24 tentacles E. carneola (Verrill, 1 928) 32 tentacles E. timida de Quatrefages, 1 842 A maximum of 28 tentacles in E. timida was recorded by Manuel (1977: 492, Fig. 4), but his illustration shows that microcnemes were absent from the dorsal exocoels, only two pairs of microcnemes occurring there instead of the usual three pairs. Up to 36 tentacles have been recorded in some species, for example E. timida ( = E. callianthus Rawlinson, 1935) and E. sipunculoides (Stimpson, 1853; Carlgren, 1931). In such instances additional pairs of microcnemes occur in the lateral exocoels or, as shown by Stephenson (1935), in the ventral exocoels. Other anomalies are known. One microcneme of a pair may develop before the other resulting in an odd number of tentacles. This has been observed in E. vivipara Carlgren, 1950c. Dixon (1886) also noted anomalies, in E. timida. The second pattern of development of microcnemes results in a quite different arrangement. After the eight macrocnemes have formed the four single microcnemes do not develop in the lateral and ventral exocoels but six pairs of microcnemes arise, one pair in each exocoel (Fig. 4f). This ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 217 results in 20 tentacles. It is not yet known whether all microcnemes appear at the same time, as seems probable, or whether they develop in certain exocoels first. This type of arrangement was first noted by Faurot (1895) in E. adenensis Faurot, 1895. Bourne (1916) drew attention to it when describing E. rakaiyae Bourne, 1916, and discussed a possible interpretation of the sequence of development of the tentacles and micromesenteries. Stephenson (1935: fig. 108A) found that same arrangement in an edwardsiid from Low Island on the Great Barrier Reef. The species was subsequently named E. stephensoni by Carlgren (19506), but is here referred to Edwardsianthus pudica (Klunzinger, 1877) comb. nov. All recorded specimens of E. pudica have six pairs of microcnemes. Similarly, I have examined 57 specimens of E. gilbertensis (Carlgren, 1931) and all had six pairs. This arrangement of microcnemes separates the two species E. pudica and E. gilbertensis from those that develop mesenteries in the normal way. Table 1 The species of Edwardsia de Quatrefrages, 1 84 1 . Species Dimensions of Number of Nemathybomes nematocysts from Tentacles arrangement nemathybomes (urn) Source of data E. annamensis Carlgren, 1943 16 rows 41 -0-60-6 x 4-2-5-6 Carlgren, 1943: 19 42-3-60-6 x 2-0-2-8 '£. beautempsi de Quatrefages, 1842 16 rows 39-0-65-0 x 4-5-6-0 Manuel, 1977: 487 45-0-64-0x2- 5-3-0 E. calif ornica McMurrich, 1913 16 rows 115-153x6-3-7-0 Carlgren, 1936: 18 72-0-77-0x2-5 E. capensis Carlgren, 1938 18 rows 68.0-86.0 x 5.5 6-0 Carlgren, 1938: 19 67-0-72-0 x 2- 5 E. claparedi (Panceri, 1869) 16 rows 97-0-154x5-0-7-0 Manuel, 1981: 198; 69-0-95-0 x 2-5-3-0 1977:487 E. elegans Verrill, 1869c 16 rows 67-0-101 x 6-5-7-0 Carlgren, 1931: 16-18 67-0-84-0x2-5-3.0 E. hantuensis sp. nov. 16 rows 62-7-105-6 x 4-6^6-6 p. 222 54-8-68-2 x 2-6 2E.ivelli Manuel, 1975 16 rows 24-0-33-0 x 2-5-3-0 Manuel, 1975:705-711 20-0-30-0 x 2-0-3-0 E. longicornis Carlgren, 1921 16 rows 36-0-84-0 x 3-5-5-0 Manuel, 1977:487 36-0-65-0 x 2-0-2-5 E. maroccana Carlgren, 1931 16 rows 131-156x6-0 Carlgren, 1931: 15-16 74-0-84-0 x 2-0-2- 5 E. octoradiata Carlgren, 1 93 1 16 rows 55-0-79-0 x 3-5-4-0 Carlgren, 1931: 13-15 35-0-43-0 x 2-5 E. sanctaehelenae Carlgren, 1 94 1 12 rows 56-4-8 1-8x5-0-5-6 Carlgren, 1941: 1 42-3^9-3 x 2-0-2-5 E. sulcata (Verrill, 1864) 14-16 rows 84-6-104-3x7-0 Carlgren, \950d: 22 52-2-67-7x4-2 E. tintrix Annandale, 1915 16 rows 65-0-89-0x4.5 Carlgren, 1925: 19 46-0-53-0 x 2-0-2-5 E. tuberculata Diiben & Koren, 1847 16 rows (72)110-190x4-0-7-0 Carlgren, 1921:29 60-0-96-0 x 2-5 E. carlgreni (Carlgren, 1921) 16 scattered 62-0-74-0 x 5-0 Carlgren, 192 1:3; see 36-0-53-0 x 2-5 Williams, 1981:351 E. danica Carlgren, 1921 14-20 scattered 46-0-72-0 x 4-5 Carlgren, 1921:37 24-0-43-0 x 2-0-3-5 E. novazelanica Farquhar, 1898 16-24 scattered 41 -0^56-0 x 3-5^-5 Carlgren, 1924: 184-187 34-0-41-0x2-0 (E. tricolor Stuckey, 1908) E. sipunculoides (Stimpson, 1853) 18-30 scattered 65-0-74-0x6-0 Carlgren, 1931:22-23 41-0-50-0x3-5 'Type species of genus. 2 possibly Edwardsioides. 218 K. W. ENGLAND Using these two characters species whose nemathybomes have only basitrichs, that is, just one type of nematocyst, can be removed from the genus Edwardsia and subdivided into two new generic groupings based on the mesenterial arrangements. The species in which relevant data could be found in the literature are listed in Tables 1-3. The generic groupings are given in Figure 5. Danielssen (1890) introduced the genus Edwardsioides to accommodate E. vitrea Danielssen, 1 890. This is the earliest generic name available for the group comprising those species with only one nematocyst type in the nemathybomes and four microcnemes, paired with four macrocnemes. The final group (Table 3) comprises those species with one type of nematocyst in the nemathy- bomes but no microcnemes in the first cycle of mesenteries. The generic name Edwardsianthus gen. nov. is proposed here with Edwardsia pudica Klunzinger, 1877, as the type species. Further subdivision of the genera Edwardsia, Edwardsioides and Edwardsianthus is possible according to whether the nemathybomes are arranged in rows or scattered on the column. Subdivi- sion is unnecessary in Edwardsianthus which has only two species and in Edwardsioides in which only one out of fourteen species has the nemathybomes arranged in rows. However, the 19 species referred to the genus Edwardsia can be divided with some advantage. They fall into two groups, 4 Table 2 The species of Edwardsioides Danielssen, 1 890. Species Number of Tentacles Nemathybome arrangement Dimensions of nematocysts from nemathybomes (urn) Source of data E. andresi Danielssen, 1890 12 scattered 48-0-67-0 x 3-5^-0 Carlgren, 1921:49 E. arctica Carlgren, 1921 16 rows/ 38-0-60-0 x 4-0-5-0 Nemathybomes scattered scattered in proximal part. Carlgren, 1921:39 E. duodecemtentaculata Carlgren, 12 scattered 33-0-48-0 x 2-5-3-0 Carlgren, 1931:4 1931 E.finmarchica Carlgren, 1921 26 scattered 36-0-48-0 x 3-0-4-0 Carlgren, 1921:55 E.fusca Danielssen, 1890 12 scattered 31 -0-36-0 x 2-0-2- 5 Carlgren, 1921:28 E. islandica Carlgren, 1921 16 scattered 36-0-48-0 x 2-0-2-5 Carlgren, 1921:47 E.japonica Carlgren, 1931 16 scattered 74-0-101 x 2-5-3-5 Carlgren, 1931: 12 E. kameruniensis Carlgren, 1927 20 scattered 45-0-63-0 x 3-5-4-5 Carlgren, 1927:478 E. meridionalis Williams, 1 98 1 16(18) scattered 29-0-50-7 x 2-6-4-7 Williams, 1981:325 E. norvegica Carlgren, 1942 16 scattered 47-0-58-0 x 3.0 Carlgren, 1942: 60 E. timida de Quatrefages, 1 842 32 scattered 32-0-75-0 x 3-5-7-0 Manuel, 1977: 491 E. vegae Carlgren, 1921 16 scattered 84-0-100x3-0 Carlgren, 1921:53 '£. vitrea Danielssen, 1890 16 scattered 36-0-42-0 x 3-0-3-5 Carlgren, 1921:49 E. vivipara Carlgren, 1950c 16 scattered 26-8-38-1 x 3-5^-5 Carlgren, 1950c: 1 E.jonesi Seshaiya & Cutress, 12 rows 48-0-72-0 x 4-5-5-5 Seshaiya & Cutress, 1969 1969: 73 'Type species of genus. Table 3 The species of Edwardsianthus gen. nov. Species Number of Nemathybome Tentacles arrangement Dimensions of nematocysts from nemathybomes (urn) Source of data '£. pudica Klunzinger, 1877 E. gilbertensis Carlgren, 193 1 20 16-20 scattered rows 37-0-44.0x2-5 3 1-0-4 1-0x2-0-3-0 Carlgren, 1931: 18-21 (see p. 227) Carlgren, 1931:10-11 'Type species of genus. ACTINI ARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 2 1 9 species having the nemathybomes scattered on the column and 15 species having them in rows. This division has not been formalized here because it may be advantageous to subdivide the genus at a later date on a different character, namely the types of nematocyst in the nemathybomes (see p. 219). Species of Edwardsia listed by Carlgren (1949: 23-24) and other authors but not included here have been omitted either because they have been shown to be conspecific with the species listed or because insufficient data is available to make it possible to place them in one of the three genera proposed. Williams ( 1 98 1 ) listed all species referred to the genus Edwardsia with comments on their validity and synonyms. EDWARDS I I DAE With nemathybomes sunk in the mesogloea of scapus, physa present One type of Two types of nematocyst in nematocyst in nemathybomes nemathybomes Four microcnemes Microcnemes Four microcnemes of first cycle of first of first cycle paired with cycle absent paired with macrocnemes macrocnemes Edwardsioides Edwardsianthus Edwardsia Danielssen, 1890 gen. nov. Quatrefages, 1841 Fig. 5 Edwardsiidae. Distinctions between the genera Edwardsia, Edwardsioides and Edwardsianthus gen. nov. Nematocyst types in the nemathybomes o/Edwardsia. Manuel (1977: 488, Fig. 2) illustrated two types of nematocyst that he termed microbasic b-mastigophores from E. beautempsi, and two other types from E. timida. The types illustrated under E. beautempsi were not identical, but were similar to those described by Carlgren (19400: 23, Fig. V, figs 7-8, 12-13) from E. longicornis. One type of nematocyst (Manuel 1977, Fig. 2E) differs from the normal b-mastigophore (or basitrich) in that the armature of the thread can be divided into four distinct regions. A similar nematocyst has been studied, both discharged and undischarged, from specimens referred to E. tuberculata Duben & Keren, 1847: 267, from off Rame Head, South Devon (Fig. 6a-c). The basal part of the thread is folded and bears a few short pointed spines arranged spirally, corresponding to Schmidt's 'faltstuck' (Schmidt 1969: 294, Fig. 5). Then follows a length of thread with a feature so far unique in nematocyst morphology (Fig. 6b). It is a heavily armed portion having many thick close-set spines that are not arranged spirally around the thread but lie in three rows along it. They recall in appearance the flight feathers of an arrow (Fig. 6b), rather than conforming to Carlgren's impression of a membrane. Above this extraordinary region there are a few large spines arranged spirally, reducing sharply in size to those of the longest and distalmost part of the thread which is armed with minute spines arranged spirally throughout its length. The basal part of the thread is no more than 1| times wider in diameter than the remainder. This nematocyst type, therefore, incor- porates features of both haplonemes and mastigophores. Shaft and thread are of nearly the same diameter, resembling the basitrich in morphology, but a 'faltstuck' is present and the spines are set close together as in the mastigophore. The undischarged capsule resembles in appearance that of a basitrich, apart from the folded part of the thread near the tip of the capsule. The straight part of the thread, in the undischarged capsule, does not show evidence of the spines except at the bottom, but has a thin line running along its length in the centre. The bottom of the capsule contains most of the coiled part of the thread, only a small section of the thread extending into the upper part, as a loose loop (Fig. 6a). The armature of the coiled thread is clearly visible. 220 K. W. ENGLAND Fig. 6 Nematocysts of the nemathybomes of Edwardsia tuberculata (diagrammatic), (a) Undischarged pterotrich. (b) Discharged thread of same, showing types of spines and their arrangements on the thread, (c) Part of same enlarged, (d) Undischarged microbasic telotrichous mastigophore (t-mastigophore), with folded shaft, (e) Part of same, discharged, showing position of large spines at the head of the shaft, (f) Part of same showing alternative, unfolded shaft, undischarged. Scale: (a,d) = 20 urn, remainder not to scale. Since the lower part of the thread is only slightly wider than the remainder, and because the undischarged capsule resembles a basitrich more than a mastigophore, I propose to classify this nematocyst type among the haplonemes. The name 'pterotrich' is suggested. The undischarged capsule can be recognised from the 'faltstuck' and the few large protruding spines at the end of the straight portion of the thread. The other type of nematocyst (Carlgren 1940« Fig. V, figs 7-8; Manuel 1977, fig. 2F) is a mastigophore in that when discharged it shows a definite shaft with parallel sides and a distinct thread (Fig. 6d-f). The diameter of the shaft is some four times that of the thread, not with shaft and thread of the same diameter as illustrated by Carlgren. It cannot, however, be classified as a ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 221 b-mastigophore since the shaft has spines only at the forward end (Fig. 6e). Although the dis- charged capsule resembles WeilFs (1934) microbasic euryteles there is no swelling at the end of the shaft and before discharge the capsule resembles rather a basitrich with a thickened portion in the centre of the straight part of the thread. Weill might possibly have termed this nematocyst a microbasic telotrichous mastigophore, for which the shortened form microbasic t-mastigophore is proposed. Manuel (1977, figs 2G-H) illustrated two nematocysts of one type from the species E. timida which appear to be basitrichs, but from his illustrations it is not possible to be certain. If the two nematocysts are the same and two clearly defined size ranges are present then the nematocysts of this species differ from those of other Edwardsia species and a new genus would be appropriate to accommodate E. timida. Alternatively, if all nematocysts from the nemathybomes fall into a single range then the species might be transferred to the genus Edwardsioides. The figures quoted by Manuel (1977: 491) suggest that the species should be so referred. Other species currently included in Edwardsia (Table 2) may also present the same problem, the size-ranges of nematocysts overlapping in both length and diameter, for example E. ivelli Manuel, 1975: 705-71 1 . octoradiata Carlgren, 1931 tintrix Annandale, 1915 suloota Verrill, 1864 alaperedi-i Panceri , 1869 tuberoulata DUben & Koren, 1847 maroacana Carlgren, 1931 ivelli Manuel , 1975 longioomis Carlgren, 1921 beautempsi de Quatrefages, 1842 annamensis Carlgren, 1943 elegans Verrill, 1869b 20 50 1 00 150 |j Fig. 7 Comparison of nematocyst size ranges in nemathybomes of certain edwardsiids. (Sources of data as in Table 1.) A bar chart (Fig. 7) of the size ranges of nematocysts in the nemathybomes of species of Edwardsia shows an interesting variation. In certain species the two size ranges are distinct (E. claparedi, E. tuberculata), whereas in others the two ranges overlap (E. beautempsi, E. longicornis). These species are cited since the types of nematocyst in their nemathybomes are well known. It is too early to say if there is any significance in this variation, but it will at least assist in identifying species of Edwardsia. Edwardsia hantuensis sp. nov. Figs 8-9 M ATERIAL EXAMINED. Singapore: PulauHantu, 1 specimen, 19Sep 1970,BMNH 1983.4. 8.1. Size in spirit 30 mm high, 5 mm wide. DESCRIPTION. External appearance. Column divided into physa, scapus, scapulus and capitulum. Scapus with eight rows of large prominent nemathybomes and an easily removed cuticle. Near scapulus nemathybomes close together but towards physa further apart. Tentacles up to 16, long, smooth, tapered. Anatomy. Eight macrocnemes and up to eight microcnemes in typical Edwardsia manner. Retractors strong, diffuse, with 25-28 muscle processes, some branched, arising from mesogloeal sheet (Fig. 8a). Parietal muscles large, well developed, on both sides of mesogloea, 222 K. W. ENGLAND Fig. 8 Edwardsia hantuensis sp. nov. BMNH 1983.4.8. 1 . Transverse sections of (a) retractor muscle, (b) parietal muscle, (c) nemathybome. Scale: (a-c) = 50um. (co column, f filament, m mesentery, n nematocysts. Stippled, ectoderm; black, mesogloea; white, endoderm.) cross-sectional area nearly half that of main retractor in same area of column (Fig. 8b). Microcnemes small. Nemathybomes (Fig. 8c) large, with both pterotrichs and microbasic t-mastigophores (p. 223, Fig. 9). Physa without a central pore but many cinclides present near tip of each exocoel. Cnidome. See Table 4 and Figure 9. The size ranges of the nematocysts of the nemathybomes are the overall ranges. The pterotrichs in the nemathybomes nearest the scapulus are smaller than those of the nemathybomes near the physa. For example, 62-7-75-8 x 4-0-5-3 um near scapulus, 92-4-105-6 x 5-3-6-6 um near physa. Colour. Column light sienna, nemathybomes white. RECORDED DISTRIBUTION. Recorded only from Singapore (present material). Table 4 Size ranges of cnidae of Edwardsia hantuensis sp. nov. compared with those reported from E. sulcata (Verrill) and E. elegans Verrill (in um). Location/ typeofcnida E. hantuensis sp. nov. Singapore (BMNH 1983.4.8.1) E. sulcata (Verrill) Massachusetts from Carlgren, 1950d: 23 E. elegans Verrill Maine from Carlgren, 1931: 16 Tentacle Spirocyst 1 1-8-18-9x1 -8-3-5 Basitrich 10-6-23-6x1 -5 21 -0-26-3 x 2-8-3-5 17(19)-29(31)x 2-5-3-0 Capitulum Basitrich 10-6-11-8x1-8 Nemathybome Pterotrich 62-7- 105-6x4-6-6-6 84-6-104-3 x 7-0 67-0-101 x 6-5-7-0 Microbasic t-mastigophore 54-8-69-2 x 2-6 52-2-67-7x4-2 67-0^84-0 x 2-5-3-0 Actinopharynx f 12-0-16-0 x 1-5 Basitrich 11 -8-26-0 x 2-0-3-5 1 29-0-36-0 x 2-5(3) Microbasic amastigophore 26-0-30-7 x 5-8-7-0 29-0-31-0x5-0 Filament Basitrich 12-0-17-7x2-4 Basitrich 17-7-22-4x3-5 19-0-22-6 x 2-5-2-8(3-5) Microbasic amastigophore 18-9 23-6x4- 1-5-8] 24-0-28-2 x 4-5-7-0 Microbasic f p-mastigophore 21 -2-28-3 x 3-5-4-1 J 223 HABITAT. On reef flats in sand, Mean Low Tide Level (MLTL). REMARKS. The presence of nemathybomes with both pterotrichs and microbasic t-mastigophores places this species in the genus Edwardsia. The characters of this species are similar to those of Edwardsia sulcata (Verrill, 1 864) sensu Carlgren, 1950J: 22, and E. elegans Verrill, 1869c. However, there are some differences in the fT\ Fig. 9 Edwardsia hantuensis sp. nov., nematocyst signature (see Table 4). Tentacle (a) spirocyst, (b) basitrich. Capitulum (c) basitrich. Nemathybome (see Fig. 6) (d) microbasic t-mastigophore, (e) pterotrich. Actinopharynx (f) basitrich, (g) microbasic amastigophore. Filament (h-i) basitrichs, (j) microbasic amastigophore, (k) microbasic p-mastigophore. Scale: (a-k) = 10 ^im. 224 K. W. ENGLAND nematocyst size ranges between these two species and the present material (see Table 4 for compari- son). The large difference between the size ranges of the pterotrichs at top and bottom of the column may also be significant. Such differences have not been described before, as far as can be ascertained, and investigations into possible differences within specimens of other species are required. The present material is, therefore, provisionally referred to a new species, Edwardsia hantuensis sp. nov. Genus EDWARDSIOIDES Danidssen, 1890 Edwardsioides Danielssen, 1890: 100, pi. 5, fig. 3; pi. 16, figs 4-10. Edwardsia: Carlgren, 1921: 28, 39, 43, 47, 53, 55; Carlgren, 1927: 478; Carlgren, 1931: 4-7, 12-13; Carlgren, 1942: 60; Carlgren, 1950c: 1; Seshaiya & Cutress, 1969: 73; Manuel, 1977: 491; Williams, 1981: 325. DEFINITION. Body divisible into physa, scapus, scapulus and capitulum; physa short, without nemathybomes or cuticle. Scapus long with nemathybomes sunk in mesogloea and containing only one type of nematocyst, basitrichs. Siphonoglyph weak, ventral. Mesenteries: eight perfect (macrocnemes) and at least four imperfect (microcnemes) which are minute and usually confined to region of capitulum. First four microcnemes always paired with macrocnemes. Gonads, filaments, and parietal and retractor muscles on macrocnemes only. Parietal muscles well developed: retractors strong-diffuse to restricted-reniform. Cnidome: spirocysts, basitrichs, and microbasic amastigophores. TYPE SPECIES. Edwardsioides vitrea Danielssen, 1890: 100, by present designation. REMARKS. No specimen of this genus was found in the collection, but material qf E. vivipara Carlgren (1950c: 1-2, fig. 1) from South Australia has been examined and the type of nematocyst in the nemathybomes in that species confirmed as a basitrich. Genus EDWARDSIANTHUSnov. Edwardsia: Klunzinger, 1877: 80; Faurot, 1895: 121; Carlgren, 1900: 46; Bourne, 1916: 517-518; Carlgren, 1931: 7; Carlgren, 1950ft: 428, (part). Edwardsiella Andres, 1883: 305 (part). DEFINITION. Body divisible into physa, scapus, scapulus and capitulum; physa short, with- out nemathybomes or cuticle. Scapus long with nemathybomes sunk in mesogloea; cuticle present. Nemathybomes having only one type of nematocyst, basitrichs. Tentacles usually 20. Siphonoglyph weak, ventral. Eight perfect mesenteries (macrocnemes) and six pairs of imperfect mesenteries (microcnemes), minute and restricted to distal part of column. Imperfect mesenteries never paired with perfect mesenteries. Gonads, filaments, and parietal and retractor muscles on macrocnemes only. Parietals well developed; retractors strong-diffuse to restricted-reniform. Cnidome: spirocysts, basitrichs, microbasic amastigophores. TYPE SPECIES. Edwardsia pudica Klunzinger, 1877, by original designation. Edwardsianthus pudica (Klunzinger, 1877) comb. nov. Figs 10-12 Edwardsia pudica Klunzinger, 1877: 80, pi. 6, fig. 3; Carlgren, 1931: 18-20, figs 16-17. Edwardsiella pudica Andres, 1883: 309; Carlgren, 1900: 46, pi. 1, fig. 5. Edwardsia adenensis Faurot, 1895: 121, pi. 6, fig. 5, pi. 7, fig. 6, text-figs 8-9. ?Edwardsia vermiformis Bourne, 1916: 517, pi. 51, fig. 3. Edwardsia rakaivae Bourne, 1916: 518, pi. 51, fig. 4, text-fig. 1. Edwardsia bocki Carlgren, 1931: 7-9, figs 5-6. Edwardsia stephensoni Carlgren, 1950ft: 428-429, figs 1-2. MATERIAL EXAMINED. Singapore: Pulau Jong, 1 specimen, 17 May 1970, BMNH 1983.4.8.2; Pulau Hantu, 2 specimens, 19 Sep 1970, BMNH 1983.4.8.3^; Pasir Panjang, 1 specimen, 23 Feb 1971, BMNH 1983.4.8.5. Aden: below Jebel Hadid, 5 specimens, Aug 1966, BMNH 1983.4.8.6-9. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 225 Fig. 10 Edwardsianthus pudica comb. nov. Variation in size and development of retractor muscles seen in transverse sections of two specimens: (a) Aden, BMNH 1983.4.8.6; (b) Singapore, BMNH 1983.4.8.3. Scale: (a-b) = 500 urn. (f filament, m mesentery, black, mesogloea; white, endoderm.) Madagascar, 3 part specimens, Thomassin coll., Museum National d'Histoire Naturelle, Paris. Great Barrier Reef: Edwardsia stephensoni (syntypes), 3 specimens, BMNH 1954.6.28.3-4,6. In addition a specimen was seen alive on Gan Island, Addu Atoll, Maldives, March 1970, but not collected. DESCRIPTION. Height up to 200mm, diameter up to 15mm. Divisions of column distinct. Capitulum thin walled, translucent, mesenteries within clearly visible in life; scapulus thin walled, lacking nemathybomes and cuticle, scapus thick walled, cuticle easily removed, with numerous small nemathybomes scattered irregularly over its length becoming most numerous towards physa, physa thin walled, lacking nemathybomes or terminal pore but having cinclides which are mostly endodermal evaginations; ectodermal suckers or tenaculi occasionally present (Fig. 1 Ic). Tentacles up to 20, in two cycles 8, 12; long, slender, tapered. Mouth raised on cone. Anatomy. Eight macrocnemes arranged according to Edwardsia plan, one pair of microcnemes in each of the six exocoels. Microcnemes extremely small; having appearance of a pair of lips. In sections musculature comprising a single vertical layer of fibres over the complete surface. Microcnemes not extending below capitulum. Macrocnemes with both retractor muscles and parietal muscles well developed (Figs lOa-b, lla-b); in sections retractors (Figs lOa-b) diffuse, with numerous closely packed high muscle folds, branched and unbranched, sometimes arranged in a pattern of several unbranched folds together, followed by one or sometimes 2-3 branched folds with a series 226 K. W. ENGLAND Fig. 11 Edwardsianthus pudica comb. nov. (a) Transverse section of parietal muscle, Aden, BMNH 1983.4.8.6; (b) Transverse section of parietal muscle, Singapore, BMNH 1983.4.8.3; (c) Vertical section through physa, Aden, BMNH 1983.4.8.6. Scale: (a-b)=100um, (c) = 250um. (ci cinclides, co column, cu cuticle, su suckers. Barred, cuticle; stippled, ectoderm (modified at suckers); black, mesogloea; white, endoderm.) of small curved off-shoots. Near column wall retractor is circumscribed, that is developing on both sides of extended mesogloeal sheet (Figs 10a,b). This part of mesogloeal sheet continuous with that of mesentery and sometimes itself branched (Fig. 10b). Parietal muscles well developed equally on each side of mesentery (Fig. lla-b). Macrocnemes fertile, species apparently dioecious. Longi- tudinal muscles of tentacles ectodermal with short unbranched folds. Zooxanthellae present in endoderm of disk, tentacles, capitulum and actinopharynx. Cnidome. See Table 5 and Figure 12. Colour. Column light sienna, nemathybomes white. Capitulum delicate translucent purple to white. Tentacles of two colour varieties (a) delicate magenta-pink with thin purple line running from white tip to disk, (b) light-green with thin orange line. Disk white to cream. DISTRIBUTION. Recorded from the Red Sea (Klunzinger, 1877), Aden (Faurot, 1895, as E. adenensis, and present material), Zanzibar (Carlgren, 1931), the Maldives, Singapore (present material), Papua New Guinea (Bourne, 1916 as E. rakaiyae and possibly as E. vermiformis), Fidschi Island (Carlgren, 1931 as E. bocki) and the Great Barrier Reef (Carlgren, 1950ft, as E. stephensoni). HABITAT. MLTL on reef flats or shores, in coarse or muddy sand. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 227 m od p in p T1 so m in 0 fN so so (N so 1 oo 3 2 I oo m sb fN 1 2» fN fN fN fN fN •rr fN m r|- X X X X X X X X X s: ~^ OO SO ON !/-) O 00 r- o Tj- "^ -_j — -^ — MM r-~ SO m ON — ON fN -si fN fN ro O O I fN 1 S3 3 • o ffl m Tf fN so P!M m fN O fN fN o\* p 0 m O m od r^ -si- 1 "? in J_ ?2 J m oo •— fN fN fN ^> rh — m rr ON X X X X X X X X X ' ' fN — O •rl- p i^ oo O m E •— r- ro *S~) r*^ f*) ~-> ro ob g| fN fN 1 I OO OO 1 m oo fN m 1 1 m — fN *O ,vr ' , . o ^4- ^v ob so r^ fN fN — fN fN ^ p fN OO od 1 1 4 m r- ob oo n r~ oo — 4- fN — in — m rt- ON X X X X X X X X ~" O so so m r^ — t-~ oo E O m ON ob 0 r^ o T;|- •-^ fN fN fN fN m fN | | I 1 o OO fN ON < S — — ' fN I*- ON fN — fN ob c ^ m 0 0 0 oo ^-^ od m fN m m m m in >: Tf 1 1 r- t~ (^ p p- (^ 1 1 m in fN o m oo M rn -1. •/•> -5 o ON X X X X X X X X ti} C/3 (Li • — i ON in p r~ oo p ~ ON 0 O a1 ob ON — fN Tj- fN m m m CO ON m O m Ti- 44 m i Csl 0 .y* ON "^J" ^^ ^ — 1^ ro *""* SO ^3 J t/5 5i Tj- — m — m fN 'c O o V) CD ofi C l* u u* 42 in %i 1- b-c If 42 4= || ~cT o o 'C § 'C •^'C •st 'C o '-^ *~ 'C 'C O '~ «T> o <—• -~. o _*j -~ .-^ "S -ti 5>.ti u. <« ** "%> '^ 0 %1 •^ Q S • ^" ^ ^ c/5 ^s ^ •^ ^ •'- G ' C — — re u R M CO 'S.in S ffl '•> ffl "£ 1 § & & ^. a CQ J H K u £ X s; 228 K. W. ENGLAND r\ \J e \J g \j d Fig. 12 Edwardsianthus pudica comb, nov., nematocyst signature (see Table 5). Tentacle (a) spirocyst, (b) basitrich. Capitulum (c) basitrich. Nemathybome (d) basitrich. Actinopharynx (e) basitrich, (f) microbasic amastigophore. Filament (g-h) basitrichs, (i) microbasic amastigophore. Scale: (a-i)= 10 urn. REMARKS. Five specimens from Aden resembled Edwardsia pudica Klunzinger, 1877, and sensu Carlgren (1900, 1931). That is, they had scattered small nemathybomes over the whole scapus containing only one type of nematocyst, and also 20 tentacles, corresponding to eight macrocnemes and six pairs of microcnemes, an arrangement unique in this genus. The three Singapore specimens were smaller and more delicate than those from Aden, but they had better developed retractor muscles in that the free edge of the retractor had a branched fold: however, this is a difference in degree (Fig. lOb). No other differences were found and since the other characters were similar, the specimens were referred to the present species. The mesenterial arrangement is unusual in that it does not follow the usual development to the simple Edwardsia stage (p. 215). Other nominal species having this arrangement are E. gilbertensis Carlgren, 1931, E. stephensoni Carlgren, 1950& and E. bocki Carlgren, 1931. E. gilbertensis would seem valid since its nemathybomes are arranged in eight rows and the retractors poorly developed in contrast to the arrangements in E. pudica. The descriptions of both E. stephensoni and E. bocki closely resemble that of E. pudica suggesting that the three are conspecific. In the first two of these nominal species the retractors of the macrocnemes as illustrated by Carlgren are strong and well developed and the individual muscle folds follow the same plan as in E. pudica. That is, there are several unbranched folds followed by one or more branched folds, the pattern repeating itself along the muscle. The nemathybomes in both species are small, and scattered over the scapus. Also, the nematocysts are of the same type and have the same size range as in E. pudica. Carlgren did not record the presence of microbasic amastigophores in the actinopharynx, though they were scarce. The only description of E. bocki was made before nematocysts had been classified. E. stephensoni is conspecific with E. pudica and so apparently is E. bocki. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 229 Two of the five nominal species described by Bourne (1916) from Papua New Guinea appear similar to E. pudica. These are E. vermiformis and E. rakaiyae. Bourne stated: 'E. vermiformis . . . the muscle banners which were evidently large, with complicated, much branched folds . . . character- istic edwardsian structures (nemathybomes) are scattered over the areas between the longitudinal grooves of the scapus'. 'E. rakaiyae . . . micromesenteries ... 12 in number, two each sulco-lateral, lateral, or sulculo-lateral inter-mesenterial spaces'. Although Bourne's illustrations are inadequate for identification, his descriptions clearly include characters of E. pudica. Bourne himself suggested that E. vermiformis was similar to Klunzinger's figure, but without looking at the tentacles, disk and actinopharynx of the specimen it would be difficult to be certain. The well developed retractor muscle of E. pudica is the largest known of any Edwardsia species. Therefore, the presence of much branded mesogloeal folds of the retractors and the nemathybomes scattered over the column indicate that E. vermiformis might be referred to E. pudica. From the details given by Bourne of E. rakaiyae there is little doubt that the species is identical with E. pudica. The presence of pairs of microcnemes between the main mesenteries, the 20 tentacles arranged 8+12, the enormously developed retractors, and the nemathybomes scattered over the column, are all diagnostic. Hence E. rakaiyae is referred to E. pudica. A search for Bourne's material was made by R.L. Manuel at Oxford University but only one of his original slides, of E. mamillata, was found. E. mamillata seems not closely related to E. pudica since the retractor muscle is very weak. If the material were found it might still be possible to identify the species from the nematocysts in the nemathybomes and the arrangement of mesenteries. Genus SCOLANTHUS Gosse, 1 853. Scolanthus Gosse, 1853: 157; Manuel, 1981: 266. Edwardsia: Gosse, 1860: 254; Fischer, 1888: 22; Carlgren, 1931: 2. Isoedwardsia Carlgren, 1921: 56; Carlgren, 1949: 24. Alfredus Schmidt, 1979: 212. DEFINITION. Edwardsiidae with column divided into capitulum, scapulus and scapus. Proximal part of body rounded and like rest of scapus with nemathybomes scattered or in rows. Scapus with more or less well developed cuticle. Tentacles 16 — 20. Siphonoglyph ventral, feebly developed. Mesenteries, retractors and parietal muscles as in Edwardsia. Ciliated tracts of filament sometimes discontinuous. Cnidome: spirocysts and basitrichs; possibly also microbasic amastigophores in one species. TYPE SPECIES. Scolanthus callimorphus Gosse, 1853: 159, by monotypy. REMARKS. Manuel (1981) suggested that species having microbasic amastigophores and once referred to the genus Isoedwardsia, itself referred to Scolanthus, should be placed in a separate genus. The presence or absence of a type of nematocyst is usually considered to be a character of generic or higher importance. Certainly S. armatus (see below) does not have microbasic amastigophores and, from published data on the six species of Isoedwardsia originally included, only /. nidarosiensis Carlgren, 1942: 61, has this type of nematocyst. The descriptions of the other species were published between 1920 and 1931 at a time when Carlgren and other workers did not differentiate between types of nematocysts. /. nidarosiensis is provisionally referred to Scolanthus until the type specimen can be re-examined. Scolanthus armatus (Carlgren, 1931) comb. nov. Figs 13-14 Edwardsia armata Carlgren, 1931: 2-4, figs 1-2; Carlgren, 1949: 24. MATERIAL EXAMINED. Singapore: Pulau Hantu, 2 specimens, 19 Sept 1970; Pasir Panjang and Pungol Point, 1 specimen each location, 11 Feb 1971, BMNH 1983.4.8.10-13. Great Barrier Reef: Three Isles, 1 specimen, University of Queensland Expedition, 1973. DESCRIPTION. Column divided into capitulum, scapulus and scapus. Proximal end of scapus 230 K. W. ENGLAND Fig. 13 Scolanthus armatus comb. nov. (a-b) Transverse section of retractor muscles, BMNH 1983.4.8.10-11. (c) Same specimen as (a), transverse section of parietal muscle. Scale: (a) = 100 um, (b) = 20um, (c) = 40um. (m mesentery, f filament, Stippled, ectoderm; black, mesogloea; white, endoderm.) Table 6 Size ranges of cnidae of Scolanthus armatus comb. nov. (in um). Localities of specimens Fiji Location/ Singapore Singapore Singapore Edwardsia armata Typeofcnida (BMNH 1983.4.8.10) (BMNH 19834.8.11) (BMNH 1983.4.8.13) (after Carlgren 193 1:2) Tentacle Spirocyst 18-8-27-1 x 1-2-1-8 13-0-23-6x1-8 13-0-24-8x1 -8-24 1 9-0-26-0 x 2-0-2- 5 Basitrich 47-2-87-3 x 3-0-4-1 42- 5-81 4x3-0-4-1 47-2-70-8x4- 1-4-7 52-0-72-0 x 3-0-3-5 Capitulum Basitrich 1 1-8-15-3x2-4 Nemathybome Basitrich 43-7-76-6x4- 1-5-8 47-2-70-8x4- 1-5-8 63-7-69-6x4-1 40-0-70-0x3- 5^-5 Actinopharynx Basitrich 24-8-31-9x2-4 21 -2-28-3 x 2-4-3-0 28-3-30-7 x 2-4 1 7-0-24-0 x 2-0-2- 5 Basitrich 43-7-64-9 x 2-4-4-7 43-7-51-9 x 4-1-4-7 38-9-47-2 x 4-7-5-8 34-0-41 -Ox 4-5-5-0 Filament Basitrich 23-6-33-0 x 2-4-3-0 22-4-29-5 x 3-0 27-0-36-5 x 2-4-3-0 Basitrich 39-0-48-3 x 3-5^4-7 25-0-47-2 x 3-5-4-7 33-0-49-6 x 4-7-5-8 ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 231 \J b g h Fig. 14 Scolanthus armatus comb, nov., nematocyst signature (see Table 6). Tentacle (a) spirocyst, (b) basitrich. Capitulum (c) basitrich. Actinopharynx (e-f) basitrichs. Filament (g-h) basitrichs. Scale: (a-h)=10nm. normally rounded but may be flattened by attachment to stones. Scapus with small, prominent, but irregularly scattered nemathybomes extending to proximal tip of column. Scapulus indistinct, without nemathybomes. Capitulum distinct, mesenterial insertions into endoderm visible through transparent wall. Tentacles up to 16, long, slender, capitate. Mouth at top of cone. Column in preserved specimens 45 mm high, 5 mm wide. Anatomy. Eight macrocnemes and up to eight microcnemes, developing in normal Edwardsia pattern. Microcnemes of dorsal exocoels developing last, normally in pairs but both of a pair not always arising together. Retractors of macrocnemes strong, diffuse (Fig. 1 3a-b), outer part developing as extension of mesentery whence muscle folds arise on both sides of mesogloeal sheet. Microcnemes small, confined to region of capitulum. Parietal muscles well developed. Cnidome. See Table 6 and Figure 14. Colour. Capitulum translucent with red markings, scapus light orange-brown with white nemathybomes. Tentacles translucent with white capitate tips. DISTRIBUTION. Fidschi Island, Viti Levu, Fiji (Carlgren, 1931). Singapore: Pulau Hantu, Pasir Panjang and Pungol Point; Great Barrier Reef: Three Isles (present material). HABITAT. In coarse to muddy sand, MLTL. Often with other edwardsiid species, for example Edwardsia hantuensis sp. nov., Edwardsianthus pudica comb. nov. and Edwardsianthus gilbertensis comb. nov. 232 K. W. ENGLAND REMARKS. The present material resembles Edwardsia armata Carlgren, 1931, in having irregularly scattered nemathybomes, 1 6 tentacles with prominent capitate tips bearing large nematocysts, and an identical mesenterial arrangement. The size range and types of cnidae given by Carlgren also correspond although in the actinopharynx the nematocysts of the present material were larger than in that studied by him. However, the size range of cnidae can vary a little within any species (Table 6) and the range may be usual in this species. Carlgren assumed that the rounded lower end of the column of this species was a physa, an assumption easily made when the cuticle is missing. However, sections of the present material revealed no physa. The nemathybomes and cuticle, reduced in thickness, were present almost to the base of the proximal end. Such specimens cannot be referred to the genus Edwardsia and should be placed in the genus Scolanthus ( = Isoedwardsid) (see Manuel 1981). There seems to be no other species referred to the genera Isoedwardsia and Scolanthus which these specimens resemble. None has capitate tips, and the cnidae are different in size (Table 7). The present specimens are, therefore, referred to S. armatus. I am grateful to Mrs U. Canfield for drawing my attention to the absence of a physa in this species. Table 7 The species of Scolanthus. Dimensions of Number of Nemathybome nematocysts from Species Tentacles arrangement nemathybomes (um) Source of data 5. armatus (Carlgren, 1931) 16 scattered 43-7-76-6x4- 1-5-8 p. 230 '5. callimorphus Gosse, 1853 16 scattered 60-0-90-0 x 3-0-5-0 Manuel, 1981:265 S. curacaoensis Pax, 1924 16 scattered 38(43)-53 x 2-5(3) Carlgren, 1931:26 5. ignotus (Carlgren, 1920) 17-20 scattered 34-0-48-0 x 3-0-3-5 Carlgren, 1920: 149 S. ingolft (Carlgren, 1921) 16 scattered 50-0-60-0 x 4-0-5-0 Carlgren, 1921:56 ? S. nidarosiensis (Carlgren, 1942) 16 scattered 62-0-67-0 x 2-5-3-0 Carlgren, 1942:61 'Type species of genus. Tribe THEN ARIA Carlgren, 1898 Family ALICIIDAE (Duerden, 1 895; Duerden, 1 897) Genus 77?//fC77,S Klunzinger, 1877 Triactis Klunzinger, 1877: 85; Carlgren, 1949: 44; Carlgren, 19506: 433. Viatrix: Haddon & Shackleton, 1893: 127 (part). Hoplophoria Haddon, 1898: 438 (part). Phyllodiscus: Stephenson, 1922: 280 (part); Carlgren, 1940a: 31. DEFINITION, (sensu Carlgren, 1949: 44). Aliciidae with well developed pedal disk. Column divided into capitulum and scapus, sometimes scapus with small vesicles below margin. Margin with stalked outgrowths (peduncles) which in young specimens occur sparingly and are little branched, in older ones close together and dichotomously branched. Hemispherical mastigophoral spherules on stalks near branches of peduncles and in fosse. Stalks few, with longitudinal weak bands of endodermal muscle. Sphincter weak or absent. Tentacles hexamerously arranged, distally with spots as on column. Longitudinal muscles of tentacles and muscles of oral disk ectodermal. Two siphonoglyphs. Six pairs of perfect mesenteries and several imperfect pairs. Two pairs of directives. Retractor and parietobasilar muscles weak. Cnidome: spirocysts, basitrichs, microbasic and macrobasic amastigophores, possibly also microbasic b-mastigophores. TYPE SPECIES. Triactis producta Klunzinger, 1877, by monotypy. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 233 REMARKS. Carlgren (1 949: 44) stated that the outgrowths of the column occurred in the middle, but later he stated that 'the upper part of the column (the capitulum) above the outgrowths is more thin walled than the lower part, and somewhat narrower just above them, and has ectodermal muscles and groups of spirocysts' (Carlgren, 19506: 433). Two comments on Carlgren's views can be made. Firstly, the peduncles are outgrowths of the margin and not of the middle of the column, so that Carlgren's later statement is correct. The difference in thickness between the column wall and the capitulum is shown in Figure 15. Secondly, the presence of spirocysts in the capitulum was not confirmed. Examination of serial sections from two specimens did not reveal any spirocysts in this region and they occurred only in the tentacle tips. The nematocysts of the capitulum are microbasic amastigophores and basitrichs (microbasic b-mastigophores?). However, the presence of weak ectodermal muscle in the capitulum is confirmed. The generic definition of Triactis is modified from that given by Carlgren (1949) to reflect the position of the peduncles and the presence of a capitulum. His reference to vesicles on the pedicels and peduncles is amended to spherules (see p. 210). Carlgren (1949) specified no sphincter but in one of the present specimens a weak sphincter was apparent, recalling that depicted by Haddon (1898: 439, text fig.) in Hoplophoria cincta. However, this may have been due to the influence of other muscles on the degree of folding of the circular muscles of the column. In a second specimen no sphincter was found. Triactis producta Klunzinger, 1877 Figs 15- 16 Triactis producta Klunzinger, 1877: 85, p. 6, fig. 8. Viatrix cincta Haddon & Shackleton, 1893: 127. Hoplophoria cincta Haddon, 1898: 438-439, pi. 23, figs 1 1-15. Phyllodiscus indicus Stephenson, 1921: 561, fig. 18; Stephenson, 1922: 280; Stephenson, 1928: 18, fig. 1 1. Phyllodiscus cinctus Stephenson, 1922: 280; Stephenson, Stephenson & Tandy, 1931 : 38; Carlgren, 19400: 31, fig. 8.2. Triactis cincta: Carlgren, 1945: 7; Carlgren, 1947: 14; Carlgren, 19506: 433. MATERIAL EXAMINED. Aden: 14 specimens, 17 Jul 1966, BMNH 1983.4.8.27-30. Singapore: Pulau Semakau, 42 specimens, 22 Feb 1970, BMNH 1983.4.8.14-24; Pulau Biola, 2 specimens, 13 Sept 1970, BMNH 1983,4.8.25-26. DESCRIPTION. Column with scapus and capitulum, definite margin, and deep fosse. Margin with peduncles each branched into several pedicels, the latter each terminating in or capped by a small mastigophoral spherule. Large mastigophoral spherules on upper surfaces of peduncles close to branches. A second mastigophoral spherule may occur at foot of each peduncle, in fosse, especially near directive axis (Fig. 15.) Major outgrowths up to twelve, corresponding to primary and secondary mesenterial cycles. Between major outgrowths smaller ones may be present. Column mostly smooth but small vesicles may occur over 1st- and 2nd-cycle endocoels though their appearance may be due to contraction. Capitulum either withdrawn so that tentacles lie over or among the outgrowths, or fully extended to nearly same height as rest of column. Tentacles up to 48, long, slender, tips blunt, surface granular due to small groups of nematocysts, their size increasing thickness of ectoderm. Disk small, central mouth on top of cone. Height of column to 7 mm, diameter across peduncles 8 mm. Anatomy. Mesenteries hexamerously arranged in three cycles, two pairs of directives supporting siphonoglyphs. 1st cycle perfect, fertile. Retractors weak, diffuse, mesogloeal branches thick, with few muscle folds. Parietobasilar muscles weak absent. Basilar muscles present. Weak sphincter found in one specimen but indistinct or absent in all others. Longitudinal muscles of tentacles and radial muscles of disk ectodermal; in sections folds short, mostly unbranched. Ectodermal longitudinal muscle processes on capitulum. Zooxanthellae present in endoderm of column, disk, peduncles and spherules. Cnidome. See Table 8 and Figure 16. The size ranges of certain types of nematocysts were found to differ considerably between groups of specimens collected from a single locality, but the cnidome was constant. Nematocysts in tentacles, column and actinopharynx labelled 'b-mastigophores' may be a form of basitrich. Colour. Column translucent pale brown with a much darker overlay of small 234 K. W. ENGLAND n Fig. 15 Triactis producta, BMNH 1983.4.8.27. Vertical section through part of column and crown showing two mastigophoral spherules, one on a pedicel from the margin (to left) and one in the fosse (to right), a tentacle, the disk and adjacent actinopharynx. Scale: = 500 urn. (act actinopharynx, ca capitulum, fo base of fosse, ma margin, n nematocysts, sph sphincter, sp spherules, t tentacle. Stippled, ectoderm; black, mesogloea; white, endoderm.) spots on its lower part grading to continuous beneath peduncles. Capitulum dark brown with pale translucent vertical lines. Tentacles translucent mottled with white, disk pale brown, mouth orange-brown. Upper surfaces of peduncles and fosse with iridescent blue-green sheen over brown ground colour. Spherules white on outer surface, but golden orange within from zooxanthellae. DISTRIBUTION. Known from several widely spaced localities: Red Sea (Klunzinger, 1 877), Maldives (Stephenson, 1921, 1922, as Phyllodiscus indicus), Torres Straits (Haddon & Shackleton, 1893 as Viatrix cincta), Great Barrier Reef (Carlgren, \95Qb as Triactis cincta), Aden and Singapore (present material). HABITAT. Usually reported on scleractinian coral, both living and dead. The specimens from Singapore were on Pavona frondifera Lamarck (det. Dr S. H. Chuang, University of Singapore). Those from Aden were on similar coral fronds, at depth 2 m, there being several specimens on each. REMARKS. The presence of branched outgrowths with large mastigophoral spherules on the peduncles (p. 211), and of similar spherules in the fosse, suggests that these specimens should be referred to a species of Triactis. The specimens from Aden and Singapore differed slightly, especially in nematocyst size, but the differences are considered too small to justify recognition of two taxa. Klunzinger ( 1 877) proposed the genus Triactis to accommodate T. producta, a species from the ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 235 Red Sea. His description and illustrations of the species were clear and the specimens from Aden can be referred to it with confidence. Haddon & Shackleton (1893) collected similar specimens from the Torres Strait and pro- visionally referred them to the genus Viatrix Duchassaing & Michelotti, 1866, under the specific name V. cincta Haddon & Shackleton, 1893. Later, following statements by McMurrich (1893, 1896), Haddon (1898) referred the species to the genus Hoplophoria Wilson, 1890. According to \J \J h f\ q m Fig. 16 Triactis producta, nematocyst signature (see Table 8). Tentacle (a) spirocyst, (b) basitrich, (c) microbasic b-mastigophore, (d-e) microbasic amastigophore. Column (f) basitrich, (g) microbasic b-mastigophore, (h) microbasic amastigophore. Spherules, major (i) basitrich, (j) microbasic b-mastigophore, (k-1) microbasic amastigophores, (m) macrobasic amastigophore. Spherules, minor (n) basitrich, (o) microbasic b-mastigophore, (p) microbasic amastigophore. Actinopharynx (q-r) microbasic amastigophores. Filament (s-u) microbasic amastigophores. Scale: (a-u)= 10 urn. 236 K. W. ENGLAND m (N ' Ooosor^soso^ OO SO O r^ O^Osjl22stiS oo (N | 1 ~i (N m o 1 l~"'"2sI~'"r'4AA •— — f*"} ON ^^ — E.1 fN — i so sO r*^ so irj rt (N ON O v^ ^ ^ ccrr^^-dNoc cb^IT a| Tl- *~> f> fS| ^hrnrNlrnrnooON ON XX X X l~~ 1^ X X X X X P— X c sb -5 u oT 91 ON o c — O 0 p so O so r- j_ CL. oo t i •^o v^i m •"t x •* -ri- •* sb — ro ^^ 10 so O so O O O 'S (SI c*5 00 fN •— r^i rn^o — ^rSfNtN^sbob SJ "o ON XX X X sO x ^^ X X X X X X •S a> ~" SO ON O sot — ^CN| ONONm^-int — r^ •~ X t — ON fN SO sb "" ^r ON ON f*"l VN — ON ^ "re a si II ~T so O f> J- re bo >j fr.y x> -s •^ o — S re reOcsO ^reoore^^-creooreO-aOS1^ .CjOO^OO uji-— j 00, L uri---r\ BOo m -" 60 •— o - re w 8 S -1 f- a •& j -4 ? s on O3 2 xi ^ U 'i & b'l §"'1 & 'i o 's o 'i §* • — _o 003.0 oox) oo>-xi soja 00^1 °° 1/5 ° '5 "s. ° ''S ° ''S s ° '5 ° '« ° ''S f -y 1 1 -1 1 -1 1 I -I 1 1 1 'I 1 •X ca * « 238 K. W. ENGLAND Haddon the name Hoplophoria was introduced, but not defined, by Wilson ( 1 890) to accommodate an actinian collected in the Bahamas. McMurrich, however, wrote: 'It seems fairly certain that this species is identical with Viatrix globulifera originally described by Duchassaing & Michelotti (I860)'. If McMurrich's statement were correct then the genus Viatrix Duchassaing & Michelotti, 1 866, would take precedence over Hoplophoria Wilson, 1 890, so why Haddon ( 1 898) employed the later name for the genus is not clear. Still, Haddon's description was adequate to refer his material to the genus Triactis (sensu Carlgren, 1947) and to identify the specimens from Singapore, shown here to be conspecific with T. producta (see below). Carlgren ( 1 9506), after studying material from Low Island, suggested that Triactis cinctus and T. producta were conspecific. This suggestion is confirmed from the present study of specimens from Aden and Singapore. The species name T. producta Klunzinger, 1877, has priority over T. cincta (Haddon & Shackleton, 1893) and is employed here. Phyllodiscus indicus Stephenson, 1921, is here considered conspecific with Triactis producta. When first described the species was referred to Phyllodiscus Kwietniewski, 1897, but only provisionally since the anatomy of Triactis was incompletely known. Stephenson included in the same genus P. semoni Kwietniewski, 1897: 407, and Hoplophoria cincta (Haddon & Shackleton, 1893: 127). Stephenson's description of/*, indicus and his illustration (Stephenson, 1921, Fig. 18) show that his concept of Phyllodiscus is identical with Triactis sensu Klunzinger. However, Phyllodiscus Kwietniewski, 1897, is regarded as valid and P. semoni is not referred to Triactis (following Carlgren, 1924: 11). Family ACTINIIDAE (Rafinesque, 1815) NOMENCLATURE. Williams, Cornelius & Clark (1982), in proposing validation of the family name Actiniidae, considered that the earliest use of this family name was by Goldfuss (1820). However, L. B. Holthuis of Leiden has since informed the International Commission on Zoological Nomenclature of an earlier use, by Rafinesque (1815: 155) (Holthuis, Opinion 1295, Bulletin of Zoological Nomenclature 42: 34-36). The family name has been validated by the Commission. Dating of the Rafinesque work follows Iredale (1911). Genus ANTHOPLEURA Duchassaing & Michelotti, 1860 Actinia (Isacmaea) Ehrenberg, 1834: 34 (part). Tarastephanus Brandt, 1835: 1 (part). Monostephanus Brandt, 1835: 10 (part). Cereus: Milne Edwards, 1857: 266. Bunodes: Gosse, 1855: 274; Gosse, 1860: 195, 198; Johnson, 1861: 302; Klunzinger, 1877: 77; Andres, 1881: 318; Andres, 1883: 423; Pennington, 1885: 167; Dixon, 1889: 310; Duerden, 1898: 454; Carlgren, 1898: 19; Carlgren, 1900: 66; Stuckey, 1908: 368; Fischer, 1889: 301 (part). Isacmaea: Milne Edwards, 1857: 288. Anthopleura Duchassaing & Michelotti, 1860: 49; Duchassaing & Michelotti, 1866: 126; Andres, 1883: 440. Aegean Gosse, 1865:41. Evactis Verrill, 19696: 470; Andres, 1883: 446. Actinioides Haddon & Shackleton, 1893: 126; Haddon, 1898: 424; Kwietniewski, 1897: 33 (389); Duerden, 1898: 453; Carlgren, 1900: 63; Carlgren, 1938: 32; Pax, 1907: 79; Pax, 1908: 490. Isactinia Carlgren, 1900: 53. Cribrina: McMurrich, 1901: 18; McMurrich, 1904: 287; Pax, 1908: 474. Bunodactis: Pax, 1920: 31; Pax, 1926: 23; Stephenson, 1921: 529; Stephenson, 1922: 271. Tealiopsis: Verrill, 1922: 1 10; Verill, 1928: 26. AnthostellaCar\gren, 1938:38. NOMENCLATURE. The generic name Anthopleura Duchassaing & Michelotti, 1860, is used here following the practice of Carlgren (1949), Stephenson (1935) and Manuel (1981). DEFINITION. Actiniidae with well developed pedal disk. Column with adhesive verrucae arranged in more or less distinct vertical rows especially in the upper part. Heterotrichs present in ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 239 column ectoderm. Atrichal marginal spherules present or atrichal spherules in fosse. Atrichs sup- pressed here in some specimens. Sphincter weak to strong, endodermal, diffuse to circumscribed. Tentacles simple, hexamerously or irregularly arranged, longitudinal muscles ectodermal or meso-ectodermal. Numerous perfect mesenteries, all stronger ones fertile. About the same number of mesenteries distally and proximally but, since mesenteries grow from base upwards in some species, a few more may be found proximally. Retractors of stronger mesenteries diffuse, sometimes restricted. Cnidome: spirocysts, atrichs, heterotrichs, basitrichs and microbasic amastigophores. TYPE SPECIES. A. krebsi Duchassaing & Michelotti, 1860, by monotypy. REMARKS. The characteristic nematocysts of the marginal spherules are a distinctive feature of the genus Anthopleura. They are usually referred to as atrichs but are possibly of two types, atrichs and heterotrichs. Anthopleura spp. lacking atrichs and heterotrichs in the spherules are provisionally assigned to the genus. Examples are Anthopleura dixoniana (Haddon & Shackleton, 1893) and A. handi Dunn, 1977, redescribed here. Species lacking atrichs in the spherules but having other characters of Anthopleura were often mistakenly assigned to the genus Bunodactis (now referred to the genus Aulactinia, see Dunn, Fu-Shiang Chia & Levine, 1980: 2078, also p. 255). The best character by which Anthopleura can be distinguished from Bunodactis is the presence of heterotrichs in the ectoderm of the column. They are not found in species referred to Bunodactis, Aulactinia or Gyractis. Their reported occurrence is listed in Table 9. It is interesting that Schmidt (1972: 91-92) did not find them in A. balli (Cocks, 1851) from the Mediterranean, nor Carlgren when first describing A. sanctaehelenae Carlgren, 1941: 4, and the generic affinity of these species needs re-appraisal. The distinctive heterotrichs of the column (see Figs 19h, 21h, 23h, and 25h) should not be confused with the atrichs that sometimes occur only at the limbus in certain Table 9 Anthopleura spp. having heterotrichs in column ectoderm. Species Nematocyst referred to as: Source of data A, annae Carlgren, 1940ft A. artemisia (Pickering in) Dana, 1846: 38 A. asiatica Uchida & Muramatsu, 1958 A . aureoradiata (Stuckey, 1 909) A. dixoniana (Haddon & Shackleton, 1893) A. elegantissima (Brandt, 1835) A. handi Dunn, 1977 A. hermaphroditica Carlgren, 1898 A. insignis Carlgren, 1940ft A. krebsi Duchassaing & Michelotti, 1860 A. kurogane Uchida & Muramatsu, 1958 A.michaelseni(Pa\, 1920) A. midori Uchida & Muramatsu, 1958 A. mortenseni Carlgren, 1941 A. nigrescens (Verrill, 1928) A. orientalis Averincev, 1967 A. pacifica Uchida 1938 A, pannikkarri Parulekar, 1968 A . rubripunctata (Grube, 1 840) A.stellula(Ehrenberg, 1834) A. varioarmataWatzl, 1922 A.waridi (Carlgren, 1900) A. xanthogrammica (Brandt, 1835) Atrichs Atrichs Holotrichs Atrichs Heterotrichs Atrichs Atrichs Atrichs Atrichs (limbus only?) Atrichs Holotrichs Atrichs Holotrichs Atrichs Heterotrichs Atrichs Basitrichs/heterotrichs Holotrichs Haplonemes Haplonemes Atrichs Heterotrichs Atrichs Carlgren, 1940ft: 4 Hand, 1955:65, Fig. 15 Uchida & Muramatsu, 1958: 1 18, Fig. 5 Carlgren, 1950c:4 England (see p. 242) Hand, 1955:59, Fig. 13 Dunn, 1977: 8 Carlgren, 1959:22 Carlgren, 1940ft: 4 Carlgren & Hedgepeth, 1952: 154 Uchida & Muramatsu, 1958: 1 14, Fig. 2 Carlgren, 1938:45 Uchida & Muramatsu, 1958: 1 12, Fig. 1 Carlgren, 1941:3 England (see p. 249) Averincev, 1967: 70 Uchida & Maramatsu, 1958: 116, Fig. 3 Parulekar, 1968: 590-595 Schmidt, 1972: 89 Schmidt, 1970:24 Carlgren, 1952:374 England (see p. 252) Hand, 1955: 52, Fig. 11 *Uchida & Maramatsu list basitrichs but illustrate heterotrichs. 240 K. W. ENGLAND anemones, but usually in large numbers, as for example in Epiactis prolifera Verrill, 18696 (personal observation), Bunodactis maculosa Carlgren, 1954, and Isanemonia australis Carlgren, 1950c. Anthopleura dixonlana (Haddon & Shackleton, 1893) Figs 17-19 Actinioides dixoniana Haddon & Shackleton, 1893: 126; Haddon, 1898: 424-426, pi. 22, fig. 6, pi. 27, figs 1-2. Actinioides papuensis Haddon, 1898: 426-428, pi. 22, fig. 7, pi. 27, figs 3-7. Actiniogeton papuensis: Carlgren, 1949: 62. Anthopleura dixoniana: Carlgren, 1938: 32; Carlgren, 1949: 54. MATERIAL EXAMINED. Singapore: Changi Creek, 50 specimens, 16Feb 1970 BMNH 1983.8.4.1-50; Pungol Point, 26 specimens, 19 May 1970, BMNH 1983.8.4.51-77; Pasir Panjang, 9 specimens, 24 Jun 1970, BMNH 1983.8.4.76-85. Torres Strait: Mabuaing, 2 specimens, coll. A. C. Haddon, Cambridge University Zoological Museum, Co 177 (in jar labelled Actinioides sesere). Specimens also examined from Hong Kong and Gan, Addu Atoll, Maldives. DESCRIPTION. Column tall, height about twice diameter when extended, upper part with vertical rows of small prominent verrucae, each row topped at margin by a large spherule that may or may not be atrichal; but some specimens lacking atrichal marginal spherules. Spherules simple or compound with nematocyst batteries, when present, located on top and inner faces. Verrucae extending onto outer surfaces of spherules. Tentacles varied in number, up to 60, slender, tapered, in two circlets, inner longer than outer. Disk flat; base present. Anatomy. 3-8 siphonoglyphs seen, each supported by directives. Number of mesenteries varied but most are perfect; number of tentacles same as number of mesenteries at base of column (Table 10). Typical arrangements of mesenteries in apparent cycles are shown in Figure 17. Mesenteries apparently in three cycles but not hexamerously arranged, all or most of the 1st cycle being directives. Retractor muscles weak- diffuse; parietobasilar muscles weak with free inner edge. Sphincter weak-diffuse to circumscribed and varied between specimens (Fig. 18). Marginal spherules endocoelic; mostly atrichal, some lacking atrichs interspersed irregularly with atrichal spherules. Zooxanthellae present. Largest specimen, height 10 mm, diameter 10 mm. Cnidome. See Table 1 1 and Figure 19. Colour. Column either pale green to yellowish green near limbus, becoming grey towards margin, covered with small yellow spots; or deep brown near the limbus becoming pale greenish white above with white vertical lines. Marginal spherules green or brown with orange-brown or golden-yellow spots or patches (atrichal regions). Verrucae grey. Tentacles with brownish tint on upper surface broken by 2-4 translucent white elliptical patches. Disk brown, tinged with green and yellow around the mouth and with cream or white patches over the endocoels. Table 10 Anthopleura dixoniana. Relation between numbers of tentacles, mesenteries, directivies, siphonoglyphs and spherules. Each line of the table represents one specimen. Mesenteries Directive Tentacles at base mesenteries Siphonoglyphs Spherules 18 9 pairs 3 pairs 3 8 21 10 pairs 4 pairs 4 11 34 1 7 pairs 5 pairs 5 38 19 pairs 5 pairs 5 20 39 1 9 pairs +1 6 pairs 6 20 42 22 pairs 5 pairs 5 19 62 5 pairs 5 20 ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 241 Fig. 17 Anthopleura dixoniana Complete mesenterial arrangements of 3 specimens, (a) BMNH 1983.8.4.1; (b) BMNH 1983.8.4.2; (c) BMNH 1983.8.4.4. (black rectangles - directive mesenteries.) Fig. 18 Anthopleura dixoniana Vertical sections of two specimens to show variation of sphincter muscle, (a-b) Singapore, BMNH 1983.8.4.51; (c-d) Singapore, BMNH 1983.8.4.1. Scale: (a-b) = (c-d) = 30 nm. (Stippled, ectoderm of fosse; black, mesogloea; white, endoderm.) DISTRIBUTION. Recorded from Torres Strait (Haddon & Shackleton, 1 893, as Actinioides dixoniana and A. papuensis), the Maldives, Singapore, and Hong Kong (present material). HABITAT. In holes in rocks and stones; many specimens aggregating in a small area, but not close to each other; upper shore. REMARKS. The presence of atrichal marginal spherules, verrucae on the column, and a uniform number of mesenteries at margin and base, place this species in the genus Anthopleura. The varied 242 K. W. ENGLAND O? r- p^^sO ^ O oo TtsOO od m> fN fN r*") fN mi TJ~ in mi so •4" OO Op O OO OO 4 SO OO fN 4 O oo fN — — m — — fN m — rr fN m ON XXXX X XXXXXXX SO — fN SO f"^N f^ s^ fvj ^^ ^^ ^-j- — •- X — ob m >n m ob — ON in so ^ o *-7 fN — — — — — fN — — fN — fN 2 OfNOOO TJ- sOOsOsOfNOOO 03 fNm OfN .OO.. ONfN in ONON O fN (^ OO P^1 ^OOfNO O OO O O OO od 44 4so OfNor~- oo4 so oo4 4 o oo fNfN fNm fN — mm — — TtTl- — fN fN — fN — fN II II 111 11 1 11 1 1 OO^ OfN OOOSOOO SOfN SO OOSO O OO 03 so^t — m ooor-~oo ONON m o— fN so , , ^ oooooortooo ooooo od fN — ^h •* fN m -^l- ^t r*\ *t oo^t oot^ oofNr— r- OOTJ- \o ooso ^r so oo — fN — m — — mm — fN m — m fN m ON XX XX XXXX XX X XX X X ino fNfN oi^in* •<*• — fN fN — m — fN II II 111 Illll OfN n f^ T}~ OO TfsO O(^Or^ ^r so fNso O O ^A m oo fNfN fNm fN — fNm fN m — m m m £ ON XX XX XXXX X X XX X X a. TJO TtsO O — OOO TT TJ- fNso fN SO c X oob ^in ob — r-~ >n o O mr-- m — fN— — — fM_rnmi fN fN — fN — fN ^ o o. •g IJ L_L!1J_J 11 l i ooo O'T mso-^fro O O ooso oo oo s: CS 03 OfN fN— ONONinsO fN. OO ON— O so •S c — — — — — fNfN — . — fN — — S o i/5 .J2 E *^s o a (i: a c — u m m sosomsosom so so J a 'p u od fN •^(•fNm'^l-fNm •rt- ^t- •j; a. oo fNfN fNm fN — fN m fNfN Tf — TJ- mi m *• <*- ON XX XX XXXX XX X XX X X r t | o ™— oom t^m OTTOOOO t^in m O(^ o in Q >% MM OsO m^~ sOOr^»n r*^^- ON mso r^ — f-^_- __ ,u y-^.^;>,o-^ aoo^c* oo^O^oo — 0^ •> O *; B *- JjSrO— JolS,-^-"!-^^-^— i~ 'S l. v> O •;: o g .— t/) g 'c« *- '£-•— '">-*'£, S '£ " .y 2 ~ 'w "$> .y ^ .y 2 3 R 1 £ 5c/5CQ~~dbI Ot/50a3-^~-(DOQ2 s-SjCOfflS E^ M H — 1 f— ^ 0 § ^ u; ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 243 fi\ \J Fig. 19 Anthopleura dixoniana, nematocyst signature (see Table 11). Tentacle (a) spirocyst, (b) basitrich. Marginal spherule (c) spirocyst, (d) basitrich, (e) heterotrich, (f) atrich. Column (g) basitrich, (h) heterotrich. Actinopharynx (i-j) basitrichs, (k) microbasic amastigophore. Filament (1-m) basitrichs, (n) microbasic p-mastigophore, (o) microbasic amastigophore. Scale: (a-o) = 10 um. number of siphonoglyphs supported by directives recalls the genus Actinioides Haddon & Shackleton, 1893, proposed to accommodate the then newly described A. dixoniana, A.papuensis and A. sesere. The first two were distinguished purely on sphincter character and colour. Both species were stated by Haddon ( 1 898) to have several siphonoglyphs, all supported by directives. A . sesere was separated from the others again on the character of the sphincter, but it is not clear from the description if the A. sesere material had many siphonoglyphs. The Pasir Panjang specimens had a sphincter similar to that illustrated by Haddon (1898, pi. 27, fig. 1; see Fig. 18b) in A. dixoniana, and had several siphonoglyphs supported by directives. No other species of Anthopleura has been reported with these characters, but see A. waridi (p. 250) which differs from the present material in having the atrichal spherules in the fosse and not on the margin. It is thus apparent that the specimens are referable to A. dixoniana. The Changi Creek specimens, although smaller than the others, had a sphincter similar to that in A. papuensis and can be referred to this species. The specimens from Pasir Panjang and Pungol Point had sphincters similar to those of A. dixoniana. Examination of the nematocysts of these two species, however, showed no difference in either type 244 K. W. ENGLAND Fig. 20 Anthopleura handi. (a) Vertical section through margin showing spherule and sphincter, BMNH 1983.8.24.51. (b-c) Difference between sphincter muscles in vertical section of two other specimens, (d-e) Extremes of size of nematocyst batteries on atrichal marginal spherules. Scale: (a-c)= 10 urn, (d-f) arbitrary, (n nematocysts, p pore, sph sphincter, ve verrucae. Stippled, ectoderm; black, mesogloea; white, endoderm.) or size range (Table 1 1). The sphincter, in section, could differ between specimens. A range from weak-diffuse to fairly strong restricted sphincter muscles was found in both A. dixoniana and A. papuensis. Hence A.papuensis can be confidently referred to A. dixoniana. Haddon's original specimens labelled A. sesere, preserved in Cambridge University Zoological Museum, comprised two species in one jar. Most were certainly 'A. sesere' (now referred to Gyractis excavata, p. 255), but two specimens were identified as A. dixoniana. One of these had eight pairs of directives supporting eight siphonoglyphs among twenty-six pairs of mesentries. Carlgren (1938: 32), when placing the similar species Actiniogeton sultana, suggested that Actinioides dixoniana should be referred to Anthopleura, especially since Haddon had depicted ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 245 large acrorhagi. Carlgren stated that if the marginal spherules of A. dixoniana were true acrorhagi, and if more mesenteries were present at the base than at the margin, the species should be referred to Anthopleura and the genus Actinioides would lapse. A. dixoniana clearly possesses atrichal marginal spherules (true acrorhagi), but the number of mesenteries is about the same distally and proximally. However, since this is true of nearly all species of Anthopleura a difference of only a few more mesenteries at margin or base is of no consequence (p. 214). Hence Actinioides should be referred to Anthopleura. The present specimens are, therefore, referred to Anthopleura dixoniana with which A . papuensis is conspecific. Anthopleura handi Dunn, 1977 Figs 20-21 lAnthopleura hermaphroditica Carlgren, 1898: 23, fig. 18; Carlgren, 1921: 148; Carlgren, 1927: 32; Carlgren, 1949: 54, pi. 2, fig. 2; Carlgren, 1959: 22. Anthopleura handi Dunn, 1977: 54-64, figs 1-8. MATERIAL EXAMINED. Singapore: Pasir Ris, 15 specimens, 24 Mar 1970 BMNH 1983.8.24.51-60. DESCRIPTION. Column tall when extended, mostly smooth but with small inconspicuous verrucae near margin continuing onto marginal spherules. Marginal spherules haphazardly arranged, large to small (Figs 20a, d, e). Some spherules with atrichs but usually without. Tentacles tapered, smooth, in two circlets, inner slightly longer than outer. Disk flat; mouth large; two distinct siphonoglyphs. Base slightly smaller in diameter than disk. Anatomy. Sphincter diffuse (Fig. 20a-c). Mesenteries hexamerously arranged in up to four cycles; first three cycles perfect and fertile, including directives; 4th cycle incomplete, usually without filaments. Retractor muscles weak, diffuse, restricted in contraction; mesogloeal folds much branched. Parietobasilar muscles prominent, inner edge detached from mesentery. Basilar muscles present. Two well developed siphonoglyphs supported by directives. Usually same number of mesenteries distally and proxi- mally but sometimes with additional mesenteries at either base or disk. Four specimens examined had arrangements shown in Table 12. Although verrucae often indistinct in life, in histological Table 12 Anthopleura handi. Relation between numbers of tentacles, mesenter- ies at disk and base and spherules in four specimens (BMNH 1983.8.24.51-60). Mesenteries Tentacles Disk Base Spherules 50 50 25 50 50 58 24 54 54 58 24 69 70 64 30 section structure can be seen. Imperforate cinclides located at centre of column in primary endocoels, up to 6 cinclides in a vertical row; arising as endodermal invaginations. Dioecious; broods, two embryos found in enteron of one specimen. Zooxanthellae not found. Cnidome. See Table 13 and Figure 21. Colour. Specimens from Pasir Ris were white. See Dunn (1977: 54-55) for other colour varieties. DISTRIBUTION. Malaysia (Dunn, 1977) and Singapore (present material). HABITAT. Attached to mussels in sand, MLTL. 246 K. W. ENGLAND ob s? s ON 9 A 7 m rs 7 rs £ 2 2 ob ^ O rr ob rs ON s^ / S_^ ^^n1 ^**^ >»^ ^— ^ \^s 53 X? p •n m so P m SO o | g" rn oo f*^ so OO n m rs p rs rs SO OO ON OO rs ON a O p"I ON — so rn m rs ob so ^h r^ ON ON ••3- ob — rs —i rs rs S S)5 m ON ON SO oo so I I I rs rs 3 2 "^s .S S ^— in r<~. ^V ON SO — ON m m rn M O -"t g K Si- — — — < — — — "• . rs — _ -c a ^ s? rs in ^ SO rs T^- oo p r^ 00 p op JS rs 06 Jrr oo O rs oo oo o si rn oo -41 rs rs si r*"i in 4 o oo rs rs — m — — t**t m — rs Tt — r<-i rs ro e X X X X X X X X X X X X X X X e3 U _. ro ON •n rs so m rs O O so in P- O C o X II — rs rs rs in OO so m 1 1 77 — rs o rs O ON — • rs rs so r- o — rs 1 I oo <•<•> -s: .S a rn r~- O NO r- — ON 00 — m m oo m — in Q 1/3 xti ~~ — ^^ ™^ z rs rs oo — — rs 1 1 rs C C (U U — rs 00 O rs r- 32 oo ON UH o — i ffi _ ' , _4- ON in '. ob JO .O — o j-^ O sb — ob O J c^ -iS — — ^! "^ — rs • — > — rs ~- — t/5 a> c cs u. -Si 4J at 3 C i m t- N k. ., o a " o c/5 S3 *c ^5; -—• 5 o -^ _o 2 _y -^ •g . .y ft. i_- . x ^k ^ ^ ^ fN ^ O ^'c t« -C >% o CJ "^ ^ >i"S i; Q "S "S S M "S "S 5 Q 5 QQ ro c u K. O 'C • — o "^*? CJ *C O (-; •^ 'C ' C o *r"- ^ ' C ' C 3 ^^ O '*^ ^* 0 t_ -> 0 •= S! *— ' rtj g o ~ a! o ft, »^ <-* •— ^ ?" "« Ti •— '-is _ -^ 3 •s ° 5 8. l-a'S C C/3 03 111 rdl II .5 sfl eo 11 S S S3 _2 03 ffl i 1 s I H j H* k S^l ^ 'T c ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 247 V ) Fig. 21 Anthopleura handi, nematocyst signature (see Table 13). Tentacle (a) spirocyst, (b) basitrich. Marginal spherule (c) spirocyst, (d) basitrich, (e) heterotrich, (f) atrich. Column (g) basitrich, (h) heterotrich. Actinopharynx (i-j) basitrichs, (k) microbasic amastigophore. Filament (1-m) basitrichs, (n) microbasic p-mastigophore, (o) microbasic amastigophore. Scale: (a-o)= 10 |im. REMARKS. The verrucated column and the atrichal marginal spherules suggest that this species is best referred to Anthopleura. Of the many included species only three or four have been found to brood, A. hermaphroditica Carlgren, \898),A.aureoradiata(St\ickey, 1909; see Carlgren, 1950c:4), Anthopleura sp. Atoda, 1945: 274, and A. handi Dunn, 1977: 54). In the present material the mesenterial arrangement, sphincter characteristics and nematocyst data were as reported by Dunn in A. handi. There were in addition cinclides in the centre of the column and both atrichs and heterotrichs in the marginal spherules (Table 13 and Figure 21). The size ranges of the cnidae differed only slightly from those in Dunn's account and need not be considered further. Dunn suggested that A. handi differed from A. hermaphorditica in having the sphincter diffuse, not well developed and circumscribed. Carlgren stated that the sphincter was circumscript with a tendency to become mesogloeal near its base, but his illustration ( 1 949, pi. 2, fig. 2) appears to show a circumscript sphincter. The section was probably taken from the region where the sphincter passed through the mesentery, where the muscle folds of the sphincter and mesenteric retractor or 248 K. W. ENGLAND transverse muscles might have interwoven. This might account for Carlgren's conclusion that the sphincter tended to become mesogloeal. The plane of section might account for a lamella of mesogloea between the sphincter and the column mesogloea, the sphincter in reality being diffuse. The sphincter of A. hermaphroditica is similar to that of A. handi (Figs 20a, b, c) but appears much larger. Carlgren's specimens of A. hermaphroditica should be re-examined. Carlgren (1959: 22) gave the nematocyst size ranges in two specimens of A. hermaphroditica. In one specimen they were close to those found by Dunn and myself, but the ranges of his second specimen were so different, particularly in tentacles and column, as to suggest a distinct species. Comparison of the size ranges of this specimen with those of Bunodactis hermaphroditica (sensu Carlgren, 1959: 23) suggests that it belongs to that species. The presence of atrichs, however, precludes Bunodactis. The question arises as to whether B. hermaphroditica is best referred to Bunodactis or should be considered within the scope of Anthopleura but with the atrichs of the marginal spherules absent or suppressed (p. 239). Carlgren stated that A . hermaphroditica is hermaphrodite but male gonads were not seen in the present specimen. Two specimens had young in the enteron as well as eggs on the gonads. The only differences recorded here between A . handi and A . hermaphorditica are the presence of cinclides in A. handi and its considerably fewer tentacles. Carlgren reported a maximum of 90 tentacles in A. hermaphroditica whereas Dunn found a maximum of 61 in 200 specimens of A. handi and a maximum of 69 was found in the present material. It seems best to recognize provisionally both A. handi and A. hermaphroditica until Carlgren's material can be re-examined and the identity of the latter established. Anthopleura nigrescens (Verrill, 1928) Figs 22-23 Tealiopsis nigrescens Verrill, 1928: 26, pi. 5A-B. Bunodactis nigrescens: Carlgren, 1949: 65. Anthopleura nigrescens: England, 1969: 5; Dunn, 1974a: 377-382, figs 1-4. MATERIAL EXAMINED. Holotype: United States National Museum of Natural History, L ^r. amsp sph Fig. 22 Anthopleura nigrescens (United States National Museum 209154). (a) Vertical section through margin showing position of an atrichal marginal spherule and strong circumscribed sphincter, (b) Vertical section of restricted diffuse sphincter of a smaller specimen. Scale: (a) = 200 urn, (b) = 100 urn. (amsp atrichal marginal spherule, sph sphincter. Stippled, ectoderm; black, mesogloea; white, endoderm.) ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 249 Washington D.C., 1925-1485. Paratype: American Museum of Natural History 1480. Hawaii: 2 specimens, coll. C. E. Cutress, 1955, USNM 209154; 3 specimens, coll. D. F. Dunn (pers. coll.). Hong Kong: 2 specimens, coll. B. Morton, 20 Mar 1981. India: Cochin, coll. Kuruvilla Matthew, 1968, BMNH 1983.8.4.86. Australia: North Queensland, 2 specimens, coll. R. Muffley, James Cook University, Townsville, Queensland. DESCRIPTION. See Dunn (19740: 377-382, figs 1-4). DISTRIBUTION. Hawaii (Verrill, 1928 as Tealiopsis nigrescens, Cutress, 1955 and Dunn, 19740); Hong Kong, Queensland, and India (present material). REMARKS. The species is readily identified by the heavily verrucated column, the prominent atrichal marginal spherules and the irregularly arranged mesenteries with several siphonoglyphs never supported by directive mesenteries. The spherules are normally large with verrucae on the outer face. The large nematocyst batteries are located on the crown of the spherule, often turned into the fosse, and are easily distinguished even in preserved material. Usually all spherules are atrichal. The sphincter varied from restricted-diffuse (Fig. 22b) to strong-circumscribed (Fig. 22a), and was frequently larger than indicated by Dunn (19740). The sphincter of the holotype is strong- circumscribed, similar to Figure 22a. Dunn recorded 80 tentacles in samples from a Hawaiian population but a specimen from the present material has as many as 167. Cnidome. See Table 14 and Figure 23. The size ranges are slightly larger than those quoted by Dunn, but the differences would seem so small as to indicate that only one taxon is involved. Table 14 Size ranges of cnidae of Anthopleura nigrescens (in n.m). Localities of specimens Location/ Type of cnida Hawaii 'Holotype' (USNM 1925-1485) Hawaii 'Paratype' (AMNH 1485) Hawaii (present material) Tentacle Spirocyst Basitrich Basitrich 24-0-3-0 18-0-21-6 x 1-2-1-8 14-1-26-4 x 1-8-3-0 15-6-20-4x1-8-2-4 14-4-28-8 x 1-8-2-4 12-0-14-4x1-2 18-0-24-Oxl -2-1 -8 1 4-4-274 x 1-8-3-0 12-0-13-2x1 -8 18-0-24-0 x 1-8 Column Basitrich Heterotrich 9-6-20-4 x 2-4-3-0 9-6-24-Oxl -8-3-0 16-8-24-0 x 3-0-4-0 1 3-2-24-0 x 2-4-3-0 19-2-21-6x3-6 Marginal spherule Spirocyst Basitrich Atrich Heterotrich 26-4x2-4 9-6-13-2 x 1-8 33-6-54-0 x 3-6-5-4 30-0-42-0 x 3-0-3-6 12-0-20-4x1-8 absent absent 9-6-13-2x1-2 33-6-54-0 x 3-6-5-4 31 -2^2-0 x 3-0-3-6 8-4-19-2x1-8 43-2-54-0 x 3-6-4-8 30-0-42-0 x 3-0-3-6 Actinopharynx Basitrich Basitrich 10-8-16-2 x 1-8 1 8-0-25-2 x 2-4-3-0 9-6-12-0x1-8 21 -6-26-4 x 2-5-3-0 10-8-16-8 x 1-8 18-0-25-2 x 2-4-3-0 10-8-13-2 x 1-8 21-6-28-8 x 2-4-3-0 Microbasic amastigophore Filament Basitrich Basitrich 19-0-20-0x3-6 13-2-16-9x2-4 22-8-3 1-2x3-6-4-8 19-2-24-0 x 3-0-3-6 10-8-18-0x1-8 27-6-3 1-2x3-6-4-2 19-2-20-4x3-6 (3 only) 13-2-18-0x2-4 22-8-3 1-2x3-6-4-8 19-2x3-6(1 only) 14-4-18-0x1-8 24-0-3 1-2x3-6-4-2 Microbasic p- mastigophore Microbasic amastigophore 1 8-2-23-4 x 3-0-4-0 18-0-21-6x3-6 144x2-4(1 only) 15-6-19-2 x 3-0-3-6 9-6-1 5-6x2-4-3-0 16-8-20-4 x 3-0-3-6 250 K. W. ENGLAND g i/ \J \ Fig. 23 Anthopleura nigrescens, nematocyst signature (see Table 14). Tentacle (a) spirocyst, (b) basitrich. Marginal spherule (c) spirocyst, (d) basitrich, (e) heterotrich, (f) atrich. Column (g) basitrich, (h) heterotrich. Actinopharynx (i-j) basitrichs, (k) microbasic amastigophore. Filament (1-m) basitrichs, (n) microbasic p-mastigophore, (o) microbasic amastigophore. Scale: (a-o)= 10 um. Anthopleura waridi (Carlgren, 1900), comb. nov. Figs 24-26 Bunodes waridi Carlgren, 1900: 66-67, pi. 1 fig. 17. Bunodactis waridi: Carlgren, 1949: 65; England 1969: 5. MATERIAL EXAMINED. Aden: Sapper Bay, 38 specimens, Sep 1966, BMNH 1983.8.4.87-106, 109-113. S. India: Mandapam Camp, 2 specimens, coll. C. E. Cutress, 15 Feb 1963, BMNH 1983.8.4.107-108. DESCRIPTION. Column tall with vertical rows of adhesive verrucae, large in upper part of column becoming progressively smaller towards the limbus and extending to it. Each row of verrucae ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 251 terminating at margin in a large spherule on the parapet. Inside fosse a separate atrichal spherule on each primary endocoel and usually on 2nd and 3rd cycle endocoels, but sometimes irregularly arranged. Atrichal spherules small to large, tentaculoid, sometimes located close to but always separated from the marginal spherules (Figs 24a, d, e). Tentacles slender, tapered, arranged in three or more cycles. Disk flat. Base present. Anatomy. Sphincter strong, circumscribed (Fig. 24b). asp msp Fig. 24 Anthopleura waridi. (a) Vertical section of margin and sphincter with an atrichal spherule in the fosse, BMNH 1983.8.4.89. (b) Sphincter of (a) enlarged, (c) Development of triple pairs of directive mesenteries, in transverse section of column of same specimen, (d-e) Variation in shape of atrichal and marginal spherules in vertical sections of margin. Scale: (a)= 135 urn, (b) = 45 urn, (c-e) not to scale, (asp atrichal spherule, fo fosse, msp marginal spherule, s siphonoglyphs, sph sphincter, ve verruca, 1,2,3 are mesenteries of 1st, 2nd and 3rd pairs respectively. Stippled, ectoderm; black, mesogloea; white, endoderm.) 252 K. W. ENGLAND Mesenteries regularly or irregularly arranged, in up to four cycles; 1st and 2nd cycles perfect, 3rd cycle imperfect, 4th cycle only rudimentary, in some specimens not reaching disk. Siphonoglyphs 1-5, always supported by directives; sometimes in groups of three, with or without additional siphonoglyphs in other areas of actinopharynx. Mesentery development starting basally and proceeding upwards, resulting in a few more mesenteries at base than at top of column in some specimens; but in others number is same (Table 1 5). Retractor muscles strong, diffuse, on first cycle mesenteries, sometimes restricted. Parietobasilar muscles well developed, basilar muscles weak. Gonads on all older mesenteries including directives. Filaments absent from some younger cycles. Largest specimen, in contraction, height 1 3 mm, disk diameter 10 mm. Cnidome. See Table 1 6 and Figure 25. Colour. Column dull olive-green with bright or crimson verrucae; tentacles and disk translucent crimson. Table 15 Anthopleura waridi. Relation between numbers of tentacles, mesenteries at disk and base, siphonoglyphs and spherules in four specimens (see BMNH 1983.87-106, 109-113). Mesenteries Tentacles Disk Base Siphonoglyphs Spherules 40 40 40 3 20 60 60 72 2 27 75 70 104 3 26 96 96 104 3 24 Table 16 Size ranges of cnidae of Anthopleura waridi (in um). Location/ Typeofcnida Specimens from Aden (BMNH 1983.8.4.84) (BMNH 1983.8.4.85) (BMNH 1983.8.4.86) Tentacle Spirocyst 14-4-24-Oxl -8-3-0 144-26-5 x 2-0-3-0 15-2-29-1 x 2-0-4-0 Basitrich 12-0-18-0 x 1-8-2-4 1 5-2-20-5 x 2-0-2-6 15-9-2 1-8x2-0-2-6 Column Basitrich 12-0-16-8 x 1-8-2-4 13-3-23-2 x 2-0-3-3 1 3-3-22-5 x 2-0-3-3 Heterotrich 18-0^21-6 x 3-0-3-6 1 5-2-29-5 x 4-6-6-0 20-0-26-5 x 5-3-6-0 Marginal spherule Spirocyst 1 3-3-36-3 x 2-0-4-0 17-2-33-0 x 2-0-3-3 23-2-37-0 x 3-3^-0 Basitrich 12-0-16-8 x 2-0-2-4 11-3-15-2x2-0 1 2-0-20-4 x 2-0-24 Atrich 35-0-44-2 x 5-3-6-6 35-7-47-5 x 5-3-6-6 32-4-48-0 x 4-2-6-0 Heterotrich 32-4-41 -Ox 4-6-5-3 29-8-44-2 x 4-0-5-3 32-4-444 x 3-6-4-8 Actinopharynx Basitrich 12-0-13-0 x 1-8 10-0-12-6 x 2-0-2-4 10-8-12-Oxl -8 Basitrich 21-6^24-0x2-4 21 -2-23-2 x 2-4-4-0 20-4-24-0 x 2-4 Microbasic amastigophore 18-0 204x4-2^-8 18-5 21-2 x 4-0-4-6 1 8-0-24-0 x 4-8 54 Filament Basitrich 12-Ox 1-8 Basitrich 20-4-27-6 x 3-6-4-8 23-2-33-0 x 4-6-5-3 23-0-33-0 x 4-6-8-0 Microbasic p- mastigophore 12.0-14-4x24 16-6x3-3(1 only) Microbasic amastigophore 16-8-18-0 x 3-6-4-2 1 5-9-20-0 x 3-3-4-0 1 6-6-25-2 x 4-0-5-3 ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 253 \J g \j Fig. 25 Anthopleura waridi, nematocyst signature (see Table 16). Tentacle (a) spirocyst, (b) basitrich. Spherule in the fosse (c) spirocyst, (d) basitrich, (e) heterotrich, (f) atrich. Column (g) basitrich, (h) heterotrich. Actinopharynx (i-j) basitrichs, (k) microbasic amastigophore. Filament (1-m) basitrichs, (n) microbasic p-mastigophore, (o) microbasic amastigophore. Scale: (a-o)= 10 (im. DISTRIBUTION. Zanzibar (Carlgren, 1900, as Bunodes waridi), Aden and S. India (present material). HABITAT. Aden specimens found in old shells of rock oyster (Crassostrea sp.) and barnacles. Indian specimens on pieces of stone. REMARKS. The verrucated column and atrichal spherules suggest that the material should be referred to Anthopleura. The presence of atrichal spherules in the fosse instead of on the parapet, and additional spherules on the parapet, indicate a different genus; but other characters of the species, especially the types of nematocyst, resemble those found in other species of Anthopleura. Also in some species of Anthopleura the nematocyst battery of a marginal spherule often forms a separate protuberance on the top or on the inner face (p. 210). The colour in life recalls Carlgren's illustration of 1900 (pi. 1, fig. 17). Whilst colour of a species 254 K. W. ENGLAND II II II III! II II II II Illl Illl 11 II lull ii Ilii Fig. 26 Anthopleura waridi. Variation in arrangement of mesenteries in three specimens, (a) BMNH 1983.8.4.87, (b) BMNH 1983.8.4.88, (c) BMNH 1983.8.4.89. Directive mesenteries indicated by black rectangles. is often of limited value in identification, in A. waridi it is distinctive and diagnostic. Carlgren stated that the species had forty-eight tentacles, twenty-four rows of verrucae and possessed 'pseudorandsackchen'. His illustration shows verrucae extending to the limbus. Carlgren (1949: 65) subsequently referred A. waridi to the genus Bunodactis. Examination of some poorly preserved specimens led me to identify them as Bunodactis (England, 1969: 5) (p. 209). Carlgren's description gives little additional information to assist in the identification since he did not describe the anatomy. The only published description of a species of Anthopleura matching the particular mode of increasing the number of directive mesenteries and siphonoglyphs found in A. waridi is that of A. dixoniana. It appears that when a siphonoglyph is formed the area between the directives may develop further mesenteries. The two nearest the directives have their retractors facing outwards to complete two new pairs of directives: the next pair forms between the mesenteries of the second pair, possibly also to become directives (Fig. 24c). Additional pairs (non-directives) form between pairs of directives in the usual manner. Single siphonoglyphs meanwhile divide into two or three separate siphonoglyphs (Figs 24c and 26). A. dixoniana differs from A. waridi in having atrichal spherules only on the margin and not also in the fosse; and in most of the Ist-cycle mesenteries being directives, which appear to arise in a different sequence. The sphincter of A. dixoniana is weak and varies in appearance while that of A. waridi is consistently strong-circumscribed. A. nigrescens (Verrill, 1928) also has several siphonoglyphs but they are never supported by directives as in A. waridi. The remaining species of Anthopleura are not known to have atrichal spherules in the fosse or irregularly arranged mesenteries, and need not be considered further. The present material is, therefore, referred to A. waridi comb. nov. Genus GYRACTIS Boveri, 1 893 Gyractis Boveri, 1893: 241-253. Actinioides Haddon & Shackleton, 1893: 126 (part); Haddon & Duerden, 1896: 1 (part); Haddon, 1898: 428 (part); Carlgren, 1900: 63; Carlgren, 1920: 151-153; Carlgren, 1938: 32. Actiniogeton Carlgren, 1938: 32; Carlgren, 1949: 62; Carlgren, 1954: 579-580; Dunn, 19746: 181-188. DEFINITION. Actiniidae with well developed pedal disk. Column with longitudinal rows of adhesive verrucae. Marginal spherules large, often digitate, lacking atrichs. Sphincter weak, diffuse to circumscribed. Tentacles and mesenteries regularly or irregularly arranged. Siphonoglyphs two or more, which may or may not be connected to directive mesenteries. About same number of mesenteries distally as proximally. Retractor and parietobasilar muscles weak, diffuse; basilar ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 255 muscles present. Asexual reproduction by fission probably frequent. Cnidome: spirocysts, basitrichs and microbasic amastigophores. TYPE SPECIES. G. excavata Boveri, 1893; by original designation. REMARKS. Dunn, Fu-Shiang Chia & Levine (1980: 2078) have suggested that since Aulactinia ( = Bunodactis) has been defined as being with or without marginal spherules (pseudospherules) some confusion might arise with the genus Actiniogeton, here regarded congeneric with Gyractis. Aulactinia capitata Verrill, 1 864, the type species of A ulactinia, lacks marginal spherules, as does A. incubans Dunn et al., 1980. Other species originally referred to this genus [B. reynaudi (Milne Edwards, 1857), B. mortenseni (Carlgren, 1924), B. maculosa Carlgren, 1954, and B. ( — Epiactis) novazealandica(Stephenson, 1918)] also lack marginal spherules (personal observation). Marginal cinclides are usually present, with consequent slight swelling of the margin between the mesenterial insertions, but spherules are absent (p. 210). Hence these species should not be placed in the same genus as one having prominent, often large, marginal spherules, as exist in Gyractis excavata. To resolve this problem it is suggested that Aulactinia be re-defined to include only species lacking marginal spherules. Species currently referred to Aulactinia (or Bunodactis) that have prominent marginal spherules should be transferred to Gyractis, the definition of which now accommodates both regularly and irregularly arranged species. The difference between regularly and irregularly arranged mesenteries is considered here to be a specific rather than a generic character. Carlgren's (1938, 1949) definitions of Actinioides ( = Actiniogeton), however, included only those species having irregularly arranged mesenteries. If this were the only difference between Actiniogeton and Aulactinia, then Actiniogeton could be considered congeneric with Aulactinia. Dunn et al. ( 1 980: 2077) discussed the presence of atrichs in the ectoderm of the outer tentacles of Aulactinia incubans. As they pointed out, those occur only sporadically and are about the same size as the basitrichs. It may be that such atrichs are partially discharged basitrichs in which the straight part of the thread has been discharged and then broken off. I have seen a similar artefact in species referred to other genera. It is possible that partial discharge occurs during killing and preservation, followed by the loss of the discharged portion by either chemical action or normal handling. Gyractis excavata Boveri, 1 893 Figs 27-28 Gyractis e:;cavata Boveri, 1893: 250, pi. 10, figs 3-4, 6. Gyractis pallida Boveri, 1893: 251, pi. 20, figs 1-2, 5. Actinioides sesere Haddon & Shackleton, 1893: 126; Haddon, 1898: 428, pi. 22, figs 8-9, pi. 28, figs 1-2. Actinioides sultana Carlgren, 1900: 63, pi. 1, figs 12-13; Carlgren, 1938: 33-35, figs 13-15. Actinioides rapanuensis Carlgren, 1920: 151-153, figs 8-10. Actiniogeton sesere: Carlgren, 1949: 62; Carlgren, 1954: 579-580, figs 9-11; Dunn, 19746: 181-188, figs 1-6. Actiniogeton rapanuensis: Carlgren, 1949: 62. Actiniogeton sultana: Carlgren, 1949: 62. MATERIAL EXAMINED Aden: 43 specimens, Oct 1966, BMNH 1983.8.24.1-10, 41-50; Maldives: Can, Addu Atoll, 10 specimens, 8 Mar 1970, BMNH 1983.8.24. 1 1-20. Singapore: Pulau Biola, 29 specimens, 13 Sep 1970, BMNH 1983.8.24.21-40. Hawaii: 2 specimens, coll. D. F. Dunn, 1973. Torres Straits: Mabuaing, 4 specimens, coll. A. C. Haddon, Cambridge University Zoological Museum Co 177, type material of Actinioides sesere Haddon & Shackleton, 1893. DESCRIPTION. Column short, verrucae in vertical rows, with debris and stones attached. Verrucae continuing onto large digitate marginal spherules, sometimes distorting them into frondose appearance. Verrucae small in clearly defined rows, or large irregularly arranged due to growth and distortion. Margin distinct. Fosse shallow, sometimes almost disappearing in full expansion. Disk wider than column, mouth sometimes raised on cone. Tentacles short, stout, tapered, in four or five circlets, occupying about one third of disk. Inner circlets in some specimens spaced away from third and succeeding ones. Pedal disk present. Anatomy. Siphonoglyphs 2-11. Directives absent. Mesenteries continuous from base to disk, arranged irregularly with at least 1st and 2nd 256 K. W. ENGLAND g Fig. 27 Gyractis excavata Variation of sphincter muscle within and between specimens from different localities illustrated by vertical sections, (a-d) Aden, BMNH 1983.8.24.1-10. (e-g) Can, BMNH 1983.8.24.11-20. (c-g) illustrate discontinuities noted in text (p. 258). Scale (a-d) = 25um, (e-g) = 50 um. (Stippled, ectoderm; black, mesogloea; white, endoderm; circles, zooxanthellae.) cycles perfect and usually a few of 3rd cycle; first three cycles fertile, 4th usually without filaments or gonads. Although gonads present, species habitually reproducing by fission, from base upwards. Retractor muscles diffuse to restricted-diffuse; parietobasilar muscles prominent, with- out free edge. Basilar muscles well developed (Dunn, 1974c). Sphincter muscle varied even within a single specimen, from weak-diffuse to weak-circumscribed (Fig. 27). Zooxanthellae present in ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 257 (N ~j so O ,:)• sO SO oo fN fN fN od ro r*1 44 fN i 4 i p sp 1 1 oo fN fN fN fN TT fN rn rn rn Os X X X X X X X X X "• fN sO fN OO oo rr OO sO rj- X *j~) .' i Os ob fN -~t ~~t >s~i sb z c ^ fN fN SO O 4 fN 4 fN Os 1 1 Tt OO 1 fN fN OO at UO fN fN ; ob sb O ob rh 0 **"7 O O SO fN p fN od 44 fN i 4 2 O <""> "c I oo Os fN fN X X fN X fN X X fN f> X X o Ci X "" Tj- I/-) p — r~ m •^- Tj- o X sb Os ob •^f- T^J- Tt" fN fN sb z c 5 fN — 44 in fN fN fN oo ~T i oo o X fN fN C3 .y. "~t~ ^t~ ob '. ob o OS sb ob ob O S fN fN M SO so fN r^i rn f*1 od 44 ^ O SO O so ^ rn OO fN fN fN fN fN 4- fN fN fN rr X X X X X X X X X X 1 — ' IT) fN so p r^ 00 O •* O oo X •— ' O fN OO fN >/•> fN fN fN ^_i — < fN fN ~^ r'l »— I fN S s 44 «A 1 1 OO fN I OO 1 1 oo oo (i 1 "O -£. m fN oo Os 00 0 Os O rh Os fN fN rsi p p fN Tt so p fN fN m r^ fN ro fi od 44 p sO 44 4 oo fN fN fN rn rh fN rn fN rn Os X X X X X X X X X — O fN O 0 ^ oo •^- rj- oo X ij- Os ob ^c sb sb fN •^c fN 3. g| fN — 44 fN fN fN fN OO OO OO fN fN I fN c <£ Tt fN rn . Os fN fN O fN — fN — Os a a c/) w C , — k %i g ~ SO O ^ O fN OO •2 fN rn m fN m tj O. od i 4 fN Tf 2 C/) u- f! OO fN fN fN fN ^h fN fN fN rn •j^> o OS X X X X X X X X X o U5 "~ o p O OO •<* OO •* Tj- fN o 'S X •<3- ob fN — ob fN fN sb fN sb fN — m 2; fN u CJ u Jj? vi OO SO 1 1 OO -T i CD rs '£ O <. 0 x3 O O p, Wj o o Si J3 .5? „ o ^ ^ 8'5 s: i "£, « — O '^ 2 '.n O -4-1 — 1 H. L) C -H 2'S.S C 01 03 5 'S3 s ^ III II ii c/5 c/5 S: ca ns ^ 03 09 si 2 rt H O >> -l H K u X s; 258 K. W. ENGLAND \J b \J c n\ u \J i Fig. 28 Gyractis excavata, nemotocyst signature (see Table 17). Tentacle (a) spirocyst, (b) basitrich. Column (c) basitrich. Actinopharynx (d-e) basitrichs, (f) microbasic amastigophore. Filament (g-h) basitrichs, (i) microbasic p-mastigophore, (j) microbasic amastigophore. Scale: (a-j)= 10 urn. endoderm of tentacles, disk, upper column and filaments. Cnidome. See Table 17 and Figure 28. Basitrichs of spherules of same size range as those of column. Microbasic amastigophores of actinopharynx scarce. Colour. Column flesh at base grading upwards through green to grey near margin. Verrucae bright green or pink. Disk and tentacles of three colour morphs: (a) disk grey, tentacles brown with white subterminal band; (b) disk green, tentacles brownish, lacking subterminal band; (c) disk and tentacles patterned green with irregular pink markings. DISTRIBUTION. Reported from Torres Strait (Haddon & Shackleton, 1893, as Actinioides sesere), Sri Lanka (Boveri, 1893), Zanzibar and Durban (Carlgren, 1900, 1938, as Actinioides sultana), W Australia (Carlgren, 1954, as Actiniogeton sesere), Hawaii (Dunn, \914b as Actiniogeton sesere), Aden, Gan and Singapore (present material) and Fiji (J. S. Ryland, pers. comm., del. D. Fautin). HABITAT. Numerous in rock crevices on exposed outer edges of reefs, and many carpet large areas. At the Aden and Gan localities many specimens packed tightly together into cracks in a red granite or basalt. In Singapore on dead coral. REMARKS. The specimens from Aden and Gan seemed to be of two species. Those from Aden, while having the same coloration and general form as those from Gan and occupying a similar habitat, were more delicate. The verrucae of the Aden population were smaller and arranged in vertical rows. The spherules too were smaller and not distorted by verrucae; and the sphincters were much weaker (Fig. 27). On the other hand the sphincter was of the same pattern throughout and the apparent interpopulation differences could be found in a single specimen. The muscle folds of the sphincters from both localities were often discontinuous (Fig. 27c-g), but this was not considered an artefact. All specimens of the samples examined had several siphonoglyphs without directive ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 259 fN 1 o" O O oo 1/i "O tN s: ON -— fN fN fN "2 1 -a S X X X X 'C vX» ;2 C p p O O £s ^ S a fN — r-- s cd 11 "i ^ eS p p tN NO p 4 'c C ^— ^^ ^— "^ 3 Q o a *) O m ON 0 (N a i*. rn i^~) l/~i c 1 p si 4 SO i 'gj ~ — * fN fN tN fN ^~ tN fN (N •* c § ~o c X X X X X X X X X *3 ~ M C ^ N— ' -* in o m f~l -.3- O NO m O c £?•- n ON NO ON ON ob f^i ro NO NO cC O • « U- — fN «_! tN — r^ — fN ^^ •s It OO fN oo m p I \ m m S t] cd . ^£Q O rn ON ob S fN 2 2 c tn' c« O o 1 ^j 00 O CJ fc fN ^b C ii? \ m OO up p O fN fN c o a; ON fN fN <^~i in fN -<;f -^T cd /"~v o ^^ X X X X X X X *• S 'S c O 0 0 O en O O •d a. f^ f^ ,3. .1. ob ob ob CJ .^ •2 ab ^_ p^ fN fN — fN tN > O •S ~ \ p O \ op p 4 'S x y. : — 2: — 2 — fN '. tN O C g 13 /b c ^"* 0 /•? T3 o T3 GO . ^ D. c« C fee (S O o O & a o cd £ U cd — _e c C'.e •aa cd ^ y 2 .y -g. i s-s o If ^ — C- >i.O o 5! U n ^o U CJ Q O j^ OXJ X °° OO 0 C " •2 o 3 8. -~- o _*j 2 '5. « |i "I---S t °3 O '-3 ^ 03 03 S -s1 2 2 111! H ^ cj O H 0 >, -J H ^ a s; Z 260 K. W. ENGLAND mesenteries, and the size ranges and types of cnidae present were identical. The specimens from Singapore had the same variations as those from Aden and Gan. Boveri (1893: 250-252) described two species of sea-anemone which differed only in the type of sphincter, and because they were not bilaterally symmetrical he proposed a new family, the Holoactiniidae, to include them. He proposed the genus Gyractis to accommodate the two species and named them G. excavata and G. pallida. His definition of Gyractis included a verrucated column and digitate marginal spherules, and the tentacles were said to occupy approximately half of the area of the disk with a clear gap between the 2nd and 3rd cycles. According to Boveri's description these two species lacked directive mesenteries, as the pairs of retractor muscles always faced each other: hence the absence of bilateral symmetry. He stated that there were no siphonoglyphs but if each were little differentiated from the rest of the actinopharynx there might nevertheless have been many. Boveri's illustrations of both species (1893, pi. 10, figs 1-6) show the whole animal, the arrangement of the tentacles and of the sphincter muscles, which together with the absence of directive mesenteries conform to the description given here. The colour varieties described by Boveri encompass those recorded above. The present material is thus referred solely to Gyractis excavata Boveri, 1 893, the first of his two species. The second species, G. pallida, is here considered conspecific with G. excavata, since it has been shown that the sphincter can vary within a single specimen (Fig. 27). Under the first reviser principle the name excavata is given priority over pallida. Though the description of Actinioides sesere (Haddon, 1 898: 428) was incomplete the verrucated column, the capitulum provided with well defined conical acrorhagi, and the absence of gonadial grooves, show that the species is similar to G. excavata. Examination of four of Haddon's specimens held at Cambridge University Zoological Museum (by kind permission of the curator Dr C. B. Goodhart) confirmed that the species was identical with that of Boveri. There were no directive mesenteries but there were many siphonoglyphs, 7, 8, 9 and 9 being recorded. The types and size ranges of the cnidae taken from one of the specimens were close to those found in the present material (Table 18). Actinioides sesere is thus considered conspecific. Actinioides sultana Carlgren, 1900: 63, was based on material from Zanzibar, and later redescribed (Carlgren, 1938: 33) from material collected from Durban. Carlgren's illustrations (1938, fig. 13) and description, including the size ranges of cnidae (Table 18), conform to the range of G. excavata given here. The coloration described by Carlgren resembles closely that which I have seen. Also, the sites at Aden and Gan were identical to that described by Carlgren ( 1 938: 34) as 'low water, numerous specimens, congregated closely together and forming carpets on flat rocks'. Thus A. sultana is also considered conspecific with G. excavata. Carlgren (1938: 23) also commented on the genus Actinioides Haddon & Shackleton, 1893, which Stephenson ( 1 922: 27 1 ) had supposed was congeneric with Bunodactis. Carlgren pointed out that if Actinioides dixoniana, the type species of the genus Actinioides, were shown to have acrorhagi proper (that is, with atrichs), and if there were more mesenteries at the base than at the margin, then the genus Actinioides would indeed lapse. Since the species A . sultana (Carlgren, 1 938: 33) certainly belonged neither to Bunodactis nor to Anthopleura, Carlgren said a new genus was required and he proposed Actiniogeton (Carlgren, 1938: 32). It has been shown that Actinioides dixoniana is best referred to Anthopleura (p. 240-245), and Actiniogeton might thus become a valid genus. However, for the reasons stated (p. 255) Actiniogeton is here considered to be congeneric with Gyractis. Actinioides rapanuensis Carlgren, 1920: 151, was said to lack siphonoglyphs and directives. Its sphincter varies as in G. excavata and the size ranges of the cnidae reported resemble those of the present material (Table 18). It is thus considered that siphonoglyphs are probably present, although the species was first thought to lack them as were A. sesere and A. sultana. Thus A. rapanuensis is regarded conspecific with G. excavata. Actiniogeton sesere (Dunn, 19746: 181-188), based on material from Hawaii, has all the characters but one of G. excavata. The exception, recorded by Dunn, was a single atypical structure having the appearance of siphonoglyph. But in two specimens she kindly lent me 3 and 5 siphonoglyphs were found. In her description the presence of microbasic p-mastigophores in the ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 261 column ectoderm is mentioned. The present specimens from Gan seemed to lack them, but in some specimens from Aden occasional microbasic amastigophores were seen. The ratio of microbasic amastigophores to basitrichs was 1 : 466 (4 amastigophores, 1860 basitrichs and 2 spirocysts). In one of Dunn's specimens 1035 basitrichs were counted but no amastigophores and it seems possible that in Dunn's paper the types and size ranges of the cnidae in the column and in the actinopharynx were inadvertently transposed. Revised data on type and size range of cnidae from one of Dunn's specimens are given in Table 18. A. sesere sensu Dunn, 19746, is also considered conspecific with G. excavata. Carlgren (1947: 14-15) examined some specimens which he believed to have been studied by Boveri. In a personal communication to Carlgren, Boveri stated that he had placed specimens of both G. excavata and G. pallida in a jar without a label. Carlgren found artrichal marginal spherules in the specimens and he thought they were Gyractis. He concluded that Gyractis was congeneric with Anthopleura. Carlgren also stated that several other specimens in small jars from a Dr Ondaajte's collection, the source of Boveri's material, similarly had atrichal marginal spherules; and some of these specimens were accompanied by figures showing a rich variety of colour. Several species of Anthopleura which occur in the Indian Ocean have bright colours but the only one likely to be confused anatomically with Gyractis is A. nigrescens. It too has several siphonoglyphs and no directive mesenteries (p. 248-250) but it can be distinguished from Gyractis by its prominent atrichal spherules with conspicuous nematocyst batteries, and in contrast its colour is usually dull brown to black. From Carlgren's remarks there seems no evidence to connect the specimens he examined with Boveri's Gyractis material, and in view of Boveri's excellent description it is doubtful that Carlgren's specimens were Gyractis. Of the four remaining species originally referred to Actiniogeton Carlgren, 1949, A. papuensis Haddon, 1898: 426, is conspecific with Anthopleura dixoniana (p. 240-245). From the original description Actinioides spenceri Haddon & Duerden, 1896: 159, is unlikely to be referrable to Gyractis, since it is reported as having more mesenteries at the base than at the margin and possibly atrichal marginal spherules. However, when Carlgren (1938: 33) examined Haddon & Duerden's slides of A. spenceri he could not detect the presence of atrichs though he reported spherules to be present. Actinioides ambonensis Kwietniewski, 1897: 389, needs to be re-examined before a decision can be made: it was described as having regularly arranged mesenteries, two siphonoglyphs and 24 conical spherules but Carlgren (1938: 33) was doubtful of these details. Genus MES ACTINIA nov. DEFINITION. Actiniidae with wide pedal disk. Column smooth, short, with scapus and capitulum. Margin with prominent perforate spherules, with one or more small suckers on the outer surface. Siphonoglyphs varied in number. Mesenteries regularly or irregularly arranged, with or without directives; growing from disk downwards, but same number distally as proximally. Some mesenteries of last cycle weak near base. Many perfect mesenteries; at least 1st and 2nd cycle fertile. Sphincter diffuse-weak. Asexual reproduction common. Retractor muscles diffuse, parietobasilar and basilar muscles well defined. Cnidome: spirocysts, basitrichs, microbasic amastigophores and microbasic p-mastigophores. TYPE SPECIES. Mesactinia ganensis gen. nov., sp. nov., by monotypy. REMARKS. A new genus is proposed to include a species of actiniid having the above characters. Carlgren (1949) listed only two genera having spherules (pseudospherules) and having the same number of mesenteries distally and proximally: Tealianthus Carlgren, 1927: 38, had the 1st cycle sterile, and Isotealia Carlgren, 1898: 25, had the first two cycles sterile. The species which Mesactinia is proposed to accommodate has the 1st and 2nd cycles fertile and thus differs from species usually referred to either of the other genera. Although the type species of Mesactinia usually shows an irregular arrangement due to asexual reproduction, the definition allows for the presence of species with regular arrangements of mesenteries and tentacles. 262 K. W. ENGLAND Fig. 29 Mesactinia ganensis gen. nov., sp. nov. (a) General appearance, (b) A specimen dividing into two from the base, (c) Vertical section of margin, BMNH 1984.2.9.90. (d-g) Variation in sphincter muscle in vertical sections of BMNH 1984.2.9.1 1 1-1 19. (h) Vertical section of margin of specimen from Singapore, BMNH 1984.2.9.83. (i-j) Vertical section showing variation in sphincter between two specimens from Singapore. Scale: (a-b) = 4 mm, (d-j) = 90 urn. (sp marginal spherule, sph sphincter, su sucker. Stippled, ectoderm; black, mesogloea; white, endoderm.) ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC Mesactinia ganensis gen. nov., sp. nov. Figs 29-30 263 TYPE LOCALITY AND MATERIAL. Holotype: Gan Island, Addu Atoll, Maldives, 10 Feb 1970, BMNH 1 984.2.9.90. Paratypes: Gan Island, 1 0 specimens, lOFeb 1970, BMNH 1984.2.9.2-8; c. 50 specimens, 7 Mar 1970, BMNH 1984.2.9.10-69. Singapore: Pulau Biola, 20 specimens, 13 Sep 1970, BMNH 1984.2.9.88-1 10; Pungol Point, 10 specimens, 11 Feb 1971, BMNH 1984.2.9.111- 1 19. Malaysia: Pulau Tioman, 25 specimens, BMNH 1984.2.9.120-145. DESCRIPTION. Column slightly taller than wide or about equal in height and diameter, smooth with regularly or irregularly arranged large perforate marginal spherules on clearly defined parapet (Fig. 29a, b). Spherules with one or two small suckers on outer surface (Fig. 29c, h). Fosse shallow, capitulum distinct. Tentacles up to two hundred, long, slender, with rounded tips which are often dilated in life, but not capitate since special types or concentrations of nematocysts absent; in several cycles, inner about halfway between mouth and margin, a clear gap between this cycle and next outer one. Oral disk wider than column, sometimes lobed; when contracted forming long funnel into body. Frequently reproducing by longitudinal fission, commencing at pedal disk (Fig. 29b). Anatomy. Mesenteries irregularly arranged, in four to five cycles, without directives. Sometimes one mesentery of a pair smaller than the other. Siphonoglyphs up to five, small to almost inconspicuous, or prominent. About same number of mesenteries distally as proximally, occasionally a few more distally since they grow from margin to base (p. 214). Near base mesenteries of last cycle extremely small and only visible as line on column wall, but apparent in sections. At least older mesenteries perfect and fertile, later cycles lacking filaments. Retractor muscles present. Oral and marginal stomata present. Sphincter varied, weak-diffuse (Figs 29d-i). rr\ \j b Fig. 30 Mesactinia ganensis gen. nov., sp. nov., nematocyst signature (see Table 19). Tentacle (a) spirocyst, (b) basitrich. Column (c) basitrich. Actinopharynx (d-e) basitrichs, (f) microbasic amastigo- phore. Filament (g-h) basitrichs, (i) microbasic p-mastigophore, (j) microbasic amastigophore. Scale: 264 K. W. ENGLAND 5 rr m p O so m « ON CN rN rn rn 4 rs t m rn osi rn p w ~- ^ T« oo rs rs rs rs /-) rs -«:r rn rn .S ON X X X X X X X X X o — Cu i i rs o 10 O 10 rN SO 10 OO 10 L **— 10 O sb rs TT 10 O ob !•) ^j- 11 rs */"> ob o r^ rn 0 u~> rn »i* ^^ Cu C' ~^ ^^ ~^ rN — rs ^^ cs 1 c > oo o oo O rn m p p so c ON rs 32 SO m 10 so O rn m t m ^ oo rs rs CS rh rfc rs rn rn rn jj1 03 ON x x x X X X X X X o O ~~ rN "~i rs »o CN m r~ op rs c 3E — ob rn rN rn rn m c o ob 55 c s: a •2 c u •S a O so p p P m so |> 'o 1 "o ti sO ^^ so O rn m O rn rn t m 69 CM ^ rs cs CS CS rn rn rs rn CN rn w 0 3 r x x x X X X X X X X (Z c« | S. >o p o SO 00 op •0 O S cs o CS CS ob O TT — rN s O ON — • rN cs rN PS S3 O i O 5 "5 m rN 2 m rN 1 O rs so rs -o 'c o 0? rn >o o cs ON ON £ O vt — rs ± ON o "o c/5 u 00 c 03 Uc u '« H ON "e" <•> ^ tT-c O S! *-- 53 U CJ J2 00 o '~ 4O ^y O .Er o '-s U .2 *o ** *• "8 §• -^ -^ _b ^ 5j -^ -^ o t« .y ^ 15 CS CS . Actiniaria and Corallimorpharia. Great Barrier Reef Expedition 1928-29. Scientific Reports 5 (7): 427-457. - 1950c. Actiniaria and Zoantharia from South Australia. Kungl. Fysiografiska Sdllskapets I Lund Forhandlingar 20 ( 10): 1-15. — \95Qd. A revision of some Actiniaria described by A. E. Verrill. Journal of the Washington Academy of Sciences 40(1): 22-28. - 1952. Actiniaria from North America. Arkivfor Zoologi (2) 3 (30): 373-390. — 1954. Actiniaria and Zoantharia from South and West Australia with comments upon some Actiniaria from New Zealand. Arkivfor Zoologi (2) 6 (34): 571-595. — 1959. Corallimorpharia and Actiniaria with description of a new genus and species from Peru. Reports of the Lund University Chile Expedition 1948-49, 38. Lunds Universitets Arsskrift. (N.F.) Avd 2, 56 (6): 1-39. & Hedgpeth, J. W. 1952. Actiniaria, Zoantharia and Ceriantharia from Shallow Water in the North- western Gulf of Mexico. Publications of the Institute of Marine Science 2 (21): 143-172, pis 1-4. Cocks, W. P. 1 85 1 . Actiniae (or sea-anemones) procured in Falmouth and its neighbourhood by W. P. Cocks, Esq., from 1843-1849. 19th Annual Report of the Royal Cornwall Polytechnic Society: 3-1 1, pis 1-2. Cutress, C. E. 1955. An interpretation of the structure and distribution of the cnidae in the Anthozoa. Systematic -oology 4 (3): 1 20-1 37. Dalyell, J. G. 1 848. Rare and remarkable animals of Scotland, represented from living subjects; with practical observations on their nature. 2: iv + 322, 56 pis. (Actiniaria, ch. 10: 195-240, pis 45-49). Dana, J. D. 1846, 1849. United States Exploring Expedition during the years 1838-42. Under the command of Charles Wilkes, USN. Philadelphia. Vol. 1, text (1846); Vol. 2, atlas (1849). — 1 859. Synopsis of the report of zoophytes of the United States Exploring Expedition. New Haven; publ. by author. Danielssen, D. C. 1980. Actinida. Den Norske Nordhavs-Expedition 1876-1878 (Zoology) 19: 1-184, pis 1-25. Dixon, G. Y. 1886. Notes on two Irish specimens of Edwardsia timida Quatrefages. Proceedings of the Royal Irish Academy 5: 100-106. — & A. F. 1 889. Notes on Bunodes thallia, Bunodes verrucosa and Tealia crassicornis. Scientific Proceedings of the Royal Dublin Society 6: 310-326, pis 4-5. Drayton. See Dana, 1846. Diiben, M. W. & Koren, J. 1 847. Om nogle norske Actinier. Forhandlinger vedde skandinaviske Naturforskeres (4): 266-268. Duchassaing, P. & Michelotti, J. 1860. Memoires sur les coralliaires des Antilles. Memoire della Reals Accademia delle Scienze di Torino (2) Pt. 19. (Actinians: 3 1 3-325; (explanation of plates: 364-365), pis 6-7.) — & 1 866. Supplement aux memoires sur les coralliaires des Antilles. Memoire della Reale Accademia delle Scienze di Torino (2) Pt. 23: 97-206, pis 1-11. (Actiniaria, pp. 120-134, pis 5-6). Duerden, J. E. 1895. On the genus Alicia (Cladactis) with an anatomical description of A. costae Panceri. Annals and Magazine of Natural History (6) 15: 213-218, pi. 9. — 1897. The actiniarian family Aliciidae. Annals and Magazine of Natural History (6) 20: 1-15, pi. 1 . — 1898. The Actiniaria around Jamaica. Journal of the Institute of Jamaica 2: 449-465. Dunn, D. F. \914a. Redescription of Anthopleura nigrescens (Coelenterata, Actiniaria) from Hawaii. Pacific Science 28: 377-382. — 19746. Actiniogeton sesere (Coelenterata, Actiniaria) in Hawaii. Pacific Science. 28: 181-188. — 1977. Anthopleura handi n. sp. (Coelenterata, Actiniaria), an internal brooding, intertidal sea anemone from Malaysia. Wasmann Journal of Biology 35 (1): 54-64. — 198 1 . The clownfish sea-anemones: Stichodactylidae (Coelenterata, Actiniaria) and other sea anemones symbiotic with pomacentrid fishes. Transactions of the American Philosphical Society 71 (1): 1-1 15. -, Fu-Shiang Chia & Levine, R. 1980. Nomenclature of Aulactinia ( — Bunodactis), with description of Aulactinia incubans n. sp. (Coelenterata, Actiniaria), an internally brooding sea anemone from Puget Sound. Canadian Journal of Zoology 58: 2071-2080. Ehrenberg, C. G. 1834. Die Corallenthiere des rothen Meeres, physiologisch Untersucht und systematisch Verzeichnet Berlin. Ellis, J. 1768. An account of the Actinia sociata, or clustered animal-flower, lately found on the sea coasts of the newly-ceded Islands: In a letter from John Ellis, Esquire, F.R.S., to the Right Honourable The Earl of Hillsborough, F.R.S. Philosophical Transactions of the Royal Society of London (1767) 57: 428-437. England, K. W. 1969. Anthopleura elatensis n. sp. (Actiniidae: Actiniaria) from the Red Sea. Israel Journal of Zoology 18: 1-7. - 1971. Actiniaria from Mururoa Atoll, Tuamoto, Polynesia (Hormathiidae: Calliactis polypus, Sagartiidae: Verrillactis n. gen. paguri). Cahiers du Pacifique 15: 23-40, pis 1-4. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 287 Farquhar, H. 1898. Preliminary account of some New-Zealand Actiniaria. Journal of the Linnean Society of London (Zoology) 26: 527-536, pi. 36. Faurot, L. 1895. Etudes sur Tanatomie, Thistologie et le developpement des actinies. Archives de Zoologie experimental et generate (3) 3: 43-262, 1 2 pis. Fischer, P. 1888. Descriptions d'une nouvelle espece du genre Edwardisa Quatrefages. Bulletin de la Societe Zoologique de France 13: 22-23. - 1889. Nouvelle contribution 1'actinologie francaise. Actes de la Societe Linneene de Bordeaux (5) 43: 251-311. Forsskal, P. 1 775. Descriptions animalium, . . . quae in itinere orientali observavit. Hauniae. Goldfuss, G. A. 1820. Handbuch der Zoologie. Vol. 1, Niirnberg. Gosse, P. H. 1 853. Notes on new or little known marine animals. Annals and Magazine of Natural History (2) 12:124-128,153-159. — 1 855. On Peachia hastata with observations on the family Actiniidae. Transactions of the Linnean Society of London 21: 267-276. — 1858. Characters and descriptions of some new British sea-anemones. Annals and Magazine of Natural History (3) 1:414-419. - 1859. Characters and descriptions of some British sea-anemones. Annals and Magazine of Natural History (3) 3: 46-50. - 1860. Actinologia Britannica: A history of the British sea-anemones and corals. London. xl + 362pp., 1 1 pis. 1865. On Aegon alfordi a new British sea-anemone. Annals and Magazine of Natural History (3) 16: 41-44, pi. 7. (•rube, A. E. 1840. Actinien, Echinodermen und Wurmer des Adriatischen und Mittelmeers. Konigsberg. 92 pp, Ipl. Haddon, A. C. 1889. A revision of the British Actiniae. Part I. The Scientific transactions of the Royal Dublin Society (2) 4: 297-361, pis 31-37. — 1898. The Actiniaria of Torres Straits. Scientific transactions of the Royal Dublin Society (2) 6: 393-498, pis 22-32. — & Duerden, J. E. 1896. On some Actiniaria from Australia and other districts. Scientific transactions of the Royal Dublin Society (2) 6 (6): 139-164, pis 7-10. & Shackleton, A. M. 1893. Description of some new species of Actiniaria from Torres Straits. Scientific proceedings of the Royal Dublin Society (N.S.) 8: 116-131. Hand, C. 1955-1956. The sea anemones of central California. Part 1. The corallimorpharian & athenarian anemones. Wasmann journal of biology 12 (3): 345-375. Part 2. The endomyarian and mesomyarian anemones. Wasmann journal of biology 13(1): 37-99. Part 3. The acontiarian anemones. Wasmann journal of biology 13 (2): 189-251. Hartog, J. C. den 1980. Caribbean shallow water Corallimorpharia. Zoologische verhandelingen 176: 1-83, pis 1-14. Hertwig, R. 1 882. Report on the scientific results of the voyage of H. M.S. Challenger during the years 1873-1876 (Zoology) 6(1) Actiniaria: 1-136, pis 1-14. Iredale, T. 191 1. On some misapplied molluscan generic names. Proceedings of the Malocological Society 9: 261-262. Johnson, J. Y. 1 86 1 . Notes on the sea anemones of Madeira. Proceedings of the Zoological Society of London (1861): 298. Klunzinger, C. B. 1877. Die Korallthiere des rothen Meeres. Berlin, vii + 98 pp, 8 pis. Kwietniewski, C. R. 1896. Actiniaria von Ternate nach den Sammlungen von Herrn Prof. Dr. W. Kiikenthal. Zoologischer Anzeiger 19(512): 388-39 1 . - 1898. Actiniaria von Ambon und Thursday Island. [In Semon, R., Zoologische Forschungsreisen in Australien unddem malayischen Archipel. 5 (Lief. 4)] Denkschriften der Medizinisch-Naturwissenschaftlichen Gesellschaft zu Jena 8: 385-430, pis 25-30. Lager, E. 191 1. Actiniaria. In Michaelson & Hertmeyer, Fauna Sudwest-Austr -aliens 3 (8): 215-249. Lee, A. B. 1921. Microtomists vade-mecum. 8th edn. London. x + 594pp. Lutkens, C. 1 860. Nogle bemarkninger om de ved de danske kyster iagttagne arter af actiniernus gruppe. Videnskabelige Meddelelserfra den Naturhistoriske Forening i Kjobenhavn 1-4: 184-200. Manuel, R. L. 1975. A new sea anemone from a brackish lagoon in Sussex, Edwardsia ivelli, sp. nov. Journal of natural history 9: 705-7 1 1 . - 1977. A redescription of Edwardsia beautempsi and E. timida (Actiniaria: Edwardsidae). Cahiers de biologie marine 18: 483-497. 288 K. W. ENGLAND — 1 98 1 . On the identity of the sea anemone Scolanthus callimorphusGosse, 1 853 ( Actiniaria: Edwardsiidae). Journal of natural history 15: 265-276. McMurrich, J. P. 1893. Report of the Actiniae collected by the U.S. Fish Commission steamer 'Albatross' during the winter of 1887-8. Proceedings of the United States National Museum 16: 119-216. — 1 897. Notes on some actinians from the Bahamas Islands, collected by the late Dr. J. I. Northrop. Annals of the New York Academy of Sciences 9 (1896): 18 1-194, pi. 17. - 1901. Report on the Hexactiniae of the Columbia University Expedition to Puget Sound during the summer of 1896. Annals of the New York Academy of Science 14: 1-52, 3 pis. - 1904. The Actiniae of the Plate Collection. Zoologische Jahrbiicher, Jena. (Supplement 6) (Fauna Chilensis 3): 215-306, pis 14-19. 1913. On two new actinians from the coast of British Columbia. Proceedings of the Zoological Society (19 13): 963-972, pi. 89. Milne Edwards, H. 1 857. Histoire naturelle des coralliaires ou polypes proprement dits Paris, 1 . xxxiv + 324 pp. — & Haime, J. 1851. Monographic des polypiers fossiles des terrains palaeozo'iques, precedee d'un tableau general de la classification des polypes. Archives du Museum d' Histoire Naturelle, Paris. 5: 1-502, pis 1-20. Panceri, P. 1 869. Intorno a due nuovi polypi. Societe reals di Napoli, attidell'accademia delle Science fisichie e matematiche. 4(11): 1-11. Panikkar, N. K. 1 939. Studies on Peachia from Madras. Proceedings of the Indian Academy of Sciences. 7B (4): 182-205, pi. 24, text-figs 1-14. Parrty, G. 1951. The Actiniaria of New Zealand: a check list of recorded and a new species, a review of the literature and a key to the commoner forms. Part 1 . Records of the Canterbury Museum 6(1): 83-1 19. Parulaker, A. 1 968. On a new species of sea anemone from Maharashtra, India. Journal of the Bombay Natural History Society 65 (3): 590-595. Pax, F. 1907. Vorarbeiten zu einer Revision der Familie Actiniidae. Inaugural dissertation, Breslau. — 1908. Anthozoa: Die Actinienfauna westafrikas. Denkschriften der Medizinisch-Naturwissenschaftlichen Gesellschaft zu Jena 30: 463-504, pi. 25. — 1 920. Zoantharia and Actiniaria. In Michaelsen, J. W. Beitrdge zur kenntnis der Meeresfauna westafrikas 3: 23-33. 1926. Die Actinien der Deutschen Siidpolar-Expedition 1901-1903. Deutsche Siidpolar- Expedition 18 (zool. 10): 1-62. Peacock, H. A. 1966. Elementary microtechnique. 3rd edn. xi + 547 pp. Pennant, T. 1777. The British zoology. 4th edn. London. Pennington, A. S. 1885. British zoophytes: an introduction to the Hydroida, Actinozoa, and Poly zoa found in Great Britain, Ireland, and the Channel Islands. London, xvi + 363 pp, 24 pis. Pickering. See Dana, 1846. Quatrefages, A. de 1841 . Cotes de la Manche. Institut 9: 427. - 1 842. Memoires sur les edwardsies (Edwardsia Nob.) nouveau genre de la famille des actinies. Annales des Sciences naturelles (2) Zoology 18: 65-109, pis 1-2. Quoy, J. R. C. & Gaimard, P. 1833. Voyage de decouvertes de I' Astrolabe. Zoology. 4 (Zoophytes). Paris. Rafinesque, C. S. 1815. Analyse de la nature ou tableau de I'univers et des corps organises. Palermo. Rawlinson, R. 1935. Edwardsia callianthus sp. nov. A new British species from Menai Straits. Journal of the Marine Biological Association of the United Kingdom 20: 129-146. Riemann-Ziirneck, K. 1975. Actiniaria des Siidwestatlantik 2. Sagartiidae und Metridiidae. Helgoldnder Wissenschaftliche Meeresuntersuchungen 27: 70-95. Ross, D. M. & Wada, T. 1975. The behaviour of actinians in symbiotic associations with pagurids in Japan. Canadian journal of zoology 53: 1 735-1 748. Rueppell, E. & Leuckart, F. S. 1828. Atlas zu der Reise im nordlichen Afrika von Eduard Ruppell. Neue Wirbellose Thiere des rothen Meeres. Frankfurt am Main. pp. 1-6. Schmidt, J. 1969. Die Nesselkapseln der Aktinien und ihre differentialdiagnostische Bedeutung. Helgoldnder Wissenschaftlichen Meeresuntersuchungen 19: 284-317. - 1970. Anthopleura stellula (Actiniaria, Actiniidae) and its reproduction by transverse fission. Marine Biology, Berlin, 5: 245-255. - 1972. Prodromus zu einer Monographic der mediterranean Aktinien. Zoologica, Stuttgart 121: 1-146. 1979. Beitrage zue Differentialdiagnose, Morphologic und Evolution der Edwardsiidae (Actiniaria, Anthozoa) 1. Die Gattung Alfredus nov. gen. mil der Typusart A. lucifugus (Fischer, 1888). Zeitschrift fur die zoologische Systematik und Evolutionsforschung 17: 21 1-220. Seshaiya, R. V. & Cutress, C. E. 1969. Edwardsia jonesiin. sp. (Actiniaria, Edwardsiidae) from Porto Novo, S. India. Journal of the Marine Biological Association of India 11: 73-77. ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 289 Simon, J. A. 1892. Ein Beitrdg zur Anatomic und Systematik der Hexactinien. Inaugural Dissertation, Miinchen. Pp. 1-106. Stephenson, T. A. 1918. Coelenterata. Part 1 — Actiniaria. British Antarctic (Terra Nova) Expedition, 1910. Natural History Report. (Zoology) 5(1): 1-68. — 1 920. On the classification of Actiniaria 1 . Quarterly journal of the Microscopical Society 64: 425-574. — 1 92 1 . On the classification of Actiniaria 2. Quarterly journal of the Microscopical Society 65: 493-576. — 1922. On the classification of Actiniaria 3. Quarterly journal of the Microscopical Society 66: 247-319. — 1928. The British sea anemones Vol. 1 . Ray Society, London, xii + 148 pp, pis 1-14. — 1935. The British sea anemones Vol. 2. Ray Society, London, ix + 426 pp, pis 15-33. — , Stephenson, A., Tandy, G. & Spender, M. 1931. The structure and ecology of Low Isles and other reefs. Great Barrier Reef Expedition, 1928-29. Scientific Reports 3 (2): 17-112, pis 1-27. Stimpson, W. 1853. Synopsis of marine Invertebrata of Grand Manan: or the region about the mouth of the Bay of Fundy, New Brunswick. Smithsonian contributions to knowledge 6: 1-66, pis 1-3. (Actiniidae: 7-8, pi. 1). Stuckey, W. 1908. Notes on a New Zealand actinian, Bunodes aureoradiata. Transactions of the New Zealand Institute. 41: 367-369, pi. 17. — 1909. A review of New Zealand Actiniaria known to science, together with a description of twelve new species. Transactions of the New Zealand Institute 41: 374-398, pis 21-28. Uchida, T. 1938. Report of the Biological survey of Mutsu Bay. No. 33 Actiniaria of Mutsu Bay. Science Reports of the Tohoku Imperial University (4) Biology. 13: 281-317, pi. 11. - 1960. Carcinactis ichikawai, n. gen. n. sp., an actiniarian commensal with the crab Dorippe granulata. Japanese journal of zoology. 12: 595-601, pi. 1. & Maramatsu, S. 1958. Notes on some Japanese sea anemones. Journal of the Faculty of Science, Hokkaido University (6, Zoology) 14: 1 1 1-1 19. Verrill, A. E. 1 864. Revision of the polypi of the eastern coast of the United States. Memoirs of the Boston Society of Natural History 1: 14-45. — 18690. Synopsis of the polyps and corals of the North Pacific exploring expedition, under Commodore C. Ringgold and Capt. John Rodgers, U.S.N., from 1853-1856. Collected by Dr. Wm. Stimpson, naturalist to the expedition. Part 4, Actiniaria. Proceedings of the Essex Institute 6: 1-104. 2 pis. - 18696. Notes on Radiata No. 6. Review of the corals and polypes of the west coast of America. Transactions of the Connecticut Academy of Arts and Sciences, New Haven 1: 490-495. - 1869f. Description of a jellyfish and two actinians. Annals and magazine of natural history (4) 4: 160-163. 1898. Description of new American actinians with critical notes on other species. Part 1. American journal of science (4) 6: 493-498. — 1922. Alcyonaria and Actiniaria. Report of the Canadian Arctic Expedition, 1913-1918 1 (G): 1-165. 1928. Hawaiian shallow- water Anthozoa. Bulletin of the Bernice P. Bishop Museum 49: 1-30, pis 1-5. Watzl, O. 1922. Die actiniarien der Bahamainseln. Arkivfor Zoologi 14 (24): 1-89, 1 pi. Weill, R. 1 934. Contributions a 1'etude des cnidaires et de leur nematocystes. Travaux de la Station Zoologique de Wimereux 10: 1-347, 11: 347 (sic)-701. Williams, R. B. 1979. Edwardsia Costa, 1834 (Arthropoda, Crustacea): Proposed suppression under the plenary powers with conservation of Edwardsia de Quatrefages, 1841 and Edwardsiidae Andres, 1881 (Coelenterata: Actiniaria). Z.N.(S.) 2261. Bulletin of zoological nomenclature 36: 175-179. — 198 1 . A sea-anemone, Edwardsia meridionalis sp. nov., from Antarctica and a preliminary revision of the genus Edwardsia de Quatrefages, 1841 (Coelenterata: Actiniaria). Records of the Australian Museum 33: 325-360. Cornelius, P. F. S. & Clark, A. M. 1982. Proposed conservation of Actinia Linnaeus, 1767 and Actiniidae Goldfuss, 1820 (Coelenterata, Actiniaria) and Pentacta Goldfuss, 1820 (Echinodermata, Holothurioidea). Z.N.(S.) 825. Bulletin of zoological nomenclature 39: 288-292. Wilson, H. V. 1890. On a new actinia, Hoplophoria coralligens. Studies from the biological laboratory, Johns Hopkins University 4 (6): 379-387. Additional Reference: England, K. W. (in press) Redefinition and systematics of Heteractis aurora, the genera Heteractis and Radianthus, and the family Heteractiidae (Cnidaria: Actiniaria). Indo-Malayan Zoology. 290 K. W. ENGLAND Index Valid names are in Roman type, new names in bold type and invalid names in italics. New combinations are indicated by an asterisk (*). Main page references are in bold type. ABASILARIA215 ACONTIARIA 279 Actinia 2 10, 235, 268, 279 citrina 269 decorata 279, 280 polypus 279 ACTINIIDAE 238, 272, 273 Actiniogeton 254 papuensis 240, 241, 243, 244, 245 rapanuensis 255, 259, 260 sesere 208, 255, 258, 259, 260 sultana 244, 255, 258, 259 Actinioides 238, 243, 254 ambonensis 261 dixoniana 212, 240, 241, 243, 244, 245 papuensis 212, 240, 24 1 , 243, 244, 245 rapanuensis 255, 259, 260 sesere 243, 244, 255, 258, 259, 260 spenceri 26 1 sultana 255, 258, 260 Actinodendron sp. 207 Actinothoe 283, 284 Actinothoe sp. 207 Adamsia 279 decorata 279 miriam 279 adenensis, Edwardsia 217 adhaesivium, Cryptodendron 208 Aegean 238 Aiptasia 207, 208 Alfredus 229 ALICIIDAE232 ambonensis, Actinioides 261 andresi, Edwardsioides 218 Anemonia210, 268, 272 citrina 268, 269 annamensis, Edwardsia 217, 221 annae, Anthopleura 239 Antheopsis 273, 274 australiensis 273 carlgreni 276 concinnata 273 glandulosa 276 koseirensis, 207, 273, 276 kwietniewski 276 malayensis sp. nov. 263, 274-276 Anthopleura 210, 238-240 annae 239 artemisia 239 asiatica 239 aureoradiata 239 247 balli 239 dixoniana 206, 207, 208, 212, 213, 239, 240-245, 254 elegantissima 239 handi210, 239, 245-248 hermaphroditica 239, 245, 246, 247, 248 insignis 239 krebsi 239 kurogane 239 michaelseni 239 midori 239 mortenseni 239 nigrescens 213, 239, 248-250, 261 orientalis 239 pacifica 239 pannikkarri 239 rubripunctata 239 sanctaehelenae 239 stellula 207, 208, 239 varioarmata 239 waridi* 208, 209, 239, 250-254 xanthogrammica 239 Anthothoe 283 Anthostella 238, 268 arctica, Edwardsioides* 218 armata, Edwardsia 229 armatus, Scolanthus* 207, 229-232 armillatus, Calliactis 279 Artemidactis 283 artemisia, Anthopleura 239 asiatica, Anthopleura 239 ATHENARIA215 Atrichal marginal spherules 211 Aulactinia 239, 255 capitata 255 incubans 255 aureoradiata, Anthopleura 239, 247 aurora, Heteractis 272 australiensis, Antheopsis 273 australiensis, Epiactis 213 australis, Isanemonia 240 balli, Anthopleura 239 Basitrichal marginal spherules 211 beautempsi, Edwardsia 216, 217, 221 bocki, Edwardsia 224, 226, 228 boninensis, Metarhodactis 208 bryoides, Rhodactis 208 Bunodactis 238, 239 hermaphroditica 248 maculosa 240, 255 mortenseni 255 nigrescens 248 novazealandica 255 reynaudi 255 waridi 250 Bunodes 238, 274 koseirensis 274 waridi 250 californica, Edwardsia 217 Calliactis 279, 281 armillatus 279 miriam 279 polypus 2 14, 279-281 callianthus, Edwardsia 216 callimorphus, Scolanthus 229 capensis, Edwardsia 217 capitata, Aulactinia 255 Carcinactis 283 dolosa 28 1,284 ichikawai 280, 284 carlgreni, Antheopsis 276 carlgreni, Edwardsia 217 carneola, Edwardsia 216 Cassiopeia sp. 207 Cereus 238 Chandrae tinia 279 CHONDRA CTINIIDA E 279 CHONDRACTINIINAE279 Chondrophellia 279 cincta, Hoplophoria 233 cincta, Triactis 233 cincta, Viatrix 233, 234 cinctus, Phyllodiscus 233 citrina. Actinia 269 citrina, Anemonia 268, 269 citrina, Isactinia 268, 269, 270, 272 citrina, Telactinia* 208, 210, 263, 269-272 Cladonema sp. 207 claparedi, Edwardsia 217, 221 concinnata, Antheopsis 273 Condylactis 266, 268 Corynactis sp. 208 Cnidopus verater 272 Cribrina 238, 279 Cribrina polypus 279 Crassostrea 253 Cryptodendron adhaesivium 208 danica, Edwardsia 217 decorata. Actinia 279, 280 decorata, Adamsia 279 Diogenes diogenes 207 dixoniana, Anthopleura 206, 207, 208, 212,213,239,240-245,254 dixoniana, Actinioides 212, 240, 241, 243, 244, 245 dolosa, Carcinactis 281, 284 duodecemtentaculata, Edwardsia 216 duodecemtentaculata, Edwardsioides* 218 Edwardsia 2 15-2 19 adenensis 2 1 7, 224, 226 annamensis 2 17, 221 armata 229 beautempsi 2 16, 217, 221 bocki 224, 226, 228 californica 217 callianthus 216 capensis 2 1 7 carlgreni 2 1 7 carneola 216 claparedi 2 17, 221 ACTINIARIA FROM RED SEA AND TROPICAL INDO-PACIFIC 291 danica 217 duodecemtentaculata 216 elegans 2 17, 221,223 gilbertensis 217 hantuensis sp. nov. 207, 217, 221-224,231 ivelli217,221 longicornis 2 1 6, 2 1 7, 22 1 mamillata 229 maroccana 217, 221 novazelanica 217 octoradiata 217, 221 pudica207,217,224 rakaiyae2ll,224,226 sanctaehelenae 217 sipunculoides 216, 217 stephensoni 2 1 7, 224, 226, 227, 228 sulcata217,221,223 timida 216 tintrix217, 221 tricolor 217 tuberculata217, 221 vermiformis 224, 226 vivipara216 Edwardsianthus gen. nov. 207, 217, 218, 219,224 gilbertensis* 2 18, 228, 231 pudica* 2 18, 224-229, 231 Edwardsiella 224 pudica 224 EDWARDSIIDAE 215, 219 Edwardsioides 218, 219, 224 andresi 218 arctica* 218 duodecemtentaculata* 218 finmarchica* 218 fusca218 islandica* 218 japonica* 218 jonesi*218 kameruniensis* 218 meridionalis* 218 norvegica* 218 timida* 218 vegae* 218 vitrea218, 224 vivipara* 218 elegans, Edwardsia 217, 221, 223 elegans, Sagartia 277 elegantissima, Anthopleura 239 Entacmaea 212, 273, 274, 276 quadricolor 208, 274, 277-278 Epiactis212 australiensis 213 novazealandica 255 prolifera 240 Evactis 238 excavata, Gyractis 208, 212, 244, 255-261 finmarchica, Edwardsioides* 218 fusca, Edwardsioides 218 ganensis, Mesactinia, gen. nov., sp. nov. 213,261-265,274 gigantea, Stichodactyla 207 gilbertensis, Edwardsia 217 gilbertensis, Edwardsianthus* 218, 228, 231 glandulosa, Antheopsis 276 globulifera, Viatrix 238 Gyractis 210, 239, 254-255, 260 excavata 208, 210, 244, 255-261 pallida 255, 260 Habrosanthus 283 Halcampa215 Halophila ovalis 207 Haliplanella luciae 207 handi, Anthopleura 210, 239, 245-248 hantuensis, Edwardsia, sp. nov. 207, 217,221-224,231 helianthus, Stichodactyla 273 hemprichi, Heterodactyla 21 1 hermaphroditica, Anthopleura 239, 245, 246, 247, 248 hermaphroditica, Bunodactis 248 hertwigi, Paracondylactis 265 Heteractis 272, 274 aurora 272 Heteranthus verraculatis 207 Heterodactyla hemprichi 21 1 Hoplophoria 232, 235, 238 cincta 233, 238 Hormathia 279 HORMATHIIDAE 279 ichikawai, Carcinactis 280, 284 ignota, Isactinia 269, 271, 272 incubans, Aulactinia 255 indicus, Phyllodiscus 233, 238 insignis, Anthopleura 239 Isacmaea 238 Isactinia 238, 268 citrina 268, 269, 270, 272 ignota 269, 211,272 lobata 269, 21 1,272 mesembryanthemum 272 olivacea 271, 272 Isanemonia australis 240 islandica, Edwardsioides* 218 Isoedwardsia 229 nidarosiensis 229 Isotealia 261 ivelli, Edwardsia 2 17, 221 japonica, Edwardsioides* 218 jonesi, Edwardsioides* 218 kameruniensis, Edwardsioides* 218 koseirensis, Antheopsis 207, 273 koseirensis, Bunodes 274 kenti, Stichodactyla 207 krebsi, Anthopleura 239 kukenthali, Radianthus 272 kurogane, Anthopleura 239 kwietniewski, Antheopsis 276 lobata, Isactinia 269, 271 , 272 loligo, Phymanthus 208 longicornis, Edwardsia 216, 217, 221 luciae, Haliplanella 207 Macrodactyla 266, 268 macrodactylus, Radianthus 208 maculosa, Bunodactis 240, 255 magnifica, Radianthus 208 malayensis, Antheopsis sp. nov. 263, 274-276 malu, Radianthus 208 mamillata, Edwardsia 229 Marginal spherules, atrichal 21 1 Marginal spherules, basitrichal 211 Marginal spherules, mastigophoral 211 maroccana, Edwardsia 2 1 7, 22 1 mastigophore, microbasic telotrichous 221 mastigophoral marginal spherule 211 meridionalis, Edwardsioides* 218 Mesactinia gen. nov. 210, 261 ganensis sp. nov. 213, 261-265, 274 mesembryanthemum, Isactinia 272 Metapeachia tropica 207 Metarhodactis boninensis 208 michaelseni, Anthopleura 239 microbasic telotrichous mastigophore 221 midori, Anthopleura 239 miriam, Adamsia 279 miriam, Calliactis 279 Monostephanous 238 mortenseni, Anthopleura 239 mortenseni, Bunodactis 255 muscosus, Phymanthus 208 Nassarius spp. 207 Nemanthus sp. 208 Nematocysts 209, 214 in Edwardsia 219-221 microbasic telotrichous mastigophore 220, 221 pterotrich 220 signature 214 Nematosphere 21 1 Neocondylactis gen. nov. 265 singaporensis sp. nov. 207, 265-268 nidarosiensis, Isoedwardsia 229 nidarosiensis, Scolanthus 229 nigrescens, Anthopleura 213, 239, 248-250,261 nigrescens, Bunodactis 248 nigrescens, Tealiopsis 248 norvegica, Edwardsioides* 218 novazealandica, Bunodactis 255 novazealandica, Epiactis 255 novazelanica, Edwardsia 217 NYNANTHEAE215 octoradiata, Edwardsia 217, 221 olivacea, Isactinia 271, 272 orientalis, Anthopleura 239 Oulactis210 ovalis, Halophila 207 pacifica, Anthopleura 239 paguri, Sagartia 282, 283 paguri, Verrillactis 214, 279, 281-284 292 K. W. ENGLAND pallula, Gyractis 255. 260 pannikkarri, Anthopleura 239 papuensis, Actiniogeton 240, 241, 243, 244, 245 papuensis, Actinioides 2 1 2, 240, 241 , 243, 244, 245 Paracalliactis 279 Paracondylactis 207, 210, 263, 266, 268 hertwigi 265 Phelliactis 279 Phialoba212 Phyllodiscus 232 cinctus 233 indicus 233, 238 semoni 238 Phymanthus loligo 208 muscosus 208 polypus. Actinia 279 polypus, Calliactis 214, 279-281 polypus, Cribrina 279 polypus, Priapus 279 Priapus polypus 279 producta, Triactis 233-238 prolifera, Epiactis 240 Pterotrich 220 pudica, Edwardsia 207, 217, 224 pudica, Edwardsianthus* 218, 224-229, 231 pudica, Edwardsiella 224 quadricolor, Entacmaea 208, 274, 277-278 Radianthus 272, 273, 274 koseirensis 273 kukenthali 272 macrodactyla 208 magnifica 208 malu 208 rakaiyae, Edwardsia 217, 224, 226 rapanuemis, Actiniogeton 255, 259, 260 rapanuenxix, Actinioides 255. 259, 260 reynaudi, Bunodactis 255 Rhodactis 208 bryoides 208 rhodostoma 208, 209 rhodostoma, Rhodactis 208, 209 Ricordea yuma 208 rubripunctata, Anthopleura 239 Sagartia elegans 277 paguri 282, 283 Sagartianthus sp. 207, 283 Sagartiogeton 284 SAGARTIIDAE279,281 sanctaehelenae, Anthopleura 239 sanctaehelenae, Edwardsia 217 Scolanthus 229 armatus* 207, 229 232 callimorphus 229 nidarosiensis 229 semoni, Phyllodiscus 238 sesere, Actiniogeton 208, 255, 258, 259, 260 sesere, Actinioides 243, 244, 255, 258, 259, 260 Signature, nematocyst 214 singaporensis, Neocondylactis gen. nov., sp. nov. 207, 265-268 sipunculoides, Edwardsia 216, 217 spenceri, Actinioides 261 stellula, Anthopleura 207, 208, 239 slephensoni, Edwardsia 217, 224, 226, 227, 228 Stichodactyla 272, 273 gigantea 207 helianthus 273 kenti 207 STICHODACTYLIDAE 272-274 Stoichactis 272 sulcata, Edwardsia 217, 221, 223 sultana, Actiniogeton 244, 255, 258, 259 sultana, Actinioides 255, 258, 260 Tarastephanus 238 Tealianthus261 Tealiopsis 238 nigrescens 248 Telactinia gen. nov. 268 citrina* 208. 210, 263, 269-272 Telotrichous mastigophore, microbasic 221 THENARIA232 timida, Edwardsia 216 timida, Edwardsioides* 218 tintrix, Edwardsia 217, 221 Triactis 2 10, 232 cine la 233 producta 233-238 tricolor, Edwardsia 217 Tristemma 279 tropica, Metapeachia 207 tuberculata, Edwardsia 217, 221 varioarmata, Anthopleura 239 vegae, Edwardsioides* 218 verater, Cnidopus 272 vermiformis, Edwardsia 224, 226 verraculatis, Heteranthus 207 Verrillactis 281 Verrillactis paguri 214, 279, 281-284 Vialrix 232, 235 cincta 233, 234 globulifera 238 vitrea, Edwardsioides 218, 224 vivipara, Edwardsia 216 vivipara, Edwardsioides* 216 waridi, Anthopleura* 208, 209, 239, 250-254 waridi, Bunodactis 250 waridi, Bunodes 250 xanthogrammica, Anthopleura 239 yuma, Ricordea 208 British Museum (Natural History) The birds of Mount Nimba, Liberia Peter R. Colston & Kai Curry-Lindahl For evolution and speciation of animals Mount Nimba in Liberia, Guinea and the Ivory Coast is a key area in Africa representing for biologists what the Abu Simbel site in Egypt signified for archaeologists. No less than about 200 species of animals are endemic to Mount Nirnba. Yet, this mountain massif, entirely located within the rain-forest biome, is rapidly being destroyed by human exploitation. This book is the first major work on the birds of Mount Nimba and surrounding lowland rain-forests. During 20 years (1962-1982) of research at the Nimba Research Laboratory in Grassfield (Liberia), located at the foot of Mount Nimba, scientists from three continents have studied the birds. In this way Mount Nimba has become the ornithologically most thoroughly explored lowland rain-forest area of Africa. The book offers a comprehensive synthesis of information on the avifauna of Mount Nimba and its ecological setting. During the 20 years period of biological investigations at Nimba this in 1962 intact area was gradually opened up by man with far-reaching environmental consequences for the rain-forest habitats and profound effects on the birds. Therefore, the book provides not only a source of reference material on the systematics, physiology, ecology and biology of the birds of Mount Nimba and the African rain-forest, but also data on biogeography in the African context as well as conservation problems. Also behaviour and migration are discussed. At Nimba a number of migrants from Europe and/or Asia meet Afrotropical migratory and sedentary birds. Professor Kai Curry-Lindahl has served as Chairman of the Nimba Research Laboratory and Committee since its inception in 1962. Peter Colston is from the Subdepartment of Ornithology, British Museum (Natural History), Tring, and Malcolm Coe is from the Animal Ecology Research Group, Department of Zoology, Oxford. 1 986, 1 29pp. Hardback. 0 565 00982 6 £1 7.50. Titles to be published in Volume 53 Puellina (Bryozoa; Cheilostomata; Cribrilinidae) from British and adjacent waters By J. D. D. Bishop & B. C. Househam Miscellanea The genera of pelmatochromine fishes (Teleostei, Cichlidae). A phylogenetic review. By P. H. Greenwood Certain Actiniaria (Cnidaria, Anthozoa) from the Red Sea and Tropical Indo- Pacific Ocean. By K. W. England Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset